Home¶

1. Please feel free to edit this page.
2. TEST

Tools¶

• Ling Pipe
  • commercial strength NLP (with sentiment analysis) java API used by technocrati, buzzmetrics, westlaw, etc.

Papers¶

General Sentiment Analysis¶

1. Rev Rank: a Fully Unsupervised Algorithm for Selecting the Most Helpful Book Reviews
2. http://www.scienceforseo.com/opinion-mining/sentiment-analysis-in-text/

Graph-based techniques¶

1. The small world of human language * A very nice paper to start reading. I am thinking of modeling the sentiments as ordered graphs.
2. "TextRank: Bringing Order Into Texts": http://belobog.si.umich.edu/clair/anthology/query.cgi?type=Paper&id=W04-3252 * graph based method for keyword extraction. nouns/adjectives as nodes, and co-occurences (within a window of 2-10 words) as edges. A x x B => there is a undirected link from A to B (vice-versa). * graph based method to extract summary sentences. sentences as nodes. similarity between any two sentences is used as link weight. * Unsupervised! yet obtain comparable performances to supervised method. *

   Intuitively, TextRank works well because it does not only rely on the local context of a text unit (vertex),
   but rather it takes into account information recursively drawn from the entire text (graph).
   

   * "TextRank Java code": http://github.com/sharethis/textrank/tree/master
3. Graph-Based Text Representation For Novelty Detection

Others¶

1. InfluenceRank