def __init__(
self,
name: str,
model_params: Dict[str, Any],
) -> None:
super().__init__(name, model_params)
#self._fit_params = model_params.pop('fit_params')
self._fit_params = {}
self._model = StructuredDataClassifier(**model_params)
class AutoKerasBaselineModel(Model):
def __init__(
self,
name: str,
model_params: Dict[str, Any],
) -> None:
super().__init__(name, model_params)
#self._fit_params = model_params.pop('fit_params')
self._fit_params = {}
self._model = StructuredDataClassifier(**model_params)
def _force_fit(self, X: np.ndarray, y: np.ndarray) -> None:
self._model.fit(X, y, **self._fit_params)
def fit(self, X: np.ndarray, y: np.ndarray) -> None:
if not isinstance(self._model, StructuredDataClassifier):
raise RuntimeError(
'Due to AutoKeras being unsaveable, saving this' +
' means only the best keras model was' +
' saved. Calling fit will fit this pipeline' +
' rather than the AutoKeras algorithm. If this' +
' is desired behaviour, please use ' + '`_force_fit` instead')
self._force_fit(X, y)
def save(self, path: str) -> None:
# Get the best model that can be saved
best_model = self._model.export_model()
best_model.save(path)
# Also store the model wrapper after removing autokeras model
# as it can't be saved
self._model = None
wrapper_path = path + '.wrapper'
with open(wrapper_path, 'wb') as file:
pickle.dump(self, file)
@classmethod
def load(cls, path: str):
wrapper_class = None
# Have to load model wrapper and the underlying keras model seperatly
# as the autokeras model can't be saved
wrapper_path = path + '.wrapper'
with open(wrapper_path, 'rb') as file:
wrapper_class = pickle.load(file)
wrapper_class._model = keras.models.load_model(path)
return wrapper_class
def predict(self, X: np.ndarray) -> np.ndarray:
if isinstance(self._model, StructuredDataClassifier):
return self._model.predict(X)
else: # Already been exported,
return np.round(self._model.predict(X))
def predict_proba(self, X: np.ndarray) -> np.ndarray:
if isinstance(self._model, StructuredDataClassifier):
exported_model = self._model.export_model()
return exported_model.predict(X)
else: # Already been exported
return self._model.predict(X)
def auto_ml_classify(url, x_names, y_name, max_trials, test_size, X_new):
# load dataset
dataframe = read_csv(url)
X = dataframe[x_names].values
y = dataframe[y_name].values
print(X.shape, y.shape)
# basic data preparation
X = X.astype('float32')
y = LabelEncoder().fit_transform(y)
# separate into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=test_size,
random_state=1)
print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)
# define the search
search = StructuredDataClassifier(max_trials=max_trials)
# perform the search
search.fit(x=X_train, y=y_train, verbose=0)
# evaluate the model
loss, acc = search.evaluate(X_test, y_test, verbose=0)
print('Accuracy: %.3f' % acc)
# use the model to make a prediction
yhat = search.predict(X_new)
print('Predicted: %.3f' % yhat[0])
# get the best performing model
model = search.export_model()
# summarize the loaded model
model.summary()
# save the best performing model to file for next time you dont have to train
model.save('model_sonar.h5')
# set the seeds for reproducible results with TF (wont work with GPU, only CPU)
np.random.seed(2)
session_conf = tf.compat.v1.ConfigProto(
intra_op_parallelism_threads=1,
inter_op_parallelism_threads=1)
# Force Tensorflow to use a single thread
sess = tf.compat.v1.Session(
graph=tf.compat.v1.get_default_graph(),
config=session_conf)
tf.compat.v1.keras.backend.set_session(sess)
# define the search
search = StructuredDataClassifier(max_trials=max_trials)
# perform the search
search.fit(x=X_train, y=y_train, verbose=0, epochs=epochs)
y_pred_train = search.predict(X_train)
y_pred_test = search.predict(X_test)
# get the best performing model
model = search.export_model()
# Model summary
s = io.StringIO()
model.summary(print_fn=lambda x: s.write(x + '\n'))
model_summary = s.getvalue()
s.close()
# basic data preparation
#X = X.astype('float32')
y = LabelEncoder().fit_transform(y)
# separate into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=1)
print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)
# define the search
search = StructuredDataClassifier(max_trials=100,
column_types={
'Pclass': 'categorical',
'Sex': 'categorical',
'Age': 'numerical',
'SibSp': 'categorical',
'Parch': 'numerical',
'Ticket': 'categorical',
'Fare': 'numerical',
'Cabin': 'categorical',
'Embarked': 'categorical'
})
# perform the search
search.fit(x=X_train, y=y_train, verbose=0)
# evaluate the model
loss, acc = search.evaluate(X_test, y_test, verbose=0)
