Car Recommendation System for Dealers in Different European Countries

Recommendation systems are widely used in all walks of life, This paper is about the design of car recommendation of car rating for different European countries using collaborative filtering(CF). Compared with content-based filtering, collaborative filtering has the advantage that it can recommend items according to the historical information of each user, which has nothing to do with the content attributes of the item itself, which is the reason why we choose CF as our model to develop the recommendation system. The user model is designed using the brand, model, car sales, and country. In this paper, the model gives the similarity between the business pairs. The full model successfully gives the prediction of the top 5 sales cars in each country. The mean square error (MSE) of the whole model is 8.086, while the MSE of both predicts and tests with a rank larger than 3 is 0.4s241379. Without enough data, the MSE of the full model is a bit large. In the future, we could find more countries' car data sets to improve the prediction accuracy and can also try to make a prediction for new markets.