A Novel Cloud Theory Based Time-series Predictive Method for Middle-term Electric Load Forecasting

The middle-term electric load forecasting is an existing difficult work and often has a large error. To address the problem, this paper proposes a novel cloud theory based time-series predictive method for middle-term electric load forecasting. In this method, the time series of daily maximum load is partitioned into two parts, historical dataset and current tendency dataset, backward cloud algorithm is applied to the two datasets to form the historical cloud and the current cloud, and the corresponding rule sets are mined. Then the historical cloud and current cloud is integrated to created predictive cloud through synthesized cloud. Finally, via cloud reasoning, the forecast result can be obtained. This predictive method effectively integrates quasi-periodical regularity and current tendency of time-series data, and has a simple computing model. The case study shows that the proposed method is accurate and practical.