This course is designed to upgrade your skills as a practical machine learning development expert by implementing the machine learning model of the Home Credit Default Risk competition on Kaggle.
Upgrade your practical machine learning implementation skills by tackling Kaggle's practical competition problems
Upgrade implementation capabilities to the level where machine learning can be applied in practice
How to improve the performance of machine learning models
Improving Data Analysis Skills for Machine Learning
Specific implementation methods of machine learning feature engineering
Machine learning, with practical implementation capabilities! Implement your own contest machine learning model.
Lecture Introduction 🤖
hello,
I am Cheolmin Kwon, author of The Complete Guide to Python Machine Learning.
To become a true machine learning expert needed in practice, you need to have not only an understanding of machine learning, but also the ability to process data and understand the application tasks. However, these abilities are difficult to obtain even if you put in a lot of time and effort if you do not have actual experience or are not trained systematically.
This time, the "Kaggle Advanced Machine Learning Practical Challenge" course was created to help you cultivate these three elements while implementing Kaggle's "Home Credit Default Risk Competition" machine learning problem together, and to upgrade your practical machine learning implementation skills and confidence.
The 'Home Credit Default Risk Competition' problem has a data model and several data sets that can be used in practical work.
This lecture will explain in detail and implement code based on the problems of this competition so that you can sufficiently cultivate the relevant capabilities in the important areas of machine learning such as data models, analysis domains, data analysis EDA, feature engineering, hyperparameter tuning, and model performance optimization .
The machine learning algorithm used in the lecture is LightGBM, which is loved by many Kagglers. Through the implementation task, you will write an implementation code that is in the top 10% of the Home Credit Default Risk competition, and through this, you will be able to gain confidence in implementing a model that optimizes performance.
Features of this course 📚
1. Improved understanding of actual implementation through detailed and easy-to-understand practical code explanations and Live Coding
Most of the lectures are about explaining practical codes, and I will explain the codes line by line in great detail. In particular, for the important implementation parts, I will make it so that you can do Live Coding with me, which will further improve your understanding of the implementation.
2. Improved ability to implement performance-oriented models in preparation for competitions such as Kaggle and Deacon
In this course, you will learn advanced machine learning techniques, feature engineering, and hyperparameter tuning techniques to help you achieve high scores in competitions such as Kaggle and Deacon. This will get you to a level where you can confidently take on machine learning competitions.
3. Detailed explanation of all areas of machine learning required in practice
This course provides detailed explanations of all areas of machine learning, including data models and analysis domains, data analysis EDA, feature engineering, hyperparameter tuning, and model performance optimization. Through this, you will be able to improve not only machine learning but also data processing and business domain understanding capabilities, thereby laying the foundation for becoming a machine learning expert required in the field.
This lecture is an Advanced Machine Learning Project lecture for students who have a basic understanding of machine learning. It is designed assuming that you understand the contents of Chapters 1 to 4 (Classification) of the book ' The Complete Guide to Python Machine Learning ' .
Even if you have not read the book or lecture ' The Complete Guide to Python Machine Learning ', you can take the course if you have previewed the table of contents of 'The Complete Guide to Python Machine Learning' and are familiar with the contents of the table of contents up to Chapter 4.
Practice Environment 💻
Jupyter NotebookColab
Any environment with more than 12GB of RAM memory is possible. (8GB or so may be difficult to practice due to insufficient memory in the final practice stage.) If you do not have more than 12GB of RAM, you can create a server using Google Cloud's $300 free credit or use Google Colab. The first section of the lecture explains in detail how to set up these practice environments.
The practice code is provided in the form of a Jupyter notebook, and the practice code for Google Colab is provided separately. The practice code and lecture materials can be downloaded from the lecture session materials room.
It seems like a good lecture where you can get a feel for the real world by using real Kaggle data. The first time I came across Kwon Chul-min's lecture was when I took the lecture called "Machine Learning Complete Guide", and this lecture is also really good...
Thank you so much for always making good lectures, and what I really liked the most was that if you post a Q&A, you answer it quickly... I think this is a really good point. Of course, the lecture quality goes without saying.
Thank you again.
I know the theory to some extent, but I personally felt that there were many areas where I was lacking in implementation. This lecture was very helpful because it was as easy and detailed as possible.
I thought I knew a little bit about pandas, but it turns out that simply knowing how to do something doesn't mean you'll get better results. This is a great lecture that taught me how to apply machine learning algorithms to complex data sets, and the importance of data preprocessing and analysis domains. After listening to this, I'm shocked to see why my rankings aren't going up on Kaggle. The prize money this year is also huge...