After an coding-Experiment on data of the Austrian Press Agency we did some experiements on having a prediction about (text-)codings.
Having given a set of codings on articles. The aim of this Experiment is to predict that codings. The first approach for german language articles was to use VopalWabits with no usable results (too complex for now). Than I found a sentiment aproach Experiment on that topic. This approach was used for predicting the codings. The result ist poor but astonishing. Would need an refinement.