Data Science and Machine Learning (Theory and Projects) A to Z - Introduction to Machine Learning: Regression Exercise S

The video tutorial explores the possibility of treating classification problems as regression problems using a regression API. It explains the data format required for regression, where both feature vectors and targets must be numeric. The tutorial also covers converting classification labels to numeric values for compatibility with regression APIs. However, it highlights the limitations of using classification APIs for regression data due to the need for discrete labels, leading to potential information loss when converting regression targets to discrete labels.

How can classification data be adapted for use with a regression API?

• By converting numeric values to strings

• By using only categorical data

• By converting string labels to numeric values

• By ignoring the target values

What is a key difference between regression and classification targets?

• Classification targets are always numeric

• Regression targets are numeric and continuous

• Regression targets are always integers

• Classification targets are continuous

Why is it not always possible to use a classification API for regression problems?

• Classification APIs assume targets are from a finite set

• Classification APIs only work with string labels

• Regression APIs cannot handle discrete data

• Classification APIs require continuous targets

What happens when regression targets are converted to discrete labels?

• The data becomes more accurate

• There is a loss of information

• The data becomes easier to interpret

• The targets become continuous

Can regression routines be used for classification tasks?

• No, because they require discrete labels

• Yes, without any modifications

• No, they are incompatible

• Yes, but with a loss of information