*********** Lynx.nns is a NeuNet Pro Sample File ***********

This file contains special properties that allow NeuNet Pro to recognize it
as authorized NeuNet sample data.  Anyone using the unlicensed version
of NeuNet Pro is welcome to experiment with this sample data.
Please do not modify this file, or it will lose its status as authorized sample data.

For further information about Neunet Pro and additional sample data,
please visit the NeuNet Pro website at http://www.cormactech.com/neunet

All of this data has been collected from publicly available sources.
CorMac Technologies Inc. does not guarantee the accuracy of the data.
This data is intended solely for experimental purposes.

    ***************  More about the Lynx Data *******************

         P O P U L A T I O N    O F    C A N A D I A N    L Y N X

 This file contains the annual estimate of Canadian lynx population
 for the years 1821 through 1934.  Each record contains the previous
 40 years report and this year's report.  The goal is to predict
 this year's report based on the previous 40 years of history.

 This first row of data predicts year 1861 based on year's 1821-1860.
 The file contains 74 rows of data for years 1861-1934. 

 Experiments with NeuNet Pro show that training data consisting of
 about half the total data can be modelled with an accuaracy of
 better than 1%.  However, the model performs quite poorly on
 unseen test data.

 The prediction difficulty is probably caused by the fact that
 the data is influenced by factors beyond its own autocorrelation.
 When relevent factors are not included, the neural net does not
 have a chance to account for them.  

 This data is really an excercise in "cycle fitting" and the neural
 net fits the cycles it sees very accurately.  The fact that the
 learned cycles do not extrapolate well simpy means the data structure
 may be missing other important inputs.