Improved GA based Clustering with a New Selection Method for Categorical Dental Data
Abha Sharma, Pushpendra Kumar, Denny Ben, Meet Bikhani, Rabia Musheer Aziz · 2024
Genuine diagnosis of any illness is the heart of successful treatment and dentistry has also started acquiring its place with exposure to data computation and accessibility of massive quantities of patient data. Due to the huge amount of data availability, clustering will assist in the partitioning of data, which will eventually help the researcher to analyze and categorize the attributes of data into various groups under some similarity or dissimilarity. Hence proposed an Improved GA-based Categorical data clustering algorithm with a new Selection method (IGACS) in which the searching capability of the Genetic Algorithm (GA) is utilized to produce global optimum solutions. A novel selection operator “Poor Can be Genius” (PCG) is proposed under the IGACS method which prohibits the loss of the least fit chromosome after calculating its fitness. Experimental outcomes prove that the proposed IGACS algorithm gives 35% higher accuracy than the k-modes algorithm. In addition, the proposed IGACS algorithm is capable to discover the correct number of clusters present in the dental dataset.