A Survey on Text Classification Algorithms: From Text to Predictions

In recent years, the exponential growth of digital documents has been met by rapid progress in text classification techniques. Newly proposed machine learning algorithms leverage latest advancements deep methods, allowing for automatic extraction expressive features. The swift development these methods led to a plethora strategies encode natural language into machine-interpretable data. modelling are used conjunction with ad hoc preprocessing procedures, which description is often omitted favour more detailed explanation step. This paper offers concise review models, emphasis on flow data, from raw output labels. We highlight differences between earlier and recent, learning-based both their functioning how they transform input To give better perspective landscape, we provide an overview datasets English language, as well supplying instructions synthesis two new multilabel datasets, found be particularly scarce this setting. Finally, outline experimental results discuss open research challenges posed models.