Analysis of Web Mining Patterns Using Custom-Built Apriori Algorithm

Analysis of Web Mining Patterns Using Custom-Built Apriori Algorithm

The dissertation entitled Analysis of Web Mining Patterns Using Custom-Built Apriori Algorithm has developed a custom-built apriori algorithm for the discovery of association rules in web log data. Web server log containing the information about all the web requests to the Christ university website is used for analysis. The methodology adapted by this research is a four step process, containing data preprocessing, frequent pattern discovery, analysis and developing a tool for implementing web mining. The custom built apriori takes the preprocessed weblog file as input and generates the frequent folders and the relationship among them. This thesis has also developed a tool written in java for this web usage mining process. The tool assists the user to execute the custom built apriori algorithm and to view the associations among folders based on the given support and confidence values to the tool. The web is a highly dynamic information source. Most of the organisations put information on the web because they want it to be seen by the world. Now a days the web is well beefed up with more information in an unstructured fashion. As the web and its usage continue to grow, there is an opportunity to analyse web data and extract useful knowledge from it. The objective of this research is to predict the user behaviour in interacting with the website that helps the website designer in improving the quality of website. The dissertation is organised into 5 chapters. Chapter1, Introduction starts with a brief overview of web mining and presents the objective of the study and the problem statement. Chapter 2, Literature review, discusses background work in the field of web mining and pattern discovery. Chapter 3, methodology elaborately discusses the process used for analysing the web patterns. Chapter 4 is dedicated for results and discussion. Chapter5, conclusion, summarises the inferences concluded based on the results obtained. The chapter also discusses the limitations and challenges and concludes with future scope of the study. KeyWords: Web Mining, Preprocessing, Web Server log, Frequent pattern

V Lateefa

Computer Science