Information extraction method of topic webpage based on multi-angle feature learning

Abstract

It's difficult to find topic information in the web page because it is slow to find specific information by labor in the process and the result of commonly used methods is inaccurate. This paper proposes a multi-angle feature analysis method for web information identifying. With this method, it mines the characteristics of web page information content in a comprehensive way. Focusing on the characteristics of the web page, the text is segmented, and features are extracted and quantified from multiple perspectives. The fully connected neural network deep learning model is used for training. Besides, use linear classifiers to classify web page. The final experiment shows that this method improves the F value by more than 4% compared with the keyword method and the SVM (Support Vector Machine) method.