TIME-SERIES-DATA FEATURE-AMOUNT EXTRACTION DEVICE, TIME-SERIES-DATA FEATURE-AMOUNT EXTRACTION METHOD AND TIME-SERIES-DATA FEATURE-AMOUNT EXTRACTION PROGRAM

摘要

A time-series-data feature-amount extraction device comprising: a data processing unit in which a received unevenly spaced time-series-data group is processed into an evenly spaced time-series-data group including omissions and an omission information group indicating the presence or absence of omissions on the basis of the length of received input time-series data and a received minimum observation interval;a model learning unit that learns a weighting vector of each layer of a model, using, as the error, the difference between the elements not omitted from a matrix of the evenly spaced time-series-data group which includes omissions and the elements in the output result from a model output layer, and stores the weighting vector as a model parameter in a storage unit; and a feature-amount extraction unit that receives the time-series-data to be subjected to feature-amount extraction, inputs the received time-series-data to be subjected to feature-amount extraction into the model in order to calculate a value for the middle layer of the model using the model parameters stored in the storage unit, and outputs the calculated middle layer value as the feature amount expressing change over time in the data.