Weak Feature Signals Identification Method Based on Local Projective and Wavelet Transform

A weak feature signals identification method based on local projection and wavelet transform is introduced. Experiment indicates that the local projective algorithm can separate background signals and weak feature signals into different orthogonal sub-spaces. Wavelet transform is effective for noise reduction of sharp and break signals. The algorithm which combines the local projective and wavelet transform has an excellent effect on identifying weak feature signals in nonlinear time series.