import matplotlib.pyplot as plt
import numpy as np
from mlpy.numberGenerator.bounds import Bounds
from mlpy.neuralNetwork.feedForwardNeuralNetwork import NeuralNetwork
from mlpy.neuralNetwork.structure.layer import Layer
plt.grid(1)
plt.xlabel('Iterations')
plt.ylabel('Error')
plt.ylim([0, 1])
plt.ion()
l_rate = 0.2
bounds = Bounds(-2, 2)
inputLayer = Layer(bounds,
size=2,
prev=None,
l_rate=l_rate,
bias=True,
label="Input layer")
hiddenLayer = Layer(bounds,
size=4,
prev=inputLayer,
l_rate=l_rate,
bias=True,
label="Hidden layer")
outputLayer = Layer(bounds,
size=1,
from mlpy.neuralNetwork.structure.layer import Layer
np.set_printoptions(suppress=True)
dataSetTool = DataSetTool()
fileName = '../../dataSet/pima-indians-diabetes/pima-indians-diabetes.data'
training, generalization, testing = dataSetTool.getPrimaIndiansDiabetesSets(
fileName)
plt.xlabel('Iterations')
plt.ylabel('Error')
plt.ion()
l_rate = 0.1
bounds = Bounds(-1, 1)
inputLayer = Layer(bounds,
size=len(training[0][0]),
prev=None,
l_rate=l_rate,
bias=True,
label="Input layer")
hiddenLayer = Layer(bounds,
size=20,
prev=inputLayer,
l_rate=l_rate,
bias=True,
label="Hidden layer")
outputLayer = Layer(bounds,
size=len(training[0][1]),
def getBounds(self):
return Bounds(-1.28, 1.28)
def getBounds(self):
return Bounds(-10, 10)
def getBounds(self):
return Bounds(-600, 600)
def getBounds(self):
return Bounds(-32.768, 32.768)
import math
import numpy as np
from mlpy.numberGenerator.bounds import Bounds
from mlpy.particleSwarmOptimization.pso import PSO
from mlpy.particleSwarmOptimization.structure.particle import Particle
np.set_printoptions(suppress=True)
errors = []
bounds = Bounds(-10, 10)
# Create the mlpy with the nn weights
num_particles = 7
inertia_weight = 0.729
cognitiveConstant = 1.49
socialConstant = 1.49
num_dimensions = 50
# Configure PSO
pso = PSO()
def error(position):
err = 0.0
for i in range(len(position)):
xi = position[i]
err += (xi * xi) - (10 * math.cos(2 * math.pi * xi)) + 10
return err
print("Original: ")
sequence.printFinalAlignemnt()
pso = BoundaryConstraintPSO()
def fitness(position):
error, valid = sequence.getFitness(position)
return (error[0] * sequence.maxlen) + error[1]
from boundaryConstraints import *
pso.boundaryConstraint = bc2
pso.error = fitness
pso.bounds = Bounds(0, sequence.maxlen + 1)
pso.initialPosition = Bounds(0, sequence.maxlen + 1)
pso.num_particles = NUM_PARTICLES
pso.num_dimensions = sequence.size
pso.weight = INERTIA_WEIGHT
pso.cognitiveConstant = COGNITIVE_CONSTANT
pso.socialConstant = SOCIAL_CONSTANT
trainingErrors, trainingError = pso.train(ITERATIONS, ITERATIONS_SAMPLE_SIZE)
sequence.getFitness(pso.group_best_position)
print("Result: ")
sequence.printFinalAlignemnt()
cpso_dissipative_errors = [] cpso_dissipative_error = [] cpso_dissipative_generalization_error = [] pso_errors = [] pso_error = [] pso_generalization_error = [] iterations = 5000 samples = 30 NUM_PARTICLES_Y = 5 NUM_PARTICLES_X = 5 INERTIA_WEIGHT = 0.729844 COGNITIVE_CONSTANT = 1.496180 SOCIAL_CONSTANT = 1.496180 BOUNDS = Bounds(-10, 10) INITIAL_POSITION = Bounds(-5, 5) DESC = 'Glass' DATA_SET_FUNC = dataSetTool.getGlassDataSets DATA_SET_FILE_LOC = '../problems/dataSets/glass/glass.data' HIDDEN_LAYER_NEURONS = [12] # DESC = 'Iris' # DATA_SET_FUNC = dataSetTool.getIrisDataSets # DATA_SET_FILE_LOC = 'experiments/dataSets/iris/iris.data' # HIDDEN_LAYER_NEURONS = [8] # DESC = 'Wine' # DATA_SET_FUNC = dataSetTool.getWineDataSets # DATA_SET_FILE_LOC = 'experiments/dataSets/wine/wine.data'