A medical research project uses a large anonymized dataset of brain scan images that are categorized into predefined brain haemorrhage types. You…

The scenario describes a supervised machine learning task where the goal is to predict a discrete category. The model is trained on a dataset of brain scans that are already labeled with "predefined brain haemorrhage types." The objective is to assign a new, unseen image to one of these specific categories. This process of predicting a class label (the type of hemorrhage) from a known set of categories is the definition of a classification task.

A. clustering: This is an unsupervised learning technique used to group unlabeled data into clusters based on similarity. The dataset in the scenario is explicitly labeled.

B. regression: This supervised learning technique is used to predict a continuous numerical value (e.g., a patient's age or blood pressure), not a discrete category like a hemorrhage type.

1. Microsoft Learn

AI-900: Introduction to machine learning. In the unit "Types of machine learning

" classification is defined as a supervised technique to "predict which category

or class

an item belongs to." The document contrasts this with regression

which predicts a numeric value

and clustering

which is an unsupervised method for grouping unlabeled data.

Source: Microsoft Learn

Module: "Introduction to machine learning"

Unit 3: "Types of machine learning".

2. Microsoft Learn

AI-900: Use automated machine learning in Azure Machine Learning. This module lists the primary machine learning task types. It explicitly states

"Classification: For predicting categorical labels or classes. For example

predicting if a patient has diabetes or not." This directly aligns with predicting a type of brain hemorrhage.

Source: Microsoft Learn

Module: "Use automated machine learning in Azure Machine Learning"

Unit 2: "Introduction to automated machine learning".

3. Ng

A. (2022). CS229: Machine Learning Course Notes. Stanford University. In the section on Supervised Learning

a classification problem is defined as one where "we are trying to predict a discrete output." This is contrasted with a regression problem

which predicts a "continuous output." The notes later define clustering as an unsupervised task for finding structure in unlabeled data.

Source: CS229 Course Notes

Part I: Supervised Learning

Page 3

and Part V: Unsupervised Learning

Page 19.