High dimensional model combination and classification


Forecast Combination With Heavy Tails

Forecast combination has proven to be a very important technique to obtain accurate and robust predictions. For various reasons, in many applications, forecast errors exhibit heavy tail behaviors. Unfortunately, little has been done to deal with forecast combination for such situations. The familiar forecast combination methods such as simple average, least square regression, or those based on variance-covariance of the forecasts, may perform very poorly. This project investigates two forecast combination methods to address the problem. One is specially proposed for the situations that the forecast errors are strongly believed to have heavy tails; the other is designed for relatively more general situations when there is a lack of strong or consistent evidence on the tail behaviors of the forecast errors. The computations of the methods can be computer intensive for high-dimensional applications. Both numeric and theoretical results will be sought to study the performance of the new methods relative to existing ones.

Group name: