Video: Machine Learning and AI for the Legal Profession — An Explainer on Measuring Prediction Accuracy

Machine learning and artificial intelligence are extremely helpful tools, but the newness of these tools can make them intimidating. By answering questions about key concepts, I hope to make these useful tools more approachable.

For example, if you are using a machine learning/artificial intelligence application such as Technology Assisted Review (TAR), how do you even know if the system is working properly?

The first step is understanding the accuracy of the machine learning system’s predictions. In this video, we introduce three key concepts in measuring a machine learning system’s accuracy: they are Prevalence, Recall and Precision. We also introduce the Confusion Matrix, a chart that helps you understand inaccurate predictions such as false positives and false negatives.

About DiscoveryBriefs

DiscoveryBriefs is a video channel dedicated to improving the justice system through educational content for the legal community about legal technology and eDiscovery. You can find it on Medium or YouTube. We are eDiscovery consultants passionate about changing the status quo and our company, Precision Discovery, supports us in our mission.

About The Author

Kinny Chan is the Chief Strategy Officer for Precision Discovery. He is a lawyer and eDiscovery enthusiast. He enjoys taking complex challenges and explaining them in simple and understandable terms. He is inspired by the intersection between technology, business and the law. You can connect with him on Twitter and LinkedIn.