part03/week1_170207/비정형_데이터마이닝.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 제 4장 비정형 데이터마이닝"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 제1 절 텍스트 마이닝\n",
    "\n",
    "- 텍스트 마이닝은 텍스트(text)를 이용한 데이터 마이닝이라고 할 수 있음.\n",
    "- 텍스트를 사용하여 패턴이나 관계를 추출하고 그 안에서 의미 있는 정보나 가치를 발굴하여 해석하는 일련의 과정"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. 텍스트 마이닝의 기능\n",
    "\n",
    "- 텍스트 마이닝의 4가지 기능\n",
    "    1. 문서 요약\n",
    "    2. 문서 분류\n",
    "    3. 문서 군집\n",
    "    4. 특성 추출"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. 텍스트 마이닝의 기본 프로세스\n",
    "\n",
    "- R에서 텍스트 마이닝을 수행하는데 필요한 패키지인 tm 패키지 설치"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "also installing the dependencies 'NLP', 'slam'\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'NLP' successfully unpacked and MD5 sums checked\n",
      "package 'slam' successfully unpacked and MD5 sums checked\n",
      "package 'tm' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmp8St9I7\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"tm\",  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning message:\n",
      ": package 'tm' was built under R version 3.3.2Loading required package: NLP\n",
      "Warning message:\n",
      ": package 'NLP' was built under R version 3.3.2"
     ]
    }
   ],
   "source": [
    "library( tm )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 가. 데이터 수집\n",
    "\n",
    "- 트위터에서 텍스트 데이터 수집\n",
    "\n",
    "![](data_01.jpg)\n",
    "\n",
    "![](data_02.jpg)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "also installing the dependencies 'bit', 'curl', 'openssl', 'bit64', 'rjson', 'httr'\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'bit' successfully unpacked and MD5 sums checked\n",
      "package 'curl' successfully unpacked and MD5 sums checked\n",
      "package 'openssl' successfully unpacked and MD5 sums checked\n",
      "package 'bit64' successfully unpacked and MD5 sums checked\n",
      "package 'rjson' successfully unpacked and MD5 sums checked\n",
      "package 'httr' successfully unpacked and MD5 sums checked\n",
      "package 'twitteR' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmp8St9I7\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"twitteR\",  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning message:\n",
      ": package 'twitteR' was built under R version 3.3.2"
     ]
    }
   ],
   "source": [
    "library(twitteR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1] \"Using direct authentication\"\n"
     ]
    }
   ],
   "source": [
    "api_key <- \"RKCbEl0xGqfSdw1SC\"\n",
    "api_secret <- \"Aeps3VNrrykBNYL33BLGOby5XXXVwuHTvhk\"\n",
    "access_token <- \"829270499800776704-VyJ78FDns7XjnH\"\n",
    "access_token_secret <- \"qyGyPMisIOPXtW00ezONqw\"\n",
    "\n",
    "setup_twitter_oauth(api_key, api_secret, access_token, access_token_secret  )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- [예제 1]  다음 데이터는 R에서 수집한 트위터 메시지로, 2014년 1월 1일 이후의 메시지 중 bigdata라는 단어가 등장한 트위터 메시지"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "keyword <- \"bigdata\"\n",
    "bigdata.tw <- searchTwitter( keyword, since='2017-02-01', n=1000, lang='en' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] \"jminguezc: 6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[1] \"VanRijmenam: Why retailers must think beyond traditional engagement strategies - by @DataMentors https://t.co/JPXVC2ifax #BigData\""
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[1] \"rishu_12: RT @AINewsletter: Top #Montreal The DeepLearning eBook: https://t.co/O0SygcbvGw via @KirkDBorne #General https://t.co/BcFxVVlhzs\""
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bigdata.tw[[1]]\n",
    "bigdata.tw[[2]]\n",
    "bigdata.tw[[3]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 트위터 내용을 이용해 스크린 네임 정보와 •트위터 내용 가져오기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"jminguezc\""
      ],
      "text/latex": [
       "\"jminguezc\""
      ],
      "text/markdown": [
       "\"jminguezc\""
      ],
      "text/plain": [
       "[1] \"jminguezc\""
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweet <- bigdata.tw[[1]]\n",
    "tweet$getScreenName()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/latex": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/markdown": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweet$getText()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- list형태로 되어 있는 bigdata.tw을 data.frame형태로 변환"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"list\""
      ],
      "text/latex": [
       "\"list\""
      ],
      "text/markdown": [
       "\"list\""
      ],
      "text/plain": [
       "[1] \"list\""
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/html": [
       "\"data.frame\""
      ],
      "text/latex": [
       "\"data.frame\""
      ],
      "text/markdown": [
       "\"data.frame\""
      ],
      "text/plain": [
       "[1] \"data.frame\""
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\"</li>\n",
       "\t<li>\"Why retailers must think beyond traditional engagement strategies - by @DataMentors https://t.co/JPXVC2ifax #BigData\"</li>\n",
       "\t<li>\"RT @AINewsletter: Top #Montreal The DeepLearning eBook: https://t.co/O0SygcbvGw via @KirkDBorne #General https://t.co/BcFxVVlhzs\"</li>\n",
       "\t<li>\"RT @marcusborba: 15 #Python and R Data Visualization Charts. #BigData #DataScience #DataViz via @gp_pulipaka\n",
       "https://t.co/SNxHOqSLGX https:…\"</li>\n",
       "\t<li>\"How Big Data is Advancing Women’s Health https://t.co/O9azTROiDc 🗣#Opines on #BigData\"</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\"\n",
       "\\item \"Why retailers must think beyond traditional engagement strategies - by @DataMentors https://t.co/JPXVC2ifax #BigData\"\n",
       "\\item \"RT @AINewsletter: Top #Montreal The DeepLearning eBook: https://t.co/O0SygcbvGw via @KirkDBorne #General https://t.co/BcFxVVlhzs\"\n",
       "\\item \"RT @marcusborba: 15 #Python and R Data Visualization Charts. #BigData #DataScience #DataViz via @gp_pulipaka\n",
       "https://t.co/SNxHOqSLGX https:…\"\n",
       "\\item \"How Big Data is Advancing Women’s Health https://t.co/O9azTROiDc 🗣#Opines on #BigData\"\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\"\n",
       "2. \"Why retailers must think beyond traditional engagement strategies - by @DataMentors https://t.co/JPXVC2ifax #BigData\"\n",
       "3. \"RT @AINewsletter: Top #Montreal The DeepLearning eBook: https://t.co/O0SygcbvGw via @KirkDBorne #General https://t.co/BcFxVVlhzs\"\n",
       "4. \"RT @marcusborba: 15 #Python and R Data Visualization Charts. #BigData #DataScience #DataViz via @gp_pulipaka\n",
       "https://t.co/SNxHOqSLGX https:…\"\n",
       "5. \"How Big Data is Advancing Women’s Health https://t.co/O9azTROiDc 🗣#Opines on #BigData\"\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\"               \n",
       "[2] \"Why retailers must think beyond traditional engagement strategies - by @DataMentors https://t.co/JPXVC2ifax #BigData\"                          \n",
       "[3] \"RT @AINewsletter: Top #Montreal The DeepLearning eBook: https://t.co/O0SygcbvGw via @KirkDBorne #General https://t.co/BcFxVVlhzs\"              \n",
       "[4] \"RT @marcusborba: 15 #Python and R Data Visualization Charts. #BigData #DataScience #DataViz via @gp_pulipaka\\nhttps://t.co/SNxHOqSLGX https:…\"\n",
       "[5] \"How Big Data is Advancing Women’s Health https://t.co/O9azTROiDc <ed><U+00A0><U+00BD><ed><U+00B7><U+00A3>#Opines on #BigData\"                 "
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class(bigdata.tw)\n",
    "\n",
    "bigdata.df <- twListToDF( bigdata.tw )\n",
    "\n",
    "class(bigdata.df)\n",
    "\n",
    "bigdata.text <- bigdata.df$text\n",
    "head( bigdata.text, n=5 )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 나. 데이터  전처리 및 가공\n",
