In the modern era, customer buying behavior has emerged from physical store purchase to now shopping from digital devices via use of internet technology. All the advanced e-commerce technology are now fulfilling needs effortlessly by perform a finger click on the digital devices and all data in relation to buying, browsing and transection recorded to company database in real time. To fulfil the market demands, many advanced e-commerce platforms have implemented the product matching strategy to customers and recommends any desired product in cheapest offer price. This strategy can bring positive outcomes to both customers and retailers, it can directly increase sale volume, customer retention rate, as provided function can quickly sort out customer’s desired product in best offer in matter of seconds. In this paper, we are going to discuss, explore and leverage several machine learning techniques to resolve a real-world Zalando product matching problem. There are several datasets provided which include offers training, offers testing, the scope is to find matching product models that can effectively locate any common sell items between “Zalando” and “AboutYou”, the outcome is targeted to enhance the overall F1 matching score.
Keywords : “TF-IDF”, “Cosine Similarity”, “NGrams”, “String Matching”, “Product Matching”, “Text Matching”
Brands with similar names : Zalando: SYCAMORE Skijacke Snowboardjacke oliv VS AboutYou: Skijacke oliv
