Knowledge of eye care is needed when eye disorders occur. One of the tools needed as a tool to identify eye disorders is the application of an expert system for early detection of eye diseases, which can be used as a first step in eye disorders consultations. In this study, an expert system application was designed that uses weighted parameter determination techniques with the pairwise comparison method for Case Based Reasoning (CBR) early detection of eye diseases using the KNN algorithm. In the weighting of pairwise comparisons, this study is divided into four stages of preparation, namely: 1) Compiling a pairwise comparison matrix to see the level of importance between criteria based on the prioritization scale of pairwise comparison; 2) Normalization of the decision matrix from the results of the matrix compilation; 3) Consistency analysis by comparing the criteria with other criteria which can lead to inconsistency; 4) Implementation of the KNN algorithm which is used to determine the similarity value between a consultation and an old case in the database. Determination of parameter weights using the pairwise comparison method implemented on 20 symptoms and 18 eye diseases resulted in three weight groups, namely: severe symptoms with a weight of 0.636986, moderate symptoms with a weight of 0.258285, and mild symptoms with a weight of 0.104729.
Copyrights © 2021