2017-05-04 10:55:31 -0600 | received badge | ● Student (source) |
2017-05-04 10:43:45 -0600 | asked a question | is the sigmoid kernel in SVM same as it in LIBSVM? using LIBSVM with sigmoid can achieve 86% of accuracy, but using cv can only get about 40%. what's wrong with my code? |
2017-05-04 10:32:19 -0600 | commented question | Problem training SVM with sigmoid kernel the performance of sigmoid kernel in opencv is quite worst than it in LIBSVM. but other kernels look same as them in LIBSVM |
2017-05-04 09:23:49 -0600 | commented question | 3.2.0 SVM OpenCV Error: Assertion failed (samples.cols == var_count && samples.type() == CV_32F) in predict, file C:/repo/mingw-w64-opencv/src/opencv-3.2.0/modules/ml/src/svm.cpp, line 1930 I want to test which combinations of features are more important. the varidx works when using bayes classifier. i thing they shuld be the same cause they are all the children of StatModel. anyway, i think that maybe after useing setTrainTestSplitRatio() to divied data into train and test, the varidx may work well. but unfortunately, after this it doesn't work when i call pridect(). i mean the following the steps donot work at all. 1. setTrainTestSplitRatio 2. train 3. predict i forgot the error message, it seems that data type error., but after spliting, no way to change data type. i donot know whether the ml pass the test. |
2017-05-02 04:43:20 -0600 | asked a question | 3.2.0 SVM OpenCV Error: Assertion failed (samples.cols == var_count && samples.type() == CV_32F) in predict, file C:/repo/mingw-w64-opencv/src/opencv-3.2.0/modules/ml/src/svm.cpp, line 1930 OpenCV Error: Assertion failed (samples.cols == var_count && samples.type() == CV_32F) in predict, file C:/repo/mingw-w64-opencv/src/opencv-3.2.0/modules/ml/src/svm.cpp, line 1930 the varidx for svm doesnot work, but for Bayes it works, at least the program doesnot crash. I checked some answers from google, someone said varidx should be like this {0,1,1,1,1,0,1...}, it is so strange, i cannot believe it. |