machine learning – Acodemics

Bag of Words Implementation

Bag of words is a method for reducing natural text into a representative model for use with machine learning and natural language processing.

I used it to train a network on log messages and then assign scores to known log messages, based on a bag of words representation the neural network can give a score of how well the test data matches.

class BagOfWords:

    def __init__(self):
        self.dict = {}
        self.i = 0

    def addVocab(self, line):
        for word in line.split(" "):
            if word not in self.dict.keys() and not word.isdigit():
                self.dict[word] = self.i
                self.i += 1

    def vocabScore(self, line):
        score = [0] * len(self.dict.keys())
        for word in line.split(" "):
            if word in self.dict.keys() and not word.isdigit():
                score[self.dict[word]] += 1
        return score

class BagOfWords:

def __init__(self):

self.dict = {}

self.i = 0

def addVocab(self, line):

for word in line.split(" "):

if word not in self.dict.keys() and not word.isdigit():

self.dict[word] = self.i

self.i += 1

def vocabScore(self, line):

score = [0] * len(self.dict.keys())

for word in line.split(" "):

if word in self.dict.keys() and not word.isdigit():

score[self.dict[word]] += 1

return score

Image Recognition Tool

August 13, 2014 / Jack Concanon / 0 Comments

As I’ve recently got quite into machine learning tools I’ve written a small GUI based tool that uses the PyBrains library to ‘learn’ a common theme from a folder of training data. To test new images, drag and drop them onto the tool to get a percentage of similarity to the test data.

The tool is written using Python, PyBrains, Glade2 and PyGTK so you need those libraries available.

The tool can be downloaded here – image_recognition_tool

August 4, 2014

A presentation I made for work that is an introduction to machine learning. Notes included with power point slides. Machine Learning

Image Recognition with PyBrains

June 20, 2014 / Jack Concanon / 0 Comments

After recently completing the Machine Learning course from Stanford University on Coursera I’ve been preparing to give a small introduction to machine learning at work. Part of that is showing some demos of machine learning tools.

I made a character recognition neural network using the PyBrains Python library, it’s a great library and very fast but the documentation is very poor and examples are hard to come across. With enough digging I managed to put together something very simple and short.

In this example it reads in small PNG files of letters, extracts all of the pixel values and creates a 1D array of the values, this is used to train the neural network through back propagation. I test the network on one of the inputs. Each input is classified with a number in the addSample function, this takes the flattened array and a number (unfortunately it does not take a string as a classification). If you run the application you will see that, for example when using b.png as a test, it will return a value close to 2.

You can download the images I used here – Machine Learning Training Characters.

#!/usr/bin/python

from pybrain.datasets.supervised import SupervisedDataSet 
from pybrain.tools.shortcuts import buildNetwork
from pybrain.supervised.trainers import BackpropTrainer
import cv2

def loadImage(path):
    im = cv2.imread(path)
    return flatten(im)

def flatten(x):
    result = []
    for el in x:
        if hasattr(el, "__iter__") and not isinstance(el, basestring):
            result.extend(flatten(el))
        else:
            result.append(el)
    return result

if __name__ == "__main__":

    t = loadImage('a.png')

    net = buildNetwork(len(t), len(t), 1)
    ds = SupervisedDataSet(len(t), 1)
    ds.addSample(loadImage('a.png'),(1,))
    ds.addSample(loadImage('b.png'),(2,))
    ds.addSample(loadImage('c.png'),(3,))
    ds.addSample(loadImage('d.png'),(4,))

    trainer = BackpropTrainer(net, ds)
    error = 10
    iteration = 0
    while error > 0.001: 
        error = trainer.train()
        iteration += 1
        print "Iteration: {0} Error {1}".format(iteration, error)

    print "\nResult: ", net.activate(loadImage('b.png'))

#!/usr/bin/python

from pybrain.datasets.supervised import SupervisedDataSet

from pybrain.tools.shortcuts import buildNetwork

from pybrain.supervised.trainers import BackpropTrainer

import cv2

def loadImage(path):

im = cv2.imread(path)

return flatten(im)

def flatten(x):

result = []

for el in x:

if hasattr(el, "__iter__") and not isinstance(el, basestring):

result.extend(flatten(el))

else:

result.append(el)

return result

if __name__ == "__main__":

t = loadImage('a.png')

net = buildNetwork(len(t), len(t), 1)

ds = SupervisedDataSet(len(t), 1)

ds.addSample(loadImage('a.png'),(1,))

ds.addSample(loadImage('b.png'),(2,))

ds.addSample(loadImage('c.png'),(3,))

ds.addSample(loadImage('d.png'),(4,))

trainer = BackpropTrainer(net, ds)

error = 10

iteration = 0

while error > 0.001:

error = trainer.train()

iteration += 1

print "Iteration: {0} Error {1}".format(iteration, error)

print "\nResult: ", net.activate(loadImage('b.png'))

Categories

Meta

Tagmachine learning

Bag of Words Implementation

Image Recognition Tool

Image Recognition with PyBrains

Recent Comments