I am using Indian Pines dataset available in the given weblink (mat file & ground truth file). Here there are 16 different classes (features) with number of samples. My first question is that how to define these classes in the 2 D space. My second question is that I want to design my own support vector machine (instead of using lib SVM), how to design the vector machine taking the consideration of these 16 class classification.
http://www.ehu.eus/ccwintco/index.php?title=Hyperspectral_Remote_Sensing_Scenes
I think you are using WEKA data mining software. Are you aware of about WEKA development environment, where you can test your algorithm for more than 50+ classes. Also you can call SVM with appropriate kernel parameter instead of using SMO/Lib SVM etc.
using matlab, you have to do bluck of coding , use the other software for proccessing for ease like Erdas imgine
Dear Shrish Bajpai,
Here in your data, I think you have 16 class problem data. you mentioned that this 16 class as a feature but it is not feature. Features means independent variables. Can you please give some details about the dataset.
Please refer this paper. In this work, we develop a model for 4 class problem.
https://www.researchgate.net/publication/303011477_Quality_Assessment_of_Web_Services_Using_Multivariate_Adaptive_Regression_Splines
Conference Paper Quality Assessment of Web Services Using Multivariate Adapti...
- Badges
- Science topic
- Similar topics
- Geoscience
- Geoinformatics