The complexity of the direct problem depends on its input parameters characterizing the shape, size and structure of the scattering particle. Sometimes to reduce the number of these parameters and to simplify the calculations, regular shape (spherical, spheroidal, etc.) scatterer models are applied. However, many particles of natural and artificial origin are aggregates (clusters) of subparticles well resembling fractal-like objects (for instance, cometary and interplanetary dust grains [1]). As such compound particles strongly distinguish in shape and structure from the regular shape models, their optical properties appreciably differ from those of isolated model particles [2].

Calculation of scattered radiation characteristics for fractal clusters is a time consuming procedure. Moreover, for interpretation of remote sensing results such calculations must be carried out in a wide range of the size parameters, complex refractive indexes and parameters, characterizing the cluster structure. Therefore, to solve the inverse scattering problem a database with detailed information on the scattering properties of fractal cluster is required. We produced such a database using an artificial neural network. This network when being properly trained allows one well to determine required data.