A Multi-relational Classifier for Imbalanced Database

• Main Authors: Tong-qin Wu, 吳同欽
• Other Authors: Chien-I Lee
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/31460568845673954639

碩士 === 國立臺南大學 === 資訊教育研究所碩士班 === 93 === Most of today，s structured data is stored in relational databases. Multiple relations in a relational database have to be connected via entity/relationship model. Multi-relational classification has been widely applied in many aspects, such as financial decision making, medical research etc.. Although a lot of Multi-relational Data Mining classifiers have been proposed, such as TILDE, FOIL, CrossMine etc., but in a situation that the datasets are imbalanced, these Multi-relational Data Mining classifiers are unable to accurately classify minority data (positive data). In this thesis, we propose a Multi-relational G-mean decision Tree algorithm, called Mr.G-Tree, to solve these problems mentioned above. Finally, as show in the experiment, Mr.G-Tree can accurately classify multi-relational imbalanced dataset.