This paper uses a minimum divergence framework to introduce a new way of calculating model weights that can be used to average probabilistic predictions from statistical and machine learning models. The method is general and can be applied regardless of whether the models under consideration are fit to data using frequentist, Bayesian, or some other fitting method. The proposed method is motivated in two different ways and is shown empirically to perform better than or on a par with standard model averaging methods, including model stacking and model averaging that relies on Akaike-style negative exponentiated model weighting, especially when the sample size is small. Our theoretical analysis explains why the method has a small-sample advantage.

翻译：本文采用最小散度框架，提出一种计算模型权重的新方法，用于对统计模型和机器学习模型的概率预测进行加权平均。该方法具有通用性，无论所考虑的模型是通过频率学派、贝叶斯或其他拟合方法得到，均可适用。本文从两种不同角度论证了该方法的合理性，并通过实证研究表明，该方法在性能上优于或等同于标准模型平均方法（包括模型堆叠和依赖赤池式负指数模型权重的模型平均），尤其是在样本量较小时表现更为突出。我们的理论分析揭示了该方法在小样本情况下具有优势的原因。