Latent Personality Traits Assessment From Social Network Activity Using Contextual Language Embedding

Recognizing author identity from digital footprints without having a large corpus of documents an individual is keen interest to security researchers and government agencies. Users reveal aspects their personality via the content they share with social media followers through patterns in interactions on online networking platforms. This study examines potency emerging natural language processing (NLP) methods analyzing network activity. A linguostylistic traits assessment (LPTA) system developed estimate Twitter users’ based tweets using Myers-Briggs-type indicator (MBTI) big-five scales. novel input representation mechanism proposed process by converting them into real-valued vectors frequency, co-occurrence, context (FCC) measures. Other prevalent text schemes, such as one-hot encoding, count-based vectorization, pretrained model representations are used comparators. genetic algorithm (GA) approach reduce feature set increase efficacy features extracted. The outperforms state-of-the-art research reliably estimating user’s latent while 50 or fewer per user.