Problems of identification and inference for 'non-standard' models in complex systems with special reference to finance and teletraffic

该项目涉及处理复杂系统所需的“非标准”模型，例如那些表现出标度和分维特性的模型。重点放在处理重尾分布和长程相依观测的方法上，大多数标准统计方法都不适用于这些方法，以及在金融和电信方面的应用。该项目的一个重要部分涉及对Heyde的分形活动时间几何布朗运动模型的模型验证，该模型是替代几何布朗运动范式的候选最小描述风险资产模型。

The project is concerned with 'non-standard' models needed to deal with complex systems, such as those exhibiting scaling and fractal properties. There is a focus on methods for dealing with heavy tailed distributions and long range dependent observations, for which most standard statistical methods break down, and on applications in finance and telecommunications. An important part of the project concerns model validation for Heyde's fractal activity time geometric Brownian motion model, a candidate minimal description risky asset model to replace the geometric Brownian motion paradigm.