Scribd
Upload
23 views3 pages

The ARIMA Model

Uploaded by

praneet trimukhe
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
23 views3 pages

The ARIMA Model

Uploaded by

praneet trimukhe
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

The ARIMA model (AutoRegressive Integrated Moving Average) is a simple method for predicting

future values in a time series. It uses past data to make forecasts, such as predicting next month's
sales, temperatures, or stock prices.

ARIMA has 3 Key Parts

The ARIMA model is written as ARIMA(p, d, q) where:

1. AR (AutoRegressive) – p

o Uses past values to predict the current value.

o Example: If yesterday’s temperature helps predict today’s temperature.

2. I (Integrated) – d

o Removes trends by differencing (subtracting values) to make the data stable over
time.

o Example: Today's temperature minus yesterday's temperature.

3. MA (Moving Average) – q

o Uses past prediction errors to improve the forecast.

o Example: If the last prediction was off by 2 degrees, the model adjusts for that.

Steps to Use ARIMA

1. Make Data Stationary:

o Check if the data has a trend or pattern.

o If yes, subtract consecutive values (differencing) to flatten it.

2. Choose p, d, q Values:

o Use plots to help decide how many past values (p) and errors (q) to use, and how
many times to difference (d).

3. Fit the Model:

o Use tools like Python or Excel to create the ARIMA model.

4. Check the Model:

o See if the model's predictions match reality.

5. Make Forecasts:

o Use the model to predict future values.


Example in Simple Terms

Imagine you want to predict tomorrow's temperature. You might:

1. Look at yesterday's temperature (AR part).

2. Adjust for any trends like rising or falling temperatures over time (I part).

3. Correct based on past mistakes in your predictions (MA part).

The ARIMA model combines these three ideas to give you a better forecast.

Why Use ARIMA?

 Easy to Understand: Uses past patterns to predict the future.

 Widely Used: Works well for many types of time series data.

When to Avoid ARIMA?

 When There’s Seasonality: For repeating patterns (e.g., monthly sales), use SARIMA
(Seasonal ARIMA).

 When Data is Complicated: For more complex data, you might need advanced methods like
machine learning.

STL (Seasonal-Trend Decomposition using Loess)


STL is a method for breaking down a time series into three parts:

1. Seasonal: Repeating patterns (e.g., monthly or yearly cycles).

2. Trend: The overall direction (upward or downward) over time.

3. Remainder: The random noise left after removing the trend and seasonality.

Steps in STL:

1. Decompose the time series into trend, seasonality, and remainder.

2. Analyze the components separately.

3. Reconstruct the time series without noise for better forecasting.

Extract Features from the Model

You can extract features like height, average energy, and other statistics from the generated ARIMA
or STL models to analyze and improve predictions.

Key Features to Extract:

1. Height:

o In STL, this could be the amplitude of the seasonal component.

2. Average Energy:
o The squared sum of the signal values divided by the length, indicating the strength or
intensity of the signal.

3. Trend Slope:

o The direction and speed of change in the trend component.

4. Residual Variance:

o The variability of the remainder (noise) component.

Combining ARIMA, STL, and Feature Analysis for Prediction

1. Decompose the Time Series with STL to separate trend, seasonality, and noise.

2. Apply ARIMA on the trend and seasonally adjusted series.

3. Extract Features such as:

o Seasonal height/amplitude

o Average energy

o Trend slope

4. Evaluate Forecast Accuracy using metrics like MAE, RMSE, and MAPE.

5. Analyze and Improve Predictions by incorporating the extracted features.

You might also like