My solutions for Data Science problems in Digikala Supercup (March 2022).
Q1 (NLP): Predicting the price of products based on their categorical and text features.\ Solution: Used Bidirectional GRU + MLP components to predict the price as a regression problem.\ \ Q2 (Vision): Detecting if the image of a product has watermark. \ Solution: Fine-tuned EfficientNet on the given task.\ \ Q3 (Time Series): Predicting the next time a user will buy a certain product.\ Solution: Extracted a set of features from the data (frequency of purchase, days between purchases, etc.) and used Ridge Regression to predict the next purchase date.\ \ Q4 (Algorithm): A simple algorithm problem.\ Q5 (Database): Calculating nDCG on search results.
Design and implementation of a system to recommend shops and products to customers based on their previous orders.
Solution: