微博信息分类研究
发布时间：2018-11-21 21:23
【摘要】：微博作为一种新兴的网络交互平台,包含着海量的信息。在这些以微博为载体的信息中蕴含着巨大的潜在的商机。微博作为社交网络平台,需要企业对其产品的使用者或者需求者做出及时的服务。对于微博内容中流露出的对企业产品的抱怨或者针对产品的疑问,企业有义务也有责任在第一时间及时的对产品使用者提供相关的技术支持服务；对于微博内容中表现出对企业产品的求购信息,企业也需要在第一时间及时的提供产品的相关信息,并做出相应的导购服务。 任何企业,面对海量的微博信息,如何及时的从数以亿计的微博信息中挖掘出企业需要的微博内容是当代企业获得第一手信息的关键。在获得与企业及其产品相关的微博内容的同时,对相关联的信息进行准确的情感倾向分类和做出是否需要提供导购服务或者产品技术支持服务是企业需要关注的一项重要任务。 本文将以联想集团作为企业背景,将微博文本根据是否与联想集团及其产品相关、微博文本中流露出的情感倾向性、是否需要对用户提供导购服务、是否需要对用户提供技术支持服务对微博进行分类。根据微博信息分类系统的功能性要求,设计系统的总体架构。根据微博信息分类系统需要实现的不同分类功能,设计并实现分类器,着重设计分类器的特征提取算法,并与简单特征提取算法进行对比试验。为了检测微博信息分类系统的功能及分类器的分类效果,将对分类器进行测试,以期到达可以进行微博信息分类的理想结果。 本文的最后部分,将对微博信息分类系统的分类器测试结果进行分析。在本文分类器使用的特征提取算法和简单特征提取算法进行对比试验的结果分析中,得出本文中使用的特征提取算法分类效果更好。在单独对分类器的分类效果进行分析时,会着重考虑特征的查全率、查准率和调和值,分析分类器的具体分类效果。在分析分类效果之后,会根据测试结果分析微博信息分类系统目前的不足和待改进的地方,并提出微博信息分类系统进一步的改进方案。最后,本文将是对微博信息分类系统的展望。
[Abstract]:Weibo, as a new network interactive platform, contains a great deal of information. In these Weibo as the carrier of information contains a huge potential business opportunities. Weibo, as a social network platform, needs enterprises to provide timely service to the users or demanders of their products. As for the complaints or questions about the products of the enterprises revealed in Weibo's contents, the enterprises have the obligation and the responsibility to provide the relevant technical support services to the users of the products in the first time. For Weibo's content to show the purchase information of the enterprise's products, the enterprise also needs to provide the relevant information of the product in the first time, and make the corresponding purchasing guide service. Faced with the massive Weibo information, it is the key for contemporary enterprises to obtain first-hand information from hundreds of millions of Weibo information. While obtaining Weibo's content related to enterprises and their products, it is an important task for enterprises to classify the relevant information accurately and to make whether or not they need to provide purchasing guidance services or product technical support services. This article will take the Lenovo Group as the enterprise background, according to whether Weibo text is related to the Lenovo Group and its products, and whether the emotional tendency revealed in Weibo's text, whether or not it needs to provide the purchasing guide service to the users. Whether need to provide the technical support service to the user to carry on the classification to Weibo. According to the functional requirements of Weibo information classification system, the overall structure of the system is designed. According to the different classification functions of Weibo information classification system, the classifier is designed and implemented. The feature extraction algorithm of the classifier is designed and compared with the simple feature extraction algorithm. In order to test the function of Weibo information classification system and the classification effect of the classifier, the classifier will be tested in order to reach the ideal result which can be used for the information classification of Weibo. In the last part of this paper, the test results of Weibo information classification system are analyzed. In the analysis of the result of comparison between the feature extraction algorithm and the simple feature extraction algorithm used in this paper, it is concluded that the feature extraction algorithm used in this paper has better classification effect. In the analysis of the classification effect of the classifier separately, the recall rate, precision rate and harmonic value of the feature will be considered emphatically, and the concrete classification effect of the classifier will be analyzed. After analyzing the classification effect, the paper will analyze the deficiency of Weibo information classification system and the place that need to be improved according to the test results, and put forward the further improvement scheme of Weibo information classification system. Finally, this paper will be the prospect of Weibo information classification system.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092;TP391.1
本文编号：2348289
[Abstract]:Weibo, as a new network interactive platform, contains a great deal of information. In these Weibo as the carrier of information contains a huge potential business opportunities. Weibo, as a social network platform, needs enterprises to provide timely service to the users or demanders of their products. As for the complaints or questions about the products of the enterprises revealed in Weibo's contents, the enterprises have the obligation and the responsibility to provide the relevant technical support services to the users of the products in the first time. For Weibo's content to show the purchase information of the enterprise's products, the enterprise also needs to provide the relevant information of the product in the first time, and make the corresponding purchasing guide service. Faced with the massive Weibo information, it is the key for contemporary enterprises to obtain first-hand information from hundreds of millions of Weibo information. While obtaining Weibo's content related to enterprises and their products, it is an important task for enterprises to classify the relevant information accurately and to make whether or not they need to provide purchasing guidance services or product technical support services. This article will take the Lenovo Group as the enterprise background, according to whether Weibo text is related to the Lenovo Group and its products, and whether the emotional tendency revealed in Weibo's text, whether or not it needs to provide the purchasing guide service to the users. Whether need to provide the technical support service to the user to carry on the classification to Weibo. According to the functional requirements of Weibo information classification system, the overall structure of the system is designed. According to the different classification functions of Weibo information classification system, the classifier is designed and implemented. The feature extraction algorithm of the classifier is designed and compared with the simple feature extraction algorithm. In order to test the function of Weibo information classification system and the classification effect of the classifier, the classifier will be tested in order to reach the ideal result which can be used for the information classification of Weibo. In the last part of this paper, the test results of Weibo information classification system are analyzed. In the analysis of the result of comparison between the feature extraction algorithm and the simple feature extraction algorithm used in this paper, it is concluded that the feature extraction algorithm used in this paper has better classification effect. In the analysis of the classification effect of the classifier separately, the recall rate, precision rate and harmonic value of the feature will be considered emphatically, and the concrete classification effect of the classifier will be analyzed. After analyzing the classification effect, the paper will analyze the deficiency of Weibo information classification system and the place that need to be improved according to the test results, and put forward the further improvement scheme of Weibo information classification system. Finally, this paper will be the prospect of Weibo information classification system.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.092;TP391.1
【参考文献】
相关期刊论文 前3条
1 丁国栋;白硕;王斌;;文本检索的统计语言建模方法综述[J];计算机研究与发展;2006年05期
2 李卫疆;赵铁军;王宪刚;;基于上下文的查询扩展[J];计算机研究与发展;2010年02期
3 朱嫣岚;闵锦;周雅倩;黄萱菁;吴立德;;基于HowNet的词汇语义倾向计算[J];中文信息学报;2006年01期
，本文编号：2348289
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2348289.html
