Algo_Souls - Unilever Data Science POC Use Case - Deep Tech

algo_souls - Unilever Data Science POC Use Case - Deep tech

1.Finding Drivers for given data affecting sale(EQ) 2.Forecasting for next 6 periods

Project Name:ML for POC Use Case
Team Name: Algo_Souls
Team Members: 2 (1.Bhushan Mahajan, GitHub: https://github.com/BhushanDA) & (2.Amit divekar, GitHub:@amitdivekar30)

6.Repository Link(s): 1.First code: https://github.com/BhushanDA/POC_Use_Case/blob/master/Main_Code 2.Second_Hurdle: https://github.com/BhushanDA/POC_Use_Case/blob/master/Second_Hurdle

Presentation Link: https://drive.google.com/open?id=1F_-2LyIqOwShmerO0O_lpoFiBtBBXZC0
Deep Tech - Problem Statement - 3: models are provided in code

PITCH

1.After importing data, Training data is converted into periodic with 428 periods. 2.Drivers are selected using Boruta or Randomforest. After that Train and test data consisting of drivers only is generated. These drivers are the features from data provided which are affecting EQ. 3.Using Multivariate forecasting sales(EQ) is predicted. 4.As there is downward trend of combined train and test data in last no of observations, the predicted data is following downward trend.