Implicit aspect based sentiment analysis for restaurant review using LDA topic modeling and ensemble approach

Technological advancements in e-commerce and Web 2.0 have revolutionized how customers express their opinions about services and features through reviews on various websites. This trend is particularly prominent in the travel industry, where online sources offer valuable insights into the food and accommodations of destinations. However, the abundance of reviews available online presents a challenge for travelers in filtering relevant information. To tackle this issue, aspect-based sentiment analysis (ABSA) was proposed as a technique for extracting opinions based on specific features. Topic modeling and sentiment analysis are two significant techniques employed to assist in this analysis. Topic modeling involves identifying thematic relationships among documents, while sentiment analysis aims to determine the expressed opinions in the text. This study utilized one of the leading travel websites, Tripadvisor, to gather customer reviews of different restaurants. These reviews were then subjected to aspect-based sentiment analysis using latent Dirichlet allocation (LDA) and ensemble bagging support vector machine (EBSVM) classifier techniques. The objective is to identify the most relevant aspect within the restaurant domain and enhance sentiment analysis performance. To address class imbalances in the datasets, the synthetic minority over-sampling technique (SMOTE) was implemented. The performance of LDA was evaluated using the coherence score, which indicates the quality of topics generated for restaurant reviews. The effectiveness of the EBSVM classifier was measured using metrics such as accuracy, precision, recall, and F1 score. The proposed model achieved an accuracy of 96.1%, surpassing other techniques. Overall, this study demonstrates the effectiveness of aspect-based sentiment analysis in extracting relevant opinions from a large volume of reviews. It also highlights the potential of machine learning techniques in enhancing sentiment analysis performance. The suggested approach outperforms other techniques discussed in the existing literature, contributing to an overall improvement in sentiment analysis.