Author: OTCFin – otcfin.com

OTCFin is applying Machine Learning techniques to give real-time figures of risk measures to our clients. But, what is “Machine Learning”?

Tom Mitchell, in his book Machine Learning defines the opening line of the preface:

“The field of machine learning is concerned with the question of how to construct computer programs that automatically improve with experience.”

 

Don’t worry, we will not create Skynet, and, no, our Machine Learning tools will not destroy civilization.

With a representative dataset and statistical functions, ML algorithms are capable of predicting the value of a new given data point without having to redo complex calculations every time, because it learned what the value should be.

One of our more challenging projects for this year is building a pre-trading system for non-performing loan portfolios. Our system will need to predict each loan’s price and the time to recover the investment for a particular expected return under different scenarios.

With our experts’ help on the Real Estate Market around Europe and Microsoft Azure (OTCFin is a Microsoft’s Silver Partner), we are developing an Azure’s Machine Learning Model using historical information of the Real Estate market, home price index and judicial auctions and processes.

We have tested several statistical techniques from the basic Linear Regressions to the most sophisticated artificial neural networks, comparing the results and adjusting the algorithm for better performance.

When done, Azure allows us to create a service that our application will call to get the analytics for the loans inside the portfolio.

Although Machine Learning techniques have proved to us, they will help achieve the results; it will not replace the need to understand the business.

That’s why OTCFin is here. We have a good understanding of business and technology to help our clients for better results.