Sneaker Recommendation System Based on Brand Using the Content-Based Filtering Method

Abstract

Footwear, particularly sneakers, has evolved beyond its functional role to become a vital aspect of modern lifestyle and fashion culture. Indonesia ranks as the fourth largest global consumer of footwear, consuming 806 million pairs in 2021, accounting for 3.8% of global consumption. Popular brands such as Nike, Adidas, Yeezy, and Off-White dominate this market. This study aims to develop a sneaker recommendation system utilizing a content-based filtering method to address challenges in consumer preferences and rapid trend shifts. Data collection involved sneaker attributes, including brand, price, and region, sourced from Kaggle. Preprocessing and recommendation generation utilized Python programming, employing cosine similarity to identify and suggest the most relevant brands. Results indicate that Yeezy and Off-White sneakers are the most favored across regions, particularly in Oregon. The findings underscore the importance of personalized and accurate recommendation systems to enhance user experience in identifying sneakers aligned with their preferences. Future research should expand dataset diversity and incorporate additional filtering methods to refine recommendation accuracy.