Exploring service improvement through importance-performance analysis considering the reliability of multiple online platforms
提出一种基于多平台在线评论的服务改进方法，通过考虑平台可靠性和不同评价模式，利用主题模型和情感分析提取属性重要性，构建重要性绩效分析模型，帮助酒店管理者制定更有效的服务改进策略。
Service improvement has emerged as a pivotal task for hoteliers to ensure competitive advantage. This study proposes a service improvement method based on online reviews from multiple platforms considering the reliability of online platforms and different evaluation modes, where the reliability of an online platform is defined based on the number of online reviews on that platform and the degree of review helpfulness. In our method, Latent Dirichlet Allocation model is utilized to extract keywords, and lexicon-based sentiment analysis methods are employed to analyze the sentiment of online reviews on each platform considering different evaluation modes. The importance of attributes on each platform is measured by the TextRank method. A multi-platform-oriented importance-performance analysis model is constructed based on the integrated performance and the importance of attributes, so as to classify attributes and formulate service improvement strategies. A case study about hotel service improvement is implemented to illustrate the effectiveness of the method. Results show that the attributes classification results considering the reliability of multiple platforms is more reasonable compared to the results based on a single platform, providing more effective service improvement strategy and clearer view of attribute status on various platforms for hoteliers.