The stakeholders of a machine learning model have confirmed that they understand the objective and purpose of the model, and ensured that the proposed model aligns with their business priorities. They have also selected a framework and a machine learning model that they will be using. What should be the next step to progress along the machine learning workflow?

The typical machine learning workflow follows a logical sequence. After the business problem is defined and a model architecture is selected (as stated in the scenario), the immediate next step is to acquire and prepare the data. The model cannot be trained, tuned, or evaluated without data. Data preparation involves cleaning, transforming, and structuring the data into a suitable format for the training and testing phases. This step is fundamental and must precede any training or tuning activities. The other options occur either earlier or later in the standard ML workflow.

A. Tuning the algorithm is an optimization step that occurs after an initial model has been trained and evaluated, not before data preparation.
C. Agreeing on acceptance criteria is part of the initial business problem definition phase, which has already been completed according to the scenario.
D. Evaluating the model selection happens after training, or is part of the selection process itself, not the step immediately following the selection.

ISTQB® Certified Tester AI Testing (CT-AI) Syllabus Version 1.0

Section 2.2.1 "Typical ML Workflow"

page 13. The syllabus lists "Data Acquisition and Preparation" as the second step after "Business Problem Definition".

Ng

A. (n.d.). Machine Learning Yearning. Stanford University. Chapter 5 "Basic Recipe"

page 17. This chapter outlines the iterative process

emphasizing that data is the foundation before training and tuning.