- Datasets
We collected the datasets manually. Datasets: https://docs.google.com/spreadsheets/d/1sXH_AzBdKYsi7JDkQhYGUcFnOUtJJmlHTS0Y0xDSC-A/edit?usp=sharing - Model
We use NLP (Natural Language Processing) concepts for tokenizing and pad sequences and a 1-dimensional CNN (Convolutional Neural Network) that specializes in processing text contained in the Tensorflow library. Our model is Unsupervised Learning type because based on our datasets there is no label that represents each record. - Result Training with Plot
The following are the results of the model training results, namely Loss training and Accuracy Training plots:
