The manual is outlined as follows. It starts with an introduction to the design philosophy
of SPRANNLIB (the current part) and how neural networks and statistical functions can be
programmed in it. At the end of the part an example is given of how in SPRANNLIB a simple
backpropagation network can be programmed (chapter 5. Preceedings chapters
introduce, in chronological order, the statistical pattern recognition part,
chapter 2, and the neural networks part of the library,
chapter 3. Furthermore a chapter is devoted to the interfacing of the library
to other packages, chapter 4.
Part 2 of this manual is a (complete) listing of all available functions in SPRANNLIB with
a description and how they can be interfaced. First the Numerical Recipes in C which are
used in SPRANNLIB are treated. The next chapter lists all the constants, structures and
flags in the library. Finally, the general functions, statistical pattern recognition functions
and neural network functions are listed. The appendix describes the I/O formats of SPRANNLIB
and the platforms on which the library runs.
The numbers which are present in part 1 of the manual refer to the function description which can be found in part 2. Functions in the reference manual that are marked with an (L) are local functions, i.e. function that are only visible in the main routines.