Chapter 10

Statistics

Samples, splits, estimators, bias and variance, likelihood, uncertainty, and validation.

What this chapter does

Statistics connects data to claims. This chapter teaches how samples, estimators, splits, likelihood, uncertainty, and validation help you decide what a result actually supports.

Lessons

Read these in order.

Start with the chapter introduction, then move through the topic lessons. The order is chosen so each page can reuse ideas from the pages before it.

  1. 01
    Introduction

    Why statistics gives machine learning its caution.

  2. 02
    Datasets and Samples

    Observed data, sampled evidence, and the population we care about.

  3. 03
    Train, Validation, and Test Splits

    Separating fitting, choosing, and final reporting.

  4. 04
    Estimators

    Rules that use data to estimate unknown quantities.

  5. 05
    Bias and Variance

    Systematic error and sensitivity to the sample.

  6. 06
    Likelihood

    How plausible the observed data is under a parameter choice.

  7. 07
    Maximum Likelihood

    Choosing parameters that make the observed data most plausible.

  8. 08
    Bayesian Updating

    Combining prior belief with evidence.

  9. 09
    Confidence Intervals

    Ranges that express sampling uncertainty.

  10. 10
    Hypothesis Tests

    Evidence against a baseline assumption.

  11. 11
    Cross-Validation

    Repeated held-out evaluation across folds.

  12. 12
    Resampling

    Repeated samples used to study estimate variability.

Before moving on

  • Reason about data splits, model estimates, and sampling uncertainty.
  • Understand why likelihood and maximum likelihood appear in training.
  • Separate development evidence from final evaluation evidence.
  • Read validation results with caution and context.

Where this leads

  • Evaluation
  • Probabilistic Models

Chapter progress