Launching Machine Learning: Delivering Operational Success with Gold Standard ML Leadership

Por: Coursera . en: , ,

  • MODULE 1 - Business Applications of Machine Learning
    • This module dives deeply into the business applications of machine learning – for marketing, financial services, fraud detection and more. We'll illustrate the value delivered for these domains by way of case studies and detailed examples. And we'll precisely measure the performance of the predictive models themselves, focusing on model lift, a predictive multiplier that tells you the improvement achieved by a model.
  • MODULE 2 - Scoping, Greenlighting, and Managing Machine Learning Initiatives
    • To make machine learning work, you've got to bridge what is a prevalent gap between business leadership and technical know-how. Launching machine learning is as much a management endeavor as a technical one. Its success relies on a very particular business leadership practice. This module will demonstrate that practice, guiding you to lead the end-to-end implementation of machine learning.
  • MODULE 3 - Data Prep: Preparing the Training Data
    • The greatest technical hands-on bottleneck of a machine learning project is the preparation of the training data – which is the raw material that predictive modeling software crunches, munches, and learns from. This module will guide you to prepare that data. Business priorities are front and center in the process, since they directly inform the data requirements, including the specific meaning of the dependent variable, which is the outcome or behavior your model will actually predict.
  • MODULE 4 - The High Cost of False Promises, False Positives, and Misapplied Models
    • For many machine learning projects, high accuracy is unattainable – and, besides, accuracy isn't the right metric in the first place. The first portion of this module will demonstrate how other metrics, such as the costs incurred by prediction errors, better serve to keep a machine learning project on track. Then we'll turn to the social good that can be achieved with machine learning, and we'll cover more social justice risks, including the hazards of predicting sensitive information such as pregnancy, job resignations, death, and ethnicity. We'll wrap up by examining the promise and perils of predictive policing.