Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

[RE] Modeling Personalized Item Frequency Information for Next-basket Recommendation

Sławomir Garcarz, Avik Pal, Pim Praat

TL;DR

The experimental results confirmed that the reproduced model, TIFU-KNN, outperforms the baseline model, Personal Top Frequency, on various datasets and metrics.

Abstract

This paper focuses on reproducing and extending the results of the paper: "Modeling Personalized Item Frequency Information for Next-basket Recommendation" which introduced the TIFU-KNN model and proposed to utilize Personalized Item Frequency (PIF) for Next Basket Recommendation (NBR). We utilized publicly available grocery shopping datasets used in the original paper and incorporated additional datasets to assess the generalizability of the findings. We evaluated the performance of the models using metrics such as Recall@K, NDCG@K, personalized-hit ratio (PHR), and Mean Reciprocal Rank (MRR). Furthermore, we conducted a thorough examination of fairness by considering user characteristics such as average basket size, item popularity, and novelty. Lastly, we introduced novel $β$-VAE architecture to model NBR. The experimental results confirmed that the reproduced model, TIFU-KNN, outperforms the baseline model, Personal Top Frequency, on various datasets and metrics. The findings also highlight the challenges posed by smaller basket sizes in some datasets and suggest avenues for future research to improve NBR performance.

[RE] Modeling Personalized Item Frequency Information for Next-basket Recommendation

TL;DR

The experimental results confirmed that the reproduced model, TIFU-KNN, outperforms the baseline model, Personal Top Frequency, on various datasets and metrics.

Abstract

This paper focuses on reproducing and extending the results of the paper: "Modeling Personalized Item Frequency Information for Next-basket Recommendation" which introduced the TIFU-KNN model and proposed to utilize Personalized Item Frequency (PIF) for Next Basket Recommendation (NBR). We utilized publicly available grocery shopping datasets used in the original paper and incorporated additional datasets to assess the generalizability of the findings. We evaluated the performance of the models using metrics such as Recall@K, NDCG@K, personalized-hit ratio (PHR), and Mean Reciprocal Rank (MRR). Furthermore, we conducted a thorough examination of fairness by considering user characteristics such as average basket size, item popularity, and novelty. Lastly, we introduced novel -VAE architecture to model NBR. The experimental results confirmed that the reproduced model, TIFU-KNN, outperforms the baseline model, Personal Top Frequency, on various datasets and metrics. The findings also highlight the challenges posed by smaller basket sizes in some datasets and suggest avenues for future research to improve NBR performance.
Paper Structure (20 sections, 2 equations, 5 figures, 6 tables)

This paper contains 20 sections, 2 equations, 5 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Models
  4. Personal Top Frequency
  5. TIFU-KNN
  6. $\beta$-VAE
  7. Datasets
  8. Metrics
  9. Fairness
  10. Experimental Setup/Implementation details
  11. Experimental Results
  12. RQ1: Reproduction of TIFU-KNN results
  13. RQ2: Model performance on different datasets
  14. RQ3: Evaluation beyond recall and NDCG
  15. RQ4: Fairness analysis of TIFU-KNN
  16. ...and 5 more sections

Figures (5)

  • Figure 1: Relationship between Model Performance and Basket Size. The yellow bars represent the values of Recall@10 on the left y-axis, while the blue line depicts the number of users on the right y-axis corresponding to different basket sizes along the x-axis.
  • Figure 2: Relationship between Model Performance and Item Popularity. The yellow bars represent the values of Recall@10 on the left y-axis, while the blue line depicts the number of users on the right y-axis corresponding to the percentage of popular items in users' baskets along the x-axis.
  • Figure 3: Relationship between Model Performance and Novelty. The yellow bars represent the values of Recall@10 on the left y-axis, while the blue line depicts the number of users on the right y-axis corresponding to the percentage of unseen items in the test basket along the x-axis.
  • Figure 4: Model architecture for the Next-basket Recommendation task using $\beta$-VAE and an MLP Predictor network.
  • Figure 5: Relationship between $\beta$-VAE Model Performance and Novelty. The yellow bars represent the values of Recall@10 on the left y-axis, while the blue line depicts the number of users on the right y-axis corresponding to the percentage of unseen items in the test basket along the x-axis.