NEURAL NETWORK CLASSIFICATION ALGORITHMS FOR WEB USAGE MINING AND PROPOSED SOLUTIONS FOR HUGE WEB DATA CLASSIFICATION

Abstract

Nowadays, a huge amount of data is present on web, so to extract useful knowledge and to manage those
huge files will become mandatory to obtain fruitful business analysis results. To extract useful knowledge from World
Wide Web ( WWW) is known as web mining. Web usage mining has emerging trends on network traffic control and flow
analysis, website management, personalization, etc. Neural network have capability of self organization and is also
matched with ant colony behavior and adaptive learning. Such concept is used for information retrieval from huge web
data. It is also used for complex classification, optimization and distributed control problems [1].With the help of Neural
Network algorithms for classification of web log data; it produces the best result of classification. So, in this paper we
have introduced solutions for self-organizing and growing network which helps in information retrieval from huge web
data and also discussed various neural network algorithms i.e. GNG (Growing Natural Gas), ART(Adaptive Resonance
Theory) model, LVQ(Learning Vector Quantization) and its series. Input for neural algorithms is web log files and
expected outcome would be optimal representation of network that is further used for Information extraction in web
usage mining. Such trained network is used for classification which gives effective classification of data.