mrdbourke/zero-to-mastery-ml

Features

• Machine Learning Curricula - Provides a structured educational curriculum for mastering machine learning concepts and algorithms through interactive notebooks.
• Machine Learning Education - Provides a comprehensive curriculum for learning the end-to-end process of building and evaluating machine learning models.
• Interactive Notebook Environments - Delivers an educational curriculum through computational document formats that interleave explanatory text with executable code.
• Exploratory Data Analysis - Implements workflows for loading, cleaning, visualizing, and summarizing datasets to understand their structure.
• Dataset Preparation Utilities - Provides scripts for formatting and preparing datasets, including feature-label splitting and imputation, for model training.
• Neural Network Layers - Teaches the use of pre-defined architectural building blocks, such as dense and convolutional layers, to construct deep learning models.
• Model Experimentation Frameworks - Implements processes for iterating through training, validation, and testing to ensure model generalization.
• Hyperparameter Tuning - Implements techniques for optimizing model parameters to minimize generalization error using search strategies.
• Hyperparameter Optimization - Provides automated methods for searching and selecting the best configuration parameters for a model.
• Predictive Model Development - Guides the development, tuning, and evaluation of predictive models for classification and regression tasks.
• Python Data Science Primers - Offers introductory guides covering Python programming, NumPy arrays, Pandas DataFrames, and Matplotlib for data science exploration.
• Statistical Analysis - Uses descriptive and inferential statistics to interpret datasets and identify meaningful patterns.
• Training and Testing Splits - Divides datasets into training and testing subsets to ensure model evaluation is performed on unseen data.
• Transfer Learning - Teaches techniques for adapting pre-trained models to new tasks or datasets by modifying model heads.
• Deep Learning and Computer Vision - Implements deep learning and computer vision models for complex tasks like image recognition.
• Data Manipulation - Provides packages and utilities for cleaning, tidying, and transforming datasets to prepare them for modeling.
• Python Data Science Courses - Teaches the core Python data science ecosystem, including NumPy, Pandas, and Matplotlib for numerical and tabular analysis.
• Array Inspection - Provides capabilities for retrieving metadata including data types and dimensions from numerical arrays.
• Tabular Data Manipulation Guides - Provides educational resources and practical guides for cleaning, reshaping, and aggregating structured data.
• Tabular Structure Creation - Teaches how to build dataframes and series from dictionaries or lists to organize data for analysis.
• Tabular Filtering - Implements data extraction using boolean masks and logical expressions within tabular structures.
• Deep Learning Courses - Delivers a practical guide for constructing, training, and deploying neural networks and computer vision models via TensorFlow.
• Jupyter Notebook Curricula - Delivers structured learning paths through interactive Jupyter notebooks that interleave code and narrative.
• Scikit-Learn Examples - Provides practical code demonstrations and tutorials for implementing supervised and unsupervised learning using Scikit-Learn.
• Numerical Data Plotting - Generates essential chart types including line, scatter, and bar plots from numerical arrays.
• Array Broadcasting - Implements the process of expanding smaller arrays to match the shape of larger ones during element-wise operations.
• Element-wise Array Operations - Provides high-performance operations that apply mathematical functions across every element of multi-dimensional arrays.
• Multi-Dimensional Arrays - Demonstrates how to generate vectors, matrices, and higher-dimensional arrays from lists, tuples, and numerical patterns.
• Numerical Array Operations - Implements a comprehensive guide to performing high-performance mathematical calculations and indexing on multi-dimensional arrays.
• Numerical Libraries - Provides a foundation in using optimized libraries for matrix operations and general numerical calculations.
• Classification Metrics - Calculates precision, recall, and accuracy metrics to measure a model's ability to categorize data.
• Fine-Tuning Strategies - Demonstrates how to tweak existing pre-trained models to solve specific target tasks with limited data.
• Dataset Distribution Analysis - Performs visual and statistical analysis of feature distributions and class balances within datasets.
• Dataset Statistics Analyzers - Provides tools for calculating descriptive statistics, data types, and size metrics of datasets.
• Deep Learning Architectures - Guides the assembly of deep learning models by stacking specialized layers for feature extraction and classification.
• Error Analysis Strategies - Provides methodologies for diagnosing model performance by analyzing training and test errors to identify bias.
• Feature Contribution Analysis - Implements methods for estimating the influence of individual features on model predictions via weight analysis.
• Feature Correlation Analysis - Applies statistical methods for evaluating the relationships between different data variables in a dataset.
• Feature Importance Attribution - Provides algorithms for quantifying the relative contribution of individual input variables to model predictions.
• Learning Task Selection - Provides a framework for identifying whether a business challenge requires supervised, unsupervised, or transfer learning.
• Regression Training - Trains learning algorithms on features and labels to predict continuous target variables for new data.
• Model Evaluation Metrics - Establishes success criteria using classification and regression metrics to measure trained model quality.
• Model Training Pipelines - Provides end-to-end workflows for fitting models to datasets by identifying patterns between features and labels.
