Assessment of the risk factors of type II diabetes using ACO with self-regulative update function and decision trees by evaluation from Fisher’s Z-transformation

• Published in: Medical & biological engineering & computing Vol. 60; no. 5; pp. 1391 - 1415
• Main Author: Abdullah, A. Sheik
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2022; Springer Nature B.V
• Subjects: Algorithms; Ant colony optimization; Biomedical and Life Sciences; Biomedical Engineering and Bioengineering; Biomedicine; Blood Glucose; Computer Applications; Confidence intervals; Coronary artery disease; Decision analysis; Decision Trees; Diabetes; Diabetes mellitus; Diabetes Mellitus, Type 2; Evaluation; Glucose; Heart diseases; Hemoglobin; Human Physiology; Humans; Imaging; Kidney diseases; Lower bounds; Original Article; Radiology; Risk analysis; Risk Factors; Statistical analysis; Ant colony optimization; Data classification; Fisher’s; Type II diabetes; Data analytics; Swarm intelligence; transformation; C4.5 decision trees; Fisher’s Z-transformation
• ISSN: 0140-0118; 1741-0444; 1741-0444
• DOI: 10.1007/s11517-022-02530-2

Type II diabetes is considered to be one of the persistent diseases which is the cause of death and disability in many regions. The objective of this research work is to apply an improved combination of the ant colony optimization (ACO) algorithm with decision trees (J48) for accessing and evaluating the risk factors related to type II diabetes. The model developed concerning the routine and deviated values for each attribute corresponding to type II diabetes. Experimental evaluation has been made with the modified self-regulative function of ACO with enhancement in pheromone update rule to make the ants in the search space converge at the best optimal path. In addition to this, continuous assessment has been made with probabilistic function by the construction of new solution incrementally made by the ants through feature by feature analysis. From the results, it has been observed that the risk factors corresponding to type II diabetes are postprandial plasma glucose (PPG), fasting plasma glucose (FPG), and glycosylated hemoglobin (A1c) which has selected with an improved accuracy than that of the existing methods and algorithms. The efficiency of prediction has been tested using Fisher’s Z -transformation with a 95% of confidence level for upper and lower bounds. From the inference, it has been observed that there exists a strong correlation among the risk factors PPF and FPG with significance in P -value for the risk corresponding to type II diabetes. Hence, predictive analytics with improvement in ACO with C4.5 decision tree algorithm can also be deployed for accessing the risk factors related to NCD such as cancer, heart disease, and kidney diseases. Graphical abstract