How do you determine if an organization is likely to go bankrupt? In days past it might have taken a gargantuan level of human effort, but in the near future, it’s well possible that computers will be picking up the slack. Researchers at Russian-based HSE University have recently made huge leaps forwards when it comes to calculating financial risks, and today, we’ll be providing a brief overview of the work they’ve undertaken (along with why you should take notice).
What’s The Story?
Recently reported in E&T News, researchers from the HSE Graduate School of Business, led by professor Yuri Zelenkov and Nikita Volodarskiy, have begun exploring a new methodology for bankruptcy forecasting that utilizes cutting-edge machine learning techniques:
“Researchers of the HSE Graduate School of Business have presented a new method of forecasting bankruptcies in businesses using machine learning. The method makes it possible to fully utilize information on a company’s financial state and to make more accurate predictions than traditional statistical approaches.”
The methodology in question uses a “set of historical data on successful and failed companies.” Taking this extensive info, the researchers then used a set of business performance indicators to train their artificial intelligence on how to assess the financial health of organizations. After this, they used the AI they developed to search for “complex patterns in companies’ development and their current state.”
Now, to the untrained ear, this might not sound like a big deal. Consider the following, however: this task has historically suffered due to the “problem of imbalance classification.” In other words, as the article explains, bankruptcy happens to such companies at such a low rate of occurrence (maybe five to ten percent) that training machines to recognize the precursory signs of bankruptcy has been more than just a bit difficult.
Consider some of the classification and optimization techniques of the past—discriminant analysis, logistic regression, etc.—or even market data approaches to predicting bankruptcy. While still in use, they have left some wanting more, which is where our researchers enter the scene. Their new methodology is supposedly “less sensitive to imbalances in data” and capable of selecting the best out of a number of individual algorithms. As Professor Zelenkov explains:
“We managed to build a fast algorithm that can be trained using unbalanced data to make much more accurate predictions than traditional methods. Notably, the user can manage prediction errors of each class in a visual form. Since the model is exclusively based on companies’ financial indicators, its results are still reliable even in the extreme conditions of the Covid-19 pandemic.”
As for why all of this is so important, just think about the fact that being able to pinpoint financial risks that face a business is paramount in guiding that business down a profitable path, forecasting changes in the economy, etc. On top of that, from a purely academic point of view, bankruptcy prediction has been an important and bustling field of study for well over five decades. Those who hold interest in this field will surely be excited by this recent news and eager for even more new developments when it comes to financial predictions. If you’re a business owner who thinks bankruptcy is in the horizon, it’s best if you contact a Colorado bankruptcy attorney today.