Classify with Machine Learning. How does machine learning address data… | by Diego Perez | Jul, 2022 – Medium

Imagine we want to know if a patient is prone to suffer from a known disease according to medical tests, or we want to identify the elements of a given image, this task can be easy for humans if there are few inputs, but once the amount of data grows it becomes humanly impossible to classify the results.

Thus, taking advantage of computational power, the researchers developed several machine learning algorithms capable of dealing with this situation, in particular, a branch of machine learning called supervised learning that is in charge of solving the so-called classification problems. The most common classification problems are speech recognition, object detection, and document classification.

Classification is the process of identifying the class or category for an entry based on predetermined characteristics, which will be represented by any categorical points considered relevant or simply by integers representing individual classes. Specifically, there are two types of classification problems

Types of classification

  • Binary classification: predicts the elements of a set into two groups. Given an email, is it spam or not spam?
  • Multi-class classification: predicts which class an element is in given multiple classes of possible outcomes. Given an image, is it a dog, cat, or rabbit?

Once it is known what type of classification a given problem requires, it is necessary to generate a model that is a data structure that stores a representation of a dataset, which is trained using an algorithm to predict the results. Some of the most common algorithms for classifying data, which will be described in more detail in later publications, are as follows:

  • K-Nearest Neighbor
  • Logistic Regression
  • Decision Tree
  • Random Forest
  • Artificial Neural Networks
Spread the love

Leave a Reply

Your email address will not be published.