Semi-Automatic WordNet Based Emotion Dictionary Construction

This paper describes an algorithm for semi automatically creating an emotion dictionary using WordNet. The algorithm takes as input a set of seed words that have had emotion information assigned as well as WordNet sense information. From this list an initial dictionary is automatically created using the various relations found within WordNet. Then, various correction stages are performed where parts of the dictionary are shown to the user for verification and additional information in the form of emotion polarity and probability information are assigned. Using the proposed algorithm with a set of 549 seed words, an emotion dictionary containing over 13,000 WordNet senses was created in just under 7 person-hours of time. To evaluate the created dictionary its usefulness in improving the performance of affect and sentiment classification was examined. Classification was performed using support vector machines and a baseline non-machine learning dictionary based algorithm. The results showed that the error rate is reduced when using the dictionary over when not using the dictionary.