Matrix Similarity Measurement Used in a GA-Based Feature Extraction

A GA-based feature extraction method was presented for solving non-linear regression questions. The order of regression function is determined by practical experience, so the application of this method is limited. In this paper, the angle measurement method of matrix similarity was used to determine the order of the polynomial space. By doing this, the GA-based feature extraction method became more practical. An explicit non-linear regress function can be very easy to find. The experiment shows that it achieved satisfactory results.