燃气年度负荷预测的偏最小二乘回归分析法

城市燃气的年度负荷受到多种因素的影响，如能源政策、人口和国内生产总值等，各因素之间相互制约和影响，存在着严重的多重相关性，若采用普通的回归分析会使结果产生较大的偏差，而偏最小二乘回归方法能很好地解决这一问题．采用偏最小二乘回归分析法对燃气年度负荷进行拟合和预测，实例表明，该方法具有建模简单、计算结果可靠的特点，所得到的回归方程能综合分析各因素对于负荷的影响权重，具有较强的实用性．

Annual gas load forecasting based on partial least squares method

The annual gas load is impacted by the factors such as energy policy, urban population and gross domestic product, etc. These factors restrict and influence each other, and there is muhi-correlation among them. The common regression model may distort and destabilize the results. The partial least square regressive method （PLS） can easily solve the problems with multi-correlation of independent variables. The PLS is used in the forecasting of annual gas load, the results showed that the process of modeling is simple and the precision of prediction is credible.