Linear Regression
Simple numeric prediction baseline; interpretable but assumes a mostly linear relationship.
Model Atlas
Know why a model fits before using it.
A practical model map covering what each model does, where it is useful and the limitation you should remember in an interview or project.
Logistic Regression
Strong interpretable classification baseline with probability outputs.
Decision Tree
Rule-based model that is intuitive but can overfit without controls.
Random Forest
Many randomized trees reduce variance and handle nonlinear relationships well.
XGBoost
Sequential boosted trees that are strong on structured/tabular data.
K-Nearest Neighbors
Predicts from nearby examples; simple but sensitive to scaling and dataset size.
K-Means
Partitions data around centroids; useful when compact spherical clusters are meaningful.
PCA
Projects features into directions of maximum variance for compression and visualization.
Neural Network
Learns flexible nonlinear representations but needs more data, tuning and computation.
Transformer
Attention-based architecture behind modern language models and many sequence systems.
Embedding Model
Maps text or other objects into vectors for semantic search, clustering and retrieval.
LLM + RAG
Combines generation with external retrieval to improve grounding and domain usefulness.