Recommend Anime to users
Each piece of streaming content has its own viewers and rating based on viewer preference, genre etc. When viewers enjoy/dislike the content, they provide positive/negative reviews accordingly. While thinking of what to watch next, viewers spend hours together scrolling through different content unable to find anything they enjoy. To improve the streaming experience and increase revenue and the time spent on a website, businesses use the review information to provide recommendations to their users based on their preferences and needs. Anime is a hand-drawn computer-generated animation originating from Japan. The popularity of anime is growing faster than ever, especially internationally.
The objective of this project is to create a recommendation model using unsupervised machine learning techniques and data mining tasks like clustering and frequent patterns that can recommend anime to people based on their interests.
We plan to use models like:
- Popularity based recommender system: sorting weighted rating of all anime.
- Collaborative filtering: give recommendations based on preference of other similar users or based on the similarity between items and the rating history of the user.
- Clustering techniques such as KMeans to cluster different anime based on genre, which can then be recommended to users.
The dataset we will use is from Kaggle called “Anime Recommendations Database” which consists of data collected from 76,000 users at myanimelist.net https://www.kaggle.com/datasets/CooperUnion/anime-recommendations-database