An Indirect Measurement Method for Accurate Quality of Service Assessment

Existing methods for assessing service quality rely on customer feedback, which can introduce bias and fail to capture true feelings and causes. This research proposes a novel approach that models service quality by analyzing changes in customer behavior rather than direct feedback. We construct a predictive model using gradient boosting regression techniques to analyze big data and identify how various aspects of service provider performance (the "5W1H" Who, What, When, Where, Why, and How) impact customer behavior metrics. This model provides an objective and detailed assessment of service quality, revealing complex patterns and interactions often missed by traditional feedback methods.