Using Machine Learning to Predict Driver Risk [Webinar]

Hear from our own data geeks Ayodele Odubela, Data Scientist, and Daniel Laney, Lead Data Scientist, as they speak to using machine learning as a way to predict driver risk. Whether you were able to catch the webinar live or are interested instead in watching it on-demand, check out some of the highlights from the webinar, including:

• In a nutshell, data science is a mix of discovering and massaging data, obtaining insight from data, creating data products and communicating relevant business insights from data, providing those with the data the ability to make informed decisions.
• Model training focuses on finding an optimal solution
• There are common machine learning models that exist, one of which is called neural networks.
  • Neural networks learn from large amounts of data and are able to identify underlying relationships that humans cannot.
  • These neural networks mimic the operations of a human brain to recognize the relationships between vast amounts of data.
• Deep learning is using multiple hidden layers in a neural network.
  • They are created by stacking several hidden layers and can provide data scientists with high accuracy information.
  • This deep learning works well with unstructured data, like images and more.
• There are pros and cons that go alongside everything to develop a robust risk score.
  • As a credit reporting agency, we are legally obligated to disclose how decisions are made to consumers who were to dispute an action.
    • Typically, this means that we do not use the more complex models due to the lack of explain-ability that exists.