一种神经网络文本分类器的设计与实现

论文着重介绍了一种基于神经网络的文本分类器,分类器使用神经网络作为分类工具,特征词的词频组成原始特征向量,和神经网络输入层的神经元一一对应。并引入了信息检索中的常用技术——潜在语义索引,训练过程中结合遗传算法,优化神经网络的初始权值。最后对分类器进行了开放性测试,实验表明分类器对文本分类具有较高的平均查全率和平均精度。

A Text Classifier's Design and Realization Based on Neural Networks

This paper presents a text classifier based on neural networks as the main topic.The authors have used Artificial Neural Networks(ANN) as the classifier.The recorded term frequencies form the original feature vector,matching with neurons in the input layer of ANN one by one.This classifier imports a traditional technique called Latent Semantic Indexing(LSI).This classifier employs Genetic Algorithm(GA) in the stage of training to optimize initial weights of ANN.Finally an open test is done on this classifier.As experiment results show,this classifier can reach both high precision and high recall on average.