One of the hottest topics in the layperson discourse surrounding data is machine learning. Machine learning is a special form of artificial intelligence that uses mathematical algorithms and an initial dataset to develop models that can make judgements about new data. So, let’s investigate the three main types of machine learning, and the special case of neural networks.


Types of Machine Learning

There are three general types of machine learning: supervised, unsupervised and reinforcement learning.

Supervised learning occurs when a machine is provided with an initial dataset full of inputs and correct outputs. At first, it will take random guesses to produce an output for each row, and then compare its results with the provided answers to evaluate how accurate it was. The computer will then begin to develop a system of rules to generate its outputs in a more formulaic manner. This iterative process continues until the algorithm has determined that its classifications or predictions are sufficiently close to the right answers.

Unsupervised learning occurs when a machine is provided with an initial dataset full of inputs only. Rather than generating outputs that it can evaluate as accurate or not, the computer aims to discover the underlying structure of the data, making connections and finding patterns that may persist in the population. The algorithm may ultimately determine clusters of data that share similar characteristics or observe associations between different variables that tend to move in a similar way.

Reinforcement learning occurs when a machine is placed in a complex, changing environment. The computer continually makes new decisions in response to how the environment changes and receives rewards or penalties depending on whether the programmer wants to encourage or discourage its behaviour. The algorithm continually develops its behaviour so that it can maximise its rewards and minimise its penalties.


Neural Networks

More recent developments in the theory underlying machine learning has led to the advent of neural networks.

A neural network is a series of connected nodes (representing neurons) that generate signals and transmit them to one another.

Neural networks are usually organised in layers. The initial input stimulates signals from the neurons in the first (or input) layer. These signals will in turn stimulate a response from the next layer of neurons, and this iterative process will continue, enabling the development of complex behaviour that tends to emerge from more organic, non-linear systems. The computer will make a final decision according to the concluding signals it receives from the final neural layer.

Neural networks have become popular in aiding all types of machine learning as they are able to handle big data more easily and are more effective in obtaining generalisations that are more likely to apply to the population. However, it is important to keep in mind that neural networks tend to require a lot of training data to develop their rules, and so they can risk overfitting.


Machine learning is only becoming more popular with businesses starting to appreciate the immense learning power that machines are capable of. Becoming acquainted with machine learning techniques is a must for any budding data scientist. Happy developing!

Cover image by Gordon Johnson from Pixabay

Hunter Iceton
Author: Hunter Iceton

Hunter Iceton is an enthusiastic and positive individual. He graduated from Sydney Uni in 2017 with a Bachelor of Commerce (Liberal Studies) majoring in Finance, Marketing and Quantitative Business Analytics. For the next few years, Hunter spent his time creating and releasing music, while tutoring primary and high school students in Mathematics and Business Studies. Hunter is now excited to be joining The Data School, looking forward to approaching analytics with a creative perspective. In his spare time, Hunter enjoys continuing to create music, reading philosophy and cooking plant-based dishes. Otherwise, he can usually be found at a restaurant, a bar or an art gallery.