Machine Learning

Due to rapid technological advancements, the foundation of AI solutions for decision-making, "Supervised" ML models, is now more accessible to practitioners. Mastering the most critical predictive models has become more accessible, especially with the improvement of automation tools.

This workshop provides a comprehensive overview of "supervised" Machine Learning algorithms and their role in improving predictions across various industries. To ensure practical application, it also explores models using different technologies (SAS, Alteryx, STATISTICA, PYTHON, etc.), enabling participants to become professional practitioners and expert consultants by evaluating and selecting the most suitable solution with the right technical package.

• Explore the rise of AI with IoT and technology capacities.
• Understand the true meaning of Machine Learning (ML).
• Connect Data Analysis to Machine Learning.
• Make it all with f(X) = y.
• Differentiate between Regression and Classification models.
• Validate models with p-value and Accuracy metrics.
• Improve predictions by testing different ML models.
• Fine-tune models with the stepwise methods.
• Compare models using accuracy measures.
• Understand the utility of all cross-validation techniques.
• Design dashboards for comparative models.
• Overview Ensemble Models.

Day 1:

- Introduction to Machine Learning / - Multiple Linear Regression 

Day 2: 

- Simple & Multiple Logistic Regression 

- Models Evaluation Indicators

Day 3:

-  Linear and Quadratic Discriminant Analysis

Day 4: 

- Decision Trees / Random Forest Trees /Naive Bayes

Day 5: 

- Support Vector Machines / K Nearest Neighbors

- Colored PPT documents / Videos.

- The multiple-way model generation.

- All-in-one predictive model solution.

- Quality model indicators.

- Team Competition for Best Model Finding.  

- Complete case studies from A to Z.

- Proprietary tools for data visualization.

- Methods for selecting the best model.

- ROC charts.

It is common to have data sets with multiple variables describing business topics. But how can we extract all the hidden patterns from such complex data sets? Reducing the number of variables into simple "maps" with PCA becomes essential to highlight invisible relations within the data, facilitating the correct actions to take. This workshop also explains the difference between scientific market "clustering" and simple common sense "filtering". It empowers the definition of market niches and profiles them using Data Analysis techniques. Additionally, the program covers illustrations that reveal associations between the components of multiple variables for efficient tracking of pattern evolution. The workshop also includes practical applications with two different technologies, allowing participants to become more like consultants than mere experts.

• Highlighting the role of statistical data analysis in Unsupervised Machine Learning.
• Discovering hidden patterns within data sets.
• Understanding the "Dimension Reducibility" concept.
• Mastering all "pattern finding" algorithms in AI applications.
• Mapping complex data sets of multiple variables in simple charts.
• Visualizing relationships between variables and categories.
• Evaluating the quality of reduced multidimensionality.
• Differentiating between clustering and filtering.
• Delving into all clustering techniques and their applications
• Running professional segmentation with clustering.
• Applying it all using specialized software.

Day 1:

- Matrix Factorization / Principal Component Analysis  

Day 2: 

- t-SNE and Multi-Dimensional Scaling

- Simple Correspondence Analysis

Day 3:

- Agglomerative Clustering / K Means and Medoids

Day 4: 

- Recommender Systems

Day 5: 

- Gaussian Mixture Models

- Colored PPT documents / Videos.

- Multidimensional Reducibility

- Hierarchical   vs. Divisive Clustering

- Proprietary vs. Open Source tools   solutions

- Eigenvalues and Eigenvectors

- Team Exercises

- The science behind mapping illustration.

- Quality   measurements of unsupervised models

- Demystifying the Curse of   Multidimensionality

 The "Reinforcement Learning in Practice" training is an in-depth, hands-on workshop designed to equip participants with the knowledge and practical skills needed to effectively implement Reinforcement Learning (RL) techniques. Over three days, attendees will explore the fundamental principles of RL, understand key algorithms, and apply them to real-world problems. This workshop blends theoretical insights with coding exercises and case studies, ensuring participants gain conceptual clarity and practical experience building RL models. 

- Understand the fundamentals of Reinforcement Learning (RL) 

- Explore RL key concepts: agents, environments, rewards, and policies.

- Implement basic RL algorithms like Q-Learning and SARSA.

- Apply policy-based and value-based methods for sequential decision-making.

- Balance exploration and exploitation trade-offs.

- Optimize RL models through hyperparameter tuning.

- Discuss real-world RL applications such as robotics, finance, ...

 Day 1:

- RL basics: agents, rewards, policies, environments.

- Exploration vs. exploitation.

- Q-Learning & Markov Decision Processes (MDPs).

Day 2:

- Value-based vs. policy-based methods.

- SARSA, DQN, and Policy Gradient.

Day 3:

- RL in robotics, finance, and gaming.

- Actor-critic, PPO, A3C methods.

- Simulating RL environments.

• Utilize RL environments for experimentation.
• Exploring real-world applications in robotics, finance, gaming, and automation.
• Hands-on coding sessions using Python, OpenAI Gym, TensorFlow, or PyTorch.
• Simulating RL environments to develop problem-solving strategies.
• Learning industry best practices for deploying RL models in production.