    "\n",
    "- 유의하게 사용되지 않는 부분을 제거\n",
    "- 분석에 용이하도록 변경 과정"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 1) Corpus 생성 ( 말뭉치 생성)\n",
    "\n",
    "- 텍스트 마이닝에서 문서를 관리하느 기본구조를 Corpus라고 함.\n",
    "- 이는 텍스트 문서들의 집합을 의미"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- [예제 2] tm 패키지안에 있는 데이터을 읽어오자."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"C:/Anaconda2/R/library/tm/texts/txt\""
      ],
      "text/latex": [
       "\"C:/Anaconda2/R/library/tm/texts/txt\""
      ],
      "text/markdown": [
       "\"C:/Anaconda2/R/library/tm/texts/txt\""
      ],
      "text/plain": [
       "[1] \"C:/Anaconda2/R/library/tm/texts/txt\""
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "txt <- system.file( 'texts', 'txt', package='tm' )\n",
    "txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<<VCorpus>>\n",
       "Metadata:  corpus specific: 0, document level (indexed): 0\n",
       "Content:  documents: 5"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ovid <- Corpus(  DirSource(txt),  readerControl=list(language='lat')  )\n",
    "ovid"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<<PlainTextDocument>>\n",
       "Metadata:  7\n",
       "Content:  chars: 676"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ovid[[1]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- reader종류를 확인해보고 텍스트 이외에도 Word, PDF, CSV 등의 다양한 문서형식을 읽어올 수 있음."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>\"readDOC\"</li>\n",
       "\t<li>\"readPDF\"</li>\n",
       "\t<li>\"readPlain\"</li>\n",
       "\t<li>\"readRCV1\"</li>\n",
       "\t<li>\"readRCV1asPlain\"</li>\n",
       "\t<li>\"readReut21578XML\"</li>\n",
       "\t<li>\"readReut21578XMLasPlain\"</li>\n",
       "\t<li>\"readTabular\"</li>\n",
       "\t<li>\"readTagged\"</li>\n",
       "\t<li>\"readXML\"</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item \"readDOC\"\n",
       "\\item \"readPDF\"\n",
       "\\item \"readPlain\"\n",
       "\\item \"readRCV1\"\n",
       "\\item \"readRCV1asPlain\"\n",
       "\\item \"readReut21578XML\"\n",
       "\\item \"readReut21578XMLasPlain\"\n",
       "\\item \"readTabular\"\n",
       "\\item \"readTagged\"\n",
       "\\item \"readXML\"\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. \"readDOC\"\n",
       "2. \"readPDF\"\n",
       "3. \"readPlain\"\n",
       "4. \"readRCV1\"\n",
       "5. \"readRCV1asPlain\"\n",
       "6. \"readReut21578XML\"\n",
       "7. \"readReut21578XMLasPlain\"\n",
       "8. \"readTabular\"\n",
       "9. \"readTagged\"\n",
       "10. \"readXML\"\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       " [1] \"readDOC\"                 \"readPDF\"                \n",
       " [3] \"readPlain\"               \"readRCV1\"               \n",
       " [5] \"readRCV1asPlain\"         \"readReut21578XML\"       \n",
       " [7] \"readReut21578XMLasPlain\" \"readTabular\"            \n",
       " [9] \"readTagged\"              \"readXML\"                "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "getReaders()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 트위터에서 읽어온 bigdata.text을 corpus 형태로 변환하자"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<<VCorpus>>\n",
       "Metadata:  corpus specific: 0, document level (indexed): 0\n",
       "Content:  documents: 1000"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus <- Corpus( VectorSource(bigdata.text) )\n",
    "my.corpus"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/latex": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/markdown": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus[[1]]$content"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<<PlainTextDocument>>\n",
       "Metadata:  7\n",
       "Content:  chars: 126"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus[[1]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 2)  tm_map() 함수의 적용\n",
    "\n",
    "- Corpus의 형식의 데이터들의 변형을 위한 tm_map()함수가 제공"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/latex": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/markdown": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus.new <- tm_map( my.corpus, stripWhitespace )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 특정 패턴을 가지는 단어를 제거하거나 다른 단어로 대체"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/latex": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/markdown": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… https://t.co/3DN7slkZSv\""
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus.new <- tm_map( my.corpus.new, content_transformer(gsub),  pattern='@\\\\S*', replacement='' )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… \""
      ],
      "text/latex": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… \""
      ],
      "text/markdown": [
       "\"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… \""
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming. #BigData #DataScience #Programming #Linux #Python #RStats… \""
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus.new <- tm_map( my.corpus.new, content_transformer(gsub),  pattern='http\\\\S*', replacement='' )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 문장 부호 및 구두점 제거"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 Best Linux Distributions For Programming BigData DataScience Programming Linux Python RStats \""
      ],
      "text/latex": [
       "\"6 Best Linux Distributions For Programming BigData DataScience Programming Linux Python RStats \""
      ],
      "text/markdown": [
       "\"6 Best Linux Distributions For Programming BigData DataScience Programming Linux Python RStats \""
      ],
      "text/plain": [
       "[1] \"6 Best Linux Distributions For Programming BigData DataScience Programming Linux Python RStats \""
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus.new <- tm_map( my.corpus.new, removePunctuation )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 대문자를 소문자로 변환"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 best linux distributions for programming bigdata datascience programming linux python rstats \""
      ],
      "text/latex": [
       "\"6 best linux distributions for programming bigdata datascience programming linux python rstats \""
      ],
      "text/markdown": [
       "\"6 best linux distributions for programming bigdata datascience programming linux python rstats \""
      ],
      "text/plain": [
       "[1] \"6 best linux distributions for programming bigdata datascience programming linux python rstats \""
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus.new <- tm_map( my.corpus.new, content_transformer(tolower)  )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 분석에 필요없는 특정 단어를 제거,  stop word라고 하며  불용어라고 함."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>\"i\"</li>\n",
       "\t<li>\"me\"</li>\n",
       "\t<li>\"my\"</li>\n",
       "\t<li>\"myself\"</li>\n",
       "\t<li>\"we\"</li>\n",
       "\t<li>\"our\"</li>\n",
       "\t<li>\"ours\"</li>\n",
       "\t<li>\"ourselves\"</li>\n",
       "\t<li>\"you\"</li>\n",
       "\t<li>\"your\"</li>\n",
       "\t<li>\"yours\"</li>\n",
       "\t<li>\"yourself\"</li>\n",
       "\t<li>\"yourselves\"</li>\n",
       "\t<li>\"he\"</li>\n",
       "\t<li>\"him\"</li>\n",
       "\t<li>\"his\"</li>\n",
       "\t<li>\"himself\"</li>\n",
       "\t<li>\"she\"</li>\n",
       "\t<li>\"her\"</li>\n",
       "\t<li>\"hers\"</li>\n",
       "\t<li>\"herself\"</li>\n",
       "\t<li>\"it\"</li>\n",
       "\t<li>\"its\"</li>\n",
       "\t<li>\"itself\"</li>\n",
       "\t<li>\"they\"</li>\n",
       "\t<li>\"them\"</li>\n",
       "\t<li>\"their\"</li>\n",
       "\t<li>\"theirs\"</li>\n",
       "\t<li>\"themselves\"</li>\n",
       "\t<li>\"what\"</li>\n",
       "\t<li>\"which\"</li>\n",
       "\t<li>\"who\"</li>\n",
       "\t<li>\"whom\"</li>\n",
