Machine Learning with Python
Objectives:
- Understand the fundamental principles of Machine Learning (supervised, unsupervised, reinforcement learning)
- Prepare and transform data for model training using scikit-learn
- Train and evaluate classification, regression, and clustering models
- Discover the basics of Deep Learning using TensorFlow/Keras
Objectives:
- Understand the fundamental principles of Machine Learning (supervised, unsupervised, reinforcement learning)
- Prepare and transform data for model training using scikit-learn
- Train and evaluate classification, regression, and clustering models
- Discover the basics of Deep Learning using TensorFlow/Keras
Target Audience
- Professionals pursuing an advanced career transition into data science
- Analysts who want to understand the principles and implementation of Machine Learning with Python in order to contribute to ML engineering or predictive analytics projects (real estate, finance).
Learning Objectives
- Understand the fundamental principles of Machine Learning (supervised, unsupervised, reinforcement learning)
- Prepare and transform datasets for model training with scikit-learn
- Train and evaluate regression models (linear, polynomial, regularization)
- Train and evaluate classification models (logistic regression, decision trees, Random Forest)
- Apply clustering techniques (K-Means, DBSCAN, hierarchical clustering)
- Discover the basics of Deep Learning with TensorFlow/Keras
- Build a complete end-to-end ML pipeline using real datasets
Prerequisites
- Basic knowledge of Python (variables, functions, loops, data structures)
- Basic mathematical knowledge (descriptive statistics, basic algebra)
- Completion of the Python Beginner training or equivalent level
- A computer with Python 3.x installed and internet access
Teaching Methods
- Combination of theory (40%) and practice (60%)
- Interactive lectures and live demonstrations of ML models
- Individual and team labs using scikit-learn, TensorFlow/Keras and Jupyter Notebook
- Active learning: collaborative problem solving and group analysis of results
- Digital course materials and exercise files (permanent access)
Target Certification: RS6763 – Manipulating, analyzing and visualizing data using Python Data Science modules
Detailed Program
DAY 1 — Machine Learning Fundamentals
- Introduction to ML: definition, types of learning, applications
- Supervised vs unsupervised vs reinforcement learning
- ML pipeline: data → preparation → training → evaluation → deployment
- Key concepts: features, labels, model, prediction
- Bias and variance: underfitting, overfitting, trade-offs
- Train/test split: importance, stratification, cross-validation
- Introduction to scikit-learn: installation, conventions, API
â– Hands-on labs:
→ Lab 1: Exploring a classic dataset (Iris, Titanic)
→ Lab 2: First classification model with scikit-learn
→ Lab 3: Analyzing the impact of train/test split on model performance
DAY 2 — Data Preparation and Regression
- Feature engineering: creation, selection, feature importance
- Handling missing values: imputation strategies
- Encoding categorical variables: OneHot, Label, Target encoding
- Normalization and standardization: MinMaxScaler, StandardScaler
- Linear regression: principle, equation, interpretation
- Polynomial regression and regularization (Ridge, Lasso)
- Regression metrics: MSE, RMSE, MAE, R²
â– Hands-on labs:
→ Lab 1: Complete data preparation pipeline
→ Lab 2: Real estate price prediction (linear regression)
→ Lab 3: Comparison of regularization techniques
DAY 3 — Classification and Clustering
- Logistic regression: principle, sigmoid function, decision threshold
- Decision trees: construction, criteria, visualization
- Random Forest: ensemble learning, bagging, feature importance
- Classification metrics: accuracy, precision, recall, F1-score
- Confusion matrix and ROC curve
- K-Means clustering: principle, choosing k, elbow method
- DBSCAN and hierarchical clustering: comparison
â– Hands-on labs:
→ Lab 1: Spam email classification with Random Forest
→ Lab 2: Customer segmentation using K-Means clustering
→ Lab 3: Comparing classification models on the same dataset
DAY 4 — Introduction to Deep Learning and Final Project
- Neural networks: perceptron, layers, activation functions
- MLP architecture: dense layers, forward propagation
- Introduction to TensorFlow/Keras: installation, Sequential API
- Training concepts: epochs, batch size, loss functions, optimizers
- Dimensionality reduction: PCA, t-SNE for visualization
- Transfer learning: concept and use cases
- Saving and loading models
â– Hands-on labs:
→ Lab 1: MNIST image classification with a neural network
→ Lab 2: High-dimensional data visualization with PCA/t-SNE
→ Lab 3 (Final project): Complete end-to-end ML pipeline on real datasets
Assessment
- Initial assessment test at the beginning of the training
- Formative assessments: corrected practical exercises, group analysis of results, intermediate quizzes
- Final multiple-choice test of 30 questions (Day 4) — passing score: 60%
- Certificate of completion issued
Target Audience
- Professionals pursuing an advanced career transition into data science
- Analysts who want to understand the principles and implementation of Machine Learning with Python in order to contribute to ML engineering or predictive analytics projects (real estate, finance).
