基于网络搜索数据的GDP组合预测研究  

Research on GDP Combined Forecasting Based on Web Search Data

作　　者：王书平[1] 卢子晗 冀承秀 Wang Shuping;Lu Zihan;Ji Chengxiu(School of Economics and Management,North China University of Technology,Beijing 100144,China)

机构地区：[1]北方工业大学经济管理学院,北京100144

出　　处：《黑龙江科学》2024年第8期44-48,共5页Heilongjiang Science

基　　金：北方工业大学毓优团队培养计划项目(107051360021XN083/045);北方工业大学北京城市治理研究基地开放课题(110051360023XN277-02)。

摘　　要：网络搜索数据(Web Search Data, WSD)是研究宏观经济现象的重要微观信息依据。从需求、供给与政策环境等方面选取和筛选关键词来合成网络搜索指数,采用金枪鱼群(Tuna Swarm Optimization, TSO)算法优化的最小二乘支持向量回归(Least Squares Support Vector Regression, LSSVR)模型,对GDP进行预测。结果表明,网络搜索指数与GDP具有强相关性,合成的网络搜索指数能较好地反映GDP的波动走势;网络搜索数据的加入使得预测结果呈现出强时效性,预测效果及预测精度都取决于对最优模型的选择,引入参数智能优化算法可提高模型的预测性能。提出的TSO-LSSVR&WSD模型充分利用网络搜索数据及组合预测优势,提高了GDP的预测精度和时效性,可应用于宏观经济指标预测中。Web Search Data(WSD)is an important micro-information basis for studying macroeconomic phenomena.In this paper,keywords are selected and screened from the aspects of demand,supply and policy environment to synthesize the web search index,and the Least Squares Support Vector Regression(LSSVR)model optimized by the Tuna Swarm Optimization(TSO)algorithm is used to predict GDP.The results show that the web search index has a strong correlation with GDP,and the synthetic web search index can better reflect the fluctuation trend of GDP.The addition of web search data makes the prediction results show strong timeliness,the prediction effect and prediction accuracy depend on the selection of the optimal model,and the introduction of parameter intelligent optimization algorithm can improve the prediction performance of the model.The proposed TSO-LSSVR&WSD model makes full use of the advantages of web search data and combined forecasting to improve the accuracy and timeliness of GDP forecasting,and can be applied to the field of macroeconomic index forecasting.

关 键 词：GDP预测 组合预测 网络搜索数据 金枪鱼群算法 LSSVR模型 

分 类 号：F726[经济管理—产业经济] O29[理学—应用数学]