       "\t<li>\"this\"</li>\n",
       "\t<li>\"that\"</li>\n",
       "\t<li>\"these\"</li>\n",
       "\t<li>\"those\"</li>\n",
       "\t<li>\"am\"</li>\n",
       "\t<li>\"is\"</li>\n",
       "\t<li>\"are\"</li>\n",
       "\t<li>\"was\"</li>\n",
       "\t<li>\"were\"</li>\n",
       "\t<li>\"be\"</li>\n",
       "\t<li>\"been\"</li>\n",
       "\t<li>\"being\"</li>\n",
       "\t<li>\"have\"</li>\n",
       "\t<li>\"has\"</li>\n",
       "\t<li>\"had\"</li>\n",
       "\t<li>\"having\"</li>\n",
       "\t<li>\"do\"</li>\n",
       "\t<li>\"does\"</li>\n",
       "\t<li>\"did\"</li>\n",
       "\t<li>\"doing\"</li>\n",
       "\t<li>\"would\"</li>\n",
       "\t<li>\"should\"</li>\n",
       "\t<li>\"could\"</li>\n",
       "\t<li>\"ought\"</li>\n",
       "\t<li>\"i'm\"</li>\n",
       "\t<li>\"you're\"</li>\n",
       "\t<li>\"he's\"</li>\n",
       "\t<li>\"she's\"</li>\n",
       "\t<li>\"it's\"</li>\n",
       "\t<li>\"we're\"</li>\n",
       "\t<li>\"they're\"</li>\n",
       "\t<li>\"i've\"</li>\n",
       "\t<li>\"you've\"</li>\n",
       "\t<li>\"we've\"</li>\n",
       "\t<li>\"they've\"</li>\n",
       "\t<li>\"i'd\"</li>\n",
       "\t<li>\"you'd\"</li>\n",
       "\t<li>\"he'd\"</li>\n",
       "\t<li>\"she'd\"</li>\n",
       "\t<li>\"we'd\"</li>\n",
       "\t<li>\"they'd\"</li>\n",
       "\t<li>\"i'll\"</li>\n",
       "\t<li>\"you'll\"</li>\n",
       "\t<li>\"he'll\"</li>\n",
       "\t<li>\"she'll\"</li>\n",
       "\t<li>\"we'll\"</li>\n",
       "\t<li>\"they'll\"</li>\n",
       "\t<li>\"isn't\"</li>\n",
       "\t<li>\"aren't\"</li>\n",
       "\t<li>\"wasn't\"</li>\n",
       "\t<li>\"weren't\"</li>\n",
       "\t<li>\"hasn't\"</li>\n",
       "\t<li>\"haven't\"</li>\n",
       "\t<li>\"hadn't\"</li>\n",
       "\t<li>\"doesn't\"</li>\n",
       "\t<li>\"don't\"</li>\n",
       "\t<li>\"didn't\"</li>\n",
       "\t<li>\"won't\"</li>\n",
       "\t<li>\"wouldn't\"</li>\n",
       "\t<li>\"shan't\"</li>\n",
       "\t<li>\"shouldn't\"</li>\n",
       "\t<li>\"can't\"</li>\n",
       "\t<li>\"cannot\"</li>\n",
       "\t<li>\"couldn't\"</li>\n",
       "\t<li>\"mustn't\"</li>\n",
       "\t<li>\"let's\"</li>\n",
       "\t<li>\"that's\"</li>\n",
       "\t<li>\"who's\"</li>\n",
       "\t<li>\"what's\"</li>\n",
       "\t<li>\"here's\"</li>\n",
       "\t<li>\"there's\"</li>\n",
       "\t<li>\"when's\"</li>\n",
       "\t<li>\"where's\"</li>\n",
       "\t<li>\"why's\"</li>\n",
       "\t<li>\"how's\"</li>\n",
       "\t<li>\"a\"</li>\n",
       "\t<li>\"an\"</li>\n",
       "\t<li>\"the\"</li>\n",
       "\t<li>\"and\"</li>\n",
       "\t<li>\"but\"</li>\n",
       "\t<li>\"if\"</li>\n",
       "\t<li>\"or\"</li>\n",
       "\t<li>\"because\"</li>\n",
       "\t<li>\"as\"</li>\n",
       "\t<li>\"until\"</li>\n",
       "\t<li>\"while\"</li>\n",
       "\t<li>\"of\"</li>\n",
       "\t<li>\"at\"</li>\n",
       "\t<li>\"by\"</li>\n",
       "\t<li>\"for\"</li>\n",
       "\t<li>\"with\"</li>\n",
       "\t<li>\"about\"</li>\n",
       "\t<li>\"against\"</li>\n",
       "\t<li>\"between\"</li>\n",
       "\t<li>\"into\"</li>\n",
       "\t<li>\"through\"</li>\n",
       "\t<li>\"during\"</li>\n",
       "\t<li>\"before\"</li>\n",
       "\t<li>\"after\"</li>\n",
       "\t<li>\"above\"</li>\n",
       "\t<li>\"below\"</li>\n",
       "\t<li>\"to\"</li>\n",
       "\t<li>\"from\"</li>\n",
       "\t<li>\"up\"</li>\n",
       "\t<li>\"down\"</li>\n",
       "\t<li>\"in\"</li>\n",
       "\t<li>\"out\"</li>\n",
       "\t<li>\"on\"</li>\n",
       "\t<li>\"off\"</li>\n",
       "\t<li>\"over\"</li>\n",
       "\t<li>\"under\"</li>\n",
       "\t<li>\"again\"</li>\n",
       "\t<li>\"further\"</li>\n",
       "\t<li>\"then\"</li>\n",
       "\t<li>\"once\"</li>\n",
       "\t<li>\"here\"</li>\n",
       "\t<li>\"there\"</li>\n",
       "\t<li>\"when\"</li>\n",
       "\t<li>\"where\"</li>\n",
       "\t<li>\"why\"</li>\n",
       "\t<li>\"how\"</li>\n",
       "\t<li>\"all\"</li>\n",
       "\t<li>\"any\"</li>\n",
       "\t<li>\"both\"</li>\n",
       "\t<li>\"each\"</li>\n",
       "\t<li>\"few\"</li>\n",
       "\t<li>\"more\"</li>\n",
       "\t<li>\"most\"</li>\n",
       "\t<li>\"other\"</li>\n",
       "\t<li>\"some\"</li>\n",
       "\t<li>\"such\"</li>\n",
       "\t<li>\"no\"</li>\n",
       "\t<li>\"nor\"</li>\n",
       "\t<li>\"not\"</li>\n",
       "\t<li>\"only\"</li>\n",
       "\t<li>\"own\"</li>\n",
       "\t<li>\"same\"</li>\n",
       "\t<li>\"so\"</li>\n",
       "\t<li>\"than\"</li>\n",
       "\t<li>\"too\"</li>\n",
       "\t<li>\"very\"</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item \"i\"\n",
       "\\item \"me\"\n",
       "\\item \"my\"\n",
       "\\item \"myself\"\n",
       "\\item \"we\"\n",
       "\\item \"our\"\n",
       "\\item \"ours\"\n",
       "\\item \"ourselves\"\n",
       "\\item \"you\"\n",
       "\\item \"your\"\n",
       "\\item \"yours\"\n",
       "\\item \"yourself\"\n",
       "\\item \"yourselves\"\n",
       "\\item \"he\"\n",
       "\\item \"him\"\n",
       "\\item \"his\"\n",
       "\\item \"himself\"\n",
       "\\item \"she\"\n",
       "\\item \"her\"\n",
       "\\item \"hers\"\n",
       "\\item \"herself\"\n",
       "\\item \"it\"\n",
       "\\item \"its\"\n",
       "\\item \"itself\"\n",
       "\\item \"they\"\n",
       "\\item \"them\"\n",
       "\\item \"their\"\n",
       "\\item \"theirs\"\n",
       "\\item \"themselves\"\n",
       "\\item \"what\"\n",
       "\\item \"which\"\n",
       "\\item \"who\"\n",
       "\\item \"whom\"\n",
       "\\item \"this\"\n",
       "\\item \"that\"\n",
       "\\item \"these\"\n",
       "\\item \"those\"\n",
       "\\item \"am\"\n",
       "\\item \"is\"\n",
       "\\item \"are\"\n",
       "\\item \"was\"\n",
       "\\item \"were\"\n",
       "\\item \"be\"\n",
       "\\item \"been\"\n",
       "\\item \"being\"\n",
       "\\item \"have\"\n",
       "\\item \"has\"\n",
       "\\item \"had\"\n",
       "\\item \"having\"\n",
       "\\item \"do\"\n",
       "\\item \"does\"\n",
       "\\item \"did\"\n",
       "\\item \"doing\"\n",
       "\\item \"would\"\n",
       "\\item \"should\"\n",
       "\\item \"could\"\n",
       "\\item \"ought\"\n",
       "\\item \"i'm\"\n",
       "\\item \"you're\"\n",
       "\\item \"he's\"\n",
       "\\item \"she's\"\n",
       "\\item \"it's\"\n",
       "\\item \"we're\"\n",
       "\\item \"they're\"\n",
       "\\item \"i've\"\n",
       "\\item \"you've\"\n",
       "\\item \"we've\"\n",
       "\\item \"they've\"\n",
       "\\item \"i'd\"\n",
       "\\item \"you'd\"\n",
       "\\item \"he'd\"\n",
       "\\item \"she'd\"\n",
       "\\item \"we'd\"\n",
       "\\item \"they'd\"\n",
       "\\item \"i'll\"\n",
       "\\item \"you'll\"\n",
       "\\item \"he'll\"\n",
       "\\item \"she'll\"\n",
       "\\item \"we'll\"\n",
       "\\item \"they'll\"\n",
       "\\item \"isn't\"\n",
       "\\item \"aren't\"\n",
       "\\item \"wasn't\"\n",
       "\\item \"weren't\"\n",
       "\\item \"hasn't\"\n",
       "\\item \"haven't\"\n",
       "\\item \"hadn't\"\n",
       "\\item \"doesn't\"\n",
       "\\item \"don't\"\n",
       "\\item \"didn't\"\n",
       "\\item \"won't\"\n",
       "\\item \"wouldn't\"\n",
       "\\item \"shan't\"\n",
       "\\item \"shouldn't\"\n",
       "\\item \"can't\"\n",
       "\\item \"cannot\"\n",
       "\\item \"couldn't\"\n",
       "\\item \"mustn't\"\n",
       "\\item \"let's\"\n",
       "\\item \"that's\"\n",
       "\\item \"who's\"\n",
       "\\item \"what's\"\n",
       "\\item \"here's\"\n",
       "\\item \"there's\"\n",
       "\\item \"when's\"\n",
       "\\item \"where's\"\n",
       "\\item \"why's\"\n",
       "\\item \"how's\"\n",
       "\\item \"a\"\n",
       "\\item \"an\"\n",
       "\\item \"the\"\n",
       "\\item \"and\"\n",
       "\\item \"but\"\n",
