Term
| a ___ model compares observations to be scored to a prototype |
|
Definition
|
|
Term
| name the steps for building a similarity model |
|
Definition
1. what distinguishes good records from bad ones
2. what would the ideal good record look like
3. how should distance from the ideal be measured |
|
|
Term
| the ___ coefficient measures the extent to which a change in one variable is related to a change in another |
|
Definition
|
|
Term
| What do you do if cells don't have enough data to support the required confidence level? |
|
Definition
reduce the number of splits along each dimension
reduce the number of dimensions used to define the sparse cells |
|
|
Term
|
Definition
|
|
Term
| RFM only works for ___ customers |
|
Definition
|
|
Term
| Before rolling out a new campaign, it should be tested on customers who receive the ___, under conditions that ___ the full campaign. |
|
Definition
|
|
Term
| Test campaigns should use ___ and ___ groups that include members of every cell |
|
Definition
|
|
Term
| What is the goal of incremental response modeling |
|
Definition
| identify persuadable prospects |
|
|
Term
| What do naive bayesian models accomplish? |
|
Definition
predict very specific probabilities.
I.E. Predict cancellation rate of customers in Wyoming, acquired through direct mail with credit class C, on rate plan..." |
|
|
Term
| R^2 is a number between ___ and ___ |
|
Definition
| 0 and 1. and a high R^2 value shows a good fit |
|
|
