Identification of the human-oriented factors influencing AERC from the Web services

Abstract

Abstract For both academia and industry, the exploration of new research applications to the nature of human beings from human-like intelligence is advantageous and challenging. Promoting the safety of hospitalized patients is particularly a top priority in the health-care industry. Thus, applications of a patient safety reporting system (PSRS) from the existing reports of Web services provide assistance to identify potential hazards, and to reduce medical errors from two opposite viewpoints: the medical professional, and the patient. Few studies based on emerged machine learning systems have examined the accident event reporting category (AERC) problems from the PSRS of advanced network infrastructure for data storage and resource sharing in viewing development of the next-generation Web. The related analysis of a knowledge gap against classification problems with implied patient safety issue is lacking, and must therefore be filled. This work proposes an enhanced hybrid method, including expert recommendations, a decision-tree C4.5 algorithm, an objective cumulative probability distribution method, the rough set LEM2 algorithm, and rule-filter processing to classify the event reporting categories, analyze their human-oriented factors, and provide meaningful decision rules under the two comparative contexts of medical professional and patient. Two real data sets were retrieved from the PSRS of existing hospital databases, based on various doctor-focused and patient-focused conditional attributes to illustrate significant activity throughout the study framework. Empirical evidence shows that the proposed model has advantages of removing irrelevant features, increasing classification accuracy, and providing a well-informed rule set in supporting knowledgeable decision-making to solve AERC problems, thereby benefiting interested parties.