Autocorrelation analysis in ML Python - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is autocorrelation in time series data?

Autocorrelation measures how a time series is related to a lagged version of itself. It shows if past values influence future values.

beginner

Why is autocorrelation important in machine learning?

It helps detect patterns and dependencies in data over time, which can improve forecasting models and avoid misleading assumptions of independence.

intermediate

What does a high positive autocorrelation at lag 1 indicate?

It means the current value is strongly similar to the previous value, showing a persistent pattern or trend.

beginner

How can autocorrelation be visualized?

Using an autocorrelation plot (ACF plot), which shows autocorrelation values for different lags as bars or points.

intermediate

What is the difference between autocorrelation and partial autocorrelation?

Autocorrelation measures total correlation at a lag, while partial autocorrelation measures correlation at a lag after removing effects of shorter lags.

What does autocorrelation measure in a time series?

A Difference between two unrelated series

B Relationship between current and past values

C Random noise in data

D Correlation between two different variables

Autocorrelation measures how current values relate to past values in the same series.

Which plot is commonly used to visualize autocorrelation?

A Autocorrelation function (ACF) plot

B Scatter plot

C Box plot

D Histogram

The ACF plot shows autocorrelation values at different lags.

A high positive autocorrelation at lag 1 means:

A Current value is similar to previous value

B Values are unrelated

C Values are random

D Current value is opposite to previous value

High positive autocorrelation means values are similar to their immediate past values.

Partial autocorrelation differs from autocorrelation by:

A Only measuring lag 1

B Ignoring all lags

C Measuring correlation with unrelated series

D Measuring correlation after removing effects of shorter lags

Partial autocorrelation isolates the direct effect of a lag by removing influences of shorter lags.

Why should autocorrelation be checked before building forecasting models?

A To confirm data is random

B To remove all correlations

C To detect patterns and dependencies

D To increase noise

Checking autocorrelation helps identify patterns that improve forecasting accuracy.

Explain what autocorrelation is and why it matters in analyzing time series data.

Think about how past data points influence future ones.

Describe how you would use an autocorrelation plot to understand a time series.

Imagine looking at a bar chart showing similarity over time gaps.