Since the early 1970s, decision support systems (DSS) have evolved significantly. In this paper, the design and implementation of MSMiner, a developing platform for DSS, is introduced. The system is constructed on a data warehouse and integrated with a number of data mining algorithms. It is well suited for on-line analytical processing (OLAP). The characteristics of MSMiner include the ability to support multiple data sources and data mining strategies, additional organizational flexibility in regard to data and mining strategies, and the powerful expansibility of data mining tasks.

With rapid expansion of information resources on the Internet increasingly,Web mining technology plays an important role.How to eliminate noisy information in web pages is a main part of information extraction based on Web mining.It is a preprocessing step in the Web mining.The result of Web mining lies on the step.In the paper,we firstly analyze the feature of noisy information.Then,based on our observation,using some extracting rules and statistic methods to eliminate noisy information and extract available information.

This paper introduces a branch of data mining technology——text categorization, and discusses some related key technologies of text categorization. Based on data preprocess, it implements one algorithm, K nearest neighbor algorithm. In the end, it utilizes the experiment results to prove the importance of data preprocess for clarifying capability.

In this thesis, the author uses the text-mining method of organizing all news formats datums to organize and generate special news topics. To solve the difficulties of manually categorizing news documents, the topic generation is related to the news event probing and tracking. Deeper researches on how to organize and manage generated news events, the detailed procedure of obtaining special events as well as the retrieval of news events have been done in this thesis.With this method , we can improve system efficiency and accuracy of organizing news topics.

在大量分析网络新闻特点的基础上 ,借鉴Single Pass聚类思想 ,并结合新闻要素给出了一种基于动态进化模型的事件探测和追踪算法 该动态模型是基于新闻事件的生存特点提出的 ,包括 :基于时间距离的相似度计算模型、事件模板进化策略以及动态阈值设置思想 该算法可以自动对新闻资料进行组织生成新闻专题 ,进而为用户提供个性化服务 实验结果验证了算法的有效性

New word and phrase discovery is basic research in fields of NLP, IR and MT. The paper gives the method based on probabilistic techniques and rules for new word discovery via analyzing the current techniques of phrase extraction and combining the specialties of Chinese. This method includes the bi-gram probabilistic model, the statistical algorithm, the rich rules and rule-based algorithm for word filtering. Experiments show that this technique is fit for automatically and effectually extracting new words/phrases from large corpora.

To help data warehouse getting high-quality data, data preprocess is needed. Extraction-Transformation-Loading(ETL)tools can finish this work. Complexity and usability are the primary problems concerning the ETL tools. To deal with these problems we provide a uniform architecture design for ETL processes which covers the aspects of metadata pertinent to ETL modeling, ETL transformation function design, modeling of ETL tasks, and the description language of ETL task model (XTDL). According to this idea of design, we developed an ETL tool named MSETL, aiming to provide the high-quality data for our multi-strategy data mining platform (MSMiner), which provides the friendly interface to manage the metadata of ETL processes, including login and deleting of ETL transformation functions, constructing and deleting the tasks, and browsing the result of execution of task.

Rough集理论是一种新型的处理含糊和不确定性知识的数学工具,将Rough集理论应用于知识发现中的聚类分析,给出了局部不可区分关系、个体之间的局部不可区分度和总不可区分度、类之间的不可区分度、聚类结果的综合近似精度等定义,在此基础上提出了一种基于Rough集的层次聚类算法,该算法能够自动调整参数,以寻求更优的聚类结果 .实验结果验证了该算法的可行性,特别是在符号属性聚类方面有较好的聚类性能 .

How to get text information from this huge information space becomes a more important problem with the rapid growth of the Internet. A semantic indexing system GHUNT based on concept space is proposed to solve the problem. Some new technologies are integrated in GHUNT to obtain good performance. GHUNT is an all-sided solution for information retrieval on the Internet.

元数据在数据仓库中扮演着非常重要的角色,对数据仓库的设计、开发、维护和管理具有非常重要的作用.在作者开发的多策略数据挖掘平台(MSMiner)中将元数据的使用范围从数据仓库拓展到包括ETL和数据挖掘任务等的整个系统,使之成为系统的核心和各个子系统之间联系的纽带.文中介绍了MSMiner的体系结构、元数据的主要内容、元数据管理实现的难点及采用的技术.

With the rapid growth of the Internet, how to get information from this huge information space becomes an even more important problem. In this paper, An Intelligence Chinese Document Semantic Indexing System; ICDSIS, is proposed. Some new technologies are integrated in ICDSIS to obtain good performance. ICDSIS is composed of four key procedures. A parallel, distributed and configurable Spider is used for information gather; a multi-hierarchy document classification approach combining the information gain initially processes gathered web documents; a swarm intelligence based document clustering method is used for information organization; a concept-based retrieval interface is applied for user interactive retrieval. ICDSIS is an all-sided solution for information retrieval on the Internet.

作为数据仓库的关键部件,支持数据抽取、清洗、转换和装载的工具集对任何数据仓库工程都是一个必不可少的成功因素.该文简单介绍了ETL技术,包括ETL的相关概念、ETL在数据仓库中的功能和重要地位以及现有的研究成果,然后重点介绍了ETL的具体设计和实现方法.