This console application is part of the practical assignment for the INFDTA02-1 course at the Rotterdam University of Applied Sciences. It will take a dataset containing at least a (UserID, ProductID, Rating) and is able to use multiple approaches (like Collaborative or Content-based filtering) to predict the rating a user would give to an item.
The application currently has the following features:
- Cosine similarity
- Euclidean similarity
- Manhattan similarity
- Pearson correlation coefficient
- Calculate nearest neighbours based on a target user, a threshold and k (amount of neighbours to consider)
- And predict a rating
- Adjusted Cosine Similarity
- And predict a rating
- Slope One
- And predict a rating
- Calculate sparcity in dataset
Setting up is easy, open and run the solution. The application will guide you through the options. Feel free to play around with different datasets and algorithms.