reduce_ocr_class_svm — Approximate a trained SVM-based OCR classifier by a reduced SVM.
reduce_ocr_class_svm reduces the classification time of an SVM based OCR classifier OCRHandle by returning a reduced copy of it in OCRHandleReduced. The parameters Method, MinRemainingSV and MaxError have the same meaning as in reduce_class_svm and are described there. Please note that classification time can also be significantly reduced with a preprocessing step in create_ocr_class_svm, which possibly introduces less errors.
Original handle of SVM-based OCR-classifier.
Type of postprocessing to reduce number of SVs.
Default value: 'bottom_up'
List of values: 'bottom_up'
Minimum number of remaining SVs.
Default value: 2
Suggested values: 2, 3, 4, 5, 7, 10, 15, 20, 30, 50
Restriction: MinRemainingSV >= 2
Maximum allowed error of reduction.
Default value: 0.001
Suggested values: 0.0001, 0.0002, 0.0005, 0.001, 0.002, 0.005, 0.01, 0.02, 0.05
Restriction: MaxError > 0.0
SVMHandle of reduced OCR classifier.
If the parameters are valid the operator reduce_ocr_class_svm returns the value 2 (H_MSG_TRUE). If necessary, an exception is raised.
do_ocr_single_class_svm, do_ocr_multi_class_svm, get_support_vector_ocr_class_svm, get_support_vector_num_ocr_class_svm
Bernhard Schölkopf, Alexander J.Smola: “Lerning with Kernels”; The MIT Press, London; 1999.