To be able to solve a problem using machine learning or AI it is important we know how to categorize the problem. Categorizing the problem helps us understand which tools we have available to help us solve problem. This article will help you understand the different types of machine learning problems, and provide examples of algorithms used to solve problems in each category.
Generally there are two main types of machine learning problems: supervised and unsupervised. Supervised machine learning problems are problems where we want to make predictions based on a set of examples. Unsupervised machine learning problems are problems where our data does not have a set of defined set of categories, but instead we are looking for the machine learning algorithms to help us organize the data.
Put in another way – supervised machine learning problems have a set of historic data points which we want to use to predict the future, unsupervised machine learning problems have a set of data which we are looking for machine learning to help us organize or understand.
Within supervised machine learning we further categorize problems into the following categorizes:
A classification problem is a problem where we are using data to predict which category something falls into. An example of a classification problem could be analyzing a image to determine if it contains a car or a person, or analyzing medical data to determine if a certain person is in a high risk group for a certain disease or not. In other words we are trying to use data to make a prediction about a discrete set of values or categorizes.
Examples of algorithms used for supervised classifications problems are:
Regression problems on the other hand are problems where we try to make a prediction on a continuous scale. Examples could be predicting the stock price of a company or predicting the temperature tomorrow based on historical data.
Examples of algorithms use for supervised regression problems are:
As mentioned above unsupervised machine learning problems are problems where we have little or no idea about the results should look like. We are basically providing the machine learning algorithms with data and asking it algorithm to look for hidden features of data and cluster the data in a way that makes sense based on the data.
Examples of unsupervised machine learning problems could be genomics. In genomics we provide an algorithm with thousands of different genes and the algorithm will then cluster the genes into groups of related genes. This could be genes related to lifespan, hair color etc.
Another example of an unsupervised machine learning algorithm could isolation sounds in audio files. We would be providing the algorithm with audio files and asking the algorithm to identify features within these audio files. These types of algorithms are able to isolate voices, music and other distinct features in an otherwise chaotic environment.
Examples of algorithms used for unsupervised machine learning problems are: