So, moving on to the next topic, we will discuss programming practice. We have talked about various aspects of artificial intelligence and machine learning, but it is crucial for you to actually try programming. Here, I will briefly explain programming. Generally, to implement artificial intelligence in a program, we can use various programming languages. Common coding languages include Java, C, C++, and others. However, in the field of artificial intelligence, there are mainly two languages that are widely used: R and Python. R and Python have slightly different purposes. R is not a general programming language; it is a statistical analysis language developed by statisticians. Therefore, it is particularly suited for statistical tasks.

Analyzing data is very useful, and there are advantages to visualization and statistical analysis. In the early days of big data, R was widely used for big data analysis. However, now that big data analysis has combined with artificial intelligence and machine learning, Python is becoming increasingly popular. Nowadays, it can be said that almost everyone is using Python. Python is a general programming language developed by computer scientists. It is somewhat more difficult than R, but it is generally considered easier than other programming languages like C, C++, and Java. Python is known for its fast processing and versatility, which makes it widely used in programming for artificial intelligence. So, when people ask why Python is used,

Python has many advantages. First, it is easy and quick for people to learn, and using it is also very simple. Because it is a general-purpose programming language, it can be used not only for machine learning but also in various fields such as game development and web development. Unlike other languages, Python incorporates many modern programming concepts, such as object-oriented programming and dynamic typing. You don't need to know all the details, but these features allow it to keep up with the latest trends. However, there are many languages that have similar characteristics; in fact, many languages besides Python are easy to use, versatile, and quick to learn. But if you ask why Python stands out among them, it is because it has a well-established ecosystem.

There are many people who use Python, and these people have created a lot of libraries and code that others can use. Because of this, people gather again to utilize these resources, and as a result, the community is growing like a snowball. Now, everyone is simply using it. So, you can think of Python as being widely used and generally accepted. Typically, we use Python for artificial intelligence, but as I mentioned earlier, we need to utilize a well-established ecosystem. This means that there is Python itself, and then there are various libraries created by others that add different functionalities on top of Python. These are program modules that serve specific purposes.

then our kids use libraries. The key is to use libraries. So, which libraries will we use in Python? For artificial intelligence, you can see a stack structure. At the very bottom, there is the core of Python itself. This is just Python itself. There is basic code provided by Python. However, since it is a general programming language, it is just a general language. But what we need to deal with, such as machine learning, is mostly mathematics. What is mathematics? Ultimately, it involves handling numbers. To handle various numbers, we need to deal with data in forms such as matrices or vectors. However, handling these data types with Python alone is somewhat inefficient. So,

To handle these numerical data effectively, a library is needed, and that library is NumPy. So, fundamentally, Python uses NumPy, and most of the machine learning and deep learning that we know happens on top of it. At the very base, we have NumPy, and then using this NumPy, we implement various fundamental algorithms widely used in science, which is precisely SciPy. SciPy is short for Scientific Python. NumPy stands for Numerical Python, and SciPy is an abbreviation for Scientific Python. So, the SciPy library builds on top of NumPy, and within SciPy, various basic algorithms are available, and on top of that, we use the Cyclone library to perform numerous tasks.

In machine learning, there are various models and algorithms that implement cycles, and we usually utilize these cycles. Therefore, if we want to engage in general machine learning, we typically go through this stack to use machine learning. However, if we want to transition from machine learning to deep learning, we find that the calculations in deep learning are quite different from those in traditional machine learning. Due to the high volume of calculations required, special hardware like GPUs is necessary, which in turn requires new libraries tailored for deep learning. Thus, libraries specifically designed for deep learning have been developed by people, and these include TensorFlow and PyTorch, among others.

There are modules that fundamentally operate on top of NumPy. So, if we want to do deep learning, we need to follow stacks like TensorFlow built on NumPy. This is where the deep learning part comes in, which is my area of expertise. Therefore, if you want to delve into deep learning, it is essential to learn how to use TensorFlow and similar tools. If you're simply interested in traditional machine learning, then learning Scikit-Learn would suffice. However, in addition to machine learning and deep learning, one of the key methods for analyzing data is statistical analysis methods. In the past, statistical analysis methods often relied on tools like R, but this led us to...

I wanted to bring various statistical methods into the field of deep learning. To achieve this, people have created what is called the step model. In general data analysis, we usually work with numerical data using NumPy. However, for typical data analysis, we also need to represent various forms of data, such as gender—male or female—and regions like Seoul, Gyeonggi, and Incheon. These forms of data, including gender and addresses, can be represented in a variety of ways, which we typically refer to as a DataFrame. We need a data format that can express this DataFrame, as well as tools for statistical analysis. This series of requirements led to the development of Pandas and the step model. Therefore, Pandas is designed to handle not only numerical data but also data we will encounter later.

This is designed to easily express and visualize categorical data. So, we created Pandas and implemented a statistical methodology called Statsmodels on top of it. If you want to conduct statistical analysis, you can use this technology, which falls under statistical analysis. Therefore, if you are interested in artificial intelligence or working with data using Python, learning libraries with this structure will be beneficial. Now, regarding how to use Python, there are two main methods: one is to install Python directly on your computer.

There are several ways to install it directly. However, the simplest method is to download a package commonly referred to as Anaconda. You can just download it and install it. Once you do that, it will install Python, the necessary libraries, and a user-friendly program called Jupyter Notebook, which makes it easy to use Python. After installing it, you can run Jupyter Notebook, and you'll be able to execute Python code. So, that's the first method. The second method is that instead of installing Python directly on your computer, you can access it online through the cloud and use another computer to run your code.

There are several companies that provide services like this, and one of the most representative is Google's Colab. What it is, is that if you simply access the webpage at this address and log in with your Google account, you can run Python without having to install it on your own computer. The Python code runs on Google's servers, not on your computer. So when you access it, a webpage appears. This webpage actually has a similar interface to Jupyter Notebook, which I mentioned earlier. You can enter Python code like a = 1, b = 2, print(a + b), and then it will output the result, which is 3. In this way, you can use Colab to perform your tasks.

So basically, during this lecture, we will continuously discuss the practical application of the program. We will primarily be using Google Colab for this purpose. You can use Colab or the Anaconda package to carry out the tasks. Up to this point, we have covered how to perform basic operations in Python. In the remaining time, I will briefly explain the fundamental concepts of Python core. This is important because, in order to complete your assignments or run your projects later, you need to understand how Python works at a basic level. I will provide a very straightforward explanation of this.

I will now explain the most basic and essential concepts. When it comes to Python, as well as other programming languages, the fundamental thing you need to understand is what variables can accept in terms of data and how to perform operations on that data. 

First, let's look at numerical data. It's quite simple. If we say a equals 1, we are assigning the value 1 to a. If we say b equals 2, we are assigning the value 2 to b. Then, if we have c equals a plus b, that means a plus b equals 3, and we store this result in c. So, a will hold the value 1, b will hold the value 2, and c will hold the value 3.

Next, when we use the print function, if we write print a plus b equals c, it will display the value of a, followed by a comma, then the plus sign, followed by the value of b, and finally the equals sign followed by the value of c. So, it will show 1 plus 2 equals 3.

So this is how it gets printed. And C is the value that results from adding A and B. It's that simple. Python can actually perform basic arithmetic operations. These operations include addition, subtraction, multiplication, division, and modulus. So if A is a number and B is 3, then A + B would equal 8, A - B would equal 2, and for multiplication, we use an asterisk, so A multiplied by B would be 15. A divided by B, which is 5 divided by 3, results in approximately 1.6666666667. Next, the percent sign represents the modulus operation. For example, when you divide 5 by 3, the remainder is 2. This way, you can easily perform arithmetic operations using numbers, and it's very intuitive. Next, we have string data. When you write A like this, it is a number, but if you write A in quotes as '1', it becomes the character '1'.

You can use double quotes to equally represent a string. So, if we have a as Hello, then a now holds the string Hello, which means it becomes a variable with a value. When we print a, it will display Hello. Next, if we use the function len, it indicates the length of the string; for a, it will show 5 since there are a total of five characters. 

Now, if we want to look at each character individually, we can do so by using a followed by square brackets and inserting a number. The numbering starts from zero, where h is the 0th character, e is the 1st, l is the 2nd, l is the 3rd, and o is the 4th. Therefore, a[0] will give us h, and a[1] will give us e. 

You can also use negative indexing. The negative numbers start from the end of the string. So, the 5th character corresponds to -1, the first l corresponds to -2, the second l corresponds to -3, and o corresponds to -4, while the string itself corresponds to -5.

