期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2013
卷号：57
期号：3
出版社：Journal of Theoretical and Applied
摘要：Now a day, Millions of web servers are available to give a huge amount of electronic content to the end users using the internet. Searching the content relevant to the need of a user is a challenging task because of the ambiguity in our natural language. Classification of web pages based on their contents is useful to the search engines to give appropriate and desired data to the user. In this paper, an optimized approach is used for classifications of web pages. Feature extraction and selection of best features play a key role in classification. In this work, features are extracted from the ontology representation of content and Inverse Document Frequency (IDF). Then the best features are selected by using genetic algorithm. Using the selected features by GA, an Artificial Neural Network (ANN) is trained to classify the web pages. Results were compared with the classification methods based on neural networks with IDF based feature extraction, neural networks with ontology based feature extraction, neural networks with combined IDF and ontology based feature selection and other three methods are based on applying genetic algorithm to select the best features to classify the web pages. The parameters such as a percentage of accuracy, precision, recall and root mean square error are considered for performance evaluation. Numerical results showed that hybrid classifier trained by multilayer neural network with GA for selecting IDF and ontology based features gave 93% of accuracy, high precision and recall and lowest RMSE when comparing to all other methods.
