The solutions are offered via the CreditRobo platform.
What is machine learning?
Machine learning is the use of powerful computers and software tools that work on large sets of data to discover patterns and make predictions. Advantages of machine learning are:
- Availability of cheap and powerful computing resources
- Mature algorithms and tools
- The existence of abundance of data
Especially the latter is important when it comes to scoring the underserved consumers. Consumers are increasingly generating tons of data. Every Facebook login, every mobile phone call, credit card swipe, even viewing this web page generates a stream of data about who looked, when, and where they came from. As more of lives and work migrate to digital information (such as accounting systems online, point of sale terminals, order tracking, even customer reviews) there’s simply more data for machine learning. And much of this data can assist in scoring creditworthiness. Often even better and more accurate than only looking at the financial data and -history of a person.
Why is CreditRobo better?
Traditional credit score models or default models use a handful of variables. Methods used include logistic regression, decision trees and traditional statistical analysis. AdviceRobo tackles credit scoring and default prediction from an entirely different approach, by using machine learning and large-scale big data analytics on structured and unstructured data and combining this with our deep financial knowledge and behavioural science research. This leads to advanced feature development and the basis to create algorithms via machine learning. This has many advantages:
- Better scores and better predictions through multiple data sets: using unstructured data created through the Internet-of-Things data explosion
- Actionable insights: scores and predictions on an individual customer level
- Continuous improvement through machine learning
This leads to actual, individual, and more accurate credit scores and significant better risk modelling for default and churn.
Machine learning benefits
Discover the machine learning benefits for risk management: CreditRobo platform versus traditional risk tools (move your finger or mouse over the picture).
Why is machine learning better?
The computer handles a lot of data about a lot of people.
Machine learning can recognize patterns from one person.
And it can calculate endless scenarios for trillions of people in a split second.
Leading to predictions of individual behaviour for all these individual people.
Why is more data better?
So far, credit models at lenders use a limited amount of data points. However, adding extra data points will make the modelling more refined and more specific on an individual level, leading to more refined and accurate predictions. Therefore, CreditRobo does not only use data provided by the lender, but adds data from multiple sources: data that will support better predictions of the customer, data discovered through scientific research and extensive tests. Examples are:
- Social network data
- Mentality data
- Motivational profile data
Data security and opt-in policies are strictly followed and executed.
How the platform works?
Steps of the completely automated process of the AdviceRobo platform.
More Questions and Answers
About Model Quality
- How does AdviceRobo test stability and fitting problems?
- How does AdviceRobo validate its models?
- How can AdviceRobo ‘prove’ that the models work well?
- How do AdviceRobo models perform?
About Machine Learning
- How does AdviceRobo explain the results of Machine Learning?
- What Machine Learning techniques does AdviceRobo use?
- How does AdviceRobo validate the data quality?
- Which data does AdviceRobo use in their models?
- What does AdviceRobo do when data quality is (too) poor?
About Data Protection