A method of compressing one-dimensional time series data at uniform time intervals comprises: a first process of inputting the one-dimensional time series data and an approximation accuracy thereof; a second process of calculating node candidates of an interpolation function approximating the one-dimensional time series data; a third process of optimizing positions of the node candidates; a fourth process of judging an error between one-dimensional time series data produced using the interpolation function, based on the node candidates at optimized positions, and the original one-dimensional time series data; a fifth process of judging an error between one-dimensional time series data produced using the interpolation function on the basis of the node candidates at optimized positions and the original one-dimensional time series data, and changing a node number and returning to the second process when the error exceeds a value based on the approximation accuracy; and a sixth process of outputting data of the node candidates and corresponding time data as compressed data when it is judged in the fifth process that the error does not exceed the value based on the approximation accuracy. Therefore, compression adapted to the properties of the given one-dimensional time series data is performed. In addition, since the compression ratio is controlled by the approximation accuracy, errors after decompression can be grasped in advance.
展开▼
