An Online Housing Reputation Assessment Framework Based on Text Mining and Visualization Technologies

Learning from online comments is essential for enhancing understanding and improving online housing reputation (OHR). However, two significant issues require attention. First, analyzing online comments for reputation information extraction is a labor-intensive and time-consuming task. Second, most existing online housing information platforms lack effective visual aids, merely presenting the average comment ratings or listing comment texts without secondary interpretation. To address these challenges, this study proposes an OHR assessment framework based on text mining and visualization technologies. This study first evaluates the performance of eight sentiment analysis models for analyzing housing comments, and the attention-based BiLSTM model achieved the highest accuracy (83.57%). Additionally, a housing attribute ontology is constructed to reveal eight critical attributes influencing OHR. Finally, a reputation visualization scheme is designed to comprehensively present OHR. A case study for analyzing online comments from three construction enterprises reveals the advantages and feasibility of the proposed framework for assessing OHR. This study contributes to the body of knowledge by establishing the connection between housing comments and OHR, greatly advancing the research in the construction domain's reputation management. Furthermore, OHR analysis can facilitate decision making optimization for both consumers and managers, which has theoretical and practical significance for the healthy and sustainable development of the online housing market.