Scholarlink

<p>Effectively leveraging the extracted information from both spatial and temporal dimensions is conducive to producing accurate forecasts, especially as existing grey models are highly dependent on the temporal trends of the correlated variables and ignore the influence of dynamic spatial associations in adjacent regions. Therefore, this paper designs a novel comprehensive spatial index to measure the spatial proximity effect by improving and combining the two modules of geographical and economic distance coefficients. Subsequently, an advanced grey multivariable convolution model that embeds spatial and temporal connections is proposed to forecast economic outputs for Zhejiang, Shanghai, and Anhui province’s high-tech industries. Moreover, the Particle Swarm Optimization (PSO) algorithm is employed to support the determination of the optimal weight parameters existing in the dynamic spatial and temporal relationships across neighboring regions. The experimental results illustrate that the proposed technique outperforms a range of baseline competitors regarding MAPE, RMSE, and Improvement Rate (IR) with several forecasting horizons. Specifically, the proposed model exbibit average IR of 5.22%, 0.77%, 1.16%, 7.75%, 3.60%, 3.58%, and 5.74% over the GMC(1, 3), SPDGM(1, 3, 3), DGM(1, 3), GM(1, 1), BPNN, SVR, and ARIMA models, respectively. Furthermore, the systematic test method further confirms the effectiveness and robustness of the model, rendering it an effective tool for forecasting high-tech industry output. Consequently, this study further forecasts the future outputs of high-tech industries in Zhejiang province from 2023 to 2027 and offers policy recommendations based on these findings.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Forecasts and Analysis of Economic Outputs for Chinese High-Tech Industries: Insights from Spatial–Temporal Information Fusion

  • 期刊论文

  • 发表日期：

    2025年2月14日

查看全文
  • Song Ding,
  • Yi Wang,
  • Xingao Shen

摘要

Effectively leveraging the extracted information from both spatial and temporal dimensions is conducive to producing accurate forecasts, especially as existing grey models are highly dependent on the temporal trends of the correlated variables and ignore the influence of dynamic spatial associations in adjacent regions. Therefore, this paper designs a novel comprehensive spatial index to measure the spatial proximity effect by improving and combining the two modules of geographical and economic distance coefficients. Subsequently, an advanced grey multivariable convolution model that embeds spatial and temporal connections is proposed to forecast economic outputs for Zhejiang, Shanghai, and Anhui province’s high-tech industries. Moreover, the Particle Swarm Optimization (PSO) algorithm is employed to support the determination of the optimal weight parameters existing in the dynamic spatial and temporal relationships across neighboring regions. The experimental results illustrate that the proposed technique outperforms a range of baseline competitors regarding MAPE, RMSE, and Improvement Rate (IR) with several forecasting horizons. Specifically, the proposed model exbibit average IR of 5.22%, 0.77%, 1.16%, 7.75%, 3.60%, 3.58%, and 5.74% over the GMC(1, 3), SPDGM(1, 3, 3), DGM(1, 3), GM(1, 1), BPNN, SVR, and ARIMA models, respectively. Furthermore, the systematic test method further confirms the effectiveness and robustness of the model, rendering it an effective tool for forecasting high-tech industry output. Consequently, this study further forecasts the future outputs of high-tech industries in Zhejiang province from 2023 to 2027 and offers policy recommendations based on these findings.