Episode 4: Focus on the right prediction metrics

In this episode, Giannis Stoitsis, CTO and partner at Agroknow (Coordinator partner of the BigDataGrapes project), presents the right prediction metrics that we are calling to choose when we want to check if our prediction algorithm is working properly.

Predictive analytics sounds cool but training, deploying, and operating a reliable predictive analytics solution for food safety is not as simple as it may initially seem. The deep-understanding of predictive analytics work and the way they assist in critical-business decisions can be difficult to comprehend.  

As it has been mentioned at the 2nd episode of our video series, one of the most important aspects in predictive analytics is to define the specific problem that we want to solve from the beginning, in order to choose the right prediction metrics at the end. 

In order to be able to answer “How can we measure if a prediction algorithm is working properly?”, we have to understand that there are a plethora of prediction metrics that can be used. For this reason, in our video there are examples on metrics focused on coverage and accuracy, based on the problem that we have to solve.

Stay tuned and find out in our next video How to integrate predictions in your work flows!

These short video series are a part of the BigDataGrapes project. The project has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement number 780751.