摘要 |
<p>PROBLEM TO BE SOLVED: To optimize a discriminant function for pattern recognition.SOLUTION: A learning device of pattern recognition learns a pattern of recognition. The learning device includes: determination means that determines a discriminant function for identifying a category of each learning data from learning data for pattern recognition; and optimization means that optimizes the discriminant function while considering dispersion characteristic of the learning data in the category. In the optimization, the identification boundary may be shifted from an identification boundary of a category given by the discriminant function so as to maximize the value of the Kullback-Leibler divergence (identification capability evaluation function) which considers the dispersion up to a learning data included in the category.</p> |