Method for pattern-based prediction of time series of data points of e.g. financial market, involves determining similarity between current partial segments of time series and other partial segments

摘要

The method involves comparing the immediately preceding partial segments of time series of all historical partial segments, with partial segments of other time series. The similarity between the current partial segments of time series and other partial segments is determined according to historical curves of the partial segments. The similarity between the current partial segments of time series and other partial segments is determined for performing the normalization of partial segments of the time series.