ARIMA model basics in ML Python - Cheat Sheet & Quick Revision

Recall & Review

beginner

What does ARIMA stand for in time series forecasting?

ARIMA stands for AutoRegressive Integrated Moving Average. It is a model used to understand and predict future points in a time series by combining autoregression, differencing (integration), and moving average components.

beginner

What is the role of the 'Integrated' part in ARIMA?

The 'Integrated' part means differencing the data to make it stationary. Stationary data has a constant mean and variance over time, which helps the model make better predictions.

beginner

Explain the 'AutoRegressive' (AR) component in ARIMA.

The AR part uses past values of the time series to predict the current value. It assumes that past points have a linear relationship with the current point.

beginner

What does the 'Moving Average' (MA) component do in ARIMA?

The MA part models the error of the prediction as a combination of past errors. It helps smooth out the noise by considering the impact of previous forecast errors.

beginner

What are the three parameters of an ARIMA model and what do they represent?

The three parameters are (p, d, q): <br> - p: number of autoregressive terms (AR) <br> - d: number of times the data is differenced (Integrated) <br> - q: number of moving average terms (MA) <br> Together, they define the ARIMA model structure.

What is the purpose of differencing in an ARIMA model?

A To remove outliers from the data

B To make the time series stationary

C To smooth the data by averaging

D To increase the number of data points

Differencing helps remove trends and seasonality to make the data stationary, which is important for ARIMA modeling.

In ARIMA(p, d, q), what does 'p' represent?

A Number of autoregressive terms

B Number of times data is differenced

C Number of moving average terms

D Number of seasonal cycles

'p' is the number of autoregressive terms, meaning how many past values are used to predict the current value.

Which component of ARIMA models the relationship between past forecast errors and the current value?

A Moving Average (MA)

B Integrated (I)

C Autoregressive (AR)

D Differencing

The Moving Average (MA) component models the current value based on past forecast errors.

Why is stationarity important in ARIMA modeling?

A It increases the number of data points

B It guarantees perfect predictions

C It removes all noise from the data

D It ensures the data has a constant mean and variance over time

Stationarity means the statistical properties of the series do not change over time, which is necessary for ARIMA to work well.

If a time series is not stationary, what is the common first step before applying ARIMA?

A Use only the moving average component

B Add more data points

C Apply differencing to the data

D Ignore the problem and fit the model

Differencing is used to transform a non-stationary series into a stationary one before modeling.

Describe the three main components of an ARIMA model and their roles in time series forecasting.

Think about how past values, differencing, and past errors contribute to predictions.

Explain why making a time series stationary is important before applying an ARIMA model and how this is achieved.

Consider what changes in mean or variance over time mean for prediction.