How to Use these Machine Learning Guides
These are not intended to be comprehensive guides for machine learning. The primary goal is to best help you understand the fundamental concepts of machine learning to get started quickly with any project you may have. Furthermore, these guides will focus on Machine Learning through Python. Supplemental resources for more detailed and comprehensive guides will be provided along the way.
What is Machine Learning?
There is no widely accepted, formalized definition for machine learning yet. The generally accepted idea is that machine learning entails a computer being able to learn about a problem without being explicitly programmed to do so.
Common Types of ML Algorithms
The most common types of machine learning are supervised learning and unsupervised learning, which we will discuss further. Other common types include:
- Reinforcement learning: machine is trained to learn specific decisions. Through trial and error, the machine tries to capture the best solution
- Recommender system: predicts the “rating” or “preference” a user would give to an item
What is supervised learning?
- Formal Definition: the computer receives a set of inputs and corresponding outputs, with the goal of determining a function that can determine accurate outputs given inputs
- Key point: is that the computer is given the “correct answer” for the inputs to guide its learning
- Common types: regression, classification
- Predicting price of a product based on size
- Predicting chance of rain based on temperature
What is unsupervised learning?
- Formal Definition: the computer is given a set of data and is tasked with determining a structure from that data by clustering the data based on relationships among variables
- Key point: the computer is not given guidance on what the output should look like. Instead, it “explores” the data itself to try to draw relationships
- Common type: clustering
- Google news (sorts by topic)
- Social network analysis
- Market segmentation
- Hands on Machine Learning with Scikit-Learn & Tensorflow (Aurelien Geron)