METHOD FOR DYNAMICALLY SELECTING OPTIMAL MODEL BY THREE-LAYER ASSOCIATION FOR LARGE DATA VOLUME PREDICTION

摘要

A method for dynamically selecting an optimal model by three-layer association for large data volume prediction, including three layers, namely, a prediction model algorithm library, a weight algorithm library, an optimal weight algorithm selection algorithm, the prediction model algorithm library is the lowest layer, the weight algorithm library is arranged above the prediction algorithm model library, and the optimal weight algorithm selection algorithm is arranged above the weight algorithm library. In the method for dynamically selecting the optimal model by three-layer association for large data volume prediction, the three-layer structure has four features, that is, high expandability, prediction stability, dynamic adjustment of the model, and the non-difference of the predicted data to the model. The method uses an association algorithm to avoid the shortcomings of common algorithms, by using a method of giving a plurality of model weights, a plurality of algorithms are organically combined together, the optimal algorithm is given with a higher weight, and a relatively poor algorithm is given with a lower weight, the method ensures not only the accuracy of data prediction, but also the stability of prediction after the data length increases.