cheat_sheet.md

Numpy

Getting number total of words

len(numpy.unique(numpy.hstack(X)))

Finding number of classes

np.unique(y)

Mean and standard deviation

len(numpy.unique(numpy.hstack(X)))

Predictable numpy random

You would fixate random for reproducibility

seed = 7
numpy.ramdom.seed(seed)

Keras Data Processing

Get model summary

model.summary()

Truncate or Pad a dataset

import keras.preprocessing import sequence
sequence.pad_sequences(X_train, maxlen=max_words)

Sentiment Analysis predit new text

text = numpy.array([‘this is an excellent sentence’])
#print(text.shape)
tk = keras.preprocessing.text.Tokenizer( nb_words=2000, lower=True,split=” “)
tk.fit_on_texts(text)
prediction = model.predict(numpy.array(tk.texts_to_sequences(text)))
print(prediction)