This course will start on 12th December 2019.
About the Course - Retail Demand Prediction Using Machine Learning
Are you wondering how is machine learning being used in the industry today? Are you looking for some hands-on experience on a real-world dataset? Do you want to learn how to solve a business problem using machine learning?
If yes, then you have landed at the right place. This course is the perfect balance of theoretical learning and hands-on exercises.
The course is aimed at enabling you to build an end-to-end Machine Learning model on a real-world dataset - starting from converting a business objective into a machine learning problem, to building a complete machine learning model.
The business problem addressed as a part of the course is that of demand forecasting.
Key Takeaways from the Course
Understand the importance of demand forecasting in the industry
Learn how to Structure a Business Problem
Converting a Business Problem to Machine Learning Problem
Derive Interesting Insights from the data
We will cover the following steps in ML Lifecycle:
Exploratory Data Analysis
Data Cleaning and Preprocessing
Evaluating Model Performance
Create ML solution for real-world problems
Presenting Model results
This course assumes that you are familiar with the basic concepts of Machine learning and comfortable with coding in python.
- Introduction to the course
- Prerequisites for the course
- Defining the Business Problem
- Defining the Business Objective
- Possible Data Problems that could help achieve business objective
- Choosing the Data Problem
- Hypothesis Generation
- Datasets available & problem statement
- Creating Test Set
- Creating Validation Set
- EDA: Univariate
- EDA: Bivariate
- Insights from Data Exploration
- Preprocessing: Categorical
- Preprocessing: Numerical
- Defining Evaluation Metric
- Testing with Baseline Model (GLM)
- Feature Engineering
- Linear Model (GLM)
- Tree Based Models
- Selecting the Best model
- Hyperparameter Tuning
- Basics of Ensemble Learning
- Ensemble Techniques
- Test on holdout set
- Sharing Model Results & Report