An efficient improved parrot optimizer for bladder cancer classification

Area of Science:

• Computational intelligence
• Medical informatics
• Machine learning

Background:

• Bladder cancer (BC) presents significant morbidity and mortality risks.
• Accurate BC classification is challenging and requires expert analysis.
• Existing optimization algorithms like the Parrot Optimizer (PO) suffer from limitations such as sub-optimal convergence and high error rates.

Purpose of the Study:

• To develop an improved optimization algorithm (IPO) to overcome the limitations of the original PO.
• To enhance the accuracy and efficiency of bladder cancer classification using machine learning.
• To evaluate the performance of the proposed IPO algorithm against existing methods.

Main Methods:

• Developed the Improved Parrot Optimizer (IPO) by integrating Mirror Reflection Learning (MRL) and Bernoulli Maps (BMs).
• Evaluated IPO on CEC 2022 test functions and nine bladder cancer datasets.
• Integrated IPO with Support Vector Machine (SVM) classifier to create the IPO-SVM approach for BC classification.

Main Results:

• The IPO algorithm ranked first in optimization performance for CEC 2022 functions.
• The IPO-SVM approach demonstrated superior performance over eight other metaheuristic algorithms on BC datasets.
• IPO-SVM achieved high classification metrics: 84.11% Accuracy, 98.10% Sensitivity, 95.59% Precision, 95.98% Specificity, and 94.15% F-score.

Conclusions:

• The proposed IPO algorithm effectively avoids local optima and improves convergence speed and solution diversity.
• The IPO-SVM approach offers a promising and effective tool for accurate bladder cancer classification.
• The developed IPO algorithm has the potential to significantly aid in the early detection and management of bladder cancer.