| 摘要 |
<p>A method of reconstructing a signal from a given set of data, with a prediction function representing a predictable effect on the signal, and a noise function representing unpredictable noise. The method comprises the steps of altering the coordinate basis of the data and signal from an original coordinate basis in order to produce a prediction function having a reduced set of variables, performing a Bayesian reconstruction capable of operation of positive, negative, and complex signal values to produce a reconstruction signal, and converting the reconstruction signal back into the original coordinate basis to generate a signal.</p> |
