Spam Text Classification
This repository contains two Python scripts for spam text classification models:
- spam_text_classification_roberta.ipyn: a Kaggle notebook that demonstrates how to use a pretrained model from Hugging Face to classify spam text messages.
- fcc_sms_text_classification.py: a script that trains a recurrent neural network (RNN) with two LSTM layers and two dense layers to classify spam text messages.
Pretrained Model
The pretrained model uses a roberta-base model from Hugging Face that has been further trained on a dataset of SMS messages to differentiate between spam and non-spam texts. The RobertaTokenizer is used to tokenize text messages, and the model is trained on a GPU for faster processing. The freeCodeCamp test suite for the Neural Network SMS Text Classifier project is used as a reference for evaluating the performance of the model.
To use the pretrained model, run the notebook and call the predict_message() function, passing in the message you want to classify as an argument. The function will return the predicted label ("spam" or "ham").
Note that to successfully run the notebook, you will need access to wandb, which can be used to track your experiments and visualize your results.
LSTM Model
The LSTM model is trained on a dataset of SMS messages provided by freeCodeCamp. The model is a recurrent neural network with two LSTM layers and two dense layers. The input to the model is a sequence of characters, and the output is a probability indicating the likelihood of the message being spam. The model is trained on two versions of the input data: one with only alphabetic characters and one with alphanumeric characters.
To use the LSTM model, run the script and call the predict_message() function, passing in the message you want to classify as an argument. The function will return a list containing the probability and the label ("spam" or "ham").
Dependencies
Both models use two files provided by freeCodeCamp:
- train-data.tsv: used for training the model.
- valid-data.tsv: used for validating the model.
You can find more information about the dataset and the task of spam text classification in the Neural Network SMS Text Classifier project.