Telugu domain identification using multiple channel LSTM-CNN

Retrieving domain-oriented information from text data offers a wide range of applications in information retrieval and natural language processing due to the rapid increase of text information. The text is condensed and represented by thematic keywords. Text summarization, information extraction, question-answering, machine translation, and sentiment analysis all frequently rely heavily on domain identification. In this study, we proposed the Telugu Technical Domain Identification Multichannel LSTM-CNN approach. The ICON shared task "TechDOfication 2020" utilised and tested this architecture, and our system received a 69.9% F1 score on the test dataset and a 90.01% F1 score on the validation set.