print('Accuracy: %.3f' % acc)
# use the model to make a prediction
#row = [0.0200,0.0371,0.0428,0.0207,0.0954,0.0986,0.1539,0.1601,0.3109,0.2111,0.1609,0.1582,0.2238,0.0645,0.0660,0.2273,0.3100,0.2999,0.5078,0.4797,0.5783,0.5071,0.4328,0.5550,0.6711,0.6415,0.7104,0.8080,0.6791,0.3857,0.1307,0.2604,0.5121,0.7547,0.8537,0.8507,0.6692,0.6097,0.4943,0.2744,0.0510,0.2834,0.2825,0.4256,0.2641,0.1386,0.1051,0.1343,0.0383,0.0324,0.0232,0.0027,0.0065,0.0159,0.0072,0.0167,0.0180,0.0084,0.0090,0.0032]
#X_new = np.asarray([row]).astype('float32')
#yhat = search.predict(X_new)
#print('Predicted: %.3f' % yhat[0])
import pandas as pd #import pandas
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder
from autokeras import StructuredDataClassifier
df = pd.read_csv('/content/drive/MyDrive/breast-cancer-wisconsin.data.txt')
# basic data preparation
X = np.array(df.drop(['class'], 1)) #input
#X = X.astype('float32')
y = np.array(df['class']) #output
# integer encode
y = LabelEncoder().fit_transform(y)
# Look at the dataset again
print(f'Number of Rows: {df.shape[0]}')
print(f'Number of Columns: {df.shape[1]}')
print(df.head())
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
search = StructuredDataClassifier(max_trials=5)
search.fit(x=X_train, y=y_train, verbose=1)
loss, acc = search.evaluate(X_test, y_test, verbose=0) #classification
print('Accuracy: %.3f' % acc)
print('loss: %.3f' % loss)
y_predictions = search.predict(X_test)
model = search.export_model()
model.summary()
print("\n%s: %.2f%%" % (model.metrics_names[1], acc*100))
model.save('breast_cancer_model.tf')
from keras.utils.vis_utils import plot_model
plot_model(model, to_file='model_plot.png', show_shapes=True, show_layer_names=True)
X, y = data[:, :-1], data[:, -1]
print(X.shape, y.shape)
# basic data preparation
X = X.astype('float32')
y = LabelEncoder().fit_transform(y)
# separate into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=1)
print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)
# define the search
search = StructuredDataClassifier(max_trials=15)
# perform the search
search.fit(x=X_train, y=y_train, verbose=0)
# evaluate the model
loss, acc = search.evaluate(X_test, y_test, verbose=0)
print('Accuracy: %.3f' % acc)
# use the model to make a prediction
row = [
0.0200, 0.0371, 0.0428, 0.0207, 0.0954, 0.0986, 0.1539, 0.1601, 0.3109,
0.2111, 0.1609, 0.1582, 0.2238, 0.0645, 0.0660, 0.2273, 0.3100, 0.2999,
0.5078, 0.4797, 0.5783, 0.5071, 0.4328, 0.5550, 0.6711, 0.6415, 0.7104,
0.8080, 0.6791, 0.3857, 0.1307, 0.2604, 0.5121, 0.7547, 0.8537, 0.8507,
0.6692, 0.6097, 0.4943, 0.2744, 0.0510, 0.2834, 0.2825, 0.4256, 0.2641,
print(dataframe.shape)
# split into input and output elements
data = dataframe.values
X, y = data[:, :-1], data[:, -1]
print(X.shape, y.shape)
# basic data preparation
X = X.astype('float32')
y = LabelEncoder().fit_transform(y)
# separate into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
random_state=1)
print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)
# define the search
search = StructuredDataClassifier(max_trials=15)
# perform the search
search.fit(x=X_train, y=y_train, verbose=0)
# evaluate the model
loss, acc = search.evaluate(X_test, y_test, verbose=0)
print('Accuracy: %.3f' % acc)
# use the model to make a prediction
row = [
0.0200, 0.0371, 0.0428, 0.0207, 0.0954, 0.0986, 0.1539, 0.1601, 0.3109,
0.2111, 0.1609, 0.1582, 0.2238, 0.0645, 0.0660, 0.2273, 0.3100, 0.2999,
0.5078, 0.4797, 0.5783, 0.5071, 0.4328, 0.5550, 0.6711, 0.6415, 0.7104,
0.8080, 0.6791, 0.3857, 0.1307, 0.2604, 0.5121, 0.7547, 0.8537, 0.8507,
0.6692, 0.6097, 0.4943, 0.2744, 0.0510, 0.2834, 0.2825, 0.4256, 0.2641,
0.1386, 0.1051, 0.1343, 0.0383, 0.0324, 0.0232, 0.0027, 0.0065, 0.0159,
0.0072, 0.0167, 0.0180, 0.0084, 0.0090, 0.0032
]
print("shape of the input")
print(X.shape, y.shape)
X = X.astype('float32')
y = LabelEncoder().fit_transform(y)
# separate into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.33,
stratify=y,
random_state=42)
print("shape of the splitting Data X_train, X_test, y_train, y_test")
print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)
# define the search space
search = StructuredDataClassifier(
max_trials=30,
tuner='random',
metrics=[tf.keras.metrics.CategoricalAccuracy(), 'accuracy'])
# perform the search
search.fit(x=X_train, y=y_train, verbose=1)
# evaluate the model
loss, acc = search.evaluate(X_test, y_test, verbose=0)
print('Modell Accuracy')
print('Accuracy: %.3f' % acc)
# get the best performing model
model = search.export_model()
# summarize the loaded model
model.summary()