Beginning Anomaly Detection Using

Figure 2-49.  Training the OC-SVM model on the training data

preds = ocsvm.predict(x_test)

score = 0

for

f 

in

range(0, x_test.shape[0]):

if

(preds[f] == 1):

score = score + 1

accuracy = score / x_test.shape[0]

print

("Accuracy: {:.2%}".format(accuracy))

Figure 2-50.  Making predictions and generating the “accuracy” score

Chapter 2   traditional Methods of anoMaly deteCtion

---

68

Figure 

2-51

 shows that the accuracy is about 89.1%, which is pretty good considering 

that you assumed 10% of the data would misclassify.

Let’s run the code on x_validation this time (see Figure 

2-52

).

Figure 2-51.  The resulting output accuracy for the testing data set

SUHGV RFVYPSUHGLFW [BYDO

VFRUH