A novel model to predict the readability of online diabetes information based on fuzzy logic

The recent growth of information and communication technologies, has evolved the websites into an essential and universal intermediary for disseminating diabetes-related information. Government and non-government websites are commonly visited by people to learn about diabetes symptoms, treatment, and management; therefore, a readability examination is required. So far, numerous research has been pursued the diabetes information’s quality evaluation, but there is a lack of fuzzy-based computational model for readability prediction of online diabetes information. This study aimed to develop a fuzzy-based computational model that could predict unified reading grade level by combining six readability metrics to generate audience-specific readability score specifically for online diabetes information, differing from conventional standalone readability formulas used in previous studies. The readability of diabetes information was evaluated using six metrics, which were then used as ranked input variables in fuzzy system. The proposed model used these inputs and fuzzy rules to predict the readability based on target audience. Higher grade levels signify difficult readability, in contrast to lower grade levels, which reflect easy readability. The findings revealed that the fuzzy system predicted a readability score of 6.46 ± 1.73 (mean ± SD) for government websites and 14.40 ± 2.33 (mean ± SD) for non-government websites. This suggests that the non-government’s content is challenging for average readers and appropriate for college graduates. The proposed fuzzy model’s observations yield definite and beneficial outcomes. The study suggests that the implementation of proposed fuzzy model could support quick readability prediction, enabling the healthcare providers and editors to enhance diabetes-related content’s readability before online publication.