Course Aim

The goal of this training is to provide further detail of the main tools at our disposal to evaluate performance in the context of a classification task. Going through the training will provide the learner with an in-depth understanding of the measures and metrics that are available to quantify performance, in addition to the main aspects that can be used as control criteria for this purpose.

Course Description

The focus of this course is to understand the general principles behind assessing the classification performance of an AI system. We will explore, in more detail, the different available metrics for three different classification scenarios (binary, multi-class, and multi-label), the available statistical significance tests that are appropriate for classification, and some important computational complexity metrics.

On-demand - training that’s even more flexible

BSI’s on-demand courses are market-leading and available 24/7. Developed by top subject matter experts, they contain the same high-quality content you will find in our tutor-led training, but with the added benefit of being able to learn at your own pace and at any time.

How will I benefit?

This course will help you:

1. Awareness of the main factors affecting performance: you will gain a clear idea of what control criteria is available in the assessment of machine learning classification.
2. Understand the general process of assessing performance in a classification-based AI system.
3. Obtain deeper knowledge of performance assessment techniques: understanding of the main measures available to quantify performance in the context of classification.
4. To establish a common ground to compare and evaluate models: by incorporating these concepts early in the design lifecycle of an AI application, you’ll be able to establish clear grounds to establish how good is the model you are working on, and to compare it to other models in a sensible way.

What will I learn?

Upon completion of this course, you will be able to:

• Understand the purpose of assessing classification performance.
• Understand the general process to assess performance in classification-based AI systems.
• Apply well known measures, metrics to quantify performance when designing or evaluating AI models in general.
• Establish a common ground to compare multiple classifier models.
• Use the all the metrics and measures mentioned in the standard to evaluate classifiers in three different scenarios: binary, multi-class, and multi-label.
• Use statistical significance testing to establish comparisons with different models or versions of the same model.
• Use metrics to evaluate the computational complexity of your model.

Who should attend?

• AI managers, team-leaders, and machine learning practitioners in general.

How will I learn?

This is an online, interactive on-demand course.  Courses are available 24/7 and you can learn at any time and from any place that suits you – you just need an internet connection. You can learn as fast or as slowly as you want to. You can also take breaks at any time in the course and pick up where you left off when you are ready to continue.

During the access period, you can go back and repeat parts or all the course to refresh and reinforce what you have learned. The course content is both detailed and engaging, with explanations, activities, and knowledge checks to enhance your learning.

What will I gain?

On completion, you’ll be awarded an internationally recognized BSI training course certificate.

Prerequisites

The course assumes that delegates know some basic math concepts and have a basic understanding on machine learning, statistics, and probability.  

Modules

• Assessment of machine learning classification performance – Part 1
• Assessment of machine learning classification performance – Part 2