# Python Machine Learning Bootcamp Canonical URL: ## Overview This course begins with linear and logistic regression, the most time-tested and reliable tools for approaching a machine learning problem. The course then progresses to algorithms with a very different theoretical basis, such as k-nearest neighbors, decision trees, and random forests. This brings important statistical concepts to the forefront, including bias, variance, and overfitting. You will also learn how to measure the accuracy of your models and gain practical tips for choosing effective features and algorithms. These skills are in high demand, as machine learning algorithms now power the majority of trading on Wall Street and the product recommendation systems at major companies like Amazon, Spotify, and Netflix. This course focuses on the practical skills needed to solve real-world problems with machine learning. The mathematical foundations for each algorithm will be explained visually, but there is no formal mathematics component. Students are expected to be comfortable writing Python programs and using the NumPy and Pandas libraries. ## What you'll learn - Explore foundational techniques like linear and logistic regression for modeling numerical and categorical data - Understand the difference between regression and classification problems and when to apply each approach - Build and evaluate models using k-nearest neighbors, decision trees, and ensemble methods like random forest - Learn key concepts such as cross-validation, training vs. test sets, and performance metrics like mean squared error - Apply feature engineering techniques to improve model accuracy while managing overfitting and bias-variance tradeoffs - Use Python's essential data science libraries, NumPy, Pandas, and scikit-learn, to structure data and implement algorithms - Gain insights into how machine learning powers systems at companies like Netflix, Spotify, and Amazon - Complete a final portfolio project that demonstrates your ability to apply machine learning to solve real problems ## Prerequisites This course requires students to be comfortable with Python and its data science libraries (NumPy and Pandas). If a student has not worked in Python before, we require a student to enroll in our [Python for Data Science Bootcamp](/classes/python-data-science-bootcamp-nyc)before taking this course. ## Curriculum ### 1. Course Kick‑off & Python Refresher - Data Science tool recap - Pandas and indexing - Exploratory data analysis (EDA): standard deviations and uniform vs. normal distributions using NumPy/Pandas - Hands‑on: loading CSVs, basic plotting with Matplotlib ### 2. Data Visualization & Simple Linear Regression - Crafting clear scatterplots: labels, grids, styling - Single‑variable linear regression (attendance → concessions) - Train‑test splitting and dealing with outliers - Evaluating models with R²; interpreting residuals - Extended example: car‑sales dataset, predicting price from one feature ### 3. Binary Classification & Logistic Regression - From regression to classification: why logistic vs. linear - Implementing logistic regression on an employee “stay/leave” dataset - Classification metrics deep dive: accuracy, precision, recall, F1 score, ROC curve - Understanding variability: train‑test ratios, data shuffling, sample size effects - Confusion matrix analysis ### 4. k‑Nearest Neighbors & the Iris Dataset - Introduction to k‑NN: distance metrics, choosing k - Dataset exploration: sepal/petal measurements, plotting clusters - Preprocessing: label encoding categorical data, feature scaling - Model training, hyperparameter tuning, evaluating with confusion matrix and classification report - Brief intro to decision‑tree logic (setting up for ensembles) ### 5. Ensemble Methods & Neural Networks - Random forest classifiers on the Titanic dataset: feature engineering, importance scores - Kaggle workflow: generating predictions, submitting to competition - Neural network primer: perceptron to multilayer architectures - Hands‑on MNIST digit classification with Keras/TensorFlow in Colab ## Schedule - Jun 15, 2026 – Jun 19, 2026 — Live Online - Aug 3, 2026 – Aug 7, 2026 — Live Online - Aug 30, 2026 – Oct 11, 2026 — Live Online - Sep 8, 2026 – Oct 8, 2026 — Live Online - Sep 22, 2026 – Sep 28, 2026 — Live Online - Nov 9, 2026 – Nov 13, 2026 — Live Online - Dec 7, 2026 – Dec 11, 2026 — Live Online - Dec 29, 2026 – Feb 2, 2027 — Live Online - Jan 17, 2027 – Feb 21, 2027 — Live Online ## FAQ ### How many students are in a given class? Noble's typical class ranges from 8-12 students, but we allow up to 20 students to register for our course. ### How does this class prepare me for the job market? The classes will allow students to learn advanced topics in data science used by the most cutting edge companies such as Google, Facebook, and more. These topics will allow students to build, evaluate, and reassess forecasting models on all forms of data. ### Is there mandatory work outside of the classroom? Students are not required to complete any work outside of class. However, we provide students with bonus materials if they would like extra practice. ### What tangible skills do students leave with after the class? Students will leave with the ability to learn how to build a model from start to finish. Students will learn how to clean and balance data, apply a form of learning algorithm on the data, perform a bias test, and finally evaluate the accuracy of your model. ## Pricing **Tuition:** $1895