From: Comparing classification techniques for identification of grasped objects
Classification technique | First classifier structure scenario 1 (universal use) | ||
---|---|---|---|
Training time (ms) | Test time (ms) | Accuracy (%) | |
Bagging | 411.62 | 15.68 | 91.9 |
Decision tree (entropy) | 7.2 | 0.40 | 55.7 |
Decision tree (Gini) | 5.8 | 0.30 | 86.6 |
kNN | 1.36 | 4.75 | 91.5 |
Linear discriminant analysis | 10.43 | 0.24 | 66.8 |
SVM (linear SVC) | 144.81 | 0.53 | 66.1 |
Logistic regression | 41.49 | 0.42 | 67.8 |
Logistic regression CV | 915.77 | 0.48 | 67.8 |
MLP | 3113.93 | 0.72 | 83.5 |
Naive Bayes (Bernoulli) | 1.57 | 0.35 | 52.2 |
Naive Bayes (Gaussian) | 1.62 | 0.86 | 74.0 |
NearestCentroid | 1.19 | 0.49 | 67.9 |
Quadratic discriminant analysis | 3.09 | 0.76 | 87.0 |
Radius neighbors | 1.17 | 146.89 | 47.5 |
Random forest | 208.87 | 14.38 | 93.2 |
Ridge | 2.9 | 0.22 | 64.2 |
Ridge CV | 3.05 | 0.33 | 64.4 |
Label propagation | 56.31 | 16.38 | 93.2 |
Label spreading | 99.53 | 16.90 | 93.2 |