The study on Web product reviews mining based on an improved text categorization algorithm

With the rapid development of the Internet and e-business, more and more netizens tend to publish comments of the various products on the net. However, to obtain information from these data will face some problems: huge, emir-structured and non-structured data. In this regard, performing data mining is particularly important. This paper puts forward to a new method of classifying these reviews, and then introduces the mining process of the product comments based on the text classification algorithm; finally, verifying this new method's feasibility and accuracy with two products of IT168 website's comments. The example analysis results show that the method is practical, and also get good classification results.