基于流程的数据挖掘生成工具的设计与实现

发布时间：2018-11-03 08:42

【摘要】：随着数据挖掘知识的普及,数据挖掘已经在社会各行各业的应用迅速扩展开来。通俗来讲,数据挖掘就是从大量的数据中,提取人们感兴趣的,即正确的、未知的、有潜在价值的并最终能被用户接收和利用的知识。数据挖掘已经广泛地应用于各个领域,越来越多的学者开始关注于数据挖掘领域。随着数据挖掘相关知识在越来越多的领域开始普及,其相关的实用工具也开始受到人们的重视。数据挖掘的研究重点也慢慢向数据挖掘系统工具方向转移,越来越注重多种多样的发现策略和技术的集成。数据挖掘工具的用户主要集中在一些传统的行业中,比如银行、制造业、销售等面临转型的行业,数据挖掘工具也在不断的产生和完善。基于上述的场景和需求,本文研究并设计了一套基于流程的数据挖掘建模工具。该工具提供了一种基于web服务的流程建模方式,旨在为流程开发者提供一个统一的建模平台,减少开发者的工作量,满足不同开发者的动态需求。基于流程的数据挖掘建模方式借助于web服务,通过将第三方服务进行组装形成新的流程来满足开发者的任务需求。这种开发方式的优点在于开发者不需要过多地了解数据挖掘的相关知识,完全屏蔽了底层数据挖掘服务实现的细节,只需要调用相关的基于数据挖掘的web服务,调用服务对外提供的接口,对这些服务进行组装,就能实现数据挖掘流程的搭建。该工具可显著提升建模效率,对IT人员也可以大大减轻原有的IT支持代价,提升调度和数据运作效率。开发者可以自定义建模过程中每个步骤的细节,并能够实时跟踪流程的执行情况。为了进一步减少开发人员的工作量,本系统为开发者提供了用于流程搭建的服务库。将数据挖掘流程中常用到的数据预处理、挖掘算法、算法评估等任务进行封装,以服务模板的形式提供给开发人员调用。借助于系统提供的服务库,开发人员可以完成从数据输入、数据拼接、特征抽取、模型训练、模型评估、模型应用完整的建模过程,显著提升流程建模效率。开发人员在搭建数据挖掘流程时,可以选择导入第三方服务,也可以调用系统提供服务,极大地增加了流程开发的灵活性。本文首先阐述了数据挖掘建模系统的背景和意义,并研究系统所设计的相关技术和领域。接着从本文的整体目标出发,分析了数据挖掘建模系统的功能性需求,根据需求对系统的总体架构和模块间的交互接口进行了设计,针对核心模块的实现进行详细介绍。最后对系统进行了测试,以验证系统的合理性。
[Abstract]:With the popularization of data mining knowledge, data mining has been widely used in various industries. Generally speaking, data mining is to extract the knowledge that people are interested in, that is, correct, unknown, potentially valuable, and can be received and utilized by users from a large amount of data. Data mining has been widely used in various fields, more and more scholars begin to pay attention to the field of data mining. With the popularization of data mining related knowledge in more and more fields, the practical tools of data mining have been paid more and more attention. The research focus of data mining is gradually shifting to the tool of data mining system, and more attention is paid to the integration of various discovery strategies and technologies. Data mining tools are mainly concentrated in some traditional industries, such as banking, manufacturing, sales and other industries facing transformation, data mining tools are constantly produced and improved. Based on the above scenarios and requirements, this paper studies and designs a data mining modeling tool based on process. The tool provides a process modeling method based on web services, which aims to provide a unified modeling platform for process developers, reduce the workload of developers and meet the dynamic needs of different developers. With the help of web services, the modeling method of process-based data mining can meet the task requirements of developers by assembling third-party services into a new process. The advantage of this method is that the developer does not need to know too much about data mining, and completely hides the details of the implementation of the underlying data mining services, and only needs to call the related web services based on data mining. The data mining process can be constructed by calling the interface provided by the service and assembling these services. This tool can significantly improve the modeling efficiency, and can greatly reduce the original IT support cost for IT personnel, and improve the scheduling and data operation efficiency. Developers can customize the details of each step of the modeling process and track the execution of the process in real time. In order to further reduce the workload of developers, the system provides developers with a service library for process building. The tasks of data preprocessing, mining algorithm and algorithm evaluation are encapsulated in the process of data mining, which is provided to the developer in the form of service template. With the help of the service library provided by the system, the developer can complete the modeling process from data input, data splicing, feature extraction, model training, model evaluation, model application, and significantly improve the efficiency of process modeling. When setting up the data mining process, the developer can choose to import the third-party service or call the system to provide the service, which greatly increases the flexibility of the process development. In this paper, the background and significance of data mining modeling system are introduced, and the related technologies and fields are studied. Then, the functional requirements of the data mining modeling system are analyzed from the overall goal of this paper. According to the requirements, the overall architecture of the system and the interaction interface between modules are designed, and the implementation of the core modules is introduced in detail. Finally, the system is tested to verify the rationality of the system.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP311.13;TP311.52

【参考文献】