What is predictive modeling?

Predictive Modelling Phases Image Source

Predictive modeling is a process that uses data and statistics to predict outcomes with data models.

Often used interchangeably:

  • Predictive analytics: More like commercial applications of predictive modeling.
  • Predictive modeling: Used more generally or academically.
  • Machine learning: Use of statistical techniques to allow a computer to construct predictive models.

We as data scientists perform all these heavy-lifting of data by instructing computers step by step. But we can teach the computers to build and perfect models on their own.

So the art of teaching the machines to learn from past data, their statistics and the probabilities of the recurrence, is known as Machine Learning.

Image of statistics Image Source

Machine learning models typically fall into two categories: supervised learning and unsupervised learning.

Classification and regression algorithms are two types of supervised learning.

Image of statistics Image Source

In this crash course, we will look at three types of regression algorightms: OLS Linear Regression (which you should be already familiar to from your Statistics module!), Decision Tree Regressor and Random Forest.

The hard work begins

Regression modelling

CLICK HERE FOR THE WORKSHOP NOTES


Previous section: