International Journal of Engineering and Technology

Diabetes is a chronic disease and major problem of morbidity and mortality in developing countries. Type-2 diabetes mellitus (T2DM) is the most common type of diabetes and accounts for 90-95% of all diabetes. In the medical field a predictive data mining association algorithms is used to diagnose the disease at the earlier stage which helps the physicians in the treatment planning procedure. Recently, an improved Frequent Pattern Growth (IFP-Growth) with Hybrid Enhanced Artificial Bee Colony-Advanced Kernel Support Vector Machine (EABC-AKSVM) classification is introduced with the capability to reduce the number of rules in diagnosing the diabetes. However, the detection accuracy and robust is less. To resolve this problem, an Improved Frequent Pattern Growth (IFP-Growth) with Hybrid Modified Glowworm Swarm Optimization-Improved Relevance Vector Machine (MGSO-IRVM) Classification based Association Rule Mining (ARM) system is proposed in this work to generate effective rules. The proposed work consists of two phases: In first phase, improved FP-growth is proposed to efficiently mine frequent patterns even from uncertain medical database. This is achieved by creating additional array for each item in uncertain transactional database to keep the information of its super-item sets in IFP-tree redundant node generation through this step the computational cost is greatly reduced. In next phase, Hybrid MGSO-IRVM classifier is used to generate the association rules based on the frequent item sets, which avoids rule redundancy and conflicts during the rule mining process. Experimental results show that the proposed model is suitable and alternative model for medical classification to achieve greater accuracy, and to improve medical diagnosis.