Join Data Science Interview MasterClass (in 3 weeks) 🚀 led by FAANG Data Scientists | Just 8 slots remaining...
Approach 1 – Rank Products
The simplest design as seen below is to preprocess user query using text preprocessing (e.g. spelling check, stemming) and then encode the search query into a numerical vector (e.g. TF-IDF or Word Embedding). These along with the user’s features and millions of products will be concatenated and then input into the model to assign a ranking score per each product. The score is then used to sort the products in descending order...