Unsupervised Learning Dimensionality 

This course is designed to provide a comprehensive understanding of unsupervised learning techniques, with a focus on dimensionality analysis and clustering methods, specifically tailored for applications in Production Control and Operations (P&OC). Participants will learn how to extract meaningful insights from unlabeled data, identify hidden patterns, and develop effective strategies for data preprocessing and feature engineering in production control, inventory management, and workflow optimization contexts.

• Understand the fundamental principles of unsupervised learning and its applications in Production Control and Operations

• Apply dimensionality reduction techniques to improve model performance and data visualization in P&OC

• Implement and evaluate various clustering algorithms for workflow optimization, inventory segmentation, and anomaly detection

• Develop effective strategies for data preprocessing and feature engineering in unsupervised learning tasks for P&OC

• Leverage unsupervised learning techniques to solve real-world problems in Production Control and Operations domains

1. Introduction to Unsupervised Learning in P&OC

• Overview of unsupervised learning and its differences from supervised learning

• Types of unsupervised learning tasks and their applications in Production Control and Operations

• Challenges and considerations in unsupervised learning for production control, inventory management, and workflow optimization data

• Hands-on exercises: Exploring and visualizing unlabeled P&OC datasets

2. Dimensionality Analysis for P&OC

• The curse of dimensionality and its implications for machine learning in production control and operations

• Principal Component Analysis (PCA) for linear dimensionality reduction in P&OC data

• t-SNE and UMAP for non-linear dimensionality reduction and data visualization in P&OC

• Autoencoders and their applications in dimensionality reduction and anomaly detection for production control and operations data

• Hands-on exercises: Applying dimensionality reduction techniques to P&OC datasets

3. Clustering Methods in P&OC

• Overview of clustering and its applications in Production Control and Operations

• K-means clustering and its variations (e.g., K-medoids, Mini-batch K-means) for inventory segmentation and workflow optimization

• Hierarchical clustering (Agglomerative and Divisive) for production process analysis and resource allocation

• Density-based clustering (DBSCAN) for anomaly detection and outlier analysis in production control and operations data

• Evaluation metrics for clustering performance (e.g., Silhouette score, Calinski-Harabasz index) in P&OC

• Hands-on exercises: Implementing and evaluating clustering algorithms on P&OC case studies

4. Advanced Topics and Applications in P&OC

• Gaussian Mixture Models (GMM) for probabilistic clustering of production control and operations data

• Self-Organizing Maps (SOM) for data visualization and clustering in workflow analysis

• Combining unsupervised and supervised learning techniques (e.g., clustering for feature engineering in predictive maintenance)

• Real-world applications of unsupervised learning in Production Control and Operations (e.g., capacity planning, supply chain optimization)

• Hands-on exercises: Developing an end-to-end unsupervised learning pipeline for a P&OC problem

• Solid understanding of mathematics, including linear algebra and statistics

• Proficiency in programming with Python or R

• Familiarity with basic machine learning concepts and supervised learning algorithms

• Knowledge of production control and operations management principles