       "\\item \"if\"\n",
       "\\item \"or\"\n",
       "\\item \"because\"\n",
       "\\item \"as\"\n",
       "\\item \"until\"\n",
       "\\item \"while\"\n",
       "\\item \"of\"\n",
       "\\item \"at\"\n",
       "\\item \"by\"\n",
       "\\item \"for\"\n",
       "\\item \"with\"\n",
       "\\item \"about\"\n",
       "\\item \"against\"\n",
       "\\item \"between\"\n",
       "\\item \"into\"\n",
       "\\item \"through\"\n",
       "\\item \"during\"\n",
       "\\item \"before\"\n",
       "\\item \"after\"\n",
       "\\item \"above\"\n",
       "\\item \"below\"\n",
       "\\item \"to\"\n",
       "\\item \"from\"\n",
       "\\item \"up\"\n",
       "\\item \"down\"\n",
       "\\item \"in\"\n",
       "\\item \"out\"\n",
       "\\item \"on\"\n",
       "\\item \"off\"\n",
       "\\item \"over\"\n",
       "\\item \"under\"\n",
       "\\item \"again\"\n",
       "\\item \"further\"\n",
       "\\item \"then\"\n",
       "\\item \"once\"\n",
       "\\item \"here\"\n",
       "\\item \"there\"\n",
       "\\item \"when\"\n",
       "\\item \"where\"\n",
       "\\item \"why\"\n",
       "\\item \"how\"\n",
       "\\item \"all\"\n",
       "\\item \"any\"\n",
       "\\item \"both\"\n",
       "\\item \"each\"\n",
       "\\item \"few\"\n",
       "\\item \"more\"\n",
       "\\item \"most\"\n",
       "\\item \"other\"\n",
       "\\item \"some\"\n",
       "\\item \"such\"\n",
       "\\item \"no\"\n",
       "\\item \"nor\"\n",
       "\\item \"not\"\n",
       "\\item \"only\"\n",
       "\\item \"own\"\n",
       "\\item \"same\"\n",
       "\\item \"so\"\n",
       "\\item \"than\"\n",
       "\\item \"too\"\n",
       "\\item \"very\"\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. \"i\"\n",
       "2. \"me\"\n",
       "3. \"my\"\n",
       "4. \"myself\"\n",
       "5. \"we\"\n",
       "6. \"our\"\n",
       "7. \"ours\"\n",
       "8. \"ourselves\"\n",
       "9. \"you\"\n",
       "10. \"your\"\n",
       "11. \"yours\"\n",
       "12. \"yourself\"\n",
       "13. \"yourselves\"\n",
       "14. \"he\"\n",
       "15. \"him\"\n",
       "16. \"his\"\n",
       "17. \"himself\"\n",
       "18. \"she\"\n",
       "19. \"her\"\n",
       "20. \"hers\"\n",
       "21. \"herself\"\n",
       "22. \"it\"\n",
       "23. \"its\"\n",
       "24. \"itself\"\n",
       "25. \"they\"\n",
       "26. \"them\"\n",
       "27. \"their\"\n",
       "28. \"theirs\"\n",
       "29. \"themselves\"\n",
       "30. \"what\"\n",
       "31. \"which\"\n",
       "32. \"who\"\n",
       "33. \"whom\"\n",
       "34. \"this\"\n",
       "35. \"that\"\n",
       "36. \"these\"\n",
       "37. \"those\"\n",
       "38. \"am\"\n",
       "39. \"is\"\n",
       "40. \"are\"\n",
       "41. \"was\"\n",
       "42. \"were\"\n",
       "43. \"be\"\n",
       "44. \"been\"\n",
       "45. \"being\"\n",
       "46. \"have\"\n",
       "47. \"has\"\n",
       "48. \"had\"\n",
       "49. \"having\"\n",
       "50. \"do\"\n",
       "51. \"does\"\n",
       "52. \"did\"\n",
       "53. \"doing\"\n",
       "54. \"would\"\n",
       "55. \"should\"\n",
       "56. \"could\"\n",
       "57. \"ought\"\n",
       "58. \"i'm\"\n",
       "59. \"you're\"\n",
       "60. \"he's\"\n",
       "61. \"she's\"\n",
       "62. \"it's\"\n",
       "63. \"we're\"\n",
       "64. \"they're\"\n",
       "65. \"i've\"\n",
       "66. \"you've\"\n",
       "67. \"we've\"\n",
       "68. \"they've\"\n",
       "69. \"i'd\"\n",
       "70. \"you'd\"\n",
       "71. \"he'd\"\n",
       "72. \"she'd\"\n",
       "73. \"we'd\"\n",
       "74. \"they'd\"\n",
       "75. \"i'll\"\n",
       "76. \"you'll\"\n",
       "77. \"he'll\"\n",
       "78. \"she'll\"\n",
       "79. \"we'll\"\n",
       "80. \"they'll\"\n",
       "81. \"isn't\"\n",
       "82. \"aren't\"\n",
       "83. \"wasn't\"\n",
       "84. \"weren't\"\n",
       "85. \"hasn't\"\n",
       "86. \"haven't\"\n",
       "87. \"hadn't\"\n",
       "88. \"doesn't\"\n",
       "89. \"don't\"\n",
       "90. \"didn't\"\n",
       "91. \"won't\"\n",
       "92. \"wouldn't\"\n",
       "93. \"shan't\"\n",
       "94. \"shouldn't\"\n",
       "95. \"can't\"\n",
       "96. \"cannot\"\n",
       "97. \"couldn't\"\n",
       "98. \"mustn't\"\n",
       "99. \"let's\"\n",
       "100. \"that's\"\n",
       "101. \"who's\"\n",
       "102. \"what's\"\n",
       "103. \"here's\"\n",
       "104. \"there's\"\n",
       "105. \"when's\"\n",
       "106. \"where's\"\n",
       "107. \"why's\"\n",
       "108. \"how's\"\n",
       "109. \"a\"\n",
       "110. \"an\"\n",
       "111. \"the\"\n",
       "112. \"and\"\n",
       "113. \"but\"\n",
       "114. \"if\"\n",
       "115. \"or\"\n",
       "116. \"because\"\n",
       "117. \"as\"\n",
       "118. \"until\"\n",
       "119. \"while\"\n",
       "120. \"of\"\n",
       "121. \"at\"\n",
       "122. \"by\"\n",
       "123. \"for\"\n",
       "124. \"with\"\n",
       "125. \"about\"\n",
       "126. \"against\"\n",
       "127. \"between\"\n",
       "128. \"into\"\n",
       "129. \"through\"\n",
       "130. \"during\"\n",
       "131. \"before\"\n",
       "132. \"after\"\n",
       "133. \"above\"\n",
       "134. \"below\"\n",
       "135. \"to\"\n",
       "136. \"from\"\n",
       "137. \"up\"\n",
       "138. \"down\"\n",
       "139. \"in\"\n",
       "140. \"out\"\n",
       "141. \"on\"\n",
       "142. \"off\"\n",
       "143. \"over\"\n",
       "144. \"under\"\n",
       "145. \"again\"\n",
       "146. \"further\"\n",
       "147. \"then\"\n",
       "148. \"once\"\n",
       "149. \"here\"\n",
       "150. \"there\"\n",
       "151. \"when\"\n",
       "152. \"where\"\n",
       "153. \"why\"\n",
       "154. \"how\"\n",
       "155. \"all\"\n",
       "156. \"any\"\n",
       "157. \"both\"\n",
       "158. \"each\"\n",
       "159. \"few\"\n",
       "160. \"more\"\n",
       "161. \"most\"\n",
       "162. \"other\"\n",
       "163. \"some\"\n",
       "164. \"such\"\n",
       "165. \"no\"\n",
       "166. \"nor\"\n",
       "167. \"not\"\n",
       "168. \"only\"\n",
       "169. \"own\"\n",
       "170. \"same\"\n",
       "171. \"so\"\n",
       "172. \"than\"\n",
       "173. \"too\"\n",
       "174. \"very\"\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "  [1] \"i\"          \"me\"         \"my\"         \"myself\"     \"we\"        \n",
       "  [6] \"our\"        \"ours\"       \"ourselves\"  \"you\"        \"your\"      \n",
       " [11] \"yours\"      \"yourself\"   \"yourselves\" \"he\"         \"him\"       \n",
       " [16] \"his\"        \"himself\"    \"she\"        \"her\"        \"hers\"      \n",
       " [21] \"herself\"    \"it\"         \"its\"        \"itself\"     \"they\"      \n",
       " [26] \"them\"       \"their\"      \"theirs\"     \"themselves\" \"what\"      \n",
       " [31] \"which\"      \"who\"        \"whom\"       \"this\"       \"that\"      \n",
       " [36] \"these\"      \"those\"      \"am\"         \"is\"         \"are\"       \n",
       " [41] \"was\"        \"were\"       \"be\"         \"been\"       \"being\"     \n",
       " [46] \"have\"       \"has\"        \"had\"        \"having\"     \"do\"        \n",
       " [51] \"does\"       \"did\"        \"doing\"      \"would\"      \"should\"    \n",
       " [56] \"could\"      \"ought\"      \"i'm\"        \"you're\"     \"he's\"      \n",
       " [61] \"she's\"      \"it's\"       \"we're\"      \"they're\"    \"i've\"      \n",
       " [66] \"you've\"     \"we've\"      \"they've\"    \"i'd\"        \"you'd\"     \n",
       " [71] \"he'd\"       \"she'd\"      \"we'd\"       \"they'd\"     \"i'll\"      \n",
       " [76] \"you'll\"     \"he'll\"      \"she'll\"     \"we'll\"      \"they'll\"   \n",
       " [81] \"isn't\"      \"aren't\"     \"wasn't\"     \"weren't\"    \"hasn't\"    \n",
