Managed Futures Strategies – Support Vector Machines as a Due Diligence Tool. Thesis by Amira Roula

Summary of Masters of Finance thesis by Amira Roula, from the Stockholm School of Economics :

In this paper, we have used the Support Vector Machine (SVM) method for classifying managed futures strategies, an exiting way of automating part of the due diligence process. Previously used methods for classification have been lacking, and the ones tested (factor models) have been based on assumptions which do not accomodate for the return distributions of managed futures and the hedge fund industry overall (i.e. IDD and normal distribution assumptions, and linearity). The SVM method is not dependant on distribution assumptions, and is a process which can automate the classification part of a due diligence process. When tested on the managed futures data at hand, the SVM showed a ceratin amount of misclassification by funds tested.

Read entire article or download the PDF here ›