25.10.2016 Views

SAP HANA Predictive Analysis Library (PAL)

sap_hana_predictive_analysis_library_pal_en

sap_hana_predictive_analysis_library_pal_en

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

<strong>PAL</strong>_SINGLESMOOTH_STATISTICS_TBL:<br />

Related Information<br />

Forecast Accuracy Measures [page 365]<br />

3.5.9 Double Exponential Smoothing<br />

Double Exponential Smoothing model is suitable to model the time series with trend but without seasonality. In<br />

the model there are two kinds of smoothed quantities: smoothed signal and smoothed trend.<br />

<strong>PAL</strong> provides two methods of double exponential smoothing: Holt's linear exponential smoothing and additive<br />

damped trend Holt's linear exponential smoothing. The Holt’s linear exponential smoothing displays a<br />

constant trend indefinitely into the future. Empirical evidence shows that the Holt's linear method tends to<br />

over-forecast. A parameter that is used to damp the trend could improve the situation.<br />

Holt's Linear Exponential Smoothing<br />

Let S t and b t be the smoothed value and smoothed trend for the (t+1)-th time period, respectively. The<br />

following rules are satisfied:<br />

S 0 = x 0<br />

b 0 = x 1 – x 0<br />

S t = αx t + (1 – α) (S t-1 + b t-1 )<br />

b t = β (S t – S t–1 ) + (1 – β) b t-1<br />

Where α, β∈(0,1) are two user specified parameters. The model can be understood as two coupled Single<br />

Exponential Smoothing models, and forecast can be made by the following equation:<br />

F T+m = S T + mb T<br />

Additive Damped Trend Holt's Linear Exponential Smoothing<br />

Let S t and b t be the smoothed value and smoothed trend for the (t+1)-th time period, respectively. The<br />

following rules are satisfied:<br />

S 0 = x 0<br />

b 0 = x 1 – x 0<br />

S t = αx t + (1 – α) (S t-1 + Φb t-1 )<br />

b t = β (S t – S t–1 ) + (1 – β) Φb t-1<br />

Where α, β, Φ∈(0,1) are three user specified parameters. Forecast can be made by the following equation:<br />

F T+m = S T + (Φ + Φ 2 + ... + Φ m )b T<br />

400 P U B L I C<br />

<strong>SAP</strong> <strong>HANA</strong> <strong>Predictive</strong> <strong>Analysis</strong> <strong>Library</strong> (<strong>PAL</strong>)<br />

<strong>PAL</strong> Functions

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!