       " [86] \"haven't\"    \"hadn't\"     \"doesn't\"    \"don't\"      \"didn't\"    \n",
       " [91] \"won't\"      \"wouldn't\"   \"shan't\"     \"shouldn't\"  \"can't\"     \n",
       " [96] \"cannot\"     \"couldn't\"   \"mustn't\"    \"let's\"      \"that's\"    \n",
       "[101] \"who's\"      \"what's\"     \"here's\"     \"there's\"    \"when's\"    \n",
       "[106] \"where's\"    \"why's\"      \"how's\"      \"a\"          \"an\"        \n",
       "[111] \"the\"        \"and\"        \"but\"        \"if\"         \"or\"        \n",
       "[116] \"because\"    \"as\"         \"until\"      \"while\"      \"of\"        \n",
       "[121] \"at\"         \"by\"         \"for\"        \"with\"       \"about\"     \n",
       "[126] \"against\"    \"between\"    \"into\"       \"through\"    \"during\"    \n",
       "[131] \"before\"     \"after\"      \"above\"      \"below\"      \"to\"        \n",
       "[136] \"from\"       \"up\"         \"down\"       \"in\"         \"out\"       \n",
       "[141] \"on\"         \"off\"        \"over\"       \"under\"      \"again\"     \n",
       "[146] \"further\"    \"then\"       \"once\"       \"here\"       \"there\"     \n",
       "[151] \"when\"       \"where\"      \"why\"        \"how\"        \"all\"       \n",
       "[156] \"any\"        \"both\"       \"each\"       \"few\"        \"more\"      \n",
       "[161] \"most\"       \"other\"      \"some\"       \"such\"       \"no\"        \n",
       "[166] \"nor\"        \"not\"        \"only\"       \"own\"        \"same\"      \n",
       "[171] \"so\"         \"than\"       \"too\"        \"very\"      "
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stopwords('en')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ],
      "text/latex": [
       "\"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ],
      "text/markdown": [
       "\"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ],
      "text/plain": [
       "[1] \"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.corpus.new <- tm_map( my.corpus.new, removeWords, stopwords('en') )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 기본적인 stop word 이외에 분석에 있어 의미가 없는 단어를 추가해서 제거할 수 있음"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ],
      "text/latex": [
       "\"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ],
      "text/markdown": [
       "\"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ],
      "text/plain": [
       "[1] \"6 best linux distributions  programming bigdata datascience programming linux python rstats \""
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mystopword <- c( stopwords('en'), 'rt', 'via', 'even' )\n",
    "my.corpus.new <- tm_map( my.corpus.new, removeWords, mystopword )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 트위터에서는  RT 라는 단어는  Re-Twitter의 줄임말로  등장 빈도가 높지만 분석에는 의미가 없음."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 나. 자연어 처리\n",
    "\n",
    "##### 1) stemming\n",
    "\n",
    "- 과거형, 현재형, 미래형 동사 등을 현재형으로 변경,\n",
    "- 공통 어간을 가지는 단어를 묶음 처리"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'SnowballC' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmpkjsrvb\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"SnowballC\" ,  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "library(SnowballC)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"6 best linux distribut  program bigdata datasci program linux python rstat\""
      ],
      "text/latex": [
       "\"6 best linux distribut  program bigdata datasci program linux python rstat\""
      ],
      "text/markdown": [
       "\"6 best linux distribut  program bigdata datasci program linux python rstat\""
      ],
      "text/plain": [
       "[1] \"6 best linux distribut  program bigdata datasci program linux python rstat\""
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dic.corpus <- my.corpus.new \n",
    "my.corpus.new  <- tm_map( my.corpus.new, stemDocument )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- stemDocument()함수는 공통어간으로 변환해주어 단어가 완전하지 못함.\n",
    "- stemCompletion()함수를 사용하고, stemCompletion를 잘못하면 모든 값이 NA가 되기 때문에 이를 방지하기 위해서 아래와 같은 함수를 만들고 처리함."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\"comes comes comes comes first comes role comes bigdata comes iot comes sdn comes\""
      ],
      "text/latex": [
       "\"comes comes comes comes first comes role comes bigdata comes iot comes sdn comes\""
      ],
      "text/markdown": [
       "\"comes comes comes comes first comes role comes bigdata comes iot comes sdn comes\""
      ],
      "text/plain": [
       "[1] \"comes comes comes comes first comes role comes bigdata comes iot comes sdn comes\""
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "stemCompletion_mod <- function(x, dict) {\n",
    "     PlainTextDocument(stripWhitespace(paste(stemCompletion(unlist(strsplit(as.character(x),\" \")),dictionary=dict, type=\"first\"),sep=\"\", collapse=\" \")))\n",
    "}\n",
    "\n",
    "my.corpus.new <- dic.corpus\n",
    "my.corpus.new  <- lapply(my.corpus.new, stemCompletion_mod, dict=dic.corpus )\n",
    "my.corpus.new[[1]]$content"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "##### 2) 한글처리\n",
    "\n",
    "- KoNLP 라는 패키지를 사용해서 처리함."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "also installing the dependencies 'DBI', 'memoise', 'Rcpp', 'plogr', 'whisker', 'rstudioapi', 'git2r', 'withr', 'stringr', 'hash', 'tau', 'Sejong', 'RSQLite', 'devtools'\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'DBI' successfully unpacked and MD5 sums checked\n",
      "package 'memoise' successfully unpacked and MD5 sums checked\n",
      "package 'Rcpp' successfully unpacked and MD5 sums checked\n",
      "package 'plogr' successfully unpacked and MD5 sums checked\n",
      "package 'whisker' successfully unpacked and MD5 sums checked\n",
      "package 'rstudioapi' successfully unpacked and MD5 sums checked\n",
      "package 'git2r' successfully unpacked and MD5 sums checked\n",
      "package 'withr' successfully unpacked and MD5 sums checked\n",
      "package 'stringr' successfully unpacked and MD5 sums checked\n",
      "package 'hash' successfully unpacked and MD5 sums checked\n",
      "package 'tau' successfully unpacked and MD5 sums checked\n",
      "package 'Sejong' successfully unpacked and MD5 sums checked\n",
      "package 'RSQLite' successfully unpacked and MD5 sums checked\n",
      "package 'devtools' successfully unpacked and MD5 sums checked\n",
      "package 'KoNLP' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmpkjsrvb\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"KoNLP\" ,  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning message:\n",
      ": package 'KoNLP' was built under R version 3.3.2Checking user defined dictionary!\n",
      "\n"
     ]
    }
   ],
   "source": [
    "library(KoNLP)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Backup was just finished!\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Downloading package from url: https://github.com/haven-jeon/NIADic/releases/download/0.0.1/NIADic_0.0.1.tar.gz\n",
      "Installing NIADic\n"
     ]
    },
    {
     "ename": "ERROR",
     "evalue": "Error in if (download_method_secure()) {: TRUE/FALSE가 필요한 곳에 값이 없습니다\n",
     "output_type": "error",
     "traceback": [
      "Error in if (download_method_secure()) {: TRUE/FALSE가 필요한 곳에 값이 없습니다\n"
     ]
    },
    {
     "ename": "ERROR",
     "evalue": "Error in tryCatch({: can't install NIADic package!\n Please refer 'https://github.com/haven-jeon/NIADic' to install.\n",
     "output_type": "error",
     "traceback": [
      "Error in tryCatch({: can't install NIADic package!\n Please refer 'https://github.com/haven-jeon/NIADic' to install.\n"
     ]
    }
   ],
   "source": [
