Context Recommendations in Document Streams using Collaborative Filtering by Combining User Web Search History

摘要

On Internet, large quantity of documents are created and viewed by huge number of users. Till now, most of the researcher's point of interest is topic modeling in document streams generated by users. But Sequential pattern mining from topics in document stream is ignored. In this work, we deal with the problem of mining the sequential patterns in document stream, which will be further used to detect the user behavior on internet. This will be useful in various applications such as real time user behavior monitoring system on internet. To improve the user behavior detection process, we will make use of the user web search history and combine with twitter tweet stream document data. This will improve the accuracy of user behavior detection process by user aware sequential pattern mining. This system enhanced with topic aware recommendation to users, by using collaborative filtering approach.