面向众包服务市场的用户需求挖掘方法的研究

发布时间：2018-05-08 02:14

本文选题：众包服务 + 模糊需求　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：近年来众包服务市场持续受到人们的关注,发展如火如荼,其发展模式也有了一定的变化。从传统的以服务需求方为主导的市场模式演变为以供求双方各自为主导的服务市场模式。本文以服务供求方为主导的新型众包市场为研究背景,分析用户的模糊需求,挖掘用户的需求模式,并根据用户的需求为用户匹配相应的服务,对众包服务市场中的服务进行衡量,缩小用户搜索的可选服务集,解决新型众包服务市场中供求双方信息不平衡以及需求匹配困难的问题,为下一步的精确推荐提供可靠地基础,具有重要意义。本文构建了表达服务供求关系的需求图谱。需求图谱是通过分析众包服务市场中的交易数据,提取市场中的用户,服务提供商和服务以及之间的关系来构建的,为面向用户的需求查询以及服务的匹配提供了研究基础。在众包服务市场中,用户对于服务的需求多是模糊的文字描述,传统的搜索仅能做到字符串的匹配无法真正了解用户的需求。本文提出了用户模糊需求的处理方法,分步采用局部敏感哈希算法、关键字匹配以及引导式问答方法来明确分析用户的多重查询需求,并将其转化为查询子图在构建的需求图谱上进行搜索,使得用户的多重查询需求能够一次性满足。文章提出了GSR4FR子图匹配算法和组合服务的筛选规则算法,并对服务的质量进行了统一的衡量。其中GSR4FR算法同改进前的Nema算法相比,搜索时间有了大幅度提升,平均服务搜索质量结果较好。组合服务筛选规则算法则是根据用户的历史购买行为信息的挖掘得到的,经实验结果证明行之有效。在新型的众包服务市场中,用户的需求是模糊而复杂的,因为人类的语言本身具有一定的模糊性,在此基础上进行的需求匹配的过程也是较为困难的。本文尝试着采用处理用户模糊需求和构建需求图谱的方法来进行用户的需求匹配,并提出了相对应的子图匹配算法来匹配合适的服务以及组合服务筛选规则算法来筛选组合服务的优先级别,在新的领域问题上融合了不同的解决方法,并取得了一定的成果。最后本文基于课题所研究的内容开发了基于用户需求图谱的服务匹配插件工具,以图形化界面的形式验证了研究内容。
[Abstract]:In recent years, the crowdsourcing service market has been paid more and more attention. From the traditional service demand-oriented market model to the service market model led by both supply and demand. Based on the research background of the new crowdsourcing market, which is dominated by service supply and demand, this paper analyzes the fuzzy needs of users, excavates the demand patterns of users, and matches the corresponding services for users according to the needs of users. To measure the services in the crowdsourcing service market, to reduce the optional service set searched by users, to solve the problems of the imbalance of information between the supply and demand sides and the difficulty of matching demand in the new crowdsourcing service market. It is of great significance to provide a reliable basis for the next accurate recommendation. This paper constructs a demand map to express the relationship between supply and demand of services. The requirement map is constructed by analyzing the transaction data in the crowdsourcing service market and extracting the relationship among the users, service providers and services in the market. It provides a research basis for the user-oriented requirement query and service matching. In the crowdsourcing service market, the user's demand for the service is mostly fuzzy text description, the traditional search can only achieve string matching can not really understand the user's needs. In this paper, we propose a method to deal with the user's fuzzy requirements. The local sensitive hashing algorithm, keyword matching and guided question answering method are used to analyze the multiple query requirements of the user step by step. The query subgraph is transformed into a query subgraph which can be searched on the constructed requirement graph, so that the multiple query requirements of the user can be satisfied at one time. In this paper, a GSR4FR subgraph matching algorithm and a selection rule algorithm for composite services are proposed, and the quality of service is evaluated uniformly. Compared with the improved Nema algorithm, the search time of GSR4FR algorithm is greatly improved, and the average quality of service search results are better. The algorithm of composite service filtering rule is based on the user's historical purchase behavior information mining, and the experimental results show that it is effective. In the new crowdsourcing service market, the needs of users are fuzzy and complex, because the human language itself has a certain degree of fuzziness, and the process of requirement matching on this basis is also more difficult. In this paper, we try to use the method of dealing with the user's fuzzy requirements and constructing the requirement map to match the user's requirements. The corresponding subgraph matching algorithm is proposed to match the appropriate service and the composite service filter rule algorithm to filter the priority level of composite service. Different solutions are combined in the new domain problem and some achievements have been achieved. Finally, a service matching plug-in tool based on user requirement graph is developed based on the content studied in this paper, and the research content is verified in the form of graphical interface.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3

【参考文献】