So, if you do a - 1, you get 5. In this way, we can also see each character individually. Next, there are several operations that can be applied to strings. For example, the addition operation is used to concatenate characters. If a is hello and b is world, then a + b equals hello world. But can you do a - b? Intuitively, that doesn't work. So, that kind of operation is not possible. Can you do a multiplied by b? That also doesn't work. What about a divided by b or a modulo b? None of those work either. So, not all operations are valid; only addition works. When you add, it concatenates the strings. 

Now, what I've shown you regarding numbers and strings is that a variable holds a single value. However, there are times when you want a variable to hold multiple values. One representative example of this is a data structure called a list. A list is a data type that can hold multiple values. So, how do we...

A is represented like this with square brackets: [13.1, Hello]. In this case, it has four elements, separated by commas. If we look at these elements, the first element is just a number, the second element is also a number, the third element is a string, and the fourth element, as you can see, is the entire set enclosed in square brackets. This means that it is also a list. So, a list can contain another list as an element. Therefore, the variable A has a total of four elements. Now, how can we access each of the elements in A one by one? We do this using indices. By placing the index number within the square brackets, we can access each element.

For example, if it's a0, then since it's the first element, it will return 1. If it's a1, then it will return -3.1. What about a3? It corresponds to 0, 1, 2, 3, so it is 3. Therefore, if you access it as 3, you will get the value at that position. Right? That is exactly the a3rd element. However, since a3 is also a list, if you access the 0th element of a3, what do you get? You get its 0th element, which is 5. So that's how it works. Additionally, you can append to a by using the command a.append. After calling a.append and inserting a new item, if you add a string called edited, then when you look at a, it will originally contain 13.1, hello, and 5, and now edited will be appended to the end. So you can keep adding one by one like this. Next, there are various data types.

중에 또 쓸 수 있는 게 바로 이제 딕셔너리 형태의 데이터 타입이라는 것입니다. 이 딕셔너리는 이겁니다. 네, 이거예요. 딕셔너리입니다. 딕셔너리는 리스트와 다르게 키와 값, 우리가 이제 키와 밸류라고 표현하는데, 키와 값의 쌍을 갖는 데이터 타입이에요. 그래서 정의는 이렇게 중괄호, 우리말로 중괄호라고 불리는 컬리 브라켓으로 표현합니다. 그래서 보시면, a는 원이고, 1이에요. 그럼 원이 키가 되는 거고요, 1이 값이 되는 겁니다. 그러니까 원에는 1이라는 값이 매핑되어 있는 거고, 마찬가지로 2에는 2라는 값이 매핑되어 있는 거고, 그 다음에 여기 보시면 3, 3이라는 숫자에는 쓰리라는 값이 매핑되어 있는 거고, 4에는 4라는 값이 이렇게 매핑되어 있는 겁니다. 그래서 실제 값은 이거고, 얘를 찾기 위해서 여기에 있는 키 값들을 쓰는 거예요.

So, if we want to reference a new value, for example, 1, we can use 'a' followed by square brackets for indexing. If we put the key value of 1 inside the brackets, it will return 2. If we input a2, it will also return 2 because it corresponds to 2. If we use a3, it will return 3 because the value is 3, and if we use a4, it will return 4 because the value is 4. This is how we define data in the form of a dictionary. Now, if we want to add something new, we can use 'a' and then insert a new key value. For example, if we set 5 to 5, later when we check 'a', it will show 5 as 5. If we input 'a4' again, what will happen? Since a4 is already set to 4, right?

However, when we insert a value here, it is not the number 4. The point is that the number 4 is not being displayed correctly in the middle. So, when we change it to English, it will be different. Therefore, we can either write new values or read existing values. To reiterate, when we insert 4 into A4, A4 was originally the number 4, but now it will be erased and changed to a string. Up until now, we have discussed the most basic data types. After reviewing the data types, we need to understand how to implement algorithms that perform certain actions based on specific conditions, and how to implement loops that repeatedly execute these conditions.

It's quite simple, actually, more so than you might think. If you look at the conditional statements, we defined 'a' as 10. If you see this code, then if we use 'if'—if 'a' is less than 5, and if this condition is true, we execute this. It would print that 'a' is less than 5. However, since 'a' is 10, that is not the case. If that condition is not met, we move on to the next. Now, 'elif' stands for 'else if.' If 'a' is less than 10, then this will be executed. So we would have a statement there. But that's also not the case. If neither condition is met, we move on to 'else.' 'Else' means all other cases. So if 'a' is less than 5, we print this; if 'a' is greater than or equal to 5 but less than 10, we print that. In all other cases, if 'a' is greater than or equal to 10, we would express that.

