Discriminating the Nature of Thyroid Nodules Using the Hybrid Method

Algorithm 1

Feature selection with PCA.

(i)	Input: dimensional dataset
(ii)	Output: dataset after dimensionality reduction
(iii)	Centralize all samples
(iv)	Calculate sample covariance matrix
(v)	Eigenvalue decomposition of the matrix
(vi)	Get the eigenvectors corresponding to the largest eigenvalues
(vii)	Normalize all eigenvectors to form an eigenvector matrix
(viii)	Convert samples to new samples
(ix)	Get the output sample set
(x)	End