Scribd
Upload
32 views1 page

Timeseries

The document compares ARIMA, SARIMA, ETS, and GARCH models in terms of their purpose, use cases, assumptions, strengths, and weaknesses. ARIMA is suited for non-seasonal data, SARIMA for seasonal data, ETS for smooth short-term forecasting, and GARCH for financial data with volatility clustering. Recommendations include using ARIMA or SARIMA for forecasting based on seasonality, ETS for smooth data decomposition, and GARCH for volatility prediction.

Uploaded by

Uyên Huỳnh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as ODT, PDF, TXT or read online on Scribd
32 views1 page

Timeseries

The document compares ARIMA, SARIMA, ETS, and GARCH models in terms of their purpose, use cases, assumptions, strengths, and weaknesses. ARIMA is suited for non-seasonal data, SARIMA for seasonal data, ETS for smooth short-term forecasting, and GARCH for financial data with volatility clustering. Recommendations include using ARIMA or SARIMA for forecasting based on seasonality, ETS for smooth data decomposition, and GARCH for volatility prediction.

Uploaded by

Uyên Huỳnh
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as ODT, PDF, TXT or read online on Scribd

Model Comparison: ARIMA, SARIMA, ETS, and GARCH

Below is a comparison of the models in terms of purpose, use case, assumptions, and suitability.

Model Purpose Use Case Key Assumptions Strengths Weaknesses


- Struggles
with
- Stationarity after - Effective for seasonality
Forecasting and Non-seasonal
differencing.- short-term unless
modeling non- data with
ARIMA Linearity of forecasting.- extended.-
seasonal time series trends or
relationships.- Handles trend Requires
data. cycles.
Autocorrelation. well. differencing
for non-
stationary data.
- Parameter
- Stationarity after - Explicit selection can
Time series
Extends ARIMA to differencing.- handling of be complex.-
with
SARIMA handle seasonal Linearity.- seasonality.- Requires high
seasonality
time series data. Seasonality modeled Good for periodic data
and trends.
explicitly. data. preprocessing
and tuning.
- Assumes no
- Additive or - Simple and long-term
Exponential
Short-term multiplicative effective for clear memory
smoothing with
forecasting components.- trends and effects.-
ETS error, trend, and
for smooth Exponential seasonality.- No Struggles with
seasonality
series. smoothing stationarity complex
decomposition.
assumptions. requirement. patterns or
noise.
- Handles - Not ideal for
Models time series Financial - Volatility clusters.- volatility well.- trend or
with volatility data with Mean and variance- Captures seasonality.-
GARCH
clustering volatility stationarity heteroscedasticity Complex to
(heteroscedasticity). clustering. (conditionally). (variance implement and
dependency). tune.

Summary of Suitability
1. ARIMA: Best for simple, non-seasonal series with trends.
2. SARIMA: Useful for seasonal data with periodic behaviors.
3. ETS: Ideal for smooth, short-term forecasting with trends or seasonality.
4. GARCH: Designed for financial or similar data with volatility and clustering.

Recommendations
• Use ARIMA or SARIMA for time series forecasting based on seasonality.
• Choose ETS for clear decomposition of error, trend, and seasonality in smooth data.
• Select GARCH when focusing on volatility prediction, such as in financial markets.

You might also like