[20160812] Weekly Report of CS297

Reviewed 1 paper using Word2Vec to enhance Naive Bayes Classifier. The Title is shown below:

Screen Shot 2016-08-15 at 12.40.06 PM

 

Pros:

  • Introduced semantic analysis into text classification. Word2Vec is shown to improve the classification accuracy.
  • Applied distributed method for large-scale computation 

Cons:

  • For each class, the same corpus is used. If find corpus with each different class, the result may be more accurate

[160729] Weekly Report of CS297

Reviewed one paper using the tool (word2Vec) for determining the characteristic vocabulary.

The title of the paper is posted as below:

Screen Shot 2016-08-01 at 12.54.48 AM.png

The author proposed a work flow to detect the characteristic vocabulary of the domain in question by using 1. a crawler to gather the text information and 2. word2vec to rank the similar words. The schematic of the work flow can be seen below:

Screen Shot 2016-08-01 at 1.02.57 AM.png

Reference:

https://arxiv.org/abs/1605.09564