Text Mining : Challenges, Algorithms, Tools and Applications

Abstract

Encompassing fields such as information retrieval, extraction, classification, summarization and understanding, text mining has become indispensable across diverse application domains. Methodological advances span rule-based systems, statistical approaches, support vector machines, clustering, neural networks and most recently, deep learning, with distance and similarity estimation persisting as central challenges. This Special Issue comprises two survey papers and sixteen research articles, all addressing practical aspects of text mining. The surveys review methods for financial market prediction through social media analysis and sentiment detection, and provide a synthesis of techniques, evaluation strategies, and applications of recommender systems across domains including e-commerce, social media, and online learning. The research articles extend to text classification, sentiment analysis, summarization, and natural language understanding, as well as corpus construction and e-governance applications. Contributions include a large-scale Chinese toponym corpus, CHTopo, deep learning models for tourist behaviour prediction, semantic optimization of artifact presentation, BERT-based knowledge extraction for agricultural ontology construction, and transformer-based approaches for detecting moral features in television narratives.