摘要 |
<P>PROBLEM TO BE SOLVED: To solve a problem in an electrical apparatus working state estimating system wherein, when the number of electrical apparatuses in a house is increased, the number of combinatorial patterns increases exponentially and hence the data collection following an initial learning is extremely difficult. <P>SOLUTION: The electrical apparatus working state estimating system estimates working states of a plurality of electrical apparatuses 2a-2c used in the measured house 1 based on pattern recognition with previously stored learning data. Total load current and voltage in the measured house 1 is obtained as measured data by operating the electrical apparatuses 2a-2c one by one, the measured data is calculated every combination, a plurality of created data are created, and the learning data is set as a characteristic amount of the obtained measured data and created data. Thus, all combinatorial patterns can be extracted by calculation based on the measured data obtained when the electrical apparatuses 2a-2c operate one by one, so that the working time of the data collection following the initial learning can be reduced significantly. <P>COPYRIGHT: (C)2006,JPO&NCIPI |