Intelligent Web Mining Technique using Sequential Pattern

As organizations expand and share more information about their operations online, the website data produced by these organizations becomes an invaluable resource for studying innovations. Effectively managing this vast volume of data and presenting relevant information to users is paramount. It is not practical to analyze and retrieve data manually from large databases. Addressing this challenge requires automated extraction tools enabling users to sift through billions of web pages and unearth pertinent information. This mechanism allows individuals and organizations to analyze data patterns within web contents and page structures, facilitating the discovery of valuable insights and knowledge. It aids in predicting user behavior during their online interactions, uncovering navigation patterns, and extracting useful information from user engagements, thereby enhancing our comprehension of consumer behavior. This paper focuses on extracting patterns of web access. Generally, a weblog can be seen as a series of user identifiers and event pairs. In this paper, web log files are segmented based on mining objectives. Preprocessing techniques are employed on the original web log files to extract segments. Each segment represents a sequence of events from a single user or session, arranged in ascending timestamp order. The model interprets these segments as event sequences and identifies sequential patterns exceeding a certain support threshold. This paper presents the mining of a sequential list of papers from the Neural Information Processing Systems (NIPS) website using the PrefixSpan algorithm. The system is implemented in Matlab programming language. Matlab programming language has been used in web mining to harvest useful data from the web, such as user logs and content. The system is tested and evaluated using accuracy and accessibility.