Content Comparison

Yes, that’s one of the most powerful aspects of Symanto Brain engine. Above are just a few examples, and you can freely define the categories or labels, and the engine will immediately create new models that aim at classifying the text into one (or more) of these labels.

The pattern provides a context of the classification task to the engine. Often used patterns include “This person talks about {}”, “This person feels {}” or simply “{}”. The engine then predicts which label fits the best into the {} bracket for the input text.

A pattern can be freely defined, though some patterns tend to work better

When multi-label is disabled, the labels are mutually exclusive. Sentiment analysis belongs in this case (the sentiment of a text is either positive or negative or neutral). The predicted probabilities of all labels will sum up to 1.

When multi-label is enabled, every label is treated as a different classification task. The topic analysis is an example (a text can be related to multiple topics).

We currently offer Symanto Brain via API.