情報処理学会第85回全国大会 会期:2023年3月2日~4日 会場:電気通信大学

1V-03
Evaluating Aspect Category Sentiment Analysis Using Foreign Skiers' Review Dataset
○儀 忠躍,関 洋平(筑波大)
Measuring satisfaction at Japanese ski resorts of foreign skiers using questionnaires is very time-consuming. We attempted to measure foreign skiers’ satisfaction by using Aspect Category Sentiment Analysis (ACSA) to detect sentiment polarity in entities such as slopes, lifts, restaurants, and so on. We collected over two thousand English skiers’ reviews from Tripadvisor and manually annotated them according to their sentiment polarities towards 26 pre-defined aspect categories. Moreover, we evaluated our approach using our dataset and SemEval-2016 Task 5 dataset. The results in this study illustrated that ACSA is a promising approach to measure foreign skiers’ satisfaction.