摘要 |
PROBLEM TO BE SOLVED: To provide an active learning system for estimating target functions with a small number of experiments, when performing experiments in which the execution of all the target trials is virtually impossible; when input/output data are continuous values. SOLUTION: A boosting device (1) obtains a hypothesis from initial data to be inputted from input means (100) and the initial distribution of the data, calculates a predicted value of current data from the hypothesis, calculates error from the predicted value, and updates the distribution of the data. An input candidate data selection means (22) uses the hypothesis finally obtained by the boosting device (1) and selects input candidate data which are not contained in the past data, so that this input candidate data is instructed as a new experiment to an experimenter. A function value input means (23) receives a function value, obtained by an experiment (30) of the experimenter for the above input candidate data. A data-updating means (24) adds the obtained input candidate data and its function value, to the past data. COPYRIGHT: (C)2005,JPO&NCIPI
|