Can a new machine-learning technology transform credit assessment?

March 26, 2018 Insights Editor

Data experts are excited about its potential

Algorithms. Logistic regression. Both tried and true analytical methods.
Yet, as much as these technologies have transformed key business processes such as lending, risk and credit assessment, new developments are getting the data scientists of the world really excited, and for good reason.

Think: neural networks. It’s the next-gen step in today’s fast-moving machine learning and data analytics innovations, all of which are helping organizations efficiently and intelligently use big data to improve business outcomes. Often used within more progressive forms of machine learning known as “deep learning,” neural networks can push past the linear boundaries of logistic regression analysis, opening up a new world of opportunity with its more precise predictive potential. Here, we’ll share why this is great news for businesses and the consumers they serve.

The straight line versus an inclusive arc
When used for credit assessment, traditional logistic regression models [think: everyone above the straight line is approved and everyone below is declined] tend to dump consumers into two categories, such as prime and sub-prime. This can automatically exclude individuals who are hovering just below prime as they work toward building their credit profile. However, one reason these models are popular is because the outcomes can be easily explained to consumers, which supports compliance with consumer credit regulations that require increased lending transparency.

Neural networks, on the other hand, create a curved, non-linear arc that can help expand “approved” audiences by capturing hard-to-score consumers who are not quite prime, but are clearly trending in the right direction. This could benefit businesses by helping them more securely serve these consumers and drive growth. Likewise, it could benefit consumers by giving more people expanded access to mainstream financial services. The problem is, these networks are rather complex and the outcomes are not easily explainable to consumers, which has raised compliance concerns.

Sparking excitement in the data analytics industry
With excitement mounting around the potential of neural networks and related machine learning technologies, experts within the data science and analytics worlds have been pressing hard for a solution that’s applicable for use in credit lending decisions. As a result, many firms are producing powerful tools using deep learning techniques.

Equifax Data & Analytics has emerged with a game-changing technology that leverages the advanced, machine-learning technology of neural networks, which makes it highly configurable, granular and predictive. Yet, like traditional models, it also offers actionable reason codes that easily explain the resulting scores, which supports regulatory requirements.

The unique, patent-pending technology enables a deeper learning of consumer behavior and is applicable and deployable wherever traditional scorecards are appropriate across the commercial and consumer segments.

More to come in the near future
For data scientists within the credit and lending industries, this technology is game-changing. Businesses can finally create a high-performing model that maximizes predictiveness and accuracy, while also meeting critical regulatory guidelines. However, if recent advances in neural networking and machine learning tell us anything, it’s that we can expect even more advances in the near future. According to the team at Equifax, this is just the beginning of many exciting developments ahead.

Interested in the new Equifax neural networking technology? Check out this quick video to learn more.

The post Can a new machine-learning technology transform credit assessment? appeared first on Insights.

Previous Article
ETS Tax Intelligence: Actionable Insights Using TaxManager
ETS Tax Intelligence: Actionable Insights Using TaxManager

Situation Those charged with managing state unemployment insurance (“SUI”) tax accounts for an organization...

Next Article
March 2018 Briefly: An Unemployment Case Analysis
March 2018 Briefly: An Unemployment Case Analysis

Case Analysis: Claimant Discharged for Theft Background The claimant was discharged for theft. She was allo...