Skip to main content
You have permission to edit this article.

VIDEO | Distinguished Career Professor, Carnegie Mellon University: ML Is One of Many Disciplines Needed for Prediction Systems

  • Updated

(US and Canada) Rayid Ghani, Distinguished Career Professor at Carnegie Mellon University, speaks with Maria Espona, Professor — ArgIQ and CDO Magazine Editorial Board Member, in a video interview about using machine learning techniques for prediction and the role of accuracy and metrics in prediction.

According to Ghani, it is possible to use machine learning to predict things, but specific key questions need to be answered. They are:

  • Is it better than other methods?

  • Is it robust and accurate?

  • What actions is it trying to inform?

He explains that the goal of the prediction model is to solve a larger problem of which the prediction is a smaller component. Ghani elaborates that machine learning is one of many required disciplines. Such projects require expertise in fields like computer science, statistics, mathematics, economics, policy, and psychology.

He further stresses the accuracy aspect and says that if a model is being used to figure out the capacity of a hospital, it has to be as accurate and equitable as possible. In turn, there has to be a metric that matches the use case. The system has to do well on that metric, Ghani says.

CDO Magazine appreciates Rayid Ghani for sharing his insights and data success stories with our global community.

See more from Rayid Ghani