Emotions are closely related to sentiments. In this study entitled “Sentiment Analysis as an aid for Emotion Recognition for the Interpretations of Sacks’ Sentence Completion Testâ€, aims to develop a system that used to interpret a projective or personality test named Sacks’ Sentence Completion Test by recognizing emotion with the help of sentiment analysis. This study is significant to both fields of Computer Science and Psychology, in a way that helps to solve issues in emotion recognition and sentiment analysis and may serve as additional information to Computer-Based Test Interpretations as an innovation to their field. The researchers used the methods in determining the polarity of the sentence; first is the scoring using words from Bing Liu and the second is the machine learning through Naïve Bayes (NB) and Support Vector Machine (SVM). Word-based emotion recognition was used in emotion recognition to detect many emotions using the Plutchik’s Wheel of Emotion as a basis. Experiments were done and the data gathered were used to measure the performance of the system to compute as the accuracy of sentiment analysis in terms of: Sentiment with sarcasm, slangs, and emoticons; accuracy of the sentiment analysis when the different features or solutions are combined; accuracy of the system based on recognizing emotion using sentiment analysis; and accuracy of the system based on the interpretations of Sacks’ Sentence Completion Test. As the result, the sentiment analysis for the positive set got a fair rate of accuracy with 80.86%, negative set got a satisfactory rate of accuracy with 72.29%, and neutral set with high rate of accuracy with 13.04%. For the Emotion recognition, the accuracy ranges from 11.2% (lowest, no emotion) to 93.89% (highest, fear). The researchers found out that with this kind of result, people could use sentiment analysis as an aid for emotion recognition for Sacks’ Sentence Completion Test (SSCT).
Volume 12 | 06-Special Issue
Pages: 322-329
DOI: 10.5373/JARDCS/V12SP6/SP20201038