Kang Gyeonghun
Bayesian Linear Regression (FCB CH09, Hoff) Gyeonghun Kang @hun_learning, Yonsei ESC, Feb 06, 2021 Linear Regression Revisited Data: a target column $y\in(n\times 1)$, a design matrix $X=[x^{(1)}, x^{(2)}, …, x^{(p)}] \in (n\times p)$ Essence of LR Model = LINEAR CONDITIONAL MEAN $E[y\mid x]$ $$ E[y_i \mid x_i] = \int y ,p(y\mid x),dx =\beta^Tx_i $$ Normal LR model = linear $E[y\mid x]$ + NORMAL ERROR $\epsilon_i\sim N(0, \sigma^2)$ $$ y_i \sim N(\beta^Tx_i, \sigma^2) $$
Read more…