This is an automated email from the ASF dual-hosted git repository.
nkak pushed a commit to branch madlib2-master
in repository https://gitbox.apache.org/repos/asf/madlib.git
commit fa57c4fb40ce71f104449ec8356a3e0d53f97426
Author: Nikhil Kak <n...@vmware.com>
AuthorDate: Tue Feb 20 17:10:46 2024 -0800
PMML: Improve dev-check tests for random forest
JIRA: MADLIB-1517
This commit adds a few more random forest pmml tests that compare
forest_predict's output with pypmml's output
---
.../postgres/modules/pmml/test/pmml_rf.sql_in | 72 ++++++++++++++++++++--
1 file changed, 68 insertions(+), 4 deletions(-)
diff --git a/src/ports/postgres/modules/pmml/test/pmml_rf.sql_in
b/src/ports/postgres/modules/pmml/test/pmml_rf.sql_in
index b91db5d4..805b977a 100644
--- a/src/ports/postgres/modules/pmml/test/pmml_rf.sql_in
+++ b/src/ports/postgres/modules/pmml/test/pmml_rf.sql_in
@@ -1,3 +1,10 @@
+\i m4_regexp(MADLIB_LIBRARY_PATH,
+ `\(.*\)/lib',
+ `\1/../modules/pmml/test/pmml.setup.sql_in'
+)
+
+m4_changequote(`<!'', `!>'')
+
DROP TABLE IF EXISTS dt_golf;
CREATE TABLE dt_golf (
id integer NOT NULL,
@@ -6,7 +13,7 @@ CREATE TABLE dt_golf (
humidity double precision,
windy text,
class text
-) ;
+);
INSERT INTO dt_golf (id,"OUTLOOK",temperature,humidity,windy,class) VALUES
(1, 'sunny', 85, 85, 'false', 'Don''t Play'),
@@ -33,7 +40,7 @@ SELECT forest_train('dt_golf'::text, -- source table
'humidity, windy'::text, -- features
NULL::text, -- exclude columns
'class'::text, -- grouping
- 5::integer, -- num_trees
+ 3::integer, -- num_trees
1::integer, -- num_random_features
FALSE, -- importance
1::integer, -- num_permutations
@@ -43,9 +50,51 @@ SELECT forest_train('dt_golf'::text, -- source table
3::integer -- number of bins per continuous
variable
);
-SELECT pmml('train_output');
+DROP TABLE IF EXISTS forest_predict_output;
+SELECT forest_predict('train_output',
+ 'dt_golf',
+ 'forest_predict_output',
+ 'response');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_temperature::double
precision','predicted_(temperature::double precision)_pmml_prediction');
+
-------------------------------------------------------------------------
+-- classification, no grouping
+DROP TABLE IF EXISTS train_output, train_output_summary, train_output_group;
+SELECT forest_train('dt_golf'::text, -- source table
+ 'train_output'::text, -- output model table
+ 'id'::text, -- id column
+ '"OUTLOOK"'::text, -- response
+ 'humidity, windy'::text, -- features
+ NULL::text, -- exclude columns
+ NULL, -- grouping
+ 5::integer, -- num_trees
+ 1::integer, -- num_random_features
+ FALSE, -- importance
+ 1::integer, -- num_permutations
+ 5::integer, -- max depth
+ 1::integer, -- min split
+ 1::integer, -- min bucket
+ 3::integer -- number of bins per continuous
variable
+ );
+
+DROP TABLE IF EXISTS forest_predict_output;
+SELECT forest_predict('train_output',
+ 'dt_golf',
+ 'forest_predict_output',
+ 'response');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id',
'estimated_OUTLOOK','predicted_OUTLOOK_pmml_prediction');
+
+DROP TABLE IF EXISTS forest_predict_output;
+SELECT forest_predict('train_output',
+ 'dt_golf',
+ 'forest_predict_output',
+ 'prob');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_prob_overcast','probability_overcast');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_prob_rain','probability_rain');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_prob_sunny','probability_sunny');
+
+
-- classification, grouping
DROP TABLE IF EXISTS train_output, train_output_summary, train_output_group;
SELECT forest_train('dt_golf'::text, -- source table
@@ -65,6 +114,21 @@ SELECT forest_train('dt_golf'::text, -- source table
3::integer -- number of bins per continuous
variable
);
-SELECT pmml('train_output');
+DROP TABLE IF EXISTS forest_predict_output;
+SELECT forest_predict('train_output',
+ 'dt_golf',
+ 'forest_predict_output',
+ 'response');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id',
'estimated_OUTLOOK','predicted_OUTLOOK_pmml_prediction');
+
+DROP TABLE IF EXISTS forest_predict_output;
+SELECT forest_predict('train_output',
+ 'dt_golf',
+ 'forest_predict_output',
+ 'prob');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_prob_overcast','probability_overcast');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_prob_rain','probability_rain');
+SELECT test_pmml_output('dt_golf', 'train_output',
'forest_predict_output','id', 'estimated_prob_sunny','probability_sunny');
+
-------------------------------------------------------------------------
