News:

SMF - Just Installed!

Main Menu

Statistics with Rust, 3rd Edition

Started by BeMyLove, Today at 05:09 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

BeMyLove



Statistics with Rust, 3rd Edition | 8.02 MB

Title: Statistics with Rust, 3rd Edition
Author:  GitforGits
Category: Programming Languages, Programming, Object Oriented Programming
Language: English | 440 Pages | ISBN: 9781718504455


Description:
Designed for data analysts, statisticians and engineers to build fast and reproducible statistical analyses in Rust
In this book, we work through a real dataset to demonstrate how to grow a single Cargo workspace into a complete statistical toolkit. Starting with the installation of Rust 1.96, we will address version discrepancies across ndarray, linfa and smartcore. Next, it shows you to load and clean real data using csv, serde and Polars, and then summarise it using robust statistics, kernel density estimates and Plotters charts. After that, we will fit probability distributions with statrs, build confidence intervals, run parametric, non-parametric and permutation tests, and control error rates.
The later chapters cover OLS, LASSO, logistic regression models, Bayesian inference using Metropolis and Hamiltonian Monte Carlo methods, and multivariate methods such as PCA, CCA, LDA, ICA and MDS. The book also covers machine learning techniques such as decision trees, random forests, support vector machines and neural networks. The final chapter will cover honest evaluation using ROC curves, cross-validation, nested grid searches, bootstrap intervals, AIC, BIC and McNemar tests.

   Key Learnings
Build reproducible Rust statistics workspace with pinned crates and fixed toolchains.
Load, clean, encode and query real datasets using csv, serde and Polars pipelines.
Summarize data with robust statistics, kernel density estimates and Plotters charts.
Fit and validate probability distributions using maximum likelihood, Q-Q plots and KS distance.
Run confidence intervals, parametric, non-parametric and permutation tests with corrected error rates.
Fit and diagnose OLS, lasso and logistic regressions with robust standard errors.
Implement Metropolis and Hamiltonian Monte Carlo samplers and diagnose chains with R-hat.
Apply PCA, CCA, LDA, ICA and MDS to uncover structure in multivariate data.
Train decision trees, random forests, SVMs and neural networks with smartcore, linfa and burn.
Evaluate models honestly with ROC curves, nested cross-validation, bootstrap intervals and McNemar tests.

   Table of Content
Introduction to Rust for Statisticians
Data Handling and Preprocessing
Descriptive Statistics
Probability Distributions and Random Variables
Inferential Statistics
Regression Analysis
Bayesian Statistics
Multivariate Statistical Methods
Nonlinear Models and Machine Learning
Model Evaluation and Validation