Topic Extraction

Saplo Topics suggest suitable topics, such as "The Financial Crisis", for any given text. In contrast to automatic tagging, a topic word is not required to appear in text in order to be identified.

With the help of Saplo Topics you can automatically categorize large volumes of text. For instance you can categorize and label user-generated text such as wikis according to pre-defined business units or special interest groups (financial, legal, development etc). Topics can also be used in collaborative purposes in large organisations by linking text with similar topic labels (e.g. two persons writing about the same Topic).

Implementation

Saplo Topics is based on our contextual recognition technology, also known as Groups in Saplo Text Analysis API.

When creating a Topic, you are acctually creating a text group in our API. After that you add texts to the group so that we know what kind of texts this group (or Topic) is referring to. Now you can compare new texts to your text group and find out how similar your new text is to the group you have created. If the text is very similar to the group, then you can determine that the text is about the Topic that your group represent.