Prediction and Analysis of Next Website Request by Using Fuzzy Approach
Hardik A. Gangadwala, Ravi M. Gulati · 2023
Due to increasing amount of internet usage, mining useful information and knowledge from the proxy server log is evolving into a significant research area. Web usage mining is the method of extracting interesting patterns from Web usage log file. The browsing frequency of a user on each web site is used to analyze the behaviour. Earlier, large number of machine learning crisp methods were available for mining interesting patterns from proxy log transactions file. Earlier techniques depended on whether the website is present in the session or not. But proxy log transaction file contains numerical data. To deal with numerical data fuzzy data mining method is used for extraction of interesting patterns from numerical proxy log transaction file. In this paper, fuzzy frequent mining method is used to detect website browsing behavior from proxy log transactions file. The interesting patterns are mined out which exhibit the browsing behaviour and used to provide appropriate prediction. Web usage mining is a subfield of data mining that uses various data mining techniques to produce association rules. Data mining techniques are used to generate association rules from transaction data. Most of the time transactions are bool transactions, whereas Web usage data consists of quantitative values. To handle these real-world quantitative data, we used fuzzy data mining algorithm for extraction of association rules from quantitative Web log file. To generate fuzzy association rules first we designed a membership function. This membership function was used to transform numerical values into fuzzy terms. Experiments were carried out on different minimum support and minimum confidence. The experimental outcomes indicate that a shared pattern observer exists between the training and testing datasets, with different minimum support and minimum confidence levels.