A Machine Learning Pipeline for Predicting total spend and customer groups of customers from a set of survey data using H20 AI's GBM. This is part of a Data Science assignment for a job interview.
Firstly, clone the repository to a folder of your choice.
Next install Anaconda's distribution of Python and install the necessary libraries:
-
Create an environment with a specific version of Python and activate the environment
conda create -n <env name> python=3.7.6 -
Activate your conda environment
conda activate <env name> -
Install libraries in requirements.txt file. For Windows machine, run the following code in the command line.
for /f %i in (requirements.txt) do conda install --yes %i
The input dataset has been ignored in this repo for confidentiality reasons.
With the input dataset available, run the following file:
python src/train.py
This would run the machine learning pipeline, that processes the input data and generates predictions for both total spend and customer groups. The results of the model is as follows:
For exploration of other models run in the exploration phase, please refer to notebooks/Assignment 1 and 2.html