#Python
RJ Studio’s 130th video is conducting SMOTE, which is a popular oversampling technique using Python's Imbalanced-learn library.
◕ ◡ ◕
According to ChatGPT (June 4, 2023) 🤖, this is how SMOTE works:
1. Identify the minority class instances in the dataset.
2. Select a minority class instance at random.
3. Identify the k nearest neighbors of the selected instance from the minority class.
4. Randomly choose one of the k nearest neighbors.
5. Generate a new synthetic instance by interpolating between the selected instance and the chosen neighbor. This is done by calculating the difference between the feature vectors of the two instances and multiplying it by a random number between 0 and 1. The resulting difference is added to the selected instance to create a new synthetic instance.
6. Repeat steps 2-5 until the desired number of synthetic instances is generated.
I also demonstrate how "SMOTE" is different to "Random Oversampling" by visualizing the outcomes of the 2 techniques~! 🚀
☆ Video Chapters ☆
0:00 Intro
0:47 IMbalanced Dataset
2:10 ROS & SMOTE
3:00 ROS vs SMOTE
3:48 Outro
Please check my GitHub for code: https://github.com/rjlee13/RJStudio_GH
Python logo in the thumbnail is from https://www.python.org/community/logos/
Imbalanced-learn library logo in the thumbnail is from https://imbalanced-learn.org/stable/i...
For background music, I chose 2 good songs by J.Fla:
The Monster & The Ocean
^‿^
Hope my video taught you something new!
And feel free to correct me if I made mistakes in the comments :)
이번에는 파이썬 Imbalanced-learn을 사용해서 SMOTE이란 대표적인 오버 샘플링에 대한 비디오를 만들었습니다. ≧◡≦
또, 랜덤 오버샘플링과 SMOTE는 어떻게 다른지 데이터 시각화를 통해 확인해보았습니다.
유익한 정보이길 바랄게요~ 감사합니다.
“In teaching others, we teach ourselves.”
LinkedIn: / rj-lee
Instagram: / musicalum_rj
Music YouTube: / @musicalum
On this page of the site you can watch the video online #Python with a duration of hours minute second in good quality, which was uploaded by the user RJ Studio 03 June 2023, share the link with friends and acquaintances, this video has already been watched 536 times on youtube and it was liked by 6 viewers. Enjoy your viewing!
