An Improved Support Vector Machine Method for Harmonic and Inter-harmonic Detecting

Based on the linear regressive model of support vector machines, iterative re-weighted least squares of the support vector machines for signal spectral analysis is introduced by virtue of abstract expressions of cost functions, and the uniform mathematical analytical expression of support vector machine for solving the problem of signal spectral analysis is obtained. And a new method is presented by changing resolution, namely defining the range of spectrum by taking advantage of wide range with rough resolution at first, and then using narrow range with high resolution. The simulation has showed that this method solves the problem of large calculation and slow calculating speed