International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 05 | May 2023
p-ISSN: 2395-0072
www.irjet.net
MOVIE RECOMMENDATION SYSTEM USING COLLABORATIVE FILTERING Dr SVG Reddy1, Putchakayala Meher Sowjanya2, Annem Pavan Kumar Reddy3, Bavisetti Sai Saketh4,Lekkala Yaswanth Kumar5, Karri Viswa Abhiram Reddy6 1Associate Professor, Department of Computer Science and Engineering, GITAM University, Visakhapatnam, AndhraPradesh, 530045, India. 2,3,4,5,6B. Tech Student, Department of Computer Science and Engineering, GITAM University, Visakhapatnam, AndhraPradesh, 530045, India. ------------------------------------------------------------------------***------------------------------------------------------------------------INTRODUCTION Abstract Because of the richness of knowledge amassed up to the twenty- first century and the increasing rate at which information is gushing over the internet, there is a great deal of confusion over what to consume and what not to ingest. Even on YouTube, there are always a ton of videos available if you want to see one about a particular concept. Since the results are suitably ranked, there might not be much of a problem right now, but what if they weren't? In this situation, we would undoubtedly spend a lot of time looking for the best movie that fits us and meets our needs. When you look for something on a website, this is what happens the algorithm could be able to offer you recommendations the next time you visit a particular website without you even having to search. This feature is fascinating, isn't it? A recommender system's main responsibility is to present the user with the items that are the most pertinent. Recommendation engines are used by Amazon, Flipkart, Netflix, and YouTube to propose videos, items, movies, and other content. Regardless of what you do on these websites, a system is in place that tracks your activity and makes suggestions for activities or products that you are very likely to be interested in. This research paper addresses the logic behind movie suggestions, conventional movie recommendation systems, problems with conventional movie recommendation systems, and a suggested fix for an AI-based personalized movie recommendation system. There are already a lot of wellknown movie recommendation datasets available on Kaggle and other sites. Movie lens, the TMDB Movie Dataset, and the Netflix dataset are a few of the wellknown datasets. Websites like Netflix, Amazon Prime, and others employ movie recommendations to increase revenue or profits by eventually enhancing the user experience. In reality, in 2009 Netflix conducted a competition with a prize pool of about $1 million ($1M) for creating at least 10% upgrades to the current system. As was previously mentioned, we have access to a large amount of data, and since we are not interested in everything that is available to us, we must filter itin order to use it.
Nowadays, the recommendation system is crucial and is used by many significant applications. The proliferation of applications, the emergence of a global village, and the availability of a large amount of information are the results of the recommendation system. An overview of the approaches and techniques developed by this study introduces a recommender system based on collaborative filtering, which has evolved over time to include contentbased and hybrid approaches. The study examines various techniques used for collaborative filtering, including matrix factorization, user- based and item-based recommendation. It also serves as a guide for future research in this field. By analyzing user preferences and eating habits, we extract aspect-based ratings from reviews and recommend reviews accordingly. Furthermore, the proposed movie recommendation system is tested against multiple evaluation criteria and performs better than existing approaches. Through extensive research and review of numerous papers, it was discovered that content-based filtering typically relies on a single technique for converting text into vectors for recommendations and a single approach to detect similarity between vectors. Think of it as a hybrid strategy that only uses a content-based filtering method. The way we search for exciting items has revolutionized today's recommendation systems. It is vital to advise mobile users to download OTT movie apps. To help you choose suitable movies, it thoroughly summarizes user preferences, reviews, and[SS1] feelings. It requires absolute accuracy and timeliness, yet this information filtering technique predicts user preferences. A recommender system is a system that aims to predict or filter selections according to user choices. OTT platforms, search engines, articles, music, videos, and other similar platforms are typical applications of recommender systems. In this work, we tend to provide a movie design system based on a common approach. It's a supported collaborative filtering strategy that takes user-provided knowledge, analyses it, and then suggests the most appropriate movie for the users.
© 2023, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 154