Model attributes and parameters and their values

When switching to the scoring model, the scoring model itself is displayed in the "Scoring model" section. Namely, all the attributes with the necessary data for each of them. To open an attribute with information on it, you must click on the line with the attribute.

See below for more details on each value in the attribute tables:
1. Attribute value – the value of the attribute that was opened, in which the table itself is located.
2. Count of Scores – the number of score that the client needs to add according to the model, if this value is present in the attribute during analysis.
The model itself is these 2 values. Attribute value and score. The rest of the data is displayed for general informativeness on the analysis of the file from which the scoring model was created.
3. Count of Good – the number of good clients in the file to create a model with this attribute value.
4. Count of Bad – the number of bad clients in the file to create a model with this attribute value.
5. Rate of Good, % - proportion of good clients in the file for creating a model with this attribute value from the total number of rows with this attribute value.
6. Rate of Bad, % - proportion of bad clients in the file for creating a model with this attribute value from the total number of lines with this attribute value.
7. Count of Total – total number of clients in the file to create a model with this attribute value.
8. Population Rate of Good, % - proportion of good clients with this attribute value from the total number of good ones in %.
9. Population Rate of Bad, % - proportion of bad clients with this attribute value from the total number of bad clients in %.
10. Population Rate of Total, % - proportion of the total number of clients with this attribute value from the total number in the file in %.
11. Rate of Good from total good in Attribute – proportion of good clients with this attribute value from the total number of good ones.
12. Rate of Bad from total bad in Attribute – proportion of bad clients with this attribute value out of the total number of bad clients.
13. Measure of Probability of good – the degree of probability that a client with this attribute value will be good.
14. Weight of Evidence – how significant is the value of the attribute in the general model among all clients in the file.
15. Information Value – what informational value this attribute and its individual values carry.
To understand whether it is worth selecting an attribute and its values into the model, it is necessary to focus on the informational value. If the informational value of an attribute ends up being above 3%, that's already very good. But, if it so happened that for some attribute the information value became close to 100% or even more, then you should know that this is almost impossible, and it is necessary to double-check the sample. An informational value close to 100% should always be suspicious, it might even be better to exclude the attribute completely.