PURPOSE: To statistically process many data and acquire new knowledge by increasing the number of data to be analyzed.

CONSTITUTION: Data inputted to a data input part 1 is inputted to a preprocessing part 2, which converts respective parameters into feature quantities representing time variation. For example, when one parameter varies periodically, Fourier transformation is performed to obtain the amplitude and phases of respective frequency components. If the time variation of one parameter can be approximated by a specific function, transformation into a parameter showing features of the function. Thus, the data which is transformed into the feature quantity representing the time variation is then inputted to a statistical processing part 3, where various existent statistical processes are performed. When there is a decision on the result in the input data, it becomes what is called data with a tutor and is displayed at a result display part 4 and utilized to acquire new knowledge on the basis of it.