LSTM AE

path: keras.Sequential.LSTMSeq2Seq

description: this is a reconstruction model autoencoder using LSTM layers.

see json.

argument	type	description
parameters
X	numpy.ndarray	n-dimensional array containing the input sequences for the model
y	numpy.ndarray	n-dimensional array containing the target sequences for the model
hyperparameters
classification	bool	indicator of whether this is a classification or regression model. Default is False
epochs	int	number of epochs to train the model. An epoch is an iteration over the entire X and y data provided. Default is 35
callbacks	list	list of callbacks to apply during training
validation_split	float	float between 0 and 1. Fraction of the training data to be used as validation data. Default is 0.2
batch_size	int	number of samples per gradient update. Default is 64
window_size	int	integer denoting the size of the window per input sample
input_shape	tuple	tuple denoting the shape of an input sample
target_shape	tuple	tuple denoting the shape of an output sample
optimizer	str	string (name of optimizer) or optimizer instance. Default is keras.optimizers.Adam
loss	str	string (name of the objective function) or an objective function instance. Default is keras.losses.mean_squared_error
metrics	list	list of metrics to be evaluated by the model during training and testing. Default is [“mse”]
return_seqeunces	bool	whether to return the last output in the output sequence, or the full sequence. Default is False
layers	list	list of keras layers which are the basic building blocks of a neural network
verbose	bool	verbosity mode. Default is False
lstm_1_unit	int	dimensionality of the output space for the first LSTM layer. Default is 80
dropout_1_rate	float	float between 0 and 1. Fraction of the units to drop for the linear transformation of the inputs for the first LSTM layer. Default: 0.3
lstm_2_unit	int	dimensionality of the output space for the second LSTM layer. Default is 80
dropout_2_rate	float	float between 0 and 1. Fraction of the units to drop for the linear transformation of the inputs for the second LSTM layer. Default: 0.3
output
y	numpy.ndarray	predicted values

In [1]: import numpy as np

In [2]: from mlstars import load_primitive

In [3]: X = np.array([1] * 100).reshape(1, -1, 1)

In [4]: primitive = load_primitive('keras.Sequential.LSTMSeq2Seq',
   ...:     arguments={"X": X, "y": X, "input_shape":(100, 1), "target_shape":(100, 1),
   ...:                "window_size": 100, "batch_size": 1, "validation_split": 0, "epochs": 5})
   ...: 

In [5]: primitive.fit()
Epoch 1/5

1/1 [==============================] - ETA: 0s - loss: 0.9923 - mse: 0.9923
1/1 [==============================] - 2s 2s/step - loss: 0.9923 - mse: 0.9923
Epoch 2/5

1/1 [==============================] - ETA: 0s - loss: 0.7640 - mse: 0.7640
1/1 [==============================] - 0s 22ms/step - loss: 0.7640 - mse: 0.7640
Epoch 3/5

1/1 [==============================] - ETA: 0s - loss: 0.5618 - mse: 0.5618
1/1 [==============================] - 0s 22ms/step - loss: 0.5618 - mse: 0.5618
Epoch 4/5

1/1 [==============================] - ETA: 0s - loss: 0.3806 - mse: 0.3806
1/1 [==============================] - 0s 21ms/step - loss: 0.3806 - mse: 0.3806
Epoch 5/5

1/1 [==============================] - ETA: 0s - loss: 0.2211 - mse: 0.2211
1/1 [==============================] - 0s 23ms/step - loss: 0.2211 - mse: 0.2211

In [6]: pred = primitive.produce(X=X)

1/1 [==============================] - ETA: 0s
1/1 [==============================] - 0s 468ms/step

In [7]: pred.mean()
Out[7]: 0.7112000477046491

LSTM Dense AE