accessibility

الأبحاث وأوراق العمل

A Machine Learning Approach to Nowcasting GDP with Limited Data Span: The Case of Jordan

The paper applies a new nowcasting approach for Real Gross Domestic Product (RGDP) forecasts in Jordan using Machine learning (ML). By employing the Extreme Gradient Boosting )XGBoost( algorithm, the study generates sector-specific GDP predictions, focusing particularly on the fourth quarter of 2023. These predictions are then compared against forecasts compiled using benchmark models, including AI-based models such as Long Short-Term Memory (LSTM) and traditional models like Autoregressive Moving Average (ARIMA) and Dynamic Factor Models (DFM). Results demonstrate that XGBoost provides projections with high level of precision.The growth rate projected for the fourth quarter of 2023 stands at 2.4%, closely aligning with the observed 2.3%. Therefore, the paper advocates to add XGBoost to Central Banks’ nowcasting toolbox,  emphasizing their superior accuracy.

المرفق انقر هنا (حجم الملف 1283.01 KB، نوع الملف pdf)   
 الملخص  انقر هنا (حجم الملف 205.13KB، نوع الملف pdf)

كيف تقيم محتوى الصفحة؟

عدد زوار الصفحة:

158