Prediction and Analysis of Next Website Requests: A Hybrid Approach Integrating Fuzzy Clustering and Fuzzy Association Rule Mining

Data mining is a vital process for extracting valuable knowledge and intriguing patterns from existing databases, tailored for specific objectives. In the context of web usage mining, the increasing intricacies of web usage necessitate the development of more efficient techniques to predict user behaviour and enhance website performance. Traditional clustering approach ensures that related data points are grouped together while segregating unrelated ones. This study introduces the fuzzy clustering method, which involves categorizing a set of data points into multiple clusters. Fuzzy clustering is employed, allowing data points to possess varying degrees of membership in multiple clusters. Traditional data mining algorithms predominantly rely on binary values to identify transaction relationships. However, real-world applications frequently involve transactions with quantitative values, presenting a substantial challenge in algorithm design. Study presents a pioneering approach that leverages fuzzy clustering algorithms to predict the next website request made by users with a degree of uncertainty and vagueness. This integration of fuzziness results in more precise predictions of user behaviour. Methodologically, study enhances the conventional clustering and association rule mining method by incorporating fuzzy clustering base fuzzy association rule mining method. Paper presents a novel architecture based on fuzzy clustering client based on interest for fuzzy predictions, effectively addressing these challenges.