Skip to main content

Soldai utilities for machine learning and text processing

Project description

sutil

This repository contains a set of tools to deal with machine learning and natural language processing tasks, including classes to make quick experimentation of different classifacation models.

Dataset

This class is made to load csv styles dataset's where all the features are comma separeted and the class is in the last column. It includes functions to normalize the features, add bias, save the data to a file and load from it. Also includes functions to split the train, validation and test datasets.

from sutil.base.Dataset import Dataset

datafile = './sutil/datasets/ex2data1.txt'
d = Dataset.fromDataFile(datafile, ',')
print(d.size)

sample = d.sample(0.3)
print(sample.size)

sample.save("modelo_01")

train, validation, test = d.split(train = 0.8, validation = 0.2)
print(train.size)
print(validation.size)
print(test.size)

Regularized Logistic Regression

You can also include your own models as a Regularized Logistic Regression, implemented manually using numpy and included in the sutil.models package

import numpy as np
from sutil.base.Dataset import Dataset
from sutil.models.RegularizedLogisticRegression import RegularizedLogisticRegression

datafile = './sutil/datasets/ex2data1.txt'
d = Dataset.fromDataFile(datafile, ',')
d.xlabel = 'Exam 1 score'
d.ylabel = 'Exam 2 score'
d.legend = ['Admitted', 'Not admitted']
iterations = 400
print('Plotting data with + indicating (y = 1) examples and o indicating (y = 0) examples.\n')
d.plotData()

theta = np.zeros((d.n + 1, 1))
lr = RegularizedLogisticRegression(theta, 0.03, 0, train=1)
lr.trainModel(d)
lr.score(d.X, d.y)
lr.roc.plot()
lr.roc.zoom((0, 0.4),(0.5, 1.0))

Sklearn model

You can also embed the sklearn models in a wrapper class in order to run experiments with diferent models implemented in sklearn. In the same style you can create tensorflow, keras or pytorch models inhyereting from sutil.modes.Model class and implementing the trainModel and predict methods.

import numpy as np
from sutil.base.Dataset import Dataset
from sutil.models.SklearnModel import SklearnModel
from sklearn.linear_model import LogisticRegression

datafile = './sutil/datasets/ex2data1.txt'
d = Dataset.fromDataFile(datafile, ',')
ms = LogisticRegression()
m = SklearnModel('Sklearn Logistic', ms)
m.trainModel(d)
m.score(d.X, d.y)
m.roc.plot()
m.roc.zoom((0, 0.4),(0.5, 1.0))

Neural Network Classifer

This class let's you perform classifcation using a Neural Network, multiperceptron classifer. It wraps the sklearn MLPClassifer and implements a method to search different activations, solvers and hidden layers structures. Upu can pass your own arguments to initialize the network as you want.

from sutil.base.Dataset import Dataset
from sutil.neuralnet.NeuralNetworkClassifier import NeuralNetworkClassifier

datafile = './sutil/datasets/ex2data1.txt'
d = Dataset.fromDataFile(datafile, ',')
d.normalizeFeatures()
sample = d.sample(examples = 30)

nn = NeuralNetworkClassifier((d.n, len(d.labels)))
nn.searchParameters(sample)
nn.trainModel(d)
nn.score(d.X, d.y)
nn.roc.plot()

Experiment

The experiment class let's you perform the data split and test against different models to compare the performance automatically

import numpy as np
from sutil.base.Dataset import Dataset
from sklearn.linear_model import LogisticRegression
from sutil.base.Experiment import Experiment
from sutil.models.SklearnModel import SklearnModel
from sutil.models.RegularizedLogisticRegression import RegularizedLogisticRegression
from sutil.neuralnet.NeuralNetworkClassifier import NeuralNetworkClassifier

# Load the data
datafile = './sutil/datasets/ex2data1.txt'
d = Dataset.fromDataFile(datafile, ',')
d.normalizeFeatures()
print("Size of the dataset... ")
print(d.size)
sample = d.sample(0.3)
print("Size of the sample... ")
print(d.sample)


# Create the models
theta = np.zeros((d.n + 1, 1))
lr = RegularizedLogisticRegression(theta, 0.03, 0)
m = SklearnModel('Sklearn Logistic', LogisticRegression())
# Look for the best parameters using a sample
nn = NeuralNetworkClassifier((d.n, len(d.labels)))
nn.searchParameters(sample)

input("Press enter to continue...")

# Create the experiment
experiment = Experiment(d, None, 0.8, 0.2)
experiment.addModel(lr, name = 'Sutil Logistic Regression')
experiment.addModel(m, name = 'Sklearn Logistic Regression')
experiment.addModel(nn, name = 'Sutil Neural Network')

# Run the experiment
experiment.run(plot = True)

Project details


Release history Release notifications

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for soldai-utils, version 0.0.1
Filename, size File type Python version Upload date Hashes
Filename, size soldai_utils-0.0.1-py3-none-any.whl (27.2 kB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size soldai-utils-0.0.1.tar.gz (9.7 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page