Page 166 - {PDF Title}

Mhaske and Kumar

dataset used in this study is the Heart Disease dataset performance, whereas MD has to be lower. From
from the UCI Machine Learning Repository. 36 the graphs in Figure 4, it can be observed that the
This database contains 76 attributes, but previous proposed STI-TSA-based optimization has fulfilled this
experiments have focused on a subset of 14. The goal statement. Particularly, the IPR is higher when data are
field refers to the presence of heart disease in the patient, at 30%. For datasets with 10% and 20%, the proposed
with integer values ranging from 0 (no presence) to 4. STI-TSA-based optimization achieves a higher IPR
The names and social security numbers of the patients compared to STO, TSA, PFO, SBOA, BA, EHO-OBL,
34
35
were recently removed from the database and replaced and BFL-PSO. When data are 10%, the proposed STI-
with dummy values. One processed file containing the TSA-based optimization attained a high IPR of 0.93%;
Cleveland database is available, while the remaining with 20% data, the IPR reaches 0.94%; and with 30%
four unprocessed files are also included in this directory. data, the IPR attains 0.95%. In contrast, at 30% of data,
extant optimization schemes, such as STO, TSA, PFO,
8.2. Analysis of IPR, HFR, and MD SBOA, BA, EHO-OBL, and BFL-PSO achieved
34
35
Figure 4 shows the performance of IPR, HFR, and lower IPR values of 0.89, 0.9, 0.87, 0.9, 0.87, 0.89, and
MD for proposed STI-TSA optimization over extant 0.88, respectively.
optimization schemes, such as STO, TSA, PFO, The new STI-TSA algorithm, incorporating
SBOA, BA, EHO-OBL, and BFL-PSO. Regarding the concepts of SBO and TSA, could attain faster
34
35
fitness, the IPR and HFR have to be higher for better convergence and create high-quality solutions that
A B

C

Figure 4. Performance of privacy preservation for electronic medical records using blockchain technology.
Performance of the Siberian Tiger Integrated Tuna Swarm algorithm (STI-TSA) over extant optimization
approached for (A) information preservation ratio, (B) hiding failure rate, and (C) modification degree.
Abbreviations: BA: Bat Algorithm; BFL-PSO: Bee-foraging learning particle swarm optimization;
EHO-OBL: Elephant Herding Optimization with Opposition-based Learning; PFO: Puffer Fish Optimization;
SBOA: Secretary Bird Optimization Algorithm; STO: Siberian Tiger Optimization; TSA: Tuna Swarm algorithm.

Volume 22 Issue 1 (2025) 160 doi: 10.36922/AJWEP025040017