ML100: Understanding Machine Learning in the Enterprise
About this Course
Machine Learning is being implemented in many ways to solve a variety of different problems. In this course, we will learn what kind of questions machine learning can solve, the underlying classifications employed in solutions, the information necessary to create a solution, understanding the results, tools used to create machine learning, including low and no code methods in Azure and Power BI. Attendees will leave the course understanding what machine learning can do, what the development steps are, and how to use tools to generate a solution in your environment.
Audience Profile
At Course Completion
Outline
Introduction to Machine Learning and Data Science
- Describing Machine Learning, Data Science and AI
- The 5 different kinds of problems Machine Learning can solve
- Technology capabilities vs hype and mirrors
- Machine Learning classifications
Data Requirements for Machine Learning
- All about data selection and quality
- Understanding overfitting
- Avoiding bias in data selection
- Transparency solution development
The unagile Machine Learning Solution Process
- CRISP Method
- Understanding the data cleansing process
- Training sets and what they are for
- How the computer learns a from data
- Walk through a machine learning solution step-by-step
Interpreting Results
- Correlation and Causation
- Different result visualizations and their meaning
- Understanding Confusion Matrix
Implementation Options
- Implementation Options
- Understanding Data drift
- Ongoing Monitoring of results
- Planned obsolesce of the solution
Development Tools
- A review of common languages and libraries for creating machine learning
- No code option of development
- Incorporating Cognitive Services
- Machine Learning insights in Power BI