A Topic Recognition Method of News Text Based on Word Embedding Enhancement

Table 1

The definition and description of symbols involved in this paper.

Symbol definition	Description
	The number of tokens in document
	The vocabulary that including all words of corpus
	The number of topics when training the LDA model
	LDA model obtained by training on corpus
	The number of word vector dimension when training the word embedding model
EM	Word embedding model obtained by training on corpus
	Document-level topic distribution for document
	Word-level topic distribution for document
	Text representation based on word embedding model and doc-level topic distribution
	Text representation based on word embedding model and word-level topic distribution
	Dimension of document representation vector
	Topic distribution
	Symbol that means concatenation operation of vectors
	Symbol that means summation operation of vectors
	Corpus set consists of training corpus and test corpus