Jack-Cherish/Machine-Learning

Features

• Machine Learning Implementations - Provides from-scratch Python implementations of core supervised learning algorithms like linear regression and SVMs.
• Categorical Classifiers - Provides multiple algorithms like Naive Bayes and SVMs to classify categorical data.
• Classification Models - Implements multi-class classification capabilities using support vector models to categorize data into multiple distinct classes.
• Clustering Algorithms - Implements clustering algorithms to discover patterns and group similar unlabeled data points.
• Decision Trees - Implements recursive-partitioning decision trees using information gain and squared error minimization.
• Regression Trees - Implements decision trees for predicting continuous numerical outcomes by minimizing squared error.
• Ensemble Learning - Implements ensemble methods including bagging and boosting to improve predictive accuracy.
• Bagging Ensembles - Implements bagging ensembles to reduce model variance and overfitting through bootstrap sampling.
• K-Means Clustering - Implements k-means clustering to group unlabeled data points by minimizing distance between samples and centroids.
• Clustering Algorithms - Implements unsupervised learning algorithms like K-Means to discover patterns in unlabeled data.
• K-Nearest Neighbor Classifiers - Implements a k-nearest neighbors classifier to predict data categories based on majority vote of the closest samples.
• Linear and Logistic Regression - Implements logistic regression algorithms for predicting binary outcomes using a sigmoid function.
• Linear Regression - Implements linear regression to calculate optimal coefficients by minimizing squared errors.
• Linear Regression Models - Implements linear regression models to predict continuous numerical outcomes.
• Logistic Regression Models - Implements logistic regression models with support for regularization and class weighting.
• Recursive Partitioning Trees - Implements regression trees that split data into recursive partitions to predict continuous values.
• Boosting Algorithms - Implements boosting algorithms like AdaBoost that iteratively train models by re-weighting difficult samples.
• Binary Classification Training - Implements binary classification using a logistic function to map input features to a binary outcome probability.
• Naive Bayes Classifiers - Predicts categories using probabilistic models based on the conditional independence of features.
• SMO Algorithms - Implements the Sequential Minimal Optimization algorithm to train support vector machines.
• Supervised Learning - Provides a variety of predictive models for labeled datasets, including regression and classification.
• Support Vector Machines - Implements support vector machines to separate data classes by finding an optimal maximizing hyperplane.
• Text Classifiers - Implements text classifiers to predict document categories by calculating class probabilities from trained models.
• Unsupervised Learning - Implements unsupervised learning capabilities for discovering hidden patterns in unlabeled datasets.
• Machine Learning Educational Resources - Serves as an educational resource with practical implementations of core machine learning algorithms.
• Ordinary Least Squares - Implements ordinary least squares to find the best-fitting line between input features and target values.
• Regression Trees - Implements regression trees to estimate numerical outcomes by partitioning data into recursive segments.
• Model Evaluation - Provides tools for measuring model accuracy through error analysis and decision boundary visualization.
• Automated Feature Selection Tools - Implements feature selection by calculating information gain and Shannon entropy to identify discriminative attributes.
• Classifier Accuracy Metrics - Provides tools to measure classifier accuracy by calculating the percentage of incorrect predictions.
• Dataset Distribution Analysis - Provides visual analysis of dataset distributions to evaluate data separability.
• Decision Boundary Visualizations - Plots the separating boundaries between classes to evaluate how a classifier partitions feature space.
• Decision Stumps - Implements single-level decision stumps to serve as weak learners for ensemble algorithms.
• Pruning Techniques - Implements decision tree pruning to reduce model complexity and prevent overfitting.
• Tree Visualizers - Generates visual representations of the paths and splitting logic used by decision tree models.
• Feature Scale Normalization - Scales numeric features by mean and variance to ensure stable model convergence.
• Iterative Parameter Optimizations - Implements iterative parameter optimization using gradient ascent to maximize loss functions.
• Kernel-Based Feature Mapping - Provides kernel-based feature mapping to project data into high-dimensional spaces for non-linear classification.
• Regularization Techniques - Implements Ridge and stepwise regression to reduce model complexity and prevent overfitting.
• Text Classification - Implements a text classification workflow utilizing Naive Bayes and feature vectorization.
• Model Evaluation Metrics - Implements model evaluation metrics including confusion matrices, precision, recall, and ROC curves.
• Model Ensembling - Implements model ensembling techniques to improve predictive accuracy and robustness.
• AdaBoost Implementations - Implements AdaBoost to improve classifier performance by iteratively training weak learners.
• Text Tokenization - Segments raw text strings into lowercase word lists by removing non-alphanumeric characters.
• Gradient Ascent Algorithms - Implements gradient ascent to optimize model parameters by maximizing the defined loss function.
• Weighted Regression - Performs locally weighted regression using Gaussian kernels to reduce underfitting.
• Regression Scoring Evaluation - Evaluates regression performance by computing the total squared difference between real and predicted values.
• Stochastic Gradient Ascent - Implements stochastic gradient ascent to reduce computational complexity during model optimization.
• Text Feature Extraction - Transforms unstructured text into structured numerical features based on word frequency ranking.
• Vocabulary Generators - Implements vocabulary building to create master feature lists for text vectorization.
• Feature Vectorizations - Provides symmetric-matrix feature vectorization for transforming raw text and images into numerical formats.
• Tabular Data Preprocessing - Converts raw text files into feature matrices and label vectors for use in classifiers.
• Chinese Language Segmenters - Processes Chinese sentences into distinct words to handle the absence of whitespace during tokenization.
• Entropy Calculators - Measures dataset uncertainty by calculating the empirical entropy of class label distributions.
• Kernel-Based Feature Mapping - Uses kernel functions to project non-linear data into higher-dimensional spaces for linear separation.
• Ridge Regression - Implements ridge regression with L2 norm penalties to reduce collinearity and prevent model overfitting.
• Feature Interaction Visualizations - Provides tools to visualize feature distributions and relationships using scatter plots.

Star history