基于SOM神经网络的网络舆情信息分类模型

Classification Model of Internet Public Opinion Information Based on SOM Neural Network

研究了网络舆情信息分类方法,给出了几种分类算法的优缺点和适用性,重点分析了将SOM神经网络模型运用到网络舆情信息分类中的基本思想、网络架构、算法流程和局限性,提出了基于SOM为每个输出神经元增加一个阈值,避免出现死神经元。为输出神经元增加学习效率和邻近区域的改进方法,给出了改进后的算法流程,最后经过实验验证了改进算法的有效性,提高了网络舆情信息的查全率和查准率,为网络舆情信息分类建模提供了有益的解决方案。

This paper studied the classification method of internet public opinion information, and gave the advantages, disadvantages and applicability of several classification algorithms. The basic idea, the network architecture, algorithm flow and limitations of applying SOM neural network model to the classification of internet public opinion information was analyzed. An improved method for increasing the threshold of each output neuron based on SOM to avoid dead neurons was proposed, and an improved algorithm based on SOM for increasing learning efficiency and neighboring regions for each output neuron was proposed, and the improved algorithm flow was given. The experiment verified the effectiveness of the improved algorithm and improved the recall and precision of internet public opinion information, and provides a solution for the classification of internet public opinion information.