CLASSIFICATION OF MACHINE LEARNING

Machine Learning algorithms can be classified into 2 main categories. Here we have the two main ways to classify the algorithms used in Machine Learning:

1. Based on Learning style.
2. Based on the Functions.

1. Using Learning style to categorize

Machine Learning algorithms can be divided into 4 main categories as the followings: Supervised Learning, Unsupervised Learning, Semi-supervised Learning, and Reinforcement Learning.

- Supervised Learning

Supervised Learning is the algorithm in which the computer predicts the outcome of a new input based on known input and outcome. These sets of data are called (data, label). Supervised learning is the most common algorithm among machine learning algorithms.

Supervised learning means that when we have a input data of the variable X = {x1,x2,…,xN}  and output data Y = {y1,y2,…,yN}, xi,yi, are vectors (data, label). The known (data, label) vectors (xi,yi)∈ X × Y called the training data. Based on the training data, we should to build a function (F) from: X to the ”equivalent” Y- and F called approximated function:

Yi ≈ f(xi), ∀I = 1,2,…,N

The target of approximated function F is with the new input data X,we can estimate the value of Y with: y = f(x).

Example: In handwriting recognition, we have pictures of different letters and numbers, which are written by different people. After that, we insert this image into the algorithm and show it each input will be a letter. When given a new picture the computer has never encountered before it will predict what is written on the picture.

The example illustrated above share similarities with how human learn new words when they are young. A child is presented with the alphabet and the teacher shows them which is A, which is B. After many times of learning, the child can easily see which one is A, which is B even in new books that they are given to later.

Besides that, there are also many algorithms for facial recognition. Facebook has been using these algorithms to recognize the face of a person in a picture and ask users to tag these recognized faces with the according user account into the picture.

More on: An Introduction Of Artificial Intelligence (AI)/ Machine Learning (ML).

- Classification

A classification problem is when the output variable is a category. A classification model attempts to draw some conclusion from observed values. Given one or more inputs a classification model will try to predict the value of one or more outcomes.

Example: Just like how Gmail assigning a given email is a spam or non-spam class, credit card companies assign whether a customer could be able to pay back their loan…

- Regression

A regression problem is when the output variable is a real or continuous value.

Example: How much does a house with an area of x m2 has y rooms is z km away from the city center cost?

Recently, Microsoft revealed a gender and age prediction application base on one’s face. Predicting one’s age and gender is a regression task.

- Unsupervised Learning

Unsupervised learning is the training of machine using information that is neither classified nor labeled and allowing the algorithm to act on that information without guidance. Here the task of the machine is to group unsorted information according to similarities, patterns, and differences without any prior training of data.

Example: When dividing into groups or reducing the directions of a set of data for ease of storing and calculating the data.

- Clustering

The task of dividing the population or data points into several groups such that data points in the same groups are more similar to other data points in the same group and dissimilar to the data points in other groups.

Example: Categorize customers based on their shopping habits. Similarly, when a child is given different pieces with different shapes and colors, such as triangle, cubes, spheres in red, blue, green etc... The child will not be able to put together all of that, however, it is possible for the child to divide the pieces into groups of different colors or shape.

- Association

An association rule learning problem is where you want to discover rules that describe large portions of your data.

Example: When buying normal clothes, male customers usually buy an extra watch or a belt for them. Many viewers often watch different movies within the same genre, this will boost the shopping demands.

- Semi-supervised Learning

Where an incomplete training signal is given: a training set with some (often many) of the target outputs missing.

Example: Commonly in this group are partial images and texts that are not labeled. In most cases, these items are collected from the Internet.

- Reinforcement Learning     

The task which the computer is assigned and complete based on the circumstances to achieve the best results of that particular task.

Example: An interesting example of reinforcement learning occurs when computers learn to play video games by themselves. In this case, an application presents the algorithm with examples of specific situations, such as having the gamer stuck in a maze while avoiding an enemy. The application lets the algorithm know the outcome of actions it takes, and learning occurs while trying to avoid what it discovers to be dangerous and to pursue survival.

2. Using purpose to categorize

The second way to categorize is by using the purpose of the algorithms. We will explain this more clearly in the next issue.