/cbtbc

Condition based text binnary classification

Primary LanguagePython

CBTBC

Condition-based Text Binary Classification

Condition-based Text Binary Classification is an AutoML system and an NLP (Natural Language Processing) Python framework for text classification.

As the simplest case, our models attempt to divide a set of texts into two groups using some auto-derived conditions. For example, texts written by Tolstoy or texts written by Dostoevsky, texts mostly about nature or describing urban areas, texts related to cryptocurrency, texts related to elections, or texts related to ChatGPT etc. As more complex case a set of binary classifiers can be combined to produce classification for any number of classes.

Key Benefits

  • Requires a low amount of pre-labeled data (Ground Truth) for training productive and robust models.
  • Utilizes self-explainable models that can be used in critical areas where black-box solutions are unwelcome.
  • Need small amount of computational resources and perform very fast classification. (in comarison with ANN based soultion, for example)
  • Features AutoML algorithms - simply provide two sets of texts and the condition-based model will be generated automatically.