Getting Started - Java Data Mining: Strategy, Standard, and Practice - page 120

The table GIZMOS_RESULTS, depicted in Table 6-4, records which
customers in the trial campaign responded and placed an order for
Gizmos.

Table 6-4 Table GIZMOS_RESULTS from trial campaign

Customer_id Response

1 1

2 0

3 0

4 1

The SQL statement in the following code creates a view called
CUSTOMER_RESPONSE, which produces a data mining-ready table
based on the CUSTOMER and PURCHASES_V table.

CREATE VIEW CUSTOMER_RESPONSE as
SELECT id, name, address, city, state, zip, age, gender,
income, education, occupation, years_at_residence,
purchased_tads, purchased_zads, purchased_fads, household_size,
response
from CUSTOMER c, PURCHASES_V p, GIZMOS_RESULTS g
where c.id p.customer_id
and c.id g.customer_id

We are now ready to begin building models using this data.

6.4 Modeling

For modeling, we select the technique we will use for mining the
data. As noted above, DMWhizz will try several algorithms, each
possibly with different settings to obtain better results. However, to
get a model built quickly, and to see how well the data mining engine
(DME) can do using default settings, we leave the selection of algo-
rithm and any detailed settings to the DME. This can serve as a base-
line model against which we can measure other models.
The DMWhizz data miner will likely want to do some additional
data preparation once the initial results are obtained. For this initial
model building exercise, DMWhizz relies on the DME to prepare
the data automatically. Note that not all data mining tools support