PARAMETER ESTIMATION APPARATUS AND DATA COLLATING APPARATUS

摘要

In order to learn correlation between an observation pattern as an input and a pattern whose output is estimated from the input, a first teaching vector consisting of an observation pattern and a second teaching vector consisting of an estimation object are input. The self-correlation information on the two teaching vectors, an average vector of the two teaching vectors, and information on mutual correlation between the first and the second teaching vector are calculated. By using these to calculate a probabilistic expectation based on Bayes theory of the estimation object with respect to the input pattern, it is possible to accurately estimate a desired parameter at a low calculation cost.