Skip to content

HydraulicSheep/superStats

Repository files navigation

superStats

A developing collection of experiments in Statistics, Game Theory and Probability.

Contents:

  • Regret Matching - An introduction to the strategy-finding method of regret matching. Mathematical analysis, probabilities and simulation.

  • Linear Regression: Simulation and Calculation - Optimising Linear Regression via Simulated Annealing. Comparison with standard statistical regression calculations (Normal Equations).

  • Classification - In or Out? - Performing Logistic Regression from first principles via Gradient Ascent. Plotting of Decision Boundary and Training/Testing models.

  • Function Approximation - Outlining one Universal Approximation Theorem, writing a neural network from scratch and comparing models for a parabola. Debugging Neural Networks.

About

A developing collection of experiments in Statistics, Game Theory and Probability.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published