This document discusses using deep neural networks for semi-automatic text classification. It provides an overview of text classification applications and algorithms such as naive Bayes, random forests, and neural networks. It also presents two example use cases: classifying Spiegel Online articles by topic and Deutsche Bahn social media posts by sentiment. Evaluation shows that deep learning methods like bi-LSTM and convolutional networks outperform classic approaches like TF-IDF with naive Bayes or SVM for both short and long texts.