An Analysis of Audio Features to Develop a Human Activity Recognition Model Using Genetic Algorithms, Random Forests, and Neural Networks

Table 10

Confusion matrix of the RF-based HAR model with all MFCC features.

⁢	Brew coffee	Cooking	Use microwave oven	No activity	Take a bath	Dish washing	Hand washing	Teeth brushing	Type II error
Brew coffee	232	1	0	0	0	7	2	3	0.053
Cooking	50	76	0	0	6	0	0	0	0.424
Use microwave oven	9	0	33	0	0	0	0	0	0.214
Take a bath	9	2	0	0	398	10	0	9	0.070
Dish washing	8	7	0	0	14	100	1	4	0.254
Hand washing	0	0	0	0	2	3	60	5	0.143
Teeth brushing	8	5	2	0	11	3	1	94	0.242
No activity	4	0	0	10	0	0	1	1	0.375
Type I error	0.275	0.165	0.057	0.000	0.077	0.187	0.077	0.190