소개글
[Datamining, 데이터마이닝] Spam-mail Detection에 대한 자료입니다.
목차
1. Introduction
2. Exploring Data & Pre-Processing
3. Classification By Decision Tree
4.Classification By Artificial Neural Network
5.Classification By Bayesian Network
6. Classification By Support Vector Machine
7. Classification By Ensemble
8. Conclusion
본문내용
1 – 1. Decision Tree Using User-defined Algorithm(By SAS E-miner)
Splitting Condition : Entropy Reduction
Minimum Number of Observation : 46(1% of Data Set,
To avoid Overfitting & Underfitting)
Test Accuracy = 1266/1389 = 91.74%
Accuracy is increased little as Decision Tree used by C4.5
Because, Decision Tree is sufficiently good classifier.