Word Sense Disambiguation for Lexicon-Based Sentiment Analysis

Endang Wahyu Pamungkas, Divi Galih Prasetyo Putri

2017 · 11 citations · 27 references

Concepts

Abstract

The total number of user generated data from several media such as social media, online forums, and micro blogs is increasing rapidly. This data contains people opinion and feeling about a product, service, or even an idea. Therefore, user generated data becomes important to serveral parties to see public response and make decision. Lexicon-based sentiment analysis is one of the techniques that can be used to analyze user generated data. Recent study found that ambiguous word is becoming the one issue on the lexicon-based sentiment analysis. This paper try to handle ambiguous issue by adding Word Sense Disambiguation process. We implement several methods of Path-based semantic relatedness to find the most appropriate sense. The result shows that word sense disambiguation process is able to improve the performance of sentiment analysis. When combined with word sense disambiguation, the classification f-score is better than sentiment analysis without word sense disambiguation.

References

27