A Cross-Domain Analysis using Morphological Sentence Pattern Approach for Extracting Aspect-based Lexicon

Social media data contains people's emotions, opinions and experiences. Sentiment analysis aims to analyze the data to observe meaningful information. However, building lexicon for a lexicon based sentiment analysis is one of the biggest challenges without human-coding efforts. In this study, we proposed a cross-domain approach for building sentimental lexicon using the morphological sentence patterns for analyzing social media data. Our approach shows relatively higher F-score (79.64) than existing approaches. In addition, this approach can be used for multi-source data such as online reviews and social media data without human-coded knowledge bases.