基于用户发布文本的用户领域分析.docxVIP

PAGE IV 基于用户发布文本的用户领域分析 摘 要 当今社会迅猛发展，网络世界异彩纷呈。众多的信息充斥着我们的眼睛，文本内容的可信度也就变得扑朔迷离。为了满足用户选择性的接收网络信息的需求，对文本发布者进行领域判定是直接的想法。因为用户期望自己看到的信息是发布者自身擅长的领域，所以本文就从用户领域判定为出发点，旨在辨识文本的可信度。与以往用户领域判定的研究不同的是，本文是完全基于用户发布的文本对其进行领域判定，条件约束越少，实用性就越大。先是人工标注用户的领域，用其发布的文本训练模型，然后预测未知用户的领域。通过大量的数据集实验，结果基本符合预期，可以根据用户发布的文本对其领域进行判定。在某种程度上，这对辨识文本可信度的研究起到了至关重要的推进作用。 关键词：文本分类 卡方检验 信息增益 支持向量机 梯度提升树 Abstract At present, the rapid development of society, the network world is colorful. Because a lot of text information is full of our eyes, the trustworthiness of the text becomes whirling. In order to meet the needs of users to selectively receive the information of the network, it is a direct idea to determine the domain of the text publisher. Because the information that the user expects to see is the information that the publisher is good at, so this paper starts from the judgment of the user domain, aiming at identifying the credibility of the text. Different from the previous research, this paper determines the users field based on the users text, the less the conditions, the more practical. First, tagging the users field, training model based on the published text, and then predicting the field of unknown users. The results are in line with expectations based on a large number of experiments, so it is feasible to determine the field according to the text published by users. To a certain extent, it plays a crucial role in the research of the credibility of the identification text. Keywords: Text categorization, Chi square test, Information gain, Support vector machine, Gradient boosted tree 目 录 摘要………………………………………………..…………………………………………………………………………Ⅰ ABSTRACT ………………………………………………………………………………………………………………….Ⅱ 第1章 绪论……………………………………………………………….……………………………………………….1 1.1 课题背景及研究的目的和意义………………………………………………………………...…….1 1.2国内外的研究现状与分析……………………………………………………………………………….1 1.3本文的研究方向概览……………………………………………………………………………………….2 1.4 本文的主要研究内容……………………………………………………………………………………….3 第2章 基于实验模型的语料库构建………………………………………………………………………….4 2.