wavelet

Terms from Artificial Intelligence: humans at the heart of algorithms

A wavelet is a small function, which has finite/local extent in time or space. For example:
      fs(t)  =   1 - |t-s|   if (t-s) < 1 ,   =  0 otherwise
Wavletes can be one diennsional, for use over time series series or sequential data, 2D for images, or higher dimensions. Wavelets usually form part fo a collection with different wavelests at different sizes and locations, so that any \term{time series} or image can be expressed as a sum of the wavekets in the collection -- called a wavelet transform of the function.

Used on pages 145, 322