[MADLIB-642] SVM Classfication Performance: Classification with Kernel function can improve performance on below datasets

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Abandoned
• Fix Version/s: v1.9
• Assignee: Rahul Iyer
• Reporter: Jiali Yao
• Created: 06/Sep/12 11:57
• Updated: 22/Mar/18 18:46
• Resolved: 22/Mar/18 18:46

Description

Below data sets can not return result in several hours. It also can not return result in libsvm with similar parameter.

Data sets name TrainSize TestSize Attributes Rate(1:-1) Missing Source URL
rcv1.binary 20242 677399 47236 365951:331690 N http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/binary.html#rcv1.binary
URL Reputation Data Set 19000 3231961 6638:12362 N http://archive.ics.uci.edu/ml/machine-learning-databases/url/url.names

Test case

-- method: svm_cls_linear_ds_0_7_lsvm_classification_0
SELECT madlib.lsvm_classification
                        ( 'madlibtestdata.svm_url'::text     --input_table
                        , 'madlibtestresult.cls_model_table'::text    --model_table
                        , 'true'::boolean       --parallel
                        , 'false'::boolean        --verbose
                        , '0.1'::float8            --eta
                        , '0.001'::float8            --reg

                   ) AS q;


-- method: svm_cls_dot_ds_0_1_svm_classification_0
SELECT madlib.svm_classification
                        ( 'madlibtestdata.svm_rcv1_binary'::text     --input_table
                        , 'madlibtestresult.cls_model_table'::text    --model_table
                        , 'true'::boolean       --parallel
                        , 'madlib.svm_dot'::text    --kernel_func
                        , 'false'::boolean        --verbose
                        , '0.01'::float8            --eta
                        , '0.005'::float8             --nu
                   ) AS q;