Self-Assessment-Based Dataset for Hearing Impairment Classification

This dataset captures self-reported communication difficulties among individuals from both normal and hearing-impaired populations. It is derived from responses to a standardized self-assessment questionnaire designed to evaluate hearing handicap severity in real-life communication scenarios. Each response is categorized into four levels of difficulty: "Most of the time" (>75%), "Sometimes" (25%–75%), "Seldom" (<25%), and "Not Applicable."

The data encompasses a demographically diverse population, ensuring representation across age, gender, and socio-economic backgrounds. Statistical reliability is validated using Cronbach’s Alpha (α = 0.88), confirming strong internal consistency. Discriminant validity tests and factor correlation analysis further support the dataset's ability to distinguish between normal and impaired populations based on communication challenges.

The dataset has been successfully used to train and evaluate machine learning models for hearing impairment classification. Notably, a Neural Network model achieved superior performance with 97% accuracy, 96% precision, 98% recall, and a 97% F1-score, outperforming Logistic Regression and Decision Tree classifiers. This highlights the dataset’s value in developing data-driven auditory health assessment tools suitable for clinical and telehealth applications.