4  stats models

[1]

The process of finding the model that relates \(y\) to \(x\) and best fits the data is called regression analysis.

General Form of probabilistic Model in Regression is \[ y=E(y)+\varepsilon \]

4.1 Estimation

a statistic whose sampling distribution has the mean around the proposed value and a small variance. If the means are equal: unbiased estimator. No: biased.

4.2 build a model that fits the data

predicted value: \(\hat{y}\), \(\hat{\beta}\), …

  • model: \(y=\Exp(y)+\varepsilon\)

[2]

  • want to estimate \(f\): \(Y=f(X)+\varepsilon\)
  • want to use \(\hat{f}\) to estimate \(f\): \(\hat{Y}=\hat{f}(X)\).