Multi-modal prediction method of power grid big data based on tensor chain
In order to optimize the accuracy and computing time of big data prediction system,a multi-modal prediction method based on tensor theory for power grid was proposed.By employing the tensor and Markov theories,a multi-variate and multi-order Markov model with strong adaptability and a Markov transfer method without hypothesis were designed.On this basis,the short-term and long-term prediction algorithms based on tensor chain theory were constructed,thus the multi-modal prediction method of big data with low computational complexity was proposed.The simulation results show that the multi-modal prediction method based on tensor chain has higher prediction accuracy and lower computing time in comparison with the classical Markov prediction method.
big datatensor chainmain eigenvaluemulti-modal predictionparallel computingMarkov modelcomplexity analysisprediction accuracy
万方数据
维普
