def searchSolve(self, X, Y, updates, randomState=None):
if "max_attempts" in updates:
tempAttempts = updates["max_attempts"]
else:
tempAttempts = self.maxAttempts
if "restarts" in updates:
tempRestarts = updates["restarts"]
else:
tempRestarts = self.numRestarts
if randomState is None:
tempRandomState = self.randomState
else:
tempRandomState = randomState
if "learning_rate" in updates:
tempLearnRate = updates["learning_rate"]
else:
tempLearnRate = self.learnRate
if "max_iters" in updates:
tempMaxIters = updates["max_iters"]
else:
tempMaxIters = self.maxIters
self.model = mlrose.NeuralNetwork(hidden_nodes = [5], activation = 'relu', \
algorithm = self.algorithm, max_iters = tempMaxIters, \
bias = True, is_classifier = True, learning_rate = tempLearnRate, \
early_stopping = True, clip_max = 1e10, max_attempts = tempAttempts, \
restarts = tempRestarts, curve = True, random_state = tempRandomState)
self.model.fit(X, Y)
yPred = self.model.predict(X)
score = metrics.accuracy_score(Y, yPred)
return score
def createModel(self):
self.model = mlrose.NeuralNetwork(hidden_nodes = [5], activation = 'relu', \
algorithm = self.algorithm, max_iters = self.maxIters, \
bias = True, is_classifier = True, learning_rate = self.learnRate, \
early_stopping = False, clip_max = 1e10, max_attempts = self.maxAttempts, \
pop_size = self.popSize, mutation_prob = self.mutationProb, \
curve = True, random_state = self.randomState)
def solve(self, X, Y):
self.model = mlrose.NeuralNetwork(hidden_nodes = [5], activation = 'relu', \
algorithm = self.algorithm, max_iters = self.maxIters, \
bias = True, is_classifier = True, learning_rate = self.learnRate, \
early_stopping = False, clip_max = 1e10, max_attempts = self.maxAttempts, \
restarts = self.numRestarts, curve = True, random_state = self.randomState)
self.model.fit(X, Y)
yPred = self.model.predict(X)
score = metrics.accuracy_score(Y, yPred)
return score