Finding an efficient machine learning predictor for lesser liquid credit default swaps in equity markets

To solve challenges occurred in the existence of large sets of data, recent improvements of machine learning furnish promising results. Here to pro-pose a tool for predicting lesser liquid credit default swap (CDS) rates in the presence of CDS spreads over a large period of time, we investigate different machine learning techniques and employ several measures such as the root mean square relative error to derive the best technique, which is useful for this type of prediction in finance. It is shown that the nearest neighbor is not only efficient in terms of accuracy but also desirable with respect to the elapsed time for running and deploying on unseen data.