What is Machine Learning
I took a machine learning class this semester. Following the last semester, I took AI-related classes again, but this time, unlike last time, I think it will be more touching because it is a practice-oriented class rather than a theoretical one, and I think I can use it for projects in the future. Since today is the first class, we conducted a simple concept-oriented class. First, I learned about the difference between humans and machines about learning. Humans and machines have many similarities in the process of basically making judgments. However, there is a big difference in that the basis for judgment is that people are experiences and machines are data. And machine learning that we will learn is a discipline that gives computers the ability to learn without programming.
Supervised Learning
While conducting the class, I learned that machine learning that I knew is largely divided into three categories. First of all, I took a class on supervised learning, and this part was linked to what I learned in the summer semester. It was a process of creating a model based on the label, which is a characteristic of predicting and learning data. This supervised learning is also largely divided into two models, a regression model and a classification model.
The regression model is the process of forming a straight line as close as possible to these points when they are gathered, as shown in the graph below.
The classification model is the process of finding a straight line that divides the dataset well into two, as shown in the graph below.
Unsupervised Learning
Unsupervised learning was a relatively unfamiliar experience for me. I thought it would be very difficult to set standards because there was data to learn but no label was given, but meaningful data was output through the process of grouping each data into each other’s labels based on similarity. I thought the core of unsupervised learning was clustering, and in modern society, we encounter many things that cannot be defined, and I thought it was a very good learning method to solve such problems. In particular, I thought that exposing things that we might like on Netflix, which we usually watch, or Naver, which is a search engine that we often use, would be related to this clustering. The picture below is a table that simply shows the difference between supervised and unsupervised learning, so I attached it!
