Examples | Gain Ratio Extreme Decision Tree Scoring | Vantage Analytics Library - Examples - Vantage Analytics Library

Decision Tree that Subsequent Examples Score

call td_analyze (
  'decisiontree',
  'database = val_source;
   tablename = customer_analysis;
   columns = income, age, nbr_children;
   dependent = gender;
   min_records = 2;
   max_depth = 5;
   binning = false;
   algorithm = gainratio;
   pruning = gainratio;
   outputdatabase = val_results;
   outputtablename = cust_analysis_dt;
   operatordatabase = val_user;'
);

samplescoresize Parameter

call td_analyze (
  'decisiontreescore',
  'samplescoresize = 10;
   database = val_source;
   tablename = customer_analysis;
   modeldatabase = val_results;
   modeltablename = cust_analysis_dt;
   outputdatabase = val_results;
   outputtablename = cust_analysis_dt_score;
   retain = city_name, state_code;'
);

Produce Two Profile Tables

The two profile tables capture decision rules for each customer or prediction.

call td_analyze (
  'decisiontreescore',
  'database = val_source;
   tablename = customer_analysis;
   modeldatabase = val_results;
   modeltablename = cust_analysis_dt;
   outputdatabase = val_results;
   outputtablename = cust_analysis_dt_score;
   retain = city_name, state_code;
   profiletables = true;'
);

includeconfidence Parameter

The output table contains a column indicating how likely it is, for a particular leaf node in the tree, that the prediction is correct.

call td_analyze (
  'decisiontreescore',
  'database = val_source;
   tablename = customer_analysis;
   modeldatabase = val_results;
   modeltablename = cust_analysis_dt;
   outputdatabase = val_results;
   outputtablename = cust_analysis_dt_score_conf;
   retain = city_name, state_code;
   includeconfidence = true;'
);

targetedvalue Parameter

The output table contains a column indicating how likely it is, for a particular leaf node and binary targeted value, that the prediction is correct.

call td_analyze (
  'decisiontreescore',
  'database = val_source;
   tablename = customer_analysis;
   modeldatabase = val_results;
   modeltablename = cust_analysis_dt;
   outputdatabase = val_results;
   outputtablename = cust_analysis_dt_score_target;
   retain = city_name, state_code;
   targetedvalue = F;'
);

Confusion Matrix

A Confusion Matrix is an evaluation function.

call td_analyze (
  'decisiontreescore',
  'database = val_source;
   tablename = customer_analysis;
   modeldatabase = val_results;
   modeltablename = cust_analysis_dt_m;
   outputdatabase = val_results;
   outputtablename = cust_analysis_dt_score_m;
   retain = city_name, state_code;
   scoringmethod = scoreandevaluate;'
);