S o m e   r e m a r k s   f o r   u s e r s


The code producing data for the database has been written in two languages - Pascal and Fortran. The program files are called NeuroCluster.pas and NeuroCluster.for, respectively. The code reads some weights form the files ww_11x, ww_12x (where x$=$pas for Pascal and x$=$for for Fortran), some numbers from a file data and output results in the file a_11_12.txt.


The files ww_11x and ww_12x (e.g., ww_11pas, ww_11for, etc.) contain the weights of the artificial neural network (the two layer perceptron) which are used to calculate the expansion coefficients $a_{11}$ and $a_{12}$ of the scattering matrix elements $S_{11}$ and $S_{12}$ (please, contact the authors for a code that calculates these weight for different parameter values).


The file data sets the input vector of the perceptron and the range of the scattering angles within which the calculations have to be performed. Since the perceptron has been so far trained for the fractal dimension $D = 3$, the density $\rho = 8$ and the size parameter of cluster's grain $x = 1.5$, the free parameters in the input vector are the real and imaginary parts of refractive index $m$ and the number of grains in the cluster $N$. The structure of the input file data is as follows:


theta_min theta_max del_theta
0 180 2
 
Re(m) Im(m) N
1.70 0.001 10


Here, as it has been mentioned in the Description, theta_min, theta_max, del_theta (all in degree) are the minimum and maximum values, and the step of change of the scattering angles at which the scattering matrix elements are calculated. Re(m) (1.4 < Re(m) < 1.71) and Im(m) (0.001 < Im(m) <0.1) are the real and imaginary parts of the complex refractive index of cluster's grain, N (N < 50) is the number of grains in the cluster. Note that the files ww_11x, ww_12x (e.g., ww_11pas, ww_12pas) and data must be in the same directory as the program NeuroCluster.pas (or NeuroCluster.for)!


The output file a_11_12.txt contains the expansion coefficients of the scattering matrix element $a_{11}$ and $a_{21}$ in series of the generalized spherical function as well as the angular dependencies of the scattering matrix elements $S_{11}$, $S_{21}$ and the degree of linear polarization $P$ ( $P= - S_{21}/S_{11}$). The matrix elements are normalized in the way that integration of $S_{11}$ over all scattering angle gives the scattering effeciency $Q_{\rm sca}$.

To the top
Created by V.I.
Last modified: 31/03/03, V.I.