Table 1 Summary of the studied works.

Ahuja et al.²⁴

Develop movie recommender system

K-Means Clustering, K-Nearest Neighbor

Relies on basic clustering and similarity-based methods, potentially limiting recommendation accuracy.

Katarya and Verma²⁵

Improve movie recommendation system

Cuckoo search optimization, K-Means Clustering

Optimization-based approach might be sensitive to parameter tuning, and relies on clustering for user grouping.

Thakker et al.²⁶

Analyze CF algorithm for movie recommendation

User-based and item-based CF, Machine learning algorithms

Focuses on CF algorithms, not proposing a new method, and lacks exploration of deep learning techniques.

Gupta et al.²⁷

Enhance movie recommendation accuracy

K-NN algorithms, Collaborative filtering

Relies on traditional similarity-based methods, without incorporating rich user behavior information.

Airen and Agrawal²⁸

Explore KNN variations for movie recommendation

KNN algorithms with different similarity measures

Limited to KNN variations, not addressing broader recommendation techniques or incorporating advanced machine learning.

Chauhan et al.²⁹

Improve movie recommendation with user sentiment

Combined CF, CB filtering, sentiment analysis

Does not consider deep learning or advanced machine learning techniques, and might face challenges in handling complex user preferences.

Tahmasebi et al.³⁰

Develop hybrid movie recommender system

Deep autoencoder network, CF, CB filtering, social influence

While incorporating social influence, it might not fully capture complex user-item interactions.

Lee et al.³¹

Improve movie recommender system with sentiment and emotion

Network-based model with sentiment, emotion information

Relies on sentiment and emotion information, which might not fully capture diverse user preferences.

Sun et al.³²

Develop social-aware group recommendation

Group preference model, social-aware tolerance and altruism model

Focuses on group recommendation, not individual preferences, and might not generalize well to diverse group settings.

Yadav et al.³³

Develop movie recommender system with big data techniques

Big data clustering, PCA

While using big data techniques, the model might be complex and computationally expensive.

Abderrahmane et al.³⁴

Develop movie recommender system

KNN, Collaborative filtering

Relies on basic similarity-based methods, without incorporating rich user behavior information or advanced machine learning techniques.

Ambikesh et al.³⁵

Develop movie recommender system with Harris hawks optimization

Harris hawks optimization, K-means clustering

Optimization-based approach might be sensitive to parameter tuning, and clustering-based methods might not capture complex user preferences.

Zubi et al.³⁶

Develop efficient hybrid movie recommender system

Association rule mining, KNN classification

While using association rules, the model might not fully capture user preferences and might be limited to frequent itemsets.

Jain and Essah³⁷

Develop movie recommendation system for online streaming

Collaborative filtering (cosine similarity, KNN, SVD)

Relies on basic collaborative filtering techniques, without incorporating additional features or advanced machine learning.

Sharma et al.³⁸

Develop hybrid recommender system with sparrow clustering

Sparrow clustered recommender system

The effectiveness of the sparrow clustering algorithm for recommendation tasks is yet to be fully established.

El Alaoui et al.³⁹

Improve item recommendations using GNNs.

Deep GraphSAGE with Jumping Knowledge (JK) connections and Ordinal Aggregation Network (OAN).

Primarily item-based, may not generalize to diverse contextual data.

El Alaoui et al.⁴⁰

Compare matrix factorization (SVD) and deep MLPs for collaborative filtering.

Singular Value Decomposition (SVD) vs. Multilayer Perceptron (MLP).

Limited dataset and evaluation metrics, potential generalizability issues.

Rendle et al.⁴¹

Compare neural collaborative filtering (NCF) and matrix factorization (MF).

Learned similarities (MLP) vs. Dot products.

Computational efficiency focus may limit applicability in other contexts.

El Alaoui et al.⁴²

Find the best recommendation approach for scientific articles.

Collaborative filtering, content-based, and hybrid methods.

Evaluation relies on classical IR metrics, may not fully capture user satisfaction.

El Alaoui et al.⁴³

Improve contextual recommendations.

Dynamic Graph Attention Network with Adaptive Edge Attributes (DGAT-AEA).

Model complexity may pose real-time application challenges.

Ziaee et al.⁴⁴

Enhance movie recommendation systems.

Heterogeneous GNN and GAE-based model (MoRGH).

Focus on RMSE may not capture other recommendation qualities.

El Alaoui et al.⁴⁵

Improve social recommendation systems.

Heterogeneous Graph Attention Networks.

Focus on social recommendations may limit broader applicability.

El Alaoui et al.⁴⁶

Compare filtering methods for scientific research article recommendations.

Collaborative filtering, content-based, and hybrid methods.

Findings may be specific to the scientific literature domain.