NEUROFORECASTER/GA (IBM and compatibles) CAT.# 000-N01-00

  NeuroForcaster is a user-friendly business forecasting and analytical tool
  designed to mimic human thinking , learn and forecast engineering
  applications the movement patterns of stock , currency and property markets
  and numerous other applications . It has been widely used by many Engineers
  Investment Analysts , Economists , Fund Managers , Private Investors as
  well as prestigious Universities and Institutions as a learning tool for
  NEURAL NETWORK TECHNOLOGY .

  Forecasters can be built with any number of input and user-specified
  complexity . Besides the 12 built-in neural network models , Version 3
  supports the GENETICA Net Builder Option which evolves and optimizes
  network structures , hidden layers , hidden nodes , learning rate , tolerance
  etc..

  Based on Darwinian evolution theory , networks are combined , purged and
  reproduced to give rise to better offsprings .
  You can create powerful and sophisticated applications quickly , easily ,
  reliably and productively without spending your precious time guessing the
  network structure and control parameters . Rescaling range analysis is
  included to help the user predetermine the predictability of the input data
  and to reveal hidden cycles in the time series . Various pre-trained
  forecasters are provided in the package so that the users can compare its
  performance with other similar neural network products and built their own
  applications based on the examples given .

  Genetic Algorithm (GA) is very effective at finding optimal or near optimal
  solutions to a wide variety of problems including neural network design ,
  because it does not impose many limitations required by traditional
  methods.  Based on Darwinian evolution theory , GENETICA generates many
  possible networks , which are evaluated , purged and recombined to produce
  better offsprings which will inherit the knowledge acquired by their
  parents .  Occasionally mutation will take place to ensure that networks will
  not be trapped in local minima during the learning process .

  GENETICA in Action .
  The network display can be turned on and off , maximized or minimized to an
  icon , zoomed in and out , scrolled up and down . The total population can be
  increased or decreased beyond the default size , the learning rate and the
  tolerance level can be set to auto mode or fixed .
  The evolution process can be disabled , so that only learning of the
  current set of networks can take place . Networks are displayed in colors for
  easy recognition .

  SYSTEM REQUIREMENTS : IBM PC 386/486 or compatibles with mouse .  MS Windows
                        3.0 , 3.1 MS DOS 5.0 or higher 4 MB RAM , 5
                        MB hard disk space , 3.5 floppy drive VGA monitor or
                        better , Math coprocessor recommended.