Integrating Multidimensional Data with Neutrosophic Logic for Enhanced Diagnosis of Type 2 Diabetes

Abstract

Effective diagnostic frameworks require accounting for complex and ambivalent clinical data,
particularly given the current rise in reported cases of Type 2 Diabetes. However, traditional
approaches have operated on fixed thresholds for indicators such as BGL, BMI, and FH,
without considering the variability that occurs when other indices interact. Thus, this project
proposes a robust model that integrates apparelled detection using multi-dimensional data
and is able to process uncertainties via Neutrosophic logic for more accurate diagnosis. The
evaluation method of 100 patient cases across the three diagnostic dimensions of Truth (T),
Indeterminacy (I), and Falsity (F) also normalized both quantitative indicators (BGL, BMI)
and qualitative indicators (FH) to ensure compatibility of the data it collected. Statistical
analyses were performed on the correlativity of numerous variables; through variable plots, a
3D scatter plot was generated that expressed the suspicion of type 2 diabetes, variability and
clinical divergences of the indicators to enhance clinical diagnosis. This study endorses the
fact that Neutrosophic framework significantly outsmarts known methods by extending gains
in accuracy (92% against 80%), sensitivity (94% against 82%), and specificity (90% against
78%). With an AUC of 0.94 against as little as 0.82 for standard techniques, this model is
best for treating uncertainty. Hitherto, negligence of the true strength of maintaining proper
level of ambiguity in the structure and quality of clinical data is shocking. It could serve as
a laudable decision-support tool for physicians, thereby boasting the accuracy and reflected
confidence of the whole process of Type 2 Diabetes diagnosis.