    "useSejongDic()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'rJava' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmp2d5HTj\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"rJava\" ,  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning message:\n",
      ": package 'rJava' was built under R version 3.3.2"
     ]
    }
   ],
   "source": [
    "library(rJava)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "Sys.setenv( JAVA_HOME='D:/DOWNLOAD/JAVA/jdk7' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "extractNoun( '연습을 해보고자 한다. 명사가 잘 추출되는지 보자. 빨간색으로 글씨를 쓰고 있다.')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "#### 라. TDM 구축\n",
    "\n",
    "- 전처리된 텍스트 데이터에서 각 문서와 단어간의 사용 여부를 이용해 만들어진 matrix가 TDM( Term-Document Matrix )임.\n",
    "- TDM에서 단어의 빈도수를 보면 문서의 특성을 알 수 있음."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 1) 모든 단어\n",
    "\n",
    "- TermDocumentMatrix()함수를 실행하면, 행은 각 단어를 나타내고, 열은 트워터 메시지(문서)를 나타냄"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "my.TDM <- TermDocumentMatrix( my.corpus.new )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>1344</li>\n",
       "\t<li>1000</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 1344\n",
       "\\item 1000\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 1344\n",
       "2. 1000\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1] 1344 1000"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dim(my.TDM)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th></th><th scope=col>1</th><th scope=col>2</th><th scope=col>3</th><th scope=col>4</th><th scope=col>5</th><th scope=col>6</th><th scope=col>7</th><th scope=col>8</th><th scope=col>9</th><th scope=col>10</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><th scope=row>airlin</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>alarm</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>albert</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>alert</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>alevergara78</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>algorith</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|llllllllll}\n",
       "  & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10\\\\\n",
       "\\hline\n",
       "\tairlin & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\talarm & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\talbert & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\talert & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\talevergara78 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\talgorith & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "1. 0\n",
       "2. 0\n",
       "3. 0\n",
       "4. 0\n",
       "5. 0\n",
       "6. 0\n",
       "7. 0\n",
       "8. 0\n",
       "9. 0\n",
       "10. 0\n",
       "11. 0\n",
       "12. 0\n",
       "13. 0\n",
       "14. 0\n",
       "15. 0\n",
       "16. 0\n",
       "17. 0\n",
       "18. 0\n",
       "19. 0\n",
       "20. 0\n",
       "21. 0\n",
       "22. 0\n",
       "23. 0\n",
       "24. 0\n",
       "25. 0\n",
       "26. 0\n",
       "27. 0\n",
       "28. 0\n",
       "29. 0\n",
       "30. 0\n",
       "31. 0\n",
       "32. 0\n",
       "33. 0\n",
       "34. 0\n",
       "35. 0\n",
       "36. 0\n",
       "37. 0\n",
       "38. 0\n",
       "39. 0\n",
       "40. 0\n",
       "41. 0\n",
       "42. 0\n",
       "43. 0\n",
       "44. 0\n",
       "45. 0\n",
       "46. 0\n",
       "47. 0\n",
       "48. 0\n",
       "49. 0\n",
       "50. 0\n",
       "51. 0\n",
       "52. 0\n",
       "53. 0\n",
       "54. 0\n",
       "55. 0\n",
       "56. 0\n",
       "57. 0\n",
       "58. 0\n",
       "59. 0\n",
       "60. 0\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "              Docs\n",
       "Terms          1 2 3 4 5 6 7 8 9 10\n",
       "  airlin       0 0 0 0 0 0 0 0 0  0\n",
       "  alarm        0 0 0 0 0 0 0 0 0  0\n",
       "  albert       0 0 0 0 0 0 0 0 0  0\n",
       "  alert        0 0 0 0 0 0 0 0 0  0\n",
       "  alevergara78 0 0 0 0 0 0 0 0 0  0\n",
       "  algorith     0 0 0 0 0 0 0 0 0  0"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "as.matrix( my.TDM[55:60, 1:10] )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "##### 2) 단어 사전( dictionary )\n",
    "\n",
    "- 단어 사전은 텍스트마이닝에서 분석에 사용하고자 하는 단어들의 집합\n",
    "- 관심있는 단어들만을 추가해서 사전을 만들고 결과를 보자."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "myDict <- c('bigdata', 'data', 'big', 'amp', 'use', 'new',  'analyst', 'cloud', 'company', 'privacy', 'analytics', 'business', 'hadoop', 'datascience', 'iot' )\n",
    "my.TDM2 <- TermDocumentMatrix( my.corpus.new,  control=list( dictionary=myDict ) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<ol class=list-inline>\n",
       "\t<li>15</li>\n",
       "\t<li>1000</li>\n",
       "</ol>\n"
      ],
      "text/latex": [
       "\\begin{enumerate*}\n",
       "\\item 15\n",
       "\\item 1000\n",
       "\\end{enumerate*}\n"
      ],
      "text/markdown": [
       "1. 15\n",
       "2. 1000\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "[1]   15 1000"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th></th><th scope=col>1</th><th scope=col>2</th><th scope=col>3</th><th scope=col>4</th><th scope=col>5</th><th scope=col>6</th><th scope=col>7</th><th scope=col>8</th><th scope=col>9</th><th scope=col>10</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><th scope=row>amp</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>analyst</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>analytics</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>big</th><td>0</td><td>0</td><td>0</td><td>0</td><td>1</td><td>0</td><td>0</td><td>0</td><td>1</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>bigdata</th><td>1</td><td>1</td><td>0</td><td>1</td><td>1</td><td>1</td><td>1</td><td>1</td><td>1</td><td>1</td></tr>\n",
       "\t<tr><th scope=row>business</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>cloud</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>company</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>data</th><td>0</td><td>0</td><td>0</td><td>1</td><td>1</td><td>0</td><td>0</td><td>0</td><td>1</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>datascience</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>hadoop</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>iot</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>new</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>privacy</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "\t<tr><th scope=row>use</th><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td><td>0</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|llllllllll}\n",
       "  & 1 & 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 & 10\\\\\n",
       "\\hline\n",
       "\tamp & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tanalyst & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tanalytics & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tbig & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 1 & 0\\\\\n",
       "\tbigdata & 1 & 1 & 0 & 1 & 1 & 1 & 1 & 1 & 1 & 1\\\\\n",
       "\tbusiness & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tcloud & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tcompany & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tdata & 0 & 0 & 0 & 1 & 1 & 0 & 0 & 0 & 1 & 0\\\\\n",
       "\tdatascience & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\thadoop & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tiot & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tnew & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tprivacy & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\tuse & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/markdown": [
       "1. 0\n",
       "2. 0\n",
       "3. 0\n",
       "4. 0\n",
       "5. 1\n",
       "6. 0\n",
       "7. 0\n",
       "8. 0\n",
