The general definition assumes that the algorithm is certain actions that are supposed to solve a specific problem.

In the era of the development of artificial intelligence, the data with which we feed the system become its food. Algorithms in this situation are work instructions for his “digestive” processes – how to use the data, what should be absorbed and used for the future -. So we can say that the w algorithm is actually a kind of instruction.

Muhammad ibn Musa al-Khwarizmi was a Persian mathematician who first introduced the concept of the algorithm (Latinisation of his name) in 800. The algorithm is a finite sequence of well-defined instructions that can be implemented using a computer, usually to solve a class of problems or perform calculations (source Wikipedia). The algorithm can be as simple as a recipe for cooking an egg or as complicated as a computer program to calculate a prime number. Both of these processes have a common deterministic feature, which means that depending on the data, the initial state, and various actions, you can always predict the final result, which is predetermined by the programmer. ML algorithms “learn” from experience and provide results that are determined not only on the basis of current data but also on the basis of previous calculations, so that they “learn” and train to achieve the most accurate result.

However, in machine learning, one algorithm is not enough. If we want ML to be multi-dimensional, then many different algorithms are needed. It all depends on the data we will feed the model and what effects we want to achieve at the end of the learning process. We need here learning algorithms that determine the learning process and update the entire process and software.

Based on carefully selected algorithms and data, the ML model will not only be able to analyze information properly but will also learn from experience. If we return to the kitchen atmosphere again, then here the model can be the equivalent of a beginner cook. He has the ingredients (data) and recipes (algorithms) at his disposal. During the whole process, he learns how to mix ingredients, what to combine them with and whatnot, how much it takes to achieve the best result, and presents his “insights”. As a result, we’ll get something more than just a “cake” – the sum of the ingredients.

There are many types of them, but they can be divided into several groups, including:

Regression algorithms, e.g. linear regression – shows how two variables affect each other (e.g. time spent in a bakery up to the amount spent on shopping)

Clustering algorithms – grouping data based on their similarity

Decision tree algorithm – it is used to obtain knowledge based on examples

The deep learning algorithm is already more advanced; allows you to recognize and imitate speech or control the behavior of the robot

These are just a few examples, but today we wanted to show you what these algorithms are and why they are so important in machine learning. We will discuss individual groups in more detail soon.