Forest Type Classification Using Support Vector Machine Lab Report
CHAPTER 6
FOREST TYPE CLASSIFICATION USING SUPPORT VECTOR MACHINE
CHAPTER 6
Forest Type Classification using Support Vector Machine
In this chapter, an attempt has been made to classify the forest type using support vector machines (SVMs). SVM is regarded as a powerful technique in order to deal with the classification problems. In this work, we have explored different parameters of SVM in order to find the best possible recognition accuracy. The LibSVM tool has been used for the experimentation in this work.
There are different parameters of SVM that can be considered for the purpose of optimizing the accuracy of classification by the SVM. The first parameter that has been considered in these experiments is the type of kernel …show more content…
This has been achieved here in 8 combinations of the tolerance level and scaled interval out of the 16 combinations considered. As such, one can infer that the variations in the tolerance level and the variations in scaling the data is not having a significant impact on the recognition accuracy when we employ RBF kernel with 3-fold cross validation. Table 6.8 contains the result of experiments when RBF kernel is used with 4-fold cross validation.
Table – 6.8: Accuracy obtained using radial basis function(RBF) kernel with 4-fold cross validation
Sr. No. Value of Tolerance (ε) Scaled Interval Recognition Accuracy (in %)
1 0.0001 [-1,1] 51.8
2 0.0001 [-3,3] 58.7
3 0.0001 [1,3] 51.8
4 0.0001 [1,7] 58.7
5 0.001 [-1,1] 51.8
6 0.001 [-3,3] 58.7
7 0.001 [1,3] 51.8
8 0.001 [1,7] 58.7
9 0.01 [-1,1] 51.6
10 0.01 [-3,3] 58.8
11 0.01 [1,3] 51.6
12 0.01 [1,7]
FOREST TYPE CLASSIFICATION USING SUPPORT VECTOR MACHINE
CHAPTER 6
Forest Type Classification using Support Vector Machine
In this chapter, an attempt has been made to classify the forest type using support vector machines (SVMs). SVM is regarded as a powerful technique in order to deal with the classification problems. In this work, we have explored different parameters of SVM in order to find the best possible recognition accuracy. The LibSVM tool has been used for the experimentation in this work.
There are different parameters of SVM that can be considered for the purpose of optimizing the accuracy of classification by the SVM. The first parameter that has been considered in these experiments is the type of kernel …show more content…
This has been achieved here in 8 combinations of the tolerance level and scaled interval out of the 16 combinations considered. As such, one can infer that the variations in the tolerance level and the variations in scaling the data is not having a significant impact on the recognition accuracy when we employ RBF kernel with 3-fold cross validation. Table 6.8 contains the result of experiments when RBF kernel is used with 4-fold cross validation.
Table – 6.8: Accuracy obtained using radial basis function(RBF) kernel with 4-fold cross validation
Sr. No. Value of Tolerance (ε) Scaled Interval Recognition Accuracy (in %)
1 0.0001 [-1,1] 51.8
2 0.0001 [-3,3] 58.7
3 0.0001 [1,3] 51.8
4 0.0001 [1,7] 58.7
5 0.001 [-1,1] 51.8
6 0.001 [-3,3] 58.7
7 0.001 [1,3] 51.8
8 0.001 [1,7] 58.7
9 0.01 [-1,1] 51.6
10 0.01 [-3,3] 58.8
11 0.01 [1,3] 51.6
12 0.01 [1,7]