       "9. 0\n",
       "10. 0\n",
       "11. 0\n",
       "12. 0\n",
       "13. 0\n",
       "14. 0\n",
       "15. 0\n",
       "16. 0\n",
       "17. 0\n",
       "18. 0\n",
       "19. 0\n",
       "20. 1\n",
       "21. 0\n",
       "22. 0\n",
       "23. 0\n",
       "24. 0\n",
       "25. 0\n",
       "26. 0\n",
       "27. 0\n",
       "28. 0\n",
       "29. 0\n",
       "30. 0\n",
       "31. 0\n",
       "32. 0\n",
       "33. 0\n",
       "34. 0\n",
       "35. 0\n",
       "36. 0\n",
       "37. 0\n",
       "38. 0\n",
       "39. 0\n",
       "40. 0\n",
       "41. 0\n",
       "42. 0\n",
       "43. 0\n",
       "44. 0\n",
       "45. 0\n",
       "46. 0\n",
       "47. 0\n",
       "48. 0\n",
       "49. 0\n",
       "50. 1\n",
       "51. 0\n",
       "52. 0\n",
       "53. 0\n",
       "54. 1\n",
       "55. 0\n",
       "56. 0\n",
       "57. 0\n",
       "58. 0\n",
       "59. 0\n",
       "60. 0\n",
       "61. 0\n",
       "62. 0\n",
       "63. 0\n",
       "64. 1\n",
       "65. 1\n",
       "66. 0\n",
       "67. 0\n",
       "68. 0\n",
       "69. 1\n",
       "70. 0\n",
       "71. 0\n",
       "72. 0\n",
       "73. 0\n",
       "74. 0\n",
       "75. 0\n",
       "76. 0\n",
       "77. 0\n",
       "78. 0\n",
       "79. 0\n",
       "80. 1\n",
       "81. 0\n",
       "82. 0\n",
       "83. 0\n",
       "84. 0\n",
       "85. 0\n",
       "86. 0\n",
       "87. 0\n",
       "88. 0\n",
       "89. 0\n",
       "90. 0\n",
       "91. 0\n",
       "92. 0\n",
       "93. 0\n",
       "94. 0\n",
       "95. 1\n",
       "96. 0\n",
       "97. 0\n",
       "98. 0\n",
       "99. 0\n",
       "100. 0\n",
       "101. 0\n",
       "102. 0\n",
       "103. 0\n",
       "104. 0\n",
       "105. 0\n",
       "106. 0\n",
       "107. 0\n",
       "108. 0\n",
       "109. 0\n",
       "110. 1\n",
       "111. 0\n",
       "112. 0\n",
       "113. 0\n",
       "114. 0\n",
       "115. 0\n",
       "116. 0\n",
       "117. 0\n",
       "118. 0\n",
       "119. 0\n",
       "120. 0\n",
       "121. 0\n",
       "122. 0\n",
       "123. 0\n",
       "124. 1\n",
       "125. 1\n",
       "126. 0\n",
       "127. 0\n",
       "128. 0\n",
       "129. 1\n",
       "130. 0\n",
       "131. 0\n",
       "132. 0\n",
       "133. 0\n",
       "134. 0\n",
       "135. 0\n",
       "136. 0\n",
       "137. 0\n",
       "138. 0\n",
       "139. 0\n",
       "140. 1\n",
       "141. 0\n",
       "142. 0\n",
       "143. 0\n",
       "144. 0\n",
       "145. 0\n",
       "146. 0\n",
       "147. 0\n",
       "148. 0\n",
       "149. 0\n",
       "150. 0\n",
       "\n",
       "\n"
      ],
      "text/plain": [
       "             Docs\n",
       "Terms         1 2 3 4 5 6 7 8 9 10\n",
       "  amp         0 0 0 0 0 0 0 0 0  0\n",
       "  analyst     0 0 0 0 0 0 0 0 0  0\n",
       "  analytics   0 0 0 0 0 0 0 0 0  0\n",
       "  big         0 0 0 0 1 0 0 0 1  0\n",
       "  bigdata     1 1 0 1 1 1 1 1 1  1\n",
       "  business    0 0 0 0 0 0 0 0 0  0\n",
       "  cloud       0 0 0 0 0 0 0 0 0  0\n",
       "  company     0 0 0 0 0 0 0 0 0  0\n",
       "  data        0 0 0 1 1 0 0 0 1  0\n",
       "  datascience 0 0 0 0 0 0 0 0 0  0\n",
       "  hadoop      0 0 0 0 0 0 0 0 0  0\n",
       "  iot         0 0 0 0 0 0 0 0 0  0\n",
       "  new         0 0 0 0 0 0 0 0 0  0\n",
       "  privacy     0 0 0 0 0 0 0 0 0  0\n",
       "  use         0 0 0 0 0 0 0 0 0  0"
      ]
     },
     "execution_count": 158,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dim(my.TDM2)\n",
    "as.matrix(  my.TDM2[ , 1:10 ] )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 마. 분석 및 시각화\n",
    "\n",
    "##### 1) Association\n",
    "\n",
    "- 특정 단어와의 연관성에 따라서 단어를 조회할 수 있으며, findAssocs()함수를 통해서  연관성이 있는 단어를 찾음."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<strong>$iot</strong> = <dl class=dl-horizontal>\n",
       "\t<dt>devop</dt>\n",
       "\t\t<dd>0.3</dd>\n",
       "\t<dt>m2m</dt>\n",
       "\t\t<dd>0.3</dd>\n",
       "\t<dt>fantast</dt>\n",
       "\t\t<dd>0.28</dd>\n",
       "\t<dt>track</dt>\n",
       "\t\t<dd>0.28</dd>\n",
       "\t<dt>digitaltr</dt>\n",
       "\t\t<dd>0.27</dd>\n",
       "\t<dt>apm</dt>\n",
       "\t\t<dd>0.26</dd>\n",
       "\t<dt>blockchain</dt>\n",
       "\t\t<dd>0.26</dd>\n",
       "</dl>\n"
      ],
      "text/latex": [
       "\\textbf{\\$iot} = \\begin{description*}\n",
       "\\item[devop] 0.3\n",
       "\\item[m2m] 0.3\n",
       "\\item[fantast] 0.28\n",
       "\\item[track] 0.28\n",
       "\\item[digitaltr] 0.27\n",
       "\\item[apm] 0.26\n",
       "\\item[blockchain] 0.26\n",
       "\\end{description*}\n"
      ],
      "text/markdown": [
       "**$iot** = devop\n",
       ":   0.3m2m\n",
       ":   0.3fantast\n",
       ":   0.28track\n",
       ":   0.28digitaltr\n",
       ":   0.27apm\n",
       ":   0.26blockchain\n",
       ":   0.26\n",
       "\n"
      ],
      "text/plain": [
       "$iot\n",
       "     devop        m2m    fantast      track  digitaltr        apm blockchain \n",
       "      0.30       0.30       0.28       0.28       0.27       0.26       0.26 \n"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "findAssocs( my.TDM, 'iot', 0.25 )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "- 상관분석 패키지인 arules 패키지의 apriori()함수를 이용해 연관분석을 위한 데이터 변환이 가능함."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'arules' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmp67f15o\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"arules\" ,  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "library(arules)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "ERROR",
     "evalue": "Error in as(my.TDM, \"transactions\"): no method or default for coercing \"TermDocumentMatrix\" to \"transactions\"\n",
     "output_type": "error",
     "traceback": [
      "Error in as(my.TDM, \"transactions\"): no method or default for coercing \"TermDocumentMatrix\" to \"transactions\"\n"
     ]
    },
    {
     "ename": "ERROR",
     "evalue": "Error in .class1(object): 객체 'transcation_m'를 찾을 수 없습니다\n",
     "output_type": "error",
     "traceback": [
      "Error in .class1(object): 객체 'transcation_m'를 찾을 수 없습니다\n"
     ]
    }
   ],
   "source": [
    "transcation_m <- as( my.TDM, \"transactions\" ) \n",
    "rules.all <- apriori( transcation_m, parameter=list(supp=0.01, conf=0.5) )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 2) 워드 클라우드"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "package 'wordcloud' successfully unpacked and MD5 sums checked\n",
      "\n",
      "The downloaded binary packages are in\n",
      "\tC:\\Users\\Y.G JI\\AppData\\Local\\Temp\\Rtmp67f15o\\downloaded_packages\n"
     ]
    }
   ],
   "source": [
    "install.packages(\"wordcloud\" ,  repo=\"http://cran.nexr.com\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "library(wordcloud)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<dl class=dl-horizontal>\n",
       "\t<dt>bigdata</dt>\n",
       "\t\t<dd>938</dd>\n",
       "\t<dt>iot</dt>\n",
       "\t\t<dd>283</dd>\n",
       "\t<dt>data</dt>\n",
       "\t\t<dd>275</dd>\n",
       "\t<dt>analyt</dt>\n",
       "\t\t<dd>197</dd>\n",
       "\t<dt>big</dt>\n",
       "\t\t<dd>133</dd>\n",
       "\t<dt>machinelearn</dt>\n",
       "\t\t<dd>131</dd>\n",
       "\t<dt>datasci</dt>\n",
       "\t\t<dd>127</dd>\n",
       "\t<dt>blockchain</dt>\n",
       "\t\t<dd>126</dd>\n",
       "\t<dt>fintech</dt>\n",
       "\t\t<dd>126</dd>\n",
       "\t<dt>makeyourownlan</dt>\n",
       "\t\t<dd>113</dd>\n",
       "\t<dt>cloud</dt>\n",
       "\t\t<dd>93</dd>\n",
       "\t<dt>amp</dt>\n",
       "\t\t<dd>80</dd>\n",
       "\t<dt>artificialintellig</dt>\n",
       "\t\t<dd>77</dd>\n",
       "\t<dt>busi</dt>\n",
       "\t\t<dd>77</dd>\n",
       "\t<dt>defstar5</dt>\n",
       "\t\t<dd>75</dd>\n",
