Mathematical Statistics Lecture -

The lecture moves to estimation. The Method of Moments is introduced first—intuitive, ancient, but statistically inefficient. Then, the crown jewel: Maximum Likelihood Estimation (MLE). The professor writes:

[ \hat\theta\textMLE = \arg\max\theta \in \Theta L(\theta; x) ]

The MLE is not just a recipe; it is a theorem waiting to happen. Under regularity conditions, the lecture will sketch the proof of its consistency (as sample size grows, the estimator converges to the true value) and asymptotic normality:

[ \sqrtn(\hat\theta - \theta) \xrightarrowd N(0, I(\theta)^-1) ]

Here, ( I(\theta) ) is the Fisher information—a measure of how much information the data carry about ( \theta ). The Cramér-Rao lower bound, derived earlier, now reveals its teeth: no unbiased estimator can have variance lower than ( 1/I(\theta) ). The MLE asymptotically achieves this bound. It is, in the limit, the best possible. mathematical statistics lecture

Mathematical statistics provides powerful tools for data analysis and decision-making. Understanding probability and statistical inference are crucial steps in extracting meaningful information from data. These concepts form the backbone of more advanced statistical methods and are widely applied across various disciplines, from social sciences to medicine and engineering.

A lecture is only as good as the textbook it follows. Different universities use different bibles. Here is how to match the lecture to the text:

Pro Tip: If your professor says, "We are using Casella & Berger, Chapters 1-7," you must watch lectures twice. Once to get the idea, once to follow the math. The lecture moves to estimation

The problem: The professor defines p-value as ( P(T \geq t_obs | H_0) ), but the homework asks for a two-tailed p-value for an asymmetric distribution. The fix: Remember the strict definition: The smallest ( \alpha ) for which you would reject ( H_0 ). If the distribution is asymmetric, you must double the smaller tail, or use the likelihood ratio principle.

Introduction: The Map and the Territory

Welcome to the engine room of data science. While descriptive statistics organizes data, and probability theory models chance, Mathematical Statistics is the discipline that connects the two. It is the science of making inferences about a population based on a sample.

The fundamental problem we face is this: We observe the data (the sample), but we want to understand the reality that generated that data (the population). We have the map (the data), but we want to understand the territory (the truth). The problem: The professor defines p-value as (

This lecture breaks down the core pillars of the field: Probability Models, Estimation, and Hypothesis Testing.

A ( 100(1-\alpha)% ) confidence interval (CI) is a random interval ([L, U]) such that: [ P(\theta \in [L, U]) = 1 - \alpha ] Example (Normal, known variance): [ \barX \pm z_\alpha/2 \frac\sigma\sqrtn ]

How do we estimate $\theta$? We use an Estimator, which is simply a function of the sample data, denoted as $\hat\theta$.

Not all estimators are created equal. Mathematical statistics provides criteria to judge them.