摘要
网页分类可将信息准确筛选与呈现给用户,提高信息检索的准确率。深度学习是机器学习中一个全新的领域,其本质是一种多层的神经网络学习算法,通过逐层初始化的方法来达到极高的准确率,被多次使用在图像识别、语音识别、文本分类中。提出了基于深度学习的网页分类算法,实验数据证明该方法可有效提高网页分类的准确率。
Ahstract：Webpage classification can be used to select accurate webpage for users, which improves the accuracy of information retrieval. Deep learning is a new field in machine learning world. It＇s a multi-layer neural network learning algorithm, which achieves a very high accuracy by initializing the layer by layer. It has been used in image recognition, speech recognition and text classification. This paper uses the deep learning algorithm in webpage classification. With the experiments, it finds out that the deep learning has obvious advantages for webpage classification. Key words： Webpage Classification; Deep Learning; Stacked Auto Encoder; TFIDF
出处
《微型电脑应用》
2016年第2期25-28,共4页
Microcomputer Applications
基金
上海市基础研究重大重点项目(NO.13JC1403500)