• Sequential Workflows - Organizes the machine learning lifecycle into a linear sequence of preparation, fitting, and evaluation.
• Model Evaluation Metrics - Uses ROC curves, AUC scores, and confusion matrices to quantitatively analyze model performance.
• Hyperparameter Search Strategies - Implements algorithmic methods for finding optimal model settings, including grid search and random sampling.
• Model Performance Improvement - Provides methods to modify models and tune hyperparameters to enhance overall performance and efficiency.
• Model Performance Evaluators - Provides tools for quantifying the accuracy and reliability of machine learning models by comparing predictions against ground truth.
• Model Predictions - Provides capabilities for generating target values and predictions from trained models for new data.
• Image Labeling Engines - Provides logic for generating class labels and probabilities specifically for image data.
• Model Performance Selection - Guides the selection of models ranging from linear estimators to deep neural networks based on data constraints.
• Algorithm and Hyperparameter Selection - Teaches how to select the most effective machine learning algorithm and hyperparameters based on dataset characteristics.
• Model Persistence - Implements mechanisms for saving and retrieving trained machine learning model files from disk.
• K-Fold Cross-Validation - Uses validation techniques that partition data into k segments for iterative training and evaluation.
• Functional Model Architectures - Teaches the construction of complex neural networks using functional APIs with multiple inputs and outputs.
• Sequential Model Builders - Implements neural networks by stacking layers in a linear sequence from input to output.
• Neural Network Training Frameworks - Provides instructions for fitting neural networks to datasets by configuring optimizers and loss functions.
• Predictive Model Implementations - Implements classification and regression algorithms to demonstrate how predictive models process specific datasets.
• Prediction Visualization - Provides tools for displaying model prediction results, including the generation of confusion matrices.
• Preprocessing Pipelines - Chains imputation, encoding, and model fitting into automated workflows to streamline the ML process.
• Supervised Learning Frameworks - Provides a framework for building and training predictive models using supervised learning patterns.
• Time Series Feature Engineering - Extracts granular date components from timestamps to engineer features for temporal datasets.
• Image Augmentations - Applies random transformations like scaling and cropping to training images to improve model generalization.
• Transfer Learning Implementations - Provides practical guides for applying pre-trained neural networks to new tasks to reduce training time.
• Unsupervised Learning - Teaches algorithms for discovering patterns and clustering similar samples in unlabeled datasets.
• Cross Validation Evaluation - Implements methods for evaluating model performance using cross-validation to ensure stability across data splits.
• Categorical Encodings - Implements conversion of categorical variables into indicator columns for feature engineering in machine learning.
• Dataframe Visualizers - Integrates directly with tabular dataframes to generate visual exploration interfaces.
• Data Preprocessing Pipelines - Provides tools for cleaning and formatting raw data through reusable preprocessing pipelines for ML ingestion.
• Data Transformation Pipelines - Provides tools and workflows for cleaning, formatting, and enriching data to optimize it for model consumption.
• Image-to-Array Converters - Transforms image files into numerical arrays to enable pixel-level pattern detection by algorithms.
• Multidimensional Indexing - Teaches how to retrieve specific values and slices from multidimensional arrays using coordinate-based indexing.
• Missing Value Imputation - Employs techniques for replacing null entries using constant values or statistical measures like median imputation.
• ML Lifecycle Pipelines - Implements automated workflows that chain data preparation, model training, and evaluation into a single pipeline.
• Image-to-Tensor Loaders - Includes utilities that load image files from directories directly into numerical tensors for deep learning.
• Random Data Generators - Shows how to create arrays populated with random integers or floats using seeds for reproducible results.
• Advanced Array Indexing - Covers techniques for selecting and manipulating subsets of multidimensional data using complex indexing and slicing patterns.
• Element-wise Comparisons - Demonstrates element-wise logical comparisons between arrays and scalars to create boolean masks.
• Array Sorting and Partitioning - Provides guidance on ordering array values along specified dimensions and retrieving sorting indices.
• Array Statistical Aggregations - Provides tutorials on computing summary statistics like mean, sum, and variance across numerical arrays.
• Array Transpositions - Explains how to swap axes and change the dimensional structure of multi-dimensional arrays to align data.
• Bar Chart Generation - Visualizes quantities of similar themed items using vertical or horizontal bar plots.
• Histogram Generators - Groups numerical data into bins via histogram generators to analyze value distribution.
• Dot Product Computation - Teaches the calculation of matrix products by multiplying aligned inner dimensions of arrays.
• Machine Learning Pipelines - Teaches how to chain data cleaning, feature engineering, and model training into reusable automated pipelines.
• Line Plots - Visualizes trends over time by drawing lines between coordinate points.
• Model Accuracy Evaluators - Ships tools for measuring the correctness and performance of machine learning models against benchmark datasets.
• K-Fold Cross-Validation - Tests models across multiple different train-test splits to ensure performance is consistent and stable.
• Statistical Distribution Visualizers - Generates histograms and line plots directly from data tables to identify statistical distributions and patterns.

Star history