การวิเคราะห์อารมณ์ของประโยคโดยอาศัยความหมายของคำใช้ซัพพอร์ตเวกเตอร์แมชชีน

Abstract

The objectives of the independent study entitled “Sentence Emotion Analysis Based on Words Semantic Using Support Vecter Machine” are to develop a tool to help measure the classification of Thai sentense by applying model process, group of emotions, measure the performance, classify by Support Vector Machine, and to study the grouping of emotions, using natural language processing. The measurement of the similarity of Thai documents using natural language processing was divided into two parts, these begin: preparation of documents prior to processing, and transform the data into Attribute – Relation File Format (ARFF) using divide words and pattern matching. In total, the development tool of emotion classification using PHP, Java, and Perl The effectiveness of the similarity for Emotion Words Using Support Vecter Machine, can be tested through the accuracy and correctness of the data on different levels. The results of the test showed that the accuracy of system with Cross-Validate and the performance of system with F-measure. In conclude, there are Cross-Validate 72 percentage and F-measure more than 0.5.