The Implementation Of An Data Mining Visualization System Based On Association Rules

The explosion of information nowadays has brought the heavy demand of powerful data analyzers that can help the decision-makers get useful information hidden below data. Data miners will make great contributions to various fields like business, repository, science and medicine etc by analyzing data and discovering important patterns of data. The gulf between data and knowledge calls for powerful data analyzers, with which we can turn data-tombs into gold knowledge.As a main application field of data mining, retailing has stored huge amount of data on sales, customer shopping history etc. And the amount of data is still rising, especially with the popularity and convenience of web shopping andelectronic business.In the thesis, a data mining visualization system has been implemented for the discovery of association rules in a web transaction database. The system highly incorporates data source selection module, data pretreatment module, data exploration module, data mining module and knowledge representation module, and makes data pretreatment by taking the advantage of some tools of relational database like query, and data transformation. The platform is implemented with Visual Basic that provides a friendly interactive interface to the user, while all the transaction data is stored in the relational database. The system completely proves the data miner feasible, valuable and effective by setting an example of a web transaction data, and embeds some visualization techniques into it to realize some interesting investigation on data visualization and knowledge representation. The most important function of this system is to analysis the items in the transaction database, and to find association rules among the items on sale, then represents them in a simple and intuitionistic way to help decisions-making.