The SIMILARITY Procedure

Sequence Normalization

The working (input or target) sequence can be normalized prior to further analysis. Let ${q_{i}}$ be the original sequence with mean $\mu _{q}$ and standard deviation $\sigma _{q}$, and let ${r_{t}}$ be the normalized sequence. The normalizations are defined as follows:

  • Standard is the standard normalization

    \[  r_{i} = (q_{i} - \mu _{q}) / \sigma _{q}  \]
  • Absolute is the absolute normalization

    \[  r_{i} = (q_{i} - {min}({q}_{i})) / ({max}({q}_{i}) - {min}({q}_{i}))  \]
  • User-defined is a user-defined normalization created by the FCMP procedure.