Compiling and training the model

Now that the model is defined, it is ready to be compiled. To compile the model in Keras, we need to determine the optimizer, the loss function, and optionally the evaluation metrics. As we mentioned previously, the problem is to predict if the tweet is positive, negative, or neutral. This problem is known as a multi-category classification problem. Thus, the loss (or the objective) function that will be used in this example is the categorical_crossentropy. We will use the rmsprop optimizer and the accuracy evaluation metric.

In Keras, you can find state-of-the-art optimizers, objectives, and evaluation metrics implemented. Compiling the model in Keras is very easy using the compile function:

model.compile(optimizer='rmsprop',
          loss='categorical_crossentropy',
          metrics=['accuracy'])

We have defined the model and compiled it, and it is now ready to be trained. We can train or fit the model on the defined data by calling the fit function.

The...