Multi-label Text Classification using GloVe and Neural Network Models

This study explores multi-label text classification challenges, focusing on imbalanced datasets, variable text lengths, and diverse feature labels. We propose a hybrid approach combining the GloVe model-based bag-of-words with a CNNBiLSTM neural network. This method leverages pre-trained Globe embeddings for efficient text representation, feeding into a CNN-BiLSTM network for classification. The approach demonstrates notable efficiency, achieving 87.26% accuracy and an F1 score of 0.8737 on the test set, indicating its potential for practical application in text classification tasks.