Learning Objectives
- Understand the fundamental principles of Machine Learning (supervised, unsupervised, reinforcement learning)
- Prepare and transform datasets for model training with scikit-learn
- Train and evaluate regression models (linear, polynomial, regularization)
- Train and evaluate classification models (logistic regression, decision trees, Random Forest)
- Apply clustering techniques (K-Means, DBSCAN, hierarchical clustering)
- Discover the basics of Deep Learning with TensorFlow/Keras
- Build a complete end-to-end ML pipeline using real datasets
Prerequisites
- Basic knowledge of Python (variables, functions, loops, data structures)
- Basic mathematical knowledge (descriptive statistics, basic algebra)
- Completion of the Python Beginner training or equivalent level
- A computer with Python 3.x installed and internet access
Teaching Methods
- Combination of theory (40%) and practice (60%)
- Interactive lectures and live demonstrations of ML models
- Individual and team labs using scikit-learn, TensorFlow/Keras and Jupyter Notebook
- Active learning: collaborative problem solving and group analysis of results
- Digital course materials and exercise files (permanent access)
Target Certification: RS6763 – Manipulating, analyzing and visualizing data using Python Data Science modules
Detailed Program
DAY 1 — Machine Learning Fundamentals
- Introduction to ML: definition, types of learning, applications
- Supervised vs unsupervised vs reinforcement learning
- ML pipeline: data → preparation → training → evaluation → deployment
- Key concepts: features, labels, model, prediction
- Bias and variance: underfitting, overfitting, trade-offs
- Train/test split: importance, stratification, cross-validation
- Introduction to scikit-learn: installation, conventions, API
â– Hands-on labs:
→ Lab 1: Exploring a classic dataset (Iris, Titanic)
→ Lab 2: First classification model with scikit-learn
→ Lab 3: Analyzing the impact of train/test split on model performance
DAY 2 — Data Preparation and Regression
- Feature engineering: creation, selection, feature importance
- Handling missing values: imputation strategies
- Encoding categorical variables: OneHot, Label, Target encoding
- Normalization and standardization: MinMaxScaler, StandardScaler
- Linear regression: principle, equation, interpretation
- Polynomial regression and regularization (Ridge, Lasso)
- Regression metrics: MSE, RMSE, MAE, R²
â– Hands-on labs:
→ Lab 1: Complete data preparation pipeline
→ Lab 2: Real estate price prediction (linear regression)
→ Lab 3: Comparison of regularization techniques
DAY 3 — Classification and Clustering
- Logistic regression: principle, sigmoid function, decision threshold
- Decision trees: construction, criteria, visualization
- Random Forest: ensemble learning, bagging, feature importance
- Classification metrics: accuracy, precision, recall, F1-score
- Confusion matrix and ROC curve
- K-Means clustering: principle, choosing k, elbow method
- DBSCAN and hierarchical clustering: comparison
â– Hands-on labs:
→ Lab 1: Spam email classification with Random Forest
→ Lab 2: Customer segmentation using K-Means clustering
→ Lab 3: Comparing classification models on the same dataset
DAY 4 — Introduction to Deep Learning and Final Project
- Neural networks: perceptron, layers, activation functions
- MLP architecture: dense layers, forward propagation
- Introduction to TensorFlow/Keras: installation, Sequential API
- Training concepts: epochs, batch size, loss functions, optimizers
- Dimensionality reduction: PCA, t-SNE for visualization
- Transfer learning: concept and use cases
- Saving and loading models
â– Hands-on labs:
→ Lab 1: MNIST image classification with a neural network
→ Lab 2: High-dimensional data visualization with PCA/t-SNE
→ Lab 3 (Final project): Complete end-to-end ML pipeline on real datasets
Assessment
- Initial assessment test at the beginning of the training
- Formative assessments: corrected practical exercises, group analysis of results, intermediate quizzes
- Final multiple-choice test of 30 questions (Day 4) — passing score: 60%
- Certificate of completion issued