Detecting Auto Bot Text Content Document Based on Subspace Relative Lexicon Depth Measure Using Bigram Inverse Frequency Key Term Analyzer

The proliferation of automated text generation poses significant challenges in cybersecurity and digital communication. This paper proposes a novel approach for detecting bot-generated text content using a Subspace Relative Lexicon Depth (SRLD) measure combined with a Bigram Inverse Frequency Key Term (BIFKT) analyzer. The SRLD measure evaluates the depth and spread of word usage within a specified lexicon for effectively distinguish between human-authored and bot-generated content. BIFKT analyzer utilizes bigrams and their inverse frequency to identify key terms that are less common in human writing but frequently appear in automated content. The integration of these two techniques creates a robust framework that improves accuracy and reduces false positives compared to existing methods. The effectiveness of the proposed detection system was validated through extensive experiments on diverse datasets, including social media posts, online reviews, and news articles. The results showed a significant improvement in detection rates.