Web Personalization is defined as the process of customizing a website to the wants of specific users using their behavioral patterns and profiles collected in the Web context. This is used to improve the usability and user retention of a website. Due to the explosive growth of the Web and Web usage data, the area of Web personalization has become one of the key techniques for reducing the Web latency which has gained the momentum for improving the performance of web retrieval process. Some of the challenges involved are handling large amount of web usage data, extraction of interesting navigational patterns, building accurate user profiles, preserving users' privacy and predicting the user behavior. The objective of this research work is to design and develop a Web Personalization system that allows the users to get the content of their need from the Web site without specifying explicitly. In this thesis, we brought six significant contributions: Pre-processing, Extraction of Behavioral Patterns using Concepts as well as on Quiescent Sematic Analysis, Generation of User Profiles, Integration of User profiles & Behavioral patterns, and Probabilistic graphical models for Web Personalization. The first contribution focuses on the pre-processing of Web Usage Data (WUD), which has received a lesser attention from the Web Usage Mining (WUM) research group. A comprehensive pre-processing methodology has been proposed as a first and foremost requirement for Web Personalization Application. This involve cleaning of WUD, Users and Sessions Identification and Formatting of Data with summarization. Here, we tried to improve the quality of Data with substantial reduction in the quantity. Experiments have been conducted on WUD of academic (rnsit.ac.in) and research (NASA) websites in order to validate the efficacy of the proposed pre-processing methodology. The results show that size of the WUD has been reduced significantly by a rate of 72-83% of the initial size. Further, the cleaned WUD has been transformed into structured data for effective use in Web personalization application.