Details: SIMILARITY Procedure

The SIMILARITY procedure can be used to form time series data from transactional data.

1. accumulation

   ACCUMULATE= option

2. missing value interpretation

   SETMISSING= option

3. zero value interpretation

   ZEROMISS= option

The accumulated time series can then be transformed to form the working time series. Transformations are useful when you want to stabilize the time series before computing the similarity measures. Simple and seasonal differencing are useful when you want to detrend or deseasonalize the time series before computing the similarity measures. Often, but not always, the TRANSFORM=, DIF=, and SDIF= options should be specified in the same way for both the target and input variables.

1. time series transformation

   TRANSFORM= option

2. time series differencing

   DIF= and SDIF= option

3. time series missing value trimming

   TRIMMISSING= option

4. time series descriptive statistics

   PRINT=DESCSTATS option

After the working series is formed, it can be treated as an ordered sequence that can be normalized or scaled. Normalizations are useful when you want to compare the "shape" or "profile" of the time series. Scaling is useful when you want to compare the input sequence to the target sequence while discounting the variation of the target sequence.

1. normalization

   NORMALIZE= option

2. scaling

   SCALE= option

After the working sequences are formed, similarity measures can be computed between input and target sequences.

1. sliding

   SLIDE= option

2. warping

   COMPRESS= and EXPAND= option

3. similarity measure

   MEASURE= and PATH= option

The SLIDE= option is used to specify observation-index sliding, seasonal-index sliding, or no sliding. The COMPRESS= and EXPAND= options are used to specify the warping limits. The MEASURE= and PATH= options are used to specify how the similarity measures are computed.