Shelfion

Estimating Additional Data Needs for Robust Machine Learning Models

About the project

Based on academic papers we made a fictive dataset where we analysed how many additional (real) datapoints we would need when introducing a new random variable for our machine learning model to remain stable. This enabled Shelfion to estimate how many data they would need to gather out of their lab, as this is highly expensive and tedious data to gather.

Project members

Project Responsible (Board):

Arthur Hollevoet

Marie Verlinden

Project Leader:

Ewout Devloo

Project Members:

Henri Mertens

Baptiste Van der Eecken

Seppe Van Campe

Emile Eeckhoudt

Willem De Mesel

About Shelfion

Shelfion is your specialist in microbial shelf-life determination for your foods. Whether you are a new product developer, working on making your packages more sustainable or dealing with existing spoilage challenges, our innovative tools are there to support you.

About Crunch Analytics

Founded in 2016, Crunch Analytics is a Belgian data and AI consultancy that helps businesses harness data science, machine learning, and AI to drive smarter decision-making. Specializing in retail, e-commerce, and consumer goods, they develop tailored analytics solutions, optimize pricing strategies, and provide training to empower teams in using data effectively.