We looked at the effectiveness of supervised learning models in solving the real-world problem of web page classification. We attempted to maximize the performance of the classification model through the use of various feature sets. We observed that we were able to attain the highest accuracy by using a large training set, with a binary representation of terms, trained using the Support Vector Machine model (with the polynomial kernel filter applied). See for yourself!

PDF download: Analyzing the Impact of Web Page Textual Data Representation on the Accuracy of Selected Supervised Learning Models (0)