|
| Background |
|
|
- ÄÄÇ»ÆÃ ÆÄ¿ö¿Í ¿î¿µ°è µ¥ÀÌÅÍÀÇ Áõ°¡
- ´ë¿ë·®ÀÇ ¿î¿µ°è µ¥ÀÌÅÍÀÇ È°¿ë¹æ¾È ¹®Á¦
- ºñÁî´Ï½º °á°úÀÇ ÀÌÇØ¸¦ À§ÇÑ µ¥ÀÌÅÍ ºÐ¼®
- Àû¿ëµÇ´Â ÈÆ·Ã±â¹ý : Åë°èºÐ¼®, ÆÐÅÏÀνÄ, machine learning
|
|
| Problem Formulation |
|
|
- ºñÁî´Ï½º ¸ñÀûÀ» setting Çϰí ÀûÀýÇÑ ºÐ¼®¹æ¹ýÀ» °áÁ¤
- Database marketing, credit scoring, fraud detection, healthcare informatics¸¦ À§ÇÑ Predictive modeling Àû¿ë
|
|
| Data Difficulties |
|
|
- Data ±¸Á¶ ¹× üÁ¦
- ¿¡·¯, ÀÌ»óÄ¡, °áÃøÄ¡¿¡ ´ëÇÑ Ã³¸®¹æ¹ý
- »ùÇøµ°ú ¿À¹ö»ùÇøµ
- Dimension Reduce (Áֿ亯¼ö ¼±ÅÃ)
|
|
Introduction to Enterprise Miner |
|
|
- Workspace Components Ž»ö
- Project Setting
- ºÐ¼® flow diagram ±¸Ãà
- Ãʱ⠵¥ÀÌÅÍ Å½»ö ½ÇÇà
- Variable Selection Techniques Àû¿ë
- °áÃøÄ¡ ó¸®
|
|
| Regression |
|
|
- target marketing ¿¹Á¦¸¦ ÅëÇÑ regression ½ÇÇà
- Stepwise method ½ÇÇà
|
|
| Neural Networks |
|
|
- Multilayer Perceptron ±¸Ãà
- Visualizing network complexity
- Performing stopped training
|
|
| Decision Trees |
|
|
- Credit ScoringÀ» ÀÌ¿ëÇÑ decision tree ±¸Ãà
- Decision Tree node ±â´É ÀÌÇØ
- Multiway SplitÀ» ÀÌ¿ëÇÑ decision tree ±¸Ãà
- Pruning & Assessing
|
|
Model Evaluation and Implementation |
|
|
- Èĺ¸¸ðµ¨ ºñ±³
- °£´ÜÇÑ ensemble ¸ðµ¨ ±¸Ãà
- Score code ÀÌ¿ë ¹× ±¸Ãà
|
|
| Cluster Analysis |
|
|
- ¿µ¾÷°ü·Ã µ¥ÀÌÅ͸¦ ÀÌ¿ëÇÑ cluster analysis ½ÇÇà
- K-means cluster analysis¸¦ À§ÇÑ Cluster node »ç¿ë
- Insight node¸¦ ÀÌ¿ëÇÑ cluster visualizing
|
|