Feature Selections Using Minimal Redundancy Maximal Relevance Algorithm for Human Activity Recognition in Smart Home Environments

Table 8

The recognition accuracy rates with each features subset of HMM under the D-R criterion.

Features subset
Activity	1	2	3	4	5	6	7	8	9	10	11	12	13
0	0.900	0.900	0.800	0.800	0.833	0.867	0.833	0.833	0.700	0.700	0.733	0.667	0.667
1	0.000	0.000	0.292	0.271	0.250	0.229	0.083	0.500	0.458	0.792	0.750	0.750	0.750
2	0.000	0.005	0.184	0.256	0.483	0.657	0.768	0.860	0.908	0.918	0.928	0.942	0.961
3	0.000	0.870	0.935	0.957	0.978	0.935	0.891	0.804	0.826	0.761	0.761	0.761	0.761
4	0.000	0.000	0.524	0.524	0.714	0.952	1.000	0.952	0.929	1.000	1.000	0.976	0.976
5	0.000	0.000	0.400	0.700	0.400	0.700	0.500	0.500	0.600	0.400	0.400	0.400	0.200
6	1.000	0.986	1.000	1.000	1.000	0.971	0.971	0.986	0.986	0.986	0.986	0.986	0.971
7	0.000	0.000	0.297	0.703	0.892	0.973	0.892	0.946	0.838	0.973	0.919	0.973	0.946
8	0.090	0.119	0.239	0.373	0.388	0.701	0.701	0.701	0.701	0.731	0.776	0.806	0.821
9	0.000	0.205	0.614	0.705	0.886	0.773	0.886	0.909	0.955	0.932	0.864	0.864	0.864
Total	0.170	0.255	0.447	0.523	0.638	0.745	0.770	0.832	0.837	0.873	0.872	0.878	0.880