AI-Powered Driver Grievance Detection in Ride-Hailing Services using Machine Learning Techniques
- Version
- Download 12
- File Size 645.10 KB
- File Count 1
- Create Date 19 March 2026
- Last Updated 19 March 2026
AI-Powered Driver Grievance Detection in Ride-Hailing Services using Machine Learning Techniques
Dr. Y. Mohammed Iqbal1, A. Abubakkar2, Dr. S. Peerbasha3, Dr. M. Rajakumar4, Dr. M. Mohamed
Surputheen5
Department of Computer Science, Jamal Mohamed College (Autonomous), Affiliated to Bharathidasan
University, Trichy-20, Tamil Nadu, India
ABSTRACT - The Indian gig economy relies heavily on taxi applications, yet drivers face severe issues regarding commissions and payments, often voicing grievances on the Google Play Store. Manual analysis of these high-volume, mixed-language reviews is inefficient, leading to driver dissatisfaction and attrition.To resolve this, we propose the AI-Powered Driver Grievance Detection System. Bypassing public datasets,we scraped 11,663 raw reviews from the Play Store, resulting in 11,240 clean, high-quality records after removing duplicates and noise. Natural LanguageProcessing (NLP) techniques, including noise removal and lowercase conversion, were applied, followed byTF-IDF vectorization to extract mathematical feature representations. We evaluated four machine learning classifiers—Logistic Regression, Multinomial NaiveBayes, Random Forest, and Support Vector Machine (SVM)—using an 80-20 train-test split. The SVMnclassifier outperformed the others, achieving the highest accuracy of 90.08% in separating serious grievances from normal feedback. This automated text classification framework allows taxi platforms to rapidly identify and address specific driver problems, effectively reducing preventing driver churn. administrative delays and
Keywords: Gig Economy, Machine Learning, Driver Grievance, Natural Language Processing, Support Vector Machine, TF-IDF Vectorization, Sentiment Analysis, Text Classification.
Download