| 摘要 |
<p>The present invention relates to a method for calculating a gradient of a function model based on data by using a sub-function model based on one or a plurality of accumulated data, particularly a Gaussian process model. According to the method, provided is a model calculation unit (3), constructed to calculate a function value of the function model based on data by using an exponential function, a sum function, and a multiplication function in two loop calculations based on hardware; and the model calculation unit (3) is used to calculate the gradient of the function model based on data for a value required by a predetermined input variable.</p> |
