研究目的
To improve the accuracy of solar power forecasting through the use of ensemble techniques combining support vector regression (SVR), na?ve Bayes classifier (NBC), and hourly regression models.
研究成果
The ensemble technique significantly improves the accuracy of solar power output forecasts by combining the strengths of individual models (SVR, NBC, and hourly regression) and mitigating their weaknesses. The proposed method achieves an average forecasting error rate of 6.52% per year, demonstrating its effectiveness in enhancing forecast accuracy.
研究不足
The study focuses on short-term forecasting and uses data from a specific region (South Korea), which may limit the generalizability of the findings to other regions or time scales.
1：Experimental Design and Method Selection:
The study employs SVR, NBC, and hourly regression models for solar power forecasting. An ensemble method is proposed to combine these models' forecasts by weighting them based on the reciprocal of the standard deviation of their forecast errors.
2：Sample Selection and Data Sources:
Empirical data from solar power plants in South Korea for the year 2016, including solar radiation, temperature, humidity, and output data.
3：List of Experimental Equipment and Materials:
Not explicitly mentioned in the provided text.
4：Experimental Procedures and Operational Workflow:
The models are trained using historical data, and forecasts are made for solar power output. The ensemble method is then applied to improve forecast accuracy.
5：Data Analysis Methods:
Forecast accuracy is evaluated using normalized mean absolute error (NMAE).
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
