海洋经济的发展是我国经济发展的重要组成部分, 因此科学地对海洋经济进行预测分析具有重要的现实意义. 文章选取全国海洋生产总值作为研究对象, 分别利用灰色模型, ARIMA 模型和组合模型进行预测. 提出了一种基于交叉验证方法确定组合预测权重的新算法, 研究显示该方法相对传统组合预测具有更小的平均绝对百分比误差(MAPE).

The development of ocean economy has become an important part of China's economic development, so it is of great practical significance to predict the ocean economy scientifically. This paper analyzes the ocean gross products by grey model, ARIMA model and combined model respectively. A new algorithm is proposed based on cross-validation to determine the weights of combination forecasting. The results show that the new method has lower MAPE than traditional combination forecasting.