Foods (Mar 2023)

Enhancing Shelf Life Prediction of Fresh Pizza with Regression Models and Low Cost Sensors

  • Paul Wunderlich,
  • Daniel Pauli,
  • Michael Neumaier,
  • Stephanie Wisser,
  • Hans-Jürgen Danneel,
  • Volker Lohweg,
  • Helene Dörksen

DOI
https://doi.org/10.3390/foods12061347
Journal volume & issue
Vol. 12, no. 6
p. 1347

Abstract

Read online

The waste of food presents a challenge for achieving a sustainable world. In Germany alone, over 10 million tonnes of food are discarded annually, with a worldwide total exceeding 1.3 billion tonnes. A significant contributor to this issue are consumers throwing away still edible food due to the expiration of its best-before date. Best-before dates currently include large safety margins, but more precise and cost effective prediction techniques are required. To address this challenge, research was conducted on low-cost sensors and machine learning techniques were developed to predict the spoilage of fresh pizza. The findings indicate that combining a gas sensor, such as volatile organic compounds or carbon dioxide, with a random forest or extreme gradient boosting regressor can accurately predict the day of spoilage. This provides a more accurate and cost-efficient alternative to current best-before date determination methods, reducing food waste, saving resources, and improving food safety by reducing the risk of consumers consuming spoiled food.

Keywords