perceptron_template.py

#!/usr/bin/python
import sys
import random

def configure():
    """Reads configuration data from the file.

       returns: f, the file object
                the total number of training samples
                the number of inputs in each training sample
                the number of outputs in each training sample
    """
    try:
        script, infile = sys.argv
    except ValueError:
        print "Error, no input file specified"
        exit(-1)
    try:
        f = open(infile, "r")
    except IOError as e:
        print e
        exit(-1)
    retval = [f]
    retval.extend((int(x) for x in f.readline().split()))
    return retval


def read_data(infile, input_count):
    """Reads the data from the file into a list of inputs and outputs."""
    example_inputs = []
    example_outputs = []

    for line in infile.readlines():
        line = line.split()
        # note added bias term, a fixed input of 1.0
        example_inputs.append([int(x) for x in line[:input_count]] + [1.0])
        example_outputs.append([int(x) for x in line[input_count:]])
    return example_inputs, example_outputs


# optional
def print_weights(network, input_count, output_count):
    """Prints the weight matrix of a single layer perceptron."""
    for neuron_index in xrange(input_count + 1):
        for output_index in xrange(output_count):
            print "%s   " % network[neuron_index][output_index],
        print

def main():
    infile, example_count, input_count, output_count = configure()
    example_inputs, example_outputs = read_data(infile, input_count)

    # take the first 90% of the examples
    train_count = int(0.9 * example_count)

    alpha = 0.0000001

    pred_outputs = [0 for _ in xrange(output_count)]

    # network is a matrix of weights; rows represent input values + 1
    # bias term, columns represent output nodes; so reading down the
    # matrix, for some column, gives the weights for the perceptron
    # associated with the column's output node

    .............

    # optional: print the network
    # print_weights(network, input_count, output_count)

    # try some predictions
    for e in xrange(train_count, example_count):

        .............

        # compute loss

        .............

    # print total loss

    ............

if __name__ == "__main__":
    main()