So currently, since A is 10, the last one appears. In this way, based on the conditions of a variable, we can check if it is less than 5, less than 10, or whether it is less than, greater than, less than or equal to, or greater than or equal to. All of these conditions can be applied. Depending on these conditions, we can execute different codes. This is the basic idea. 

Next, we can apply this not only to numbers but also to characters. For example, let's say B is set to hello. Then, we check if B is equal to hello. This is done with two equal signs. If we ask if B is equal to world, we would express that as not equal. Since B is hello, it is not equal to world, so not equal comes out. Therefore, this is equal, and similarly, if we use two equal signs here, we can check for equality.

하는 거고, 이게 느낌표와 이퀄을 쓰게 되면, 얘는 같지 않느냐를 이제 테스트하는 경우가 됩니다. 그래서 이런 조건에 맞냐, 맞지 않냐에 따라서 if나 else를 써 가지고 우리가 서로 다른 명령들을 수행할 수 있고요. 그 다음에 이제 반복문입니다. 반복문은 우리가 같은 코드를 서로 다른 데이터에 대해서 똑같이 하고 싶을 때 반복하고 싶을 때 보통 쓰는데, 여러 가지 방법을 쓸 수 있지만 우리가 보통 쓰는 것은 이제 for문입니다. 이런 표현을 자주 씁니다. 그래서 뭐냐면, 예를 들면 a는 지금 리스트죠. 그래서 1과 2와 3, 4, 세 개를 갖고 있는 원소를 갖고 있습니다. 그런데 우리가 만약에 이런 코드를 썼어요. 꼭 a, 소문자 a in 대문자 A. 그 다음에는 얘는 뭐냐 하면, a에는 세 개의 원소가 있으니까 이 세 개의 원소를 돌아가면서 뽑으라는 이야기입니다. 그럼 첫 번째 루프가 돌...

a is the first element, so when we print a for the first element 1, it outputs 1. Then, when we move to the second element, which is 2, it prints 2. When we go around once more, for the third element, it outputs 3. So, having looped three times, it ends there. This is the most basic structure of a loop. The fundamental form of using it is like this, and we often use this form. For example, here we have for i in range 10. So, what does range 10 mean? Generally, range means that i starts at 0, so the first loop has i as 0. In the second loop, i becomes 1.

When the child turns 1, adding 1 to 0 makes s equal to 1, because it was originally 0. Then, when we look at the third loop, i is 2, so adding 2 to 1 means s becomes 3. This continues until the tenth loop, where i is 9. Therefore, s is the sum of all numbers from 1 to 9, which totals 45. So, when this loop is completed, s will be 45. There are many expressions used in this way because when we want to run something 10 times or 20 times, we can write it like this. A slightly modified version of this could be a code like this: for example, this one is a range from 10 to 100, meaning it starts at 10 and goes up to but does not exceed 100, so it goes up to 99.

In this way, if we go from 12 to 98 and 99, how many times will it be in total? Is it 90 times? So, it will run a total of 90 times. During these 90 iterations, we will be repeating the loop statement. What needs to be repeated inside this loop is the code. If you look closely, you can see that the elements that need to be repeated are indented. The ones that are not indented are independent and not part of the loop. So, the loop is defined by the indented section, which constitutes a single block, while everything outside of it is not. The same applies to this section; the part that is included in the loop is the one that is indented. Therefore, as you can see, when it runs, if the variable is divisible by 3 with no remainder, it means...

The remainder is what we focus on. The operation will only be performed when the remainder is zero. This means that we are adding up only those numbers for which the remainder when divided by three is zero, which are the multiples of three. 

So, this code will ultimately find the sum of numbers between 10 and 99 that are multiples of three. For example, it will select numbers like 12, 15, and 18, and the sum of these numbers will be denoted as S. The result is 1665. 

I've briefly explained the core part, and as we continue this lecture, I will explain the necessary code in more detail. 

Yes, this lecture may be a bit long, but I will conclude here for now. Thank you for listening, and I look forward to seeing you in the next lecture.

데이터 학습과 지능 1 3

데이터 학습과 지능 1 3

Scripts