Sequence based synthetic data is used to show convolution neural nets in action.
This is the source code which you can refer while going through the blog post. (http://xplordat.com/2019/01/30/sequence-based-text-classification-with-convolution-nets/)
The blog post illustrate the use of sequences for classification of text and it is being compared with the bag of words approach. Here, we are considering two models for the comaprison, first one is the Convolution Neural Network & 2nd one is traditional Naive Bayes model.
Tensorflow Keras Scikit-Learn Matplotlib Numpy
CNN is implemented using Keras with Tensorflow as its backend. Hyperparameters currently in use has been obtained after performing multiple simulations & by using different combinations of number of filters in convoution, Maxpooling layer as well the stride sizes.
Naive Bayes is implemented using Scikit-Learn.