Handling missing values in ML Python - Cheat Sheet & Quick Revision

Recall & Review

beginner

What are missing values in a dataset?

Missing values are data points that are not recorded or are absent in a dataset. They can happen due to errors, skipped questions, or data loss.

beginner

Name two simple methods to handle missing values.

Two simple methods are: 1) Removing rows or columns with missing values, 2) Filling missing values with a fixed value like mean, median, or mode.

intermediate

Why is it sometimes better to fill missing values instead of removing them?

Filling missing values keeps more data for learning, which can improve model accuracy. Removing too many rows or columns can lose important information.

beginner

What is mean imputation?

Mean imputation replaces missing values in a numeric column with the average (mean) of the available values in that column.

intermediate

What is a potential downside of using mean imputation?

Mean imputation can reduce data variability and may bias the model because it ignores the natural spread of data and relationships between features.

What does 'dropping missing values' mean?

A Replacing missing values with zeros

B Removing rows or columns that contain missing data

C Ignoring missing values during training

D Predicting missing values using a model

Which method fills missing numeric values with the average of the column?

A Mean imputation

B Median imputation

C Mode imputation

D Forward fill

Why might removing all rows with missing values be a bad idea?

A It fills missing values automatically

B It always improves model accuracy

C It can cause loss of too much data

D It changes the data types

Which of these is NOT a common way to fill missing values?

A Using mean or median

B Using mode for categorical data

C Using random noise

D Dropping the entire dataset

What is a risk of using mean imputation?

A It can bias the model by reducing variability

B It removes all missing values permanently

C It increases data variability

D It only works for categorical data

Explain why handling missing values is important before training a machine learning model.

Describe two common techniques to handle missing values and when you might use each.