Wolfram Computation Meets Knowledge

Wolfram Summer School


Zhou Li

Summer School

Class of 2014


Zhou Li is now a graduate student at Mississippi State University in Starkville, Mississippi, studying physics. With quantum transport as his research topic, he is also very interested in computer vision and pattern recognition.

Besides the above topics, he also enjoys tennis, fishing, and movies.

Project: Short Text Emotion Detection

Text emotion detection focuses on computational analysis of emotion from a text. Emotion detection in text is increasingly important considering the tasks that arise in opinion mining, affective computing, and natural language processing. One example is human-computer interaction based on verbal expressivity. This project tries to establish a system that has the functionality of automatic short text emotion tagging. WordNet-Affect is a database including six groups of synsets that relate to six basic emotions. One drawback of WordNet-Affect is that it only includes a limited number of words for each emotion category. In order to use more emotion-tagged words in text emotion detection, this project first expands the emotion words lists by using 15 books. Depending on the times of appearance in the neighbor of an emotion word in the original emotion lists, a new word has been assigned a score for each emotion. So for the new words, there will be a six-dimension vector representing the emotion related with this word. This project aims to employ the WordNet-Affect database and the new emotion-tagged words from books to figure out emotion components in short text like news headlines. Also, this project tries to figure out whether the overall emotion of a text is directly determined by the emotion components in its consisting words.