perceptron.txt

# Perceptron learning data
# 
# lines starting with # characer consedered comment
# information in this file will be parsed as case insencetive

# learning parameter eta between 0 and 1, default value = 1.0
eta = 1.0

# activation function value belongs to {signum, hard_limit}, default value = signum
activation = signum

# initial weight vector, required, parse error will occure if not exists
weight = [1,0, 0]

# patterns to be used in learning
# the order of patterns is important
# each pattern should start with the letter p followed by : and the desired value
# ex. p:+ == p:+1 == p:1   -> pattern belong to + class
#     p:- == p:-1          -> pattern belong to - class
#     any other value will be parse error
# there must be at least one pattern in this file
# pattern vector should not contain leading 1 for bias
p:+ = [1, 1]
p:+ = [1, -1]
p:+ = [0, -1]
p:- = [-1, -1]
p:- = [-1, 1]
p:- = [0, 1]