Accuracy analysis of recommendation system using singular value decomposition

Akter, Naznin; Hoque, A. H. M. Sajedul; Mustafa, Rashed; Chowdhury, Md. Shahnur Azad

doi:10.1109/iccitechn.2016.7860232

Cited by 8 publications

(3 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use SVD algorithm on e-commerce's product recommendation (Akter et al, 2017) generate two predictive matrices according to its initial values. A product catalog is a core element of all e-commerce products from which the users can find that element in two ways.…”

Section: Literature Reviewmentioning

confidence: 99%

“…The matrix factorization technique uses a matric structure, namely rows as users and columns as items. This technique factorizes 𝑚 × 𝑛 of A matrix into three matrices; the form of singular value decomposition is shown below (Akter et al, 2017;Sahoo et al, 2019).…”

Section: Singular Value Decomposition (Svd)mentioning

confidence: 99%

“…The S is an 𝑟 × 𝑟 sized diagonal matrix, possessing all singular values of matrix A's main diagonal. Stages in conducting SVD calculation are defined below (Akter et al, 2017;Bokde et al, 2015;Guan et al, 2016;Vozalis & Margaritis, 2007). 3.…”

Section: 𝑨 = 𝑼 × 𝑺 × 𝑽 𝑻mentioning

confidence: 99%

See 2 more Smart Citations

Collaborative Filtering Approach: Skincare Product Recommendation Using Singular Value Decomposition (Svd)

Nissa

Primandari

Thalib³

2023

Medstat

View full text Add to dashboard Cite

The recommendation system provides recommendations for something, be it goods, songs, or movies. The term system is not limited to a service system but concerns a model that can provide recommendations. With recent technological advances, many companies provide various skincare products because current generations are increasingly aware of self-care. With various choices, someone may experience confusion in determining the product they want to buy. Therefore, we need a system that can provide product recommendations run on any platform we use. The most common method for recommendation systems often comes with Collaborating Filtering (CF) where it relies on the past user and item dataset. The singular value decomposition (SVD) method uses a matrix factorization technique that predict the user's rating based on historical ratings. The measurement of the model's accuracy is the RMSE average of 1.01276, indicating that this value results from the best parameters. The results focus on showing skincare product recommendations to users sorted based on rating predictions.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: Singular Value Decomposition (Svd)mentioning

confidence: 99%