At the request of my friend Richaldo, in this post I'm going to explain the types of machine learning algorithms and when you should use each of them. In particular, I think that knowing the types of machine learning algorithms is like seeing the big picture of AI and what is the goal of all the things that are being done on the ground and being in a better position to analyze a real problem and design a machine learning system. In the two previous types, either there are no labels for all the observations in the data set, or there are labels for all the observations. Semi-supervised learning falls between these two.
In many practical situations, the cost of labeling is quite high, as this requires qualified human experts. Therefore, since most of the observations do not appear labels, but only in a few, semisupervised algorithms are the best candidates for building the model. These methods take advantage of the idea that, although the group membership of unlabeled data is unknown, these data contain important information about the parameters of the group. Different criteria can be used to classify types of machine learning algorithms, but I think using the learning task is fantastic for visualizing the big picture of machine learning and I think that, depending on your problem and the data you have at hand, you can easily decide if you are going to use supervised, unsupervised, or reinforced learning.
In future posts, I'll give more examples of each type of machine learning algorithm. In fact, reinforcement learning is defined by a specific type of problem, and all of its solutions are classified as reinforcement learning algorithms. It is also known as a lazy learning algorithm, as it stores all the available data sets and classifies each new case with the help of K-neighbors. The algorithms include supervised and unsupervised learning systems, as well as semi-supervised and reinforced machine learning technology.
The data points that help define the hyperplane are known as support vectors and are therefore called support vector machine algorithms.
