Research Article

An Efficient and Fast Model Reduced Kernel KNN for Human Activity Recognition

Table 4

The distribution of K neighbors in benchmarks and real-world data.

BenchmarksK number of neighbors
2345678

Iris94.7095.2091.7388.5389.4785.2689.95
Heart91.6688.1388.8988.9389.4389.4688.18
Breast82.7781.2580.6280.1379.3878.0278.14
Diabetes84.3284.2883.4584.5884.3984.4684.67
Seed96.0397.3694.0192.6693.7894.7194.02
Thyroid97.6797.0996.7196.0995.9293.0894.42
Banana94.6794.0893.3593.0893.9193.2893.16
Image99.0498.5698.2198.2497.8097.5297.87
Ringnorm92.8376.8774.8174.7474.2173.8273.56
Splice93.5191.9193.3993.3692.6893.0095.46

Human activity data3456789

HAPT85.9388.3290.5191.6090.3286.6284.34
Smartphone88.6089.2391.2892.6787.8387.9885.39