Developing statistical models to explain and forecast joint decision making

项目摘要（简要概述-一段）：虽然过去二十年来，在许多学科中，解释和预测人类行为的数学模型得到了迅速发展，但其中一个受到关注最少的过程是由多个智能体共同做出的决策。共同决定和社会影响与生活的大多数方面有关，从家庭管理到疾病的治疗选择；虽然在博弈论领域已经发展了先进的理论模型，但对实际数据集的适用性还很落后。近年来，机器学习已经成为代表决策的关键分析工具，并且在捕获代理之间的交互和联合决策方面，可以说比传统的统计方法取得了更大的进步。然而，与选择模型或博弈论不同，机器学习缺乏计量经济学和心理学基础——产出不能用于福利分析，除了能够预测结果之外，对行为过程的了解很少。目前的博士项目旨在解决这一领域的一些重要研究空白。首先，候选人将对联合和集体决策的数学模型的研究进行回顾，这在方法和应用方面都是稀疏领域急需的贡献。该项目的核心要素将围绕发展一种新的统计框架，该框架既能适应共同决策，又能确保行为的可解释性。该框架的应用领域可以从卫生到运输决策，并将取决于候选人的愿望和背景以及数据的可用性。然后将所开发方法的预测能力和福利影响与机器学习和人工智能等其他技术进行比较，从而产生一项工作，该工作不仅评估不同技术的优缺点，还反映了它们如何相互作用以分析联合决策的复杂过程。

Project Summary (brief overview - one paragraph):While the past two decades have seen rapid development in mathematical models to interpret and forecast human behaviour in many disciplines, one of the processes that has received the least attention is that of decisions made jointly by multiple agents. Joint decisions as well as social influences are relevant to most aspects of life, from household management to therapy choices in the case of illness; and while advanced theoretical models have been developed in the field of Game Theory, applicability to real datasets lags behind. In recent years, Machine Learning has emerged as a key analytical tool for representing decision making, and has arguably made greater strides than traditional statistical approaches when it comes to capturing interactions between agents and joint decision making. However, unlike Choice Modelling or Game Theory, Machine Learning lacks an econometric and psychological foundation - the outputs cannot be used for welfare analysis and little is learned about the behavioural processes beyond being able to predict outcomes. The present PhD project aims to address some important research gaps in this area. First of all, the candidate will conduct a review of the studies of mathematical models of joint and collective decision making, a much needed contribution in a sparse field both in terms of methods and applications. The core element of the project will revolve around the development of a new statistical framework which can accommodate joint decision making while ensuring behavioural interpretability. The areas of applications of this framework can range from health to transport decisions, and will depend on the aspirations and background of the candidate and data availability. The forecasting ability and welfare implications of the developed methods will then be compared to other techniques such as machine learning and AI, producing a piece of work that does not only assess strengths and weaknesses of the different techniques but also reflects on how they can interact to analyse the complex process of joint decision making.