Dissertation > Excellent graduate degree dissertation topics show

Construction of Interval Databases and Its Applications in Knowledge Discovery

Author: YinYunFei
Tutor: ZhangShiChao；YanXiaoWei
School: Guangxi Normal University
Course: Computer Software and Theory
Keywords: Interval value Association rules Knowledge Discovery Interval Clustering Bundle of goods
CLC: TP311.13
Type: Master's thesis
Year: 2005
Downloads: 91
Quote: 0
Read: Download Dissertation

Abstract

Association rule mining is an important research topic in data mining. It is the search for a strong collection of items related to a process. Digging a supermarket database, you can find the sales links between different commodities (which reflects consumer behavior), for example: bread and milk, coffee and sugar, toothpaste and toothbrush usually sales. These are common-sense knowledge. Interestingly, the association rule mining can find, like beer and diapers \This led to the association rule mining depth research and a wide range of applications. For example, it is further used to solve the inventory control (stock control), merchandising (Sales promotion), consumer behavior analysis (Customer behavior analysis). With the development of the supermarket and daily necessities industry the bundling (Binding sale) - bundled commodity (Binding commodities) sales have become an important tool for the convenience of our customers and improve profits. This is exactly the mining association rules useless. This thesis is thorough and meticulous study of this problem and proposed mining interval-valued rules: A → [B, C] ideas and methods. Bundled goods with the range of values ??(Interval values) to indicate that there are many advantages. First, the the interval value contains more information than a single specific data. Because a single data provides only a single data itself, and interval value provided by a distributed, i.e., it can take any number in the interval. Followed by the interval-valued more skills than the average, that is the interval value information entropy (Interval entropy) is greater than the average number of information entropy (Mean entropy). Furthermore, the interval-valued database mining which commodities can be found suitable for bundling, which commodities are not suitable for bundling. This has important practical value. Based on interval values ??clustering algorithm, the paper proposed two fields of traditional relational database as a new field, and one to represent the new field left endpoint domain \another new field right end point domain (the right end of the interval-valued points), thus forming the interval-valued database. Mining algorithm, the paper in-depth study of the strong association rules (relatives association rules) given the strong association rules interval function formula; interval functions value research foundation to build a complete range of grid systems, and the use of a comprehensive interval grid to meet one of nature: A ∧ C = B ∧ C and A ∨ C = B ∨ C? A = B bundle of commodities. The essence of the interval-valued association rule mining mining bundled commodities, that is, research which goods should be bundled. The main work of this paper is divided into the following four parts: (1) there are many modes omissions traditional data mining, and of great significance to study the pattern of these omissions are discussed from the perspective of physics, mathematics, biology. (2) to build a new database structure for these omissions mode to store and handle them, this new type of database called interval-valued database. (3) proposed the concept of interval-valued association rules, and in-depth study of the true meaning of the interval value rules. (4) interval valued rule mining algorithm. Finally, the paper work is summarized, and pointed out the direction for future improvements.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer software > Program design,software engineering > Programming > Database theory and systems
© 2012 www.DissertationTopic.Net Mobile