abstract |
Embodiments of the present invention provide a data sequence prediction method and computing device, the method comprising: calculating a similarity distance between each of the N objects according to historical data sequences of N objects, to obtain a similarity distance set, Wherein, the similarity distance is used to represent the similarity degree of two objects, the historical data sequence includes a plurality of data arranged according to a preset rule, and N is a positive integer greater than 1; according to the similarity distance set, clustering The algorithm divides the N objects into K prediction object classes, where K is a positive integer, and K≤N; and predicts the future data sequence of objects included in at least one prediction object class in the K prediction object classes. By implementing the embodiment of the present invention, the prediction object class is automatically defined through the clustering algorithm, which is efficient and more accurate. |