MDFT Academy Quizzes

What is the risk?

You are preparing a dataset for machine learning training. The dataset only has numerical columns. All columns have values in a range from 0 to 1, except a single price column with values ranging from 0 to 1000.

You decide to use this data for training without performing any numerical preprocessing.

What is the risk of this approach?

Choose the correct answer from the options below.

Please select at least one answer!

Congratulations, you are correct!

I'm sorry, your answer is not correct.

There is no risk. You can import this column without any preprocessing is incorrect. The prices are the largest numbers in the dataset, and may become the focus of some machine learning algorithms that prioritize large numbers.
The model will pay too much attention to the price column is correct
The model will not pay enough attention to the price column is incorrect. The prices are the largest numbers in the dataset, and may become the focus of some machine learning algorithms that prioritize large numbers.
The model may not be trained correctly because it cannot process numerical data outside a range of 0 to 1 is incorrect. Machine learning algorithms can process numerical data on any range. However, data normalized around zero tends to work best.

The process of adjusting the scale of data to ensure that all columns in a dataset are on a similar scale is called normalisation. Check out my video lesson to learn more.

Learn about data normalization

1. Data Preparation

What is the risk?