ANALYSIS AND DEVELOPMENT OF MOVIE RECOMMENDATION SYSTEMS USING ARTIFICIAL INTELLIGENCE
Vernes Vinčevića, Nešad Krnjićb, Bakir Čičakc
aUniversity of Vitez, Travnik, Bosnia and Herzegovina, vernes.vin@gmail.com
bUniversity of Vitez, Travnik, Bosnia and Herzegovina, nesad.krnjic@unvi.edu.ba
cUniversity of Vitez, Travnik, Bosnia and Herzegovina.
Abstract
The paper aims to analyze and implement an optimal hybrid movie recommendation system using artificial intelligence on a large dataset. The research question is how the combination of different artificial intelligence filtering algorithms affects the accuracy and performance of recommendations in the context of big user data. The methodological framework includes processing large datasets (Big data), implementing different recommendation algorithms, and evaluating the performance of the models using relevant metrics (e.g. RMSE, MAE, precision, responsiveness). A dataset with millions of entries was used to compare the performance of three key techniques: content-based filtering, collaborative filtering, and a hybrid model that combines the advantages of both approaches. The main findings confirm that the hybrid approach, integrating the best features of the standalone methods, provides the highest accuracy and stability in predicting user preferences in a Big data environment. Our hybrid model showed an RMSE (Root Mean Squared Error) of 0.89. A lower value is better, as it means that the model makes fewer errors. Its average RMSE (0.89) is smaller than the RMSE for KNN (0.92). The finding suggests that SVD is approximately 2.5% more accurate (better) than KNN. In terms of testing speed, SVD is 33.6 times faster than KNN, as the average testing time for SVD was 0.13 seconds, and for KNN it was 4.37 seconds, indicating a much better model based on the SVD algorithm. The results showed that the choice of algorithm and the quality of input data have an important impact on the success of the system. The contribution of the paper provides empirical confirmation of the superiority of hybrid architectures in the domain of recommender systems and modeling, but also as a practical guide for engineers and a theoretical basis for future research in the field of personalized artificial intelligence solutions, which directly addresses the challenges faced by companies such as Netflix, Amazon, YouTube, and applications in other industrial systems, and Big data analytics. The contribution of the work is also reflected in pointing out the importance of proper data processing and selection in the development of intelligent recommendation systems.
Keywords: Artificial intelligence; Machine learning algorithms; Big data; Recommendation system; Data analysis; Company management
To be cited as: Vinčević, V., Krnjić, N. and Čičak, B. (2025) Analysis and development of movie recommendation systems using artificial intelligence. International Journal of Management Courses, 27(1), pp. 34-42.