Implementations of several Entity Linking algorithms: TabEL, LIEGE and Taiyun
Python HTML
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.
JudgePage
LIEGE
MarkPage
TabEL
Taiyun
.gitignore
jist2016.pdf

```S = [] # mention集合, [   m1        ,    m2        , ...]
E = [] # entitys集合, [[e1, e2, ...], [e1, e2, ...], ...] 注意上下的对齐

id = [] # entity存储在数据库里, 所以每个entity都对应一个id, 主要是为了与人工标注的实体比较
element = {} # mention与entity编号之后的元素 {序号：mention 或者 entity}, 论文中的算法借鉴了马尔科夫链的思想, 需要构建一个矩阵, 所以先给mention, entity编号

a1 = a2 = 0.7 # 某经验系数
b1 = b2 = 0.3 # 某经验系数

contextMention = {} # 表格一行一列除掉本身的所有单元格作为mention对应的上下文, {mention : context}
contextEntity = {} # 利用abstract文件得到entity对应的上下文, {entity : context}

epsion = 0.000001 # 马尔科夫链停止的条件
upLimit = 50000 # 迭代次数上限```

```def getData(Mentions,S,E,contextMention,contextEntity, id)

def setNumber(S, E, element)

def jaccard_similarity_score(context1, context2, flag1, flag2)

flag1为True: context1为mention的上下文,
flag1为False: context1为entity的上下文,
flag2同理

def contextSim(contextMention,contextEntity,S,element,i,j)

def increase(P,Plast)

def getStateTransitionMatrix(contextMention,contextEntity,S,element)

def disambiguate(A,upLimit)

def getResultFromP(S,P,element)

def main(Mentions, nrow, ncol)

You can’t perform that action at this time.