[MXML-10-01] Gradient Boosting Method (GBM) [1/7] - Regression: Training & Prediction process

*** Dubbing: [ English ] [ 한국어 ] We will begin the GBM series, an adaptive boosting method, with the 10th module of machine learning. Let's take a look at the full table of contents of the GBM series. The GBM part also consists of regression and classification. Chapter 1 provides an overview of the GBM. Chapter 2 examines the regression algorithm. We explore the training and prediction process and analyze the regression algorithm in detail. And we will implement the GBM and SGBM regression algorithms in code. Chapter 3 examines the binary classification algorithm of GBM. We will look at the training and prediction process, and analyze the classification algorithm, and then implement it as code. And in Chapter 4, we will look at the multiclass classifications. We will get an overview of multiclass classification and implement it in code. This video is part 1 and will cover the training and prediction process of the GBM. Let's take a look at the outline of the GBM. GBM is a boosting-based ensemble learning algorithm. Train multiple weak learners, and aggregate the results of those learners to make predictions. Decision Trees are used as the weak learners. GBM was proposed by Leo Breiman in 1997 and updated by Jeremy Friedman in 1999. Since then, it has developed into XGBoost and Light GBM, etc. GBM is applicable to both regression and classification and is characterized by training on the residuals of the target data. #GradientBoosting #GradientBoostingMethod #GBM