While there have been many tourism studies exploring business opportunities from online comments, research in the context of sustainable rural tourism is rare. In particular, practitioners involved in rural destinations are usually transformed from farmers who are relatively new to information technology and therefore unable to understand the voices of tourists in a timely and effective manner. Aiming to expand the literature on sustainable rural tourism, this study cuts from tourist satisfaction and proposes a mixed methodology to investigate insights for sustaining rural destinations from online comments. Specifically, the KMeans clustering algorithm is first utilized to discover service attributes that visitors value; then Importance-Performance Analysis (IPA), an effective satisfaction analysis tool, is adopted to understand tourist satisfaction; to develop specific sustainable strategies, using an advanced opinion extraction technique this work identified dissatisfaction factors of visitors. 5832 online comments of two adjacent rural destinations (i.e. Hongcun and Xidi) were crawled to validate the proposed methodology. Results show that five service attributes were discovered, including Natural environment, Price, Food, Hospitality, and Culture & heritage; Hongcun's Natural environment and Price require urgent measures, and its Food and Hospitality are at a competitive disadvantage compared to Xidi; detailed factors dissatisfying visitors were identified through opinion extraction technique, based on which specific sustainable strategies are generated for authorities and practitioners. By excavating business opportunities from online reviews via big data analysis and text mining techniques, the proposed methodology advances the academic understanding of promoting rural tourism.

虽然有许多旅游研究从在线评论中探索商机，但在可持续乡村旅游背景下的研究却很少。尤其是乡村目的地的从业者，通常都是由农民转型而来，对于信息技术的了解相对较新，因此无法及时有效地了解游客的声音。为了扩大有关可持续乡村旅游的文献，本研究从游客满意度出发，提出了一种混合方法来调查在线评论中维持乡村目的地的见解。具体来说，首先利用KMeans聚类算法发现访问者重视的服务属性；然后是重要性绩效分析（IPA），一种有效的满意度分析工具，用于了解游客满意度；为了制定具体的可持续策略，这项工作使用先进的意见提取技术确定了访客的不满意因素。爬取了两个相邻乡村目的地（即宏村和西递）的 5832 条在线评论来验证所提出的方法。结果显示，发现了五个服务属性，包括自然环境、价格、食品、招待以及文化和遗产；宏村的自然环境和物价亟待采取措施，其餐饮和接待与西递相比处于竞争劣势；通过意见提取技术确定了参观者不满意的详细因素，并在此基础上为当局和从业者制定了具体的可持续战略。