Research On Community Content Classification Method Based On Machine Learning

With the development of the times,people are more and more willing to express their opinions and share their lives on the Internet.The popularity of smart phones makes it more and more convenient for people to post in communities or forums.However,the sharp increase of the number of posts not only causes the difficulty of website management,but also increases the difficulty of users to obtain resources.Therefore,how to classify the community content based on posts has become an essential function of the community system.According to the characteristics of community content,this paper presents a combined classification model.The whole model is composed of the following parts: first,aming at the problem of less features of posts as community content,this paper proposes a method based on word vector to obtain weighted word vector by using part of speech to enrich the features of the posts;secondly,for the problem that the single kernel function of the support vector machine cannot completely match all data distributions,the support vector machine is improved by using the mixed kernel function;finally,in order to make full use of the powerful feature extraction ability of convolutional neural network and the classification ability of support vector machine,the combined model is implemented by replacing softmax layer with support vector machine classifier.In the experimental part,this model is evaluated by using the dataset crawled from the real community website.The experimental results show that the classification accuracy of the combined model is about 1% higher than that of the convolutional neural network model and the classification results are more stable,which prove the correctness and feasibility of the combined model.