This paper uses learning methods and optimization techniques to investigate the determinants of shock propagation in the Euro area for the period 2001–2015. First, principal component analyses are used with country bond yields to identify sub-periods and country groups; second, influencing factors for country bond yields are investigated with random forest models; lastly, shock propagation among groups are examined with impulse response functions. Models in steps two and three are improved by using simulated annealing algorithm. The empirical findings achieved can be particularly relevant for both investors and policymakers. Shedding light on the determinants of financial contagion may be in fact useful for investors who can derive relevant information about countries which are less sensitive to be affected by shocks, orienting thus their investment strategies. At the same time, policymakers could draw worthwhile and preventive hedging strategies and design the most suitable crisis management policies.

本文使用学习方法和优化技术来调查2001-2015年期间欧元区激波传播的决定因素。首先，将主成分分析与国家债券收益率一起使用，以识别子时期和国家类别；其次，采用随机森林模型研究了国债收益率的影响因素。最后，利用冲激响应函数研究了各组之间的冲击传播。通过使用模拟退火算法改进了第二步和第三步中的模型。获得的经验结果对于投资者和政策制定者都可能特别重要。实际上，了解金融危机蔓延的决定因素可能对投资者有用，因为他们可以获取有关不太容易受到冲击影响的国家的相关信息，从而确定他们的投资策略。同时，决策者可以制定有价值的预防性对冲策略，并设计最合适的危机管理策略。