Concept Taxonomy¶

This document defines the categorical taxonomy for organizing the 200 machine learning concepts.

Taxonomy Categories¶

Core machine learning concepts that form the basis for all other learning. Includes fundamental terminology, data concepts, and basic ML paradigms.

Concepts related to the K-Nearest Neighbors algorithm including distance metrics, lazy learning, and KNN applications.

Concepts related to decision tree algorithms including splitting criteria, tree structure, entropy, information gain, and pruning.

Concepts related to logistic regression including sigmoid functions, binary and multiclass classification, maximum likelihood, and softmax.

Concepts related to support vector machines including margins, hyperplanes, kernel methods, and dual formulation.

Concepts related to unsupervised learning clustering algorithms, primarily k-means clustering and related metrics.

Concepts related to basic neural networks including neurons, activation functions, forward/backward propagation, and gradient descent.

Concepts specific to convolutional neural networks including convolution operations, pooling, feature maps, and CNN architectures.

Concepts related to transfer learning, pre-trained models, fine-tuning, and domain adaptation.

Concepts related to model evaluation, validation, cross-validation, confusion matrices, and performance metrics.

Concepts related to data preparation including normalization, standardization, encoding, and feature engineering.

Concepts related to optimization algorithms, regularization techniques, hyperparameter tuning, and model selection.

Concepts specific to regularization techniques to prevent overfitting, including L1, L2, dropout, and early stopping.

Other important concepts that don't fit cleanly into the above categories.