| This research has investigated and developed a method to automatically classify web pages based on task. As a part of a large research project, this phase requires us to choose three different search tasks: Health, Shopping and Education, and to classify web pages to these three kinds of tasks. We introduce information gain for the feature selection to reduce the dimension of document vector. PCA is then used to map these dimensions into a smaller dimensional space and to project web documents into the new space for the purpose of classification. The results of successive experiments show that our method is able to classify web pages efficiently for these three tasks. |
