数据挖掘 Traning and testing Tanagra.pptx

Training and testing data Tanagra;;Decision trees and Tanagra software I (concepts of training data, testing data Building and testing DM models Environment when the DM model is applied and tested in real world situation Stock trading examples. Splitting available data on training and testing sets equal, random, n%- folding, proportional to amount of data in each class). The number of alternatives and how minimize the number of subsets to be analyzed.?;Running?in class three different classification programs from Tanagra on the same test data. Comparing?confusion matrixes of these three runs. Write?a short report on comparison ?? Optional: Propose a way to improve?results??presented?in the confusion matrixes by using tools within Tanagra.? ;Review: Define DM components;Supervised learning;Review lecture 2What is the difference between DB and DM?;Review: What is Online analytical mining (OLAM)?;Review: What is supervised attribute-based data mining model ?;Review: Define DMC and DMP;Review: What is a case set?;Customer case : Example;1. Predictive Modeling (Classification/Regression);2. Segmentation (Clustering);3. Association (Data Summarization);4. Sequence and Deviation Analysis;4. Sequence, Example: DB ;Example: Discoveries on a given Web site;Deviation analysis: Discovering anomalies in data;5. Dependency Modeling