当前位置：网站首页>Chapter 9 SVM Practice

Chapter 9 SVM Practice

2022-07-31 02:46:00 【Sang Zhiwei 0208】

1 Use of SVM packages

Import svm package from sklearn

2 SVM tuning

Increased c——decreased generalization ability——increased accuracy——narrowed the transition band

kernel - kernel function, default is rbf, can be linear

gamma: kernel function parameters for rbf, poly and sigmoid.The default is auto, which will select 1/n_features

ovo:one vs one ——>Six classifiers appear in four categories

ovr (default): one vs rest --> Four classifiers appear in four categories

3 Handling of unbalanced data

Change the weight size

Then judge the most suitable classification through the graphicsdevice

4 Classifier metrics

AUC (Confusion Matrix):

$TPR=\frac{TP}{TP+FN}$

$FPR=\frac{FP}{FP+TN}$

Metrics:

Accuracy $accuracy=\frac{TP+TN}{TP+TN+FN+FP}$

Precision $precision=\frac{TP}{TP+FP}$

Recall $recall = \frac{TP}{TP+FN}$

The harmonic mean of precision and recall $F_{1}=\frac{2\cdot precision\cdot recall}{precision+recall}=\frac{2}{\frac{1}{precision}+\frac{1}{recall}}$

$F_{\beta }=\frac{(1+\beta ^{2})\cdot precision\cdot recall}{\beta ^{2}\cdot precision+recall}$ , β is small, it is important to precision, otherwise it is important to recall.