can anybody tell me how to generate confusion matrix in support vector machine? If somebody want explanation about confusion matrix then it is actually matrix which contain percent values of classification of correct classes and miscalculated classes.Like suppose I have 10 class classification and i give some input so that test values will be checked and some value will be predicted like actually and correct value in one class and wrongly predicted value in some other classes.like 80% comes in one class and rest 10 percent class 2 ,5%class 3 and 5% in class 4….like this.I want to show it in matrix form?Anybody know this how to do?
MATLAB: Can anybody tell me how to generate confusion matrix in support vector machine
libsvmmachine learningsupport vector machinesvm
Related Solutions
Hi Giansu,
Let's understand the scatter plot and confusion matrix generated by Classification Learner App for k-fold cross-validation with an example of iris dataset having 150 samples and 5-fold cross-validation.
As we choose 5-folds, the app will partition the data into 5 disjoint sets or folds cross-validation. For each fold, the app trains a model using 4 folds as training data and remaining 1-fold (i.e. held-out fold) as validation data.
It means whenever we use k-fold cross-validation, all the 150 samples will be considered as validation data or held-out fold for once. For e.g., for first iteration 1st fold will be validation and remaining 4 folds will be training data and similarly for second iteration 2nd fold will be validation and remaining 4 folds will be training data.
Scatter plot: The each prediction shown in the scatter plot is obtained when that particular observation was a part of held-out fold or validation data while model was training.
Confusion Matrix: The confusion matrix depicts how correctly the model predicted the class of the observation when that particular observation was a part of held-out fold or validation data while model was training. Hence the values are integer in confusion matrix.
Accuracy: The accuracy is calculated for each k-fold and to calculate the accuracy for the model we do average.
Following are the scatter plot and confusion matrix which I got on iris data for 5-fold cross validation:
Hope it helps!
By default, the data is divided into three subsets (trn/val/tst):
total = training + validation + testing
You can think of it two different ways:
1. total = designing + testing designing = training + validation2. total = training + nontraining nontraining = validation + testingTRAINING SUBSET: Design data used to directly determine the weights and thresholds of the model given inputs and targets (desired outputs). It's performance estimates are, obviously, highly biased.
VALIDATION SUBSET: Nontraining design data used two ways:
1. Validation stopping: Training is stopped when the error on the validation subset increases for 6(default) consecutive epochs.2. Ranking multiple designs with slightly biased performance estimates that typically differ by a. number of hidden nodes b. initial random weight assignments.TESTING SUBSET: Once designs are ranked to be satisfactory, UNBIASED PERFORMANCE ESTIMATES are obtained using the test subset.
If no unbiased performance estimate is satisfactory, the process can be repeated with the data repartitioned and/or design parameters changed.MATLAB yields separate confusion matrices for the three data division subsets in addition to the one for all of the data .
Hope this helps.
Greg
Best Answer