Automatic recommendation of web pages and clustering unanimous web users in web usage mining

Abstract

The business organizations and educational institutions wish to use personalized applications that manage huge amount of information accessible through online by creating the user profiles. Previously, the personalized filtering and rating system are used for the generation of the user profile. This process becomes more tedious, as huge amount of information is updated every day on the web. The initial phase of the research work presents a Semantic Preference Distribution Mechanism (SPDM), a conceptual model for creating a user profile with the personalized search results and checking the semantic relationship between the topic concepts. The user profile is constructed based on ontological web search. The semantic correspondence between the groups of the elements is produced by using ontology. Then, SPDM mechanism adopts the user framework to personalize search results by re-sorting the results returned from a search engine for a given query. Finally, the user profile is created with the personalized search results. The second phase, an effective recommendation engine is formed using the Depth First Search (DFS) and Genetic Algorithm (GA). The DFS approach is applied for grouping the similar navigation patterns of the users into the clusters. The performance of the proposed GA is compared with the honey bee foraging approach. The experimental result shows that the proposed GA outperforms the existing bee foraging approach in terms of time consumption, Central Processing Unit (CPU) usage, memory consumption and accuracy. The newlinethird phase proposes a Locational-Social-Topical (LST) approach for predicting newlinethe opinion of the users based on the location, social and topical factors. The location context along with the social and topical factors influences the user emotions for better prediction of user opinion. newline newline