There has been a serious issue with overload of data because of the abundance of consumer‐generated suggestions on location‐based social networks (LBSNs). The amount and fragmented nature of all these brief comments make it hard for users to efficiently gather relevant details customized to their interests, even when they provide useful, current data on venues and items. Present analysis methods frequently generate generic findings by failing to consider user preferences into consideration. The article offers a novel architecture that generates customizable tip briefs, which are constantly matched with user preferences with the goal to fill this gap in knowledge. Particular aspect significance distributions are predicted by the suggested technique’s novel integration of a user‐based collaborative filtering method, which then directs a graph‐based summarization approach to select which is most useful information. The most important novelty is the integration of these methods in a minimum description length (MDL) framework, which achieves the right balance between representativeness and conciseness to deliver summaries that are both thorough and brief. Comprehensive testing on a dataset of 10,377 real‐world proposals indicates that the proposed method, MMEQC, performs significantly superior to baseline algorithms comparable to TextRank and Opinosis. Statistical outcomes on multiple significant criteria, such as coherence (0.84), diversity (0.55), and representativeness (0.48), confirm the superior performance of the proposed approach for creating outstanding, targeted summaries. This study has presented an important step toward the extraction of customized data in online social commerce scenarios.
Al-Khiza’ay et al. (Thu,) studied this question.