Features Overview

GroupDocs.Classification main feature is an ability to classify raw text and documents with IAB-2, Documents, Sentiment or Sentiment3 taxonomies. Sentiment Classification (Analysis) supports 4 languages: English, Chinese, Spanish, and German.

GroupDocs.Classification provides flexible set of settings to customize classification process:

NameDescriptionDefault value
bestClassesCountSelect the number of results to return1
taxonomySelect taxonomy (IAB-2, Documents, Sentiment or Sentiment3).Taxonomy.Iab2
precisionRecallBalanceSelect precision/recall balance for Documents taxonomy. If the classifier is not sure of the result, it will return Other class. If precision selected, then classificator will try to return document class only if completely sure. Opposite if recall selected, then it will try to return document class even if it not sure. Default is a balanced one value.PrecisionRecallBalance.Default

Raw text classification

GroupDocs.Classification supports classifying raw text with IAB-2, Documents, Sentiment or Sentiment3 taxonomies. The classifier will not use the document structure in this case.

Document classification

GroupDocs.Classification supports classifying documents with the next format:

  • Docs
  • PDF
  • OpenOffice
  • RichText (RTF), TXT

Classification results

GroupDocs.Classification returns result in the next form:

    BestClassName: string,
    BestClassProbability: float
    BestResults :
           Name: string,
           Probability: float

BestResults will be null if the number of results to return equals 1.