I have a two dimensional feature vector, e.g., X(i,j). It contains the feature values for objects of two groups. For example, Each X(i,j) is a computed value for an object of category of P or Q. Let, if X(i,j) is computed for an object of category of P, the group value G(i) is then set to 0, i.e., G(i)=0. Else it is set to 1, i.e., G(i)=1. Thus, I have X(i,j) and G(i). Now, I want to train half of the sample data and to test the rest. The object is to find the value of how much P is predicted as Q and vise-versa. I wish to apply svmtran and svmclassify methods. I applied that methods, but, probably I am doing some mistake; therefore, it is not giving expected values. I will be grateful to you if any of you, please, explain the process of applying it with proper programming code. I need the matlab code for that task. Thanks in advance.
Before classification operation, You must prepare the data in proper form as features and labels.
then you must divide the data into two parts: training and testing.
Train the system with the training set and you use the test set to calculate the results.
To understand SVM classification models, You can look following code.
https://www.mathworks.com/matlabcentral/fileexchange/28302-svm-demo
https://www.mathworks.com/matlabcentral/fileexchange/48635-simple-example-and-generic-function-for-svm-binary-classifier
Ok, I got the matter. But still I have question. Consider that I have measured the feature values from the original images and set the label values to 0. Next, I have measured the feature values from the tempered images and set the corresponding label values to 1. Now, I have a complete set of features and labels. While dividing the features into two groups, I will apply a random chooser to separate the train and test data set. My question is here. I can do that in two ways:
1. I can separate the features and their corresponding label to a group? This means that when I am selecting a feature for the train, I am selecting the corresponding label for the train as well, i.e., if I am selecting the 10th feature set from the feature vector, I am selecting the 10th label from the labeling table.
2. I am dividing the features randomly into two groups. At the same time, I am randomly grouping the label values into two parts. That feature grouping and label grouping are performed independently. There is no correlation between these two grouping task.
Now please advice me, which is the correct process. Thanking you.
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training