Towards a more sustainable consumption using recommender systems
(2025)
Files
ThesisFINAL.pdf
Embargo Access until 2028-01-01 - Adobe PDF
- 9.74 MB
AppendicesFINAL.zip
Embargo Access until 2028-01-01 - Unknown
- 13.06 MB
Details
- Authors
- Supervisors
- Fouss, FranƧois
- Abstract
- Modern consumption patterns have severe environmental consequences, yet aligning consumer behaviour with sustainable intentions remains a challenge. Recommender systems, which significantly shape purchasing decisions, are often blamed for promoting unsustainable consumption. However, could they instead be leveraged to drive sustainability? This thesis explores how recommender systems can be adapted to encourage sustainable consumption. First, a machine learning classification model is developed to efficiently assess product sustainability based on life cycle data. Next, various in- and post-processing strategies are tested to increase the visibility of sustainable products in recommendation lists without compromising accuracy. Finally, a controlled online store experiment compares the impact of sustainability-oriented recommender systems and sustainable consumption communication on consumer behaviour. The findings reveal that sustainability-oriented recommender systems effectively increase the purchase of sustainable products by enhancing awareness and reducing search effort, whereas communicationbased nudges have a more limited effect. This research contributes to the growing field of responsible AI by demonstrating how recommender systems can be designed to align commercial objectives with environmental sustainability.
- Affiliations
Citations
Satinet, C. (2025). Towards a more sustainable consumption using recommender systems. https://hdl.handle.net/2078.5/246500