AN EFFICIENT WEB LEARNING WEB TEXT FEATURE EXTRACTION WITH EXPONENTIAL PARTICLE SWARM OPTIMIZATION
R. Dhanya, Sri Ramalinga · 2014
Due to the growth of World Wide Web the tradition of web patterns also enhanced now a days, extraction of information from web also important. For this reason web mining plays important role to discovery of individual user information and extract information from individual web log files with known text feature. Due to its extensive division, its directness and elevated dynamics, the resources occurrence the web are significantly sprinkled and they comprise no incorporated administration and arrangement. It seriously decreases the effectiveness by means of web information. Because finding text feature most important imperative problem in web mining. To conquer this problem proposed an efficient method to extract the log data to learn user profiles using Back propogation Neural Network (BPNN) and extract web text feature using Exponential Particle Swarm Optimization (EPSO). The proposed representation uses a Back propogation Neural Network (BPNN) structural design with a back propagation knowledge method to determine and investigate helpful information from the obtainable Web log data file then feature extraction is performed. Second the Web transcript feature Extraction procedure is wished-for best feature extraction and finally compares the results. In the present effort develop the best knowledge ability consequence for web log data and decrease the computation strength of an aggressive knowledge BPNN and the EPSO algorithm. Experimental results shows that the enhanced BPNN and EPSO schema extract best text features for web log information for every user.