Input
The input table is iris_test (see SVMDense Examples Input).
SQL Call
SELECT * FROM SVMDensePredict ( ON svm_iris_test AS "input" PARTITION BY ANY ON densesvm_iris_sigmoid_model AS model DIMENSION USING IDColumn ('id') InputColumns ('[1:4]') Accumulate ('id', 'species') ) AS dt ORDER BY id;
Output
id | predict_value | predict_confidence | species |
---|---|---|---|
5 | setosa | 0.746591233570276 | setosa |
10 | setosa | 0.735684055731196 | setosa |
15 | versicolor | 0.478705967446573 | setosa |
20 | setosa | 0.733761949232838 | setosa |
25 | setosa | 0.737952913158837 | setosa |
30 | setosa | 0.744839260424954 | setosa |
35 | setosa | 0.735684302199767 | setosa |
40 | setosa | 0.74610706105241 | setosa |
45 | setosa | 0.682566149898358 | setosa |
50 | setosa | 0.737950987191321 | setosa |
55 | virginica | 0.588477773284318 | versicolor |
60 | setosa | 0.650348601535809 | versicolor |
65 | versicolor | 0.428915851630996 | versicolor |
70 | setosa | 0.469605915151742 | versicolor |
75 | virginica | 0.5459264767513 | versicolor |
80 | setosa | 0.450354714402099 | versicolor |
85 | versicolor | 0.495077171256058 | versicolor |
90 | setosa | 0.541238079806321 | versicolor |
95 | versicolor | 0.452171681401104 | versicolor |
100 | versicolor | 0.503884155446139 | versicolor |
105 | virginica | 0.663559257854133 | virginica |
110 | virginica | 0.725772461767464 | virginica |
115 | versicolor | 0.462751539260052 | virginica |
120 | versicolor | 0.495331599436918 | virginica |
125 | virginica | 0.707555441131219 | virginica |
130 | virginica | 0.713214560807804 | virginica |
135 | virginica | 0.534033824524893 | virginica |
140 | virginica | 0.707555440928665 | virginica |
145 | virginica | 0.707555441266978 | virginica |
150 | virginica | 0.455661070255452 | virginica |
The prediction accuracy with the sigmoid model is 70%.