Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Green Recommender Systems: Optimizing Dataset Size for Energy-Efficient Algorithm Performance

Ardalan Arabzadeh, Tobias Vente, Joeran Beel

TL;DR

This paper investigates the potential for energy-efficient algorithm performance by optimizing dataset sizes through downsampling techniques in the context of Green Recommender Systems, suggesting that strategic dataset reduction can decrease computational and environmental costs without substantially compromising recommendation quality.

Abstract

As recommender systems become increasingly prevalent, the environmental impact and energy efficiency of training large-scale models have come under scrutiny. This paper investigates the potential for energy-efficient algorithm performance by optimizing dataset sizes through downsampling techniques in the context of Green Recommender Systems. We conducted experiments on the MovieLens 100K, 1M, 10M, and Amazon Toys and Games datasets, analyzing the performance of various recommender algorithms under different portions of dataset size. Our results indicate that while more training data generally leads to higher algorithm performance, certain algorithms, such as FunkSVD and BiasedMF, particularly with unbalanced and sparse datasets like Amazon Toys and Games, maintain high-quality recommendations with up to a 50% reduction in training data, achieving nDCG@10 scores within approximately 13% of full dataset performance. These findings suggest that strategic dataset reduction can decrease computational and environmental costs without substantially compromising recommendation quality. This study advances sustainable and green recommender systems by providing insights for reducing energy consumption while maintaining effectiveness.

Green Recommender Systems: Optimizing Dataset Size for Energy-Efficient Algorithm Performance

TL;DR

This paper investigates the potential for energy-efficient algorithm performance by optimizing dataset sizes through downsampling techniques in the context of Green Recommender Systems, suggesting that strategic dataset reduction can decrease computational and environmental costs without substantially compromising recommendation quality.

Abstract

As recommender systems become increasingly prevalent, the environmental impact and energy efficiency of training large-scale models have come under scrutiny. This paper investigates the potential for energy-efficient algorithm performance by optimizing dataset sizes through downsampling techniques in the context of Green Recommender Systems. We conducted experiments on the MovieLens 100K, 1M, 10M, and Amazon Toys and Games datasets, analyzing the performance of various recommender algorithms under different portions of dataset size. Our results indicate that while more training data generally leads to higher algorithm performance, certain algorithms, such as FunkSVD and BiasedMF, particularly with unbalanced and sparse datasets like Amazon Toys and Games, maintain high-quality recommendations with up to a 50% reduction in training data, achieving nDCG@10 scores within approximately 13% of full dataset performance. These findings suggest that strategic dataset reduction can decrease computational and environmental costs without substantially compromising recommendation quality. This study advances sustainable and green recommender systems by providing insights for reducing energy consumption while maintaining effectiveness.

Paper Structure

This paper contains 11 sections, 1 equation, 2 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Datasets & Preprocessing
  5. Data Splitting and Downsampling
  6. Algorithms and Evaluation
  7. Results & Conclusion
  8. Observations
  9. Interpretation
  10. Conclusion
  11. Acknowledgment

Figures (2)

  • Figure 1: nDCG@10 scores for each algorithm trained on varying portions of the datasets. The horizontal axis shows the percentage of the full training set, where 100% equals 80% of the total dataset, with other percentages relative to this.
  • Figure 2: Distribution of nDCG@10 Scores Across Four examined Datasets for Two Distinct Algorithm Groups at 50% and 100% Dataset Utilization.