3. Shrinkage Estimators
As a remedy for overdetermined systems and variable selection rule, I would like to cover shrinkage methods in this section. Especially, I would like to focus on ridge and lasso penalization.
» continue reading
2. Linear classifiers
In this chapter, linear classifiers will be introduced and will be compared. Specifically, logistic regression and linear discriminant analysis will be described in detail.
» continue reading
2. Univariate Kernel Density Estimation
From histogram, we will start discussing nonparametric kernel methods.
» continue reading
2. Explore-Then-Commit algorithm
Here, we continue to describe the multi-armed bandit problem in detail. The notion of regret will be introduced. Then our first bandit algorithm, explore-then-commit (ETC) will be described.
» continue reading
1. Introduction to Multi-armed Bandit
The [Bandit] series of posts is my memo on the lecture Seminar in Recent Development of Applied Statistics (Spring, 2021) by Prof. Myunghee Cho Paik. This lecture focuses on adaptive sequential decision making. To be more specific, it covers wide variants of bandit problems.
» continue reading