The Astronomical Journal (ISSN 0004-6256), vol. 109, no. 1669, p. 312-318
Serra-Ricart, Miquel; Trapero, Joaquin; Beckman, John E.; Garrido, Lluis; Gaitan, Vicens
In this paper we propose a method for interpolating multidimensional unbinned data, which could also be sparse, using artificial neural network techniques. An artificial example is first presented in order to show the reliability and potential of the neural network interpolator. A robust behavior is found. We apply the technique to the mapping of a cloud of interstellar atomic hydrogen. The cloud was mapped in H I at 21 cm and we find the neural network method ideal for interpolating the unevenly sampled data, yielding a map from which the global physical parameters of the cloud can be readily obtained.