Combining Natural Language Processing and Ensemble Models to Visualize Team-Level Imposter Syndrome Dynamics

• Adaan Ahsun

  Covenant University

  Author

Imposter Syndrome (IS) is pervasive in high-performing technology teams, posing significant threats to individual decision-making, emotional stability, and team collaboration effectiveness. Although prior research has confirmed the high prevalence of IS among software engineers, existing predictive tools predominantly focus on the individual level, lack the capacity to model team dynamics, and fail to effectively leverage psychological signals embedded in unstructured text. This study proposes and validates a framework that integrates Natural Language Processing (NLP) with ensemble machine learning for team-level IS dynamics prediction and visualization. Using the OSMI Mental Health in Tech Survey 2016 dataset, we constructed a multidimensional feature space through sentiment analysis, linguistic feature extraction, and team interaction pattern encoding, and compared the predictive performance of Random Forest, XGBoost, Support Vector Machine, and Logistic Regression models. Results indicate that the Random Forest classifier achieved optimal performance in team-level IS risk prediction with an accuracy of 89.4%, significantly outperforming traditional static assessment tools. Feature importance analysis identified "lack of psychological safety," "excessive work investment," and "negative sentiment expression density" as the top three predictors. This study provides organizational managers with an actionable early-warning framework, shifting the IS intervention window from reactive response to proactive prediction.

• pdf

Copyright (c) 2026 Adaan Ahsun (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.