Finding the right dataset
We need a good historical dataset to build our model. We will mine this dataset to build our model. To continue with the example to our fictitious company, Furnitica, we will use historical campaign response data from a previous campaign run by Furnitica.
This is synthetic data, which means it has been synthesized using a random data generation algorithm. A few sample rows in our dataset are presented in Table 2:
|
Age |
Income |
Gender |
Folder |
Response |
|
61 |
30974 |
0 |
1 |
0 |
|
42 |
38260 |
0 |
3 |
0 |
|
40 |
20135 |
0 |
4 |
0 |
|
88 |
30645 |
0 |
5 |
0 |
|
58 |
38078 |
1 |
3 |
0 |
|
73 |
20445 |
0 |
4 |
0 |
|
34 |
66198 |
0 |
3 |
0 |
|
65 |
48657 |
0 |
2 |
0 |
|
68 |
39309 |
0 |
1 |
0 |
Table 2 Sample credit card approval data
This dataset is generated from the response data of a campaign after joining it with the data in the customer relationship management system of the company. In the dataset preparation step, we have removed several features which are not directly useful in building our model. As a result, our...
