By now you’ve probably already heard of Machine Learning. If you haven’t already that’s cool too.
Machine Learning is what powers most of the cool features we have today in many of our world class applications.
- “Movie Predictions” on Netflix
- Wonderful “adaptive filters” on Snapchat
- “Content filtering” on Instagram
Have some sort of Machine Learning implementation behind them.
We are going to take a spin into the world of Machine Learning with a brief introduction to some ML basics and then a look at how to make a simple logistic regression model that predicts whether a breast cancer tumor is “malignant” or “benign”.
DISCLAIMER : This is intended for everyone. You do not need any prior knowledge of Machine Learning to follow along.
What is Machine Learning?
“Machine Learning is a field of study that gives the computer the ability to learn without being explicitly programmed”
– Arthur Samuel (1959)
In the most basic of terms, machine learning is the ability of the computer to predict future events based off of data it has collected of past events. Obviously machine learning is not just about generating predictions, machine learning also involves the understanding of data samples and categorizing them into clusters i.e grouping them and making sense off of what we have.
What is Logistic Regression?
In Machine Learning, we can classify learning into 2 broad parts
Logistic Regression falls under Supervised Learning and one of it’s main advantages it the role it plays in solving binary classification problems ( i.e either ‘a’ or ‘b’ , ‘smart’ or ‘dumb’ ).
This is made possible by the Logistic Function core behind the model. The logistic function can take any real-valued number and map it into a value between 0 and 1
hθ(x) = g(θTx)
g(x) = 1/1 + e-x
hθ(x) is the prediction function that will be used.
Here’s a link to the Github Repository containing the code and more detailed explanation of concepts.
Time to go see the implementation videos. Let’s ride 🎢🎢