[MADLIB-584] SVM: Training on small table will lead support vector failure in parallel mode

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Abandoned
• Fix Version/s: v1.9
• Assignee: Rahul Iyer
• Reporter: Jiali Yao
• Created: 23/Jul/12 04:58
• Updated: 22/Mar/18 18:46
• Resolved: 22/Mar/18 18:46

Description

0. Prepare data

drop table if exists svm_input;
CREATE TABLE svm_input(id int, ind FLOAT8[],label float8);
insert into svm_input values(1, '{1,1}', 2);
insert into svm_input values(2, '{1,2}', -2);

1. svm_novelty_detection

madlib=# SELECT madlib.svm_novelty_detection('svm_input', 'svm_model', true, 'madlib.svm_dot');
ERROR:  plpy.SPIError: plpy.SPIError: plpy.Error: No support vector model with name svm_model0 found. (plpython.c:4648)
CONTEXT:  Traceback (most recent call last):
  PL/Python function "svm_novelty_detection", line 28, in <module>
    return online_sv.svm_novelty_detection( MADlibSchema, input_table, model_table, parallel, kernel_func);
  PL/Python function "svm_novelty_detection", line 204, in svm_novelty_detection
PL/Python function "svm_novelty_detection"

2. svm_regresssion

madlib=# SELECT madlib.svm_regression('svm_input', 'svm_model', true, 'madlib.svm_dot');
ERROR:  plpy.SPIError: plpy.SPIError: plpy.Error: No support vector model with name svm_model0 found. (plpython.c:4648)
CONTEXT:  Traceback (most recent call last):
  PL/Python function "svm_regression", line 28, in <module>
    return online_sv.svm_regression( MADlibSchema, input_table, model_table, parallel, kernel_func);	
  PL/Python function "svm_regression", line 53, in svm_regression
PL/Python function "svm_regression"

3. svm_classification

madlib=# SELECT madlib.svm_classification('svm_input', 'svm_model', true, 'madlib.svm_dot');
ERROR:  plpy.SPIError: plpy.SPIError: plpy.Error: No support vector model with name svm_model0 found. (plpython.c:4648)
CONTEXT:  Traceback (most recent call last):
  PL/Python function "svm_classification", line 28, in <module>
    return online_sv.svm_classification( MADlibSchema, input_table, model_table, parallel, kernel_func);
  PL/Python function "svm_classification", line 129, in svm_classification
PL/Python function "svm_classification"