• Profanity software
  • Accessible via an admin and API interaces;
  • detect, identify, prevent and report profanity / insult at a rate of 80% and goes up to 90% including preprocessing stage.
  • identify the originator and the source,
  • Self-learning and eliminating human effort,
  • Machine Learning (ML) based,
  • Increases the accuracy rate by providing adaptation to new data over time,
  • Analyze Streaming live data
  • Via feedback forms, we detect our misclassifications and improve upon.
  • Profanity filter features
  • detecting swear, profane, bad words and beeping

Area of Use

  • Chat,
  • SMS,
  • Blog(s),
  • Feedback Form(s),
  • Forum(s),
  • Customer Message(s),
  • Complaint system(s),
  • Dating site(s)

Innovative Sides

  • By the use of Asynchronous processing, we achieve great speed:
    • 300 words : 250ms – rps: 10
    • 500words : 341ms – rps: 7
    • 1000words : 547ms – rps: 4
    • 1500words : 798ms – rps: 2
    • 3000words : 1.58s – rps: 0.8
  • Using NLP we extract profane words and their positions
  • We can eliminate special characters, URLs, emails, etc
  • Profanity software has Easy API integration with other systems
  • Finding both exact, similar, derived and suspicious word matches unlike other systems which rely on a single dictionary or exact word matches
  • Self- Learning via ML algorithms that remodels itself according to feedback and new data increasing its accuracy and adaptiveness to the customer’s needs
  • Sentence Analysis to detect insult besides profanity
  • Whitelisting and blacklisting of words
Leave a Reply

Your email address will not be published. Required fields are marked *

TURKOGLU International