Assessing the Reliability of Electronic Products Using Customer Knowledge Discovery

Abstract Reliability is an essential aspect of product quality that concerns both customers and manufacturers. Other than test-based reliability data, online reviews reflect the actual operating status of all products, where the sample is all delivered products, and the test condition is equivalent to the service condition. This manuscript proposes the framework of a reliability analysis based on online reviews, and combines the statistical reliability analysis with current text mining technology. The proposed method adopts lexicon-based text mining technology to extract the failure-related customer knowledge from product users’ online reviews. Using the information on the symptom and time of each failure experienced by customers, we classify the failure for each component and analyze the reliability using the estimated parameters of failure distributions. A comparative analysis is proposed to eliminate the uncertainty accompanying the review information. The application of the proposed framework is demonstrated by a case study of two similar mobile phone products. The results indicate that the consideration of failure distribution affects the analytical results significantly, and that the type of components, rather than the product model, has a greater impact on the product reliability.