Figure 6.

The classification performance of different regularization parameters λ. The ordinate indicates accuracy, specificity and sensibility of this method, and the abscissa denotes nine groups of different regularization parameter λ, where 1 represents that λ value is {0.1}, and 2 denotes that λ value is {0.1, 0.2}, and 3 denotes that λ value is {0.1, 0.2, 0.3}, …, 9 represents that {0.1, 0.2, …, 0.9}. Thus, when λ is {0.1, 0.2, …, 0.9}, better classification performance can be obtained, including that accuracy is 91.60%, and specificity is 93.50%, and then sensibility is 90.50%.