Presentation is loading. Please wait.

Presentation is loading. Please wait.

MIS 451 Building Business Intelligence Systems Classification (1)

Similar presentations


Presentation on theme: "MIS 451 Building Business Intelligence Systems Classification (1)"— Presentation transcript:

1 MIS 451 Building Business Intelligence Systems Classification (1)

2 2 Problem Credit rating Credit card approval  Credit rating  Rules+ applicant’s profile  Rules are learned from old data  How to learn these rules (classification) Product purchasing prediction

3 3 Data Two types of attributes: Description attribute: attribute that describes an object, such as age, income level of a customer Class label attribute: attribute that identifies the class an object belongs to.

4 4 Data Description attributes Class label attribute

5 5 Model Age <30 >=30 LowIncome Low High Low High

6 6 Rules derived from a model If age<30 then credit rating Low If age>=30 and Income level=low then credit rating Low If age>=30 and Income level=high then credit rating High

7 7 Data

8 8 Classification Entropy: is a measurement of the diversity of a data set E = - When only two classes in a data set: E = - (p1 log 2 p1 + p2 log 2 p2) The bigger E is, the more diverse the data set is

9 9 Classification Divide and Conquer Pick an attribute to divide the data set with the most entropy reduction Stop until no attribute to pick or data in all leaf nodes are pure (I.e. belong to one class)


Download ppt "MIS 451 Building Business Intelligence Systems Classification (1)"

Similar presentations


Ads by Google