Further Readings
A curated reading list for data science and machine learning: classic essays, tutorials, and books recommended by Saed Sayad.
2 min read · Updated August 8, 2026
The Data Mining Map covers the core workflow, but data science is a deep field. This page collects the readings the legacy site recommended to generations of learners — short essays and tutorials that go one level deeper than the map, grouped by theme. They pair naturally with the tutorials: read the map page for the working method, then the essay for the theory underneath.
Probability and statistical foundations
- A Short Essay on Data, Probability, Randomness and Uncertainty
- Maximum Likelihood Estimation
- Understanding the Bias-Variance Tradeoff
- Principal Components Analysis
Core machine learning methods
- Optimality of Naive Bayes — a natural companion to Naive Bayes
- Support Vector Machines — pairs with Support Vector Machine
- Random Forest and Gradient Boosting Machine — the ensemble successors to Decision Trees
- Combining Estimators to Improve Performance
- Bayesian Belief Network
- Robust Regression — pairs with Regression
Deep learning and text mining
- Deep Learning (ebook)
- Convolutional Neural Networks — extends Neural Networks
- Natural Language Processing in Python
- Mining Text and Web Data
Evaluation, visualization, and forecasting
- Receiver Operating Characteristics graphs (ROC101) — the classic ROC tutorial, companion to Evaluation: Classification
- Multivariate Visualization — extends Data Exploration
- Time Series and Forecasting
From the author and beyond
- Real Time Machine Learning — Saed Sayad’s book on learning systems that update continuously rather than in batches, and the intellectual ancestor of the Genomarker platform described on this site.
- StatQuest — Josh Starmer’s video lessons, the friendliest rigorous walkthroughs of the statistics behind nearly every method on the map.
Summary
Start with the bias-variance essay and ROC101 — they underpin Model Evaluation — then follow whichever method group you use most. Every reading here was chosen to deepen, not duplicate, the tutorials in the map.