simjoin_entitymatching.value_matcher.word2vec.Word2Vec Class Reference

Public Member Functions
	__init__ (self, inmemory_)
	train_and_save (self, blk_attr, rawtable, rawtable2=None)
	train_all_and_save (self, attrs, rawtable, rawtable2=None)
	apply_sample (self, blk_attr, tau)
	group_interchangeable (self, blk_attr, tau)
	load_blk_res (self, usage)
	load_match_res (self)
	load_model (self)

Public Attributes
str	model = ""
str	blk_res = ""
str	match_res = ""
list	setences = []
int	inmemory = inmemory_

Protected Member Functions
	_preprocess (self, blk_attr, rawtable, rawtable2=None)

Detailed Description

Word2Vec for attribute: str_eq_1w or numeric
But do numeric really needs normalization?

Constructor & Destructor Documentation

__init__()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.__init__	(		self,
			inmemory_ )

Member Function Documentation

_preprocess()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec._preprocess ( self, blk_attr, rawtable, rawtable2 = None )

protected

apply_sample()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.apply_sample	(		self,
			blk_attr,
			tau )

Apply Word2Vec for sampling and labeling cand

group_interchangeable()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.group_interchangeable	(		self,
			blk_attr,
			tau )

Apply Word2Vec for grouping interchangeable value in blocking result

load_blk_res()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.load_blk_res	(		self,
			usage )

Avoid multi-io
usage: 0 for labeler & training and 1 for value matcher
only use this method when training using a sample

load_match_res()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.load_match_res

(

self

)

load_model()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.load_model

(

self

)

train_all_and_save()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.train_all_and_save	(		self,
			attrs,
			rawtable,
			rawtable2 = None )

Train model for all attributes except id
attrs: attributes that could use word2vec

train_and_save()

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.train_and_save	(		self,
			blk_attr,
			rawtable,
			rawtable2 = None )

Member Data Documentation

blk_res

str simjoin_entitymatching.value_matcher.word2vec.Word2Vec.blk_res = ""

inmemory

int simjoin_entitymatching.value_matcher.word2vec.Word2Vec.inmemory = inmemory_

match_res

str simjoin_entitymatching.value_matcher.word2vec.Word2Vec.match_res = ""

model

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.model = ""

setences

simjoin_entitymatching.value_matcher.word2vec.Word2Vec.setences = []

---

The documentation for this class was generated from the following file:

• simjoin_entitymatching/value_matcher/word2vec.py