Figure 8.
Confusion matrix showing the performance results of each method using the 3-D acceleration sensor (Note: DSAE (Deep stacked autoencoder), NB (Naïve Bayes), SVM (Support vector machine), LDA (Linear discriminant analysis). Activities: 1—walk, 2—stand, 3—jog, 4—sit, 5—bike, 6—upstairs, 7—downstairs, 8—type, 9—write, 10—coffee, 11—talk, 12—smoke, 13—eat.