Topic Classification for Short Texts

In the context of TV and social media surveillance, constructing models to automate topic identification short texts is a key task. This paper constructs worth-to-consider for practical usage, employing Top-K multinomial classification methodology. We describe full data processing pipeline, discussing about dataset selection, text preprocessing, feature extraction, model selection learning, including hyperparameter optimization. will test compare popular methods including: standard machine deep fine-tuned BERT classification.