| Summary: | 碩士 === 國立高雄應用科技大學 === 金融資訊研究所 === 96 === Data mining applies various algorithms to find useful information and forming knowledge in great dataset. This paper developed an data mining system containing various classification algorithms including ID3, C4.5, Bayesian method and Memory-Based Reasoning (MBR) method.
To test the usability of the system, this study experiments with two cases.
The first case used datum from a bank to classify the credibility class of accommodator. In this case, the classified accuracy achieved to 70 to 90 percent.
The second case applied the system to transaction datum from Taiwan stock market to find effectively classified rules for reaping profit. In this case, the accuracy of classification was approximately 70 percent.
|
|---|
