Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Aug 25;17(8):e0273486.
doi: 10.1371/journal.pone.0273486. eCollection 2022.

Deep transfer learning with multimodal embedding to tackle cold-start and sparsity issues in recommendation system

Syed Irteza Hussain Jafri  1   2 , Rozaida Ghazali  1 , Irfan Javid  1   2 , Zahid Mahmood  3 , Abdullahi Abdi Abubakar Hassan  1
Affiliations

Deep transfer learning with multimodal embedding to tackle cold-start and sparsity issues in recommendation system

Syed Irteza Hussain Jafri et al. PLoS One. .

Abstract

Recommender systems (RSs) have become increasingly vital in the modern information era and connected economy. They play a key role in business operations by generating personalized suggestions and minimizing information overload. However, the performance of traditional RSs is limited by data sparseness and cold-start issues. Though deep learning-based recommender systems (DLRSs) are very popular, they underperform when considering rating matrices with sparse entries. Despite their performance improvements, DLRSs also suffer from data sparsity, cold start, serendipity, and generalizability issues. We propose a multistage model that uses multimodal data embedding and deep transfer learning for effective and personalized product recommendations, and is designed to overcome data sparsity and cold-start issues. The proposed model includes two phases. In the first-offline-phase, a deep learning technique is implemented to learn hidden features from a large image dataset (targeting new item cold start), and a multimodal data embedding is used to produce dense user feature and item feature vectors (targeting user cold start). This phase produces three different similarity matrices that are used as inputs for the second-online-phase to generate a list of top-n relevant items for a target user. We analyzed the accuracy and effectiveness of the proposed model against the existing baseline RSs using a Brazilian E-commerce dataset. The results show that our model scored 0.5882 for MAE and 0.4011 for RMSE which is lower than baseline RSs which indicates that the model achieved an improved accuracy and was able to minimize the typical cold start and data sparseness issues during the recommendation process.

PubMed Disclaimer

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1
Fig 1. The architecture of the proposed DTLME model for feature engineering.
Fig 2
Fig 2. Generic VGG16 model.
Fig 3
Fig 3. Pre-trained VGG16 model used for transfer learning.
Fig 4
Fig 4. Multimodal embedding for feature learning.
(a) User feature learning; (b) Item feature learning.
Fig 5
Fig 5. Sparse rating matrix for item pi for user ui.
Fig 6
Fig 6. User-item affinity matrix.
Fig 7
Fig 7. User profile, similarity calculation and top-n recommendation.
Fig 8
Fig 8. Accuracy measures for VGG-16 model.
Fig 9
Fig 9. Loss for VGG-16 model.
Fig 10
Fig 10. The average rating for products in BE-dataset.
Fig 11
Fig 11. MAE for BE-dataset @100 epochs.
Fig 12
Fig 12. MAE for BE-dataset @20 epochs.
Fig 13
Fig 13. Precision, recall and F-1 measures for DTLME model.
Fig 14
Fig 14. Comparative analysis of DTLME model with CSSVD, BPR and TF baseline RSs.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Resnick P. and Varian H. R. "Recommender system”, Communications of the ACM, vol. 40, no. 3, pp. 56–58, 1997.
    1. Xuesong Zhao. “A Study on E-commerce Recommender System Based on Big Data”, G:\UTHM Research Work\Research UTHM\Recommender Systems based on Deep Learning\A Study on E-commerce Recommender System Based on Big Data.pdfIEEE 4th International Conference on Cloud Computing and Big Data Analytics, 2019.
    1. Isinkaye F.O., Folajimi B. Y.O.A. Ojokoh. “Recommendation systems: Principles, methods and evaluation”, Egyptian Informatics Journal, Vol. 16, 2015, pp. 261–273.
    1. Da’u Aminu, Salim Naomie. “Recommendation system based on deep learning methods: a systematic review and new directions”, Springer Nature B.V., 2020.
    1. Zhang Shuai, Yao Lina, Sun Aixin and Tay Yi. “Deep Learning Based Recommender System: A Survey and New Perspectives”, ACM Comput. Surv. 52, 2019.