web-deployment : https://retail-stock-price-prediction.herokuapp.com/
We have a data of retail transactions performed over the span of two year. Apart from data analysis and visualization, a regression model is developed to predict the price of retail items belonging to different categories. Foretelling the Retail price can be a daunting task due to the huge datasets with a variety of attributes ranging from Text, Numbers(floats, integers) and DateTime. Also, outliers can be a big problem when dealing with unit price.
Evaluation metric : RMSE(Root mean squared error).
Main Topics Covered: Data Cleaning and Visualisaton, Feature Engineering, Over-sampling, Model Building, & hyper-parameter tuning. Models used: LinearRegression, Lasso, Ridge, RandomForestRegressor, XGBRegressor.
Machinehack Dataset link : https://www.machinehack.com/hackathons/retail_price_prediction_mega_hiring_hackathon/data