Design and Development of Machine Learning Based Template for Fast Prototyping in Retail Industry (Step 5 - 28/02/2018)

March 1, 2018

Report Abuse
The Step of the Project where the evaluation of the various models have been made on the basis of Parameters like MAPE, MAE and so on .....
Objectives: The objective of the Project is to build a Machine Learning based Web Service that automatically provides the weekly forecasts of the next month for each store and each product for any Ecommerce based Retail Store. Methods/Statistical analysis: The dataset used in the Project has been imported from UCI Repository and Azure Blob Storage. Once the data is available, I will perform data pre-processing, followed by Training of Time Series Models, Feature Engineering, Training of Regression Models, and Evaluation of models and finally Deployment of Different Web-Services for Time-Series Model and Regression Model. This will be followed by a comparative study of the Regression Models and the Time Series Models that I will be using in my Project. Findings: I will be using all the Regression Models and the Time Series Models for the Project Purpose. I intend to find the relationships among the various parameters of the Dataset by analysing the effects of each of the Parameters (weather/pricing) on forecasting the weekly sales for the subsequent months. I intend to make a comparative study of all the Regression Models and find out the best model that suits my data. Additionally, I would like to make an in depth study of each of the Time Series Model and figure out the best suitable model for the Purpose. To improve the results, I would like to create a hybrid model that would incorporate the best Regression and the Time Series Model. And finally I intend to deploy the model as a Web Service that can be used as a template for fast prototyping and deployment of machine learning solutions.