基于机器学习的长三角数字经济预测模型选择

Selection of a Digital Economy Prediction Model for the Yangtze River Delta Based on Machine Learning

聚焦于数字经济这一现代经济增长的核心驱动力,通过特征选择算法精炼数据,结合多种机器学习算法构建模型,并经交叉验证优化参数,依据模型性能指标筛选出最优的长三角数字经济模型,以期高效、精准地把握长三角数字经济发展趋势.研究结果显示:特征选择算法能显著提升机器学习算法的预测精度;各机器学习算法中支持向量回归的预测效果最佳;各模型组合中随机森林-支持向量回归模型为最优的数字经济预测模型,能够很好地反映出长三角数字经济稳步发展的趋势.

Focusing on the digital economy,the core driving force of modern economic growth,this study refines data through feature selection algorithms,builds models with various machine learning algorithms,and optimizes parameters through cross-validation.The optimal model for predicting the digital economy in the Yangtze River Delta is selected based on performance metrics to efficiently and accurately grasp the development trends of the digital economy in the Yangtze River Delta.The results show that feature selection algorithms can significantly improve the prediction accuracy of machine learning algorithms.Among various machine learning algorithms,support vector regression performs the best.Among all model combinations,the Random Forest-support vector regression model is the optimal digital economy prediction model,which can reflect the steady development trend of the digital economy in the Yangtze River Delta.