AniRec : Personalized Anime Recommendation System Using Hybrid Filtering

Abstract: The popularity of anime continues to rise rapidly, thus making it difficult to locate what one is interested in. With thousands of animes in a variety of genres, people tend to spend too much time looking for something worth watching, leading them to end up picking something trendy instead. AniRec, on the other hand, takes into consideration your preferences without you realizing it. Using innovative techniques, it analyses your preferences and suggests anime titles suitable for you. AniRec works on how people watch while also breaking down what shows are like. Starting off, the platform gathers details about each anime and gets them ready for deeper review. Gaps in info get filled, odd entries fixed so everything lines up right. Once tidy, things like story kind, score, or format get pulled out and turned into codes that allow the system to compare different anime effectively. By using features of each show, comparisons begin. Instead of relying on user ratings, the method checks how much one series looks like another in themes, genres, or art style. By measuring angles between data points, Cosine Similarity finds which titles sit closest in trait space. Once scored, rankings form; higher values mean stronger links to what the viewer liked before. What appears at the front of results reflects tight alignment in characteristics. Seen first in the browser view, those picks guide eyes toward fresh options shaped by personal taste. Python is used for system's backend, with Flask. Pandas handles data work while calculations rely on NumPy. Machine learning steps come alive through Scikit-learn tools. Stored inside the system's database are things like what users rate, choose, watch. Because of these details, suggestions get better bit by bit, learning exactly what each person likes. Overall, AniRec provides you personalized anime recommendation as per your preference and previous records.

Keywords: Anime Recommendation, Cosine Similarity, Machine Learning, Personalization, Hybrid Filtering.