This is an implementation of the paper
I have provided a pre processed csv file using the preprocess.py and dataFrameGen.py. Also the vector for text is generated using the vectorizer.py which uses features of the articles. (without using NLTK)
original dataset contained:
- Total lines in articles :: 10405
- Total words in articles :: 358695
- Total characters in articles :: 1889183
- Total no of unique words :: 73889
- line count
- char count
- word count
- average word size
- vowels per word
- consonants per word
- matras per word
- count of words of fize size
- count of words below size
- count of words above size
| Sr. No. | Classifier | Accuracy |
|---|---|---|
| 1 | Naive Bayes (Guassian) | 89 % |
| 2 | SVC | 70 % |
| 3 | Decision Tree | 99 % |
| 4 | K Nearest Neighbour | 98.8 % |