       "</dl>\n"
      ],
      "text/latex": [
       "\\begin{description*}\n",
       "\\item[bigdata] 938\n",
       "\\item[iot] 283\n",
       "\\item[data] 275\n",
       "\\item[analyt] 197\n",
       "\\item[big] 133\n",
       "\\item[machinelearn] 131\n",
       "\\item[datasci] 127\n",
       "\\item[blockchain] 126\n",
       "\\item[fintech] 126\n",
       "\\item[makeyourownlan] 113\n",
       "\\item[cloud] 93\n",
       "\\item[amp] 80\n",
       "\\item[artificialintellig] 77\n",
       "\\item[busi] 77\n",
       "\\item[defstar5] 75\n",
       "\\end{description*}\n"
      ],
      "text/markdown": [
       "bigdata\n",
       ":   938iot\n",
       ":   283data\n",
       ":   275analyt\n",
       ":   197big\n",
       ":   133machinelearn\n",
       ":   131datasci\n",
       ":   127blockchain\n",
       ":   126fintech\n",
       ":   126makeyourownlan\n",
       ":   113cloud\n",
       ":   93amp\n",
       ":   80artificialintellig\n",
       ":   77busi\n",
       ":   77defstar5\n",
       ":   75\n",
       "\n"
      ],
      "text/plain": [
       "           bigdata                iot               data             analyt \n",
       "               938                283                275                197 \n",
       "               big       machinelearn            datasci         blockchain \n",
       "               133                131                127                126 \n",
       "           fintech     makeyourownlan              cloud                amp \n",
       "               126                113                 93                 80 \n",
       "artificialintellig               busi           defstar5 \n",
       "                77                 77                 75 "
      ]
     },
     "execution_count": 166,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "my.TDM.m <- as.matrix(my.TDM)\n",
    "term.freq <- sort( rowSums(my.TDM.m), decreasing=T ) \n",
    "head( term.freq, 15 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "application/pdf": "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",
      "image/jpeg": "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",
      "image/png": "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",
      "text/plain": [
       "plot without title"
      ]
     },
     "metadata": {
      "image/svg+xml": {
       "isolated": true
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "wordcloud( word=names(term.freq), freq=term.freq, min.freq=15, random.order=F, color=brewer.pal(8, 'Dark2') )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### 3) 감성분석"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Warning message:\n",
      "In doRppAPICall(\"search/tweets\", n, params = params, retryOnRateLimit = retryOnRateLimit, : 1500 tweets were requested but the API can only return 130"
     ]
    }
   ],
   "source": [
    "hilton.tweets <- searchTwitter('@hilton', n=1500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "130"
      ],
      "text/latex": [
       "130"
      ],
      "text/markdown": [
       "130"
      ],
      "text/plain": [
       "[1] 130"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "length( hilton.tweets )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1] \"meghanmcinnls: RT @BrandBastion: \\\"Responding to reviews is a simple and great way to retain customers.\\\" - @leediaz24 of @hilton #SMSSummit\""
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "[1] \"BusinessWire: @BCVSocial @hilton @TapastryHilton Thanks for sharing!\""
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hilton.tweets[[1]]\n",
    "hilton.tweets[[2]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- 긍정/부정 단어 사전을 만듬."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "pos.word <- scan( \"https://gist.githubusercontent.com/mkulakowski2/4289437/raw/1bb4d7f9ee82150f339f09b5b1a0e6823d633958/positive-words.txt\"\n",
    "                 , what='character', comment.char=';' )\n",
    "neg.word <- scan( \"https://gist.githubusercontent.com/mkulakowski2/4289441/raw/dad8b64b307cd6df8068a379079becbb3f91101a/negative-words.txt\"\n",
    "                 , what='character', comment.char=';' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "pos.word <- c(pos.word, 'upgrade' )\n",
    "neg.word <- c(neg.word, 'wtf', 'wait', 'waiting', 'epicfail', 'mechanical' )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "score.sentiment = function(sentences, pos.words, neg.words, .progress='none')\n",
    "{\n",
    "\trequire(plyr)\n",
    "\trequire(stringr)\n",
    "\t\n",
    "\t# we got a vector of sentences. plyr will handle a list or a vector as an \"l\" for us\n",
    "\t# we want a simple array of scores back, so we use \"l\" + \"a\" + \"ply\" = laply:\n",
    "\tscores = laply(sentences, function(sentence, pos.words, neg.words) {\n",
    "\t\t\n",
    "\t\t# clean up sentences with R's regex-driven global substitute, gsub():\n",
    "\t\tsentence = gsub('[[:punct:]]', '', sentence)\n",
    "\t\tsentence = gsub('[[:cntrl:]]', '', sentence)\n",
    "\t\tsentence = gsub('\\\\d+', '', sentence)\n",
    "\t\t# and convert to lower case:\n",
    "\t\tsentence = tolower(sentence)\n",
    "\n",
    "\t\t# split into words. str_split is in the stringr package\n",
    "\t\tword.list = str_split(sentence, '\\\\s+')\n",
    "\t\t# sometimes a list() is one level of hierarchy too much\n",
    "\t\twords = unlist(word.list)\n",
    "\n",
    "\t\t# compare our words to the dictionaries of positive & negative terms\n",
    "\t\tpos.matches = match(words, pos.words)\n",
    "\t\tneg.matches = match(words, neg.words)\n",
    "\t\n",
    "\t\t# match() returns the position of the matched term or NA\n",
    "\t\t# we just want a TRUE/FALSE:\n",
    "\t\tpos.matches = !is.na(pos.matches)\n",
    "\t\tneg.matches = !is.na(neg.matches)\n",
    "\n",
    "\t\t# and conveniently enough, TRUE/FALSE will be treated as 1/0 by sum():\n",
    "\t\tscore = sum(pos.matches) - sum(neg.matches)\n",
    "\n",
    "\t\treturn(score)\n",
    "\t}, pos.words, neg.words, .progress=.progress )\n",
    "\n",
    "\tscores.df = data.frame(score=scores, text=sentences)\n",
    "\treturn(scores.df)\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>\n",
       "<thead><tr><th></th><th scope=col>score</th><th scope=col>text</th></tr></thead>\n",
       "<tbody>\n",
       "\t<tr><th scope=row>1</th><td>1</td><td>I love you</td></tr>\n",
       "\t<tr><th scope=row>2</th><td>-1</td><td>I hate you</td></tr>\n",
       "\t<tr><th scope=row>3</th><td>1</td><td>What a wonderful day!</td></tr>\n",
       "\t<tr><th scope=row>4</th><td>2</td><td>I love you. I like you.</td></tr>\n",
       "</tbody>\n",
       "</table>\n"
      ],
      "text/latex": [
       "\\begin{tabular}{r|ll}\n",
       "  & score & text\\\\\n",
       "\\hline\n",
       "\t1 & 1 & I love you\\\\\n",
       "\t2 & -1 & I hate you\\\\\n",
       "\t3 & 1 & What a wonderful day!\\\\\n",
       "\t4 & 2 & I love you. I like you.\\\\\n",
       "\\end{tabular}\n"
      ],
      "text/plain": [
       "  score                    text\n",
       "1     1              I love you\n",
       "2    -1              I hate you\n",
       "3     1   What a wonderful day!\n",
       "4     2 I love you. I like you."
      ]
     },
     "execution_count": 181,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample <- c('I love you', 'I hate you', 'What a wonderful day!', 'I love you. I like you.' )\n",
    "result <- score.sentiment ( sample, pos.word, neg.word)\n",
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "R",
   "language": "R",
   "name": "ir"
  },
  "language_info": {
   "codemirror_mode": "r",
   "file_extension": ".r",
   "mimetype": "text/x-r-source",
   "name": "R",
   "pygments_lexer": "r",
   "version": "3.3.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}