STATISTICAL FORECASTING OF DEMOGRAPHIC PROCESSES AND LABOUR MARKET BALANCE: AN INTEGRATION OF HOLT EXPONENTIAL SMOOTHING AND CANONICAL CORRELATION ANALYSIS (CCA) METHODS

The article presents a medium-term scenario forecast of labour market and demographic indicators of the Republic of Karakalpakstan for 2026–2030. The study is based on a two-stage approach that integrates C.C.Holt’s two-parameter exponential smoothing method with Canonical Correlation Analysis (CCA). At the first stage, a baseline forecast is constructed using the Holt method; at the second stage, it is adjusted on the basis of multivariate canonical relationships between demographic-digital factors and labour market indicators. The forecasting accuracy of the model corresponds to international standards and has been empirically validated through backtesting and leave-one-out cross-validation procedures. Three scenarios — pessimistic, baseline, and optimistic — have been developed, with the differences between them decomposed into contributing factors using the Laspeyres-Paasche-Fisher index decomposition. The findings empirically substantiate the priority role of digital skills development and labour migration management policies for the region.本文针对卡拉卡尔帕克斯坦共和国2026年至2030年的劳动力市场与人口指标开展中期情景预测研究。本研究采用两阶段研究框架，将C.C.霍尔特（C.C.Holt）双参数指数平滑法与典型相关分析（Canonical Correlation Analysis，CCA）相结合。第一阶段，通过霍尔特指数平滑法构建基准预测结果；第二阶段，基于人口-数字因素与劳动力市场指标间的多变量典型相关关系，对基准预测结果进行调整修正。该模型的预测精度符合国际标准，并通过回溯检验与留一交叉验证（leave-one-out cross-validation）程序完成实证验证。本研究共构建悲观、基准、乐观三类预测情景，并借助拉斯佩尔-帕舍-费雪（Laspeyres-Paasche-Fisher）指数分解法，将不同情景间的差异拆解至各贡献因素。本研究结果实证证实了数字技能培养与劳动力迁移管理政策在该区域发展中的优先地位。