Some columns, such as tot_income, ck_avg_bal, and transaction count columns, have values in different ranges. For optimization algorithms like gradient descent, it is important to normalize values to the same scale for faster convergence, scale consistency, and improved model performance.
We will use the TD_ScaleFit function to normalize values across different scales.
CREATE VIEW td_analytics_functions_demo.scale_fit_joined_table_input AS (
SELECT * FROM TD_ScaleFit(
ON td_analytics_functions_demo.joined_table AS InputTable
USING
TargetColumns('tot_income', 'q1_trans_cnt', 'q2_trans_cnt', 'q3_trans_cnt', 'q4_trans_cnt', 'ck_avg_bal', 'sv_avg_bal', 'ck_avg_tran_amt', 'sv_avg_tran_amt', 'cc_avg_tran_amt')
ScaleMethod('RANGE')
) AS dt
);