An evolutionary machine learning for multiple myeloma using Runge Kutta Optimizer from multi characteristic indexes.

Area of Science:

• Computational biology
• Medical informatics
• Machine learning

Background:

• Multiple myeloma (MM) is a significant hematological malignancy, representing 1.8% of all cancers in the US.
• Accurate and early diagnosis of MM is crucial for effective patient management.

Purpose of the Study:

• To develop and evaluate a machine learning framework for diagnosing multiple myeloma (MM) using multi-characteristic indexes.
• To enhance the performance of the kernel extreme learning machine (KELM) through an optimized slime mould Runge Kutta Optimizer (MSRUN).

Main Methods:

• A novel slime mould learning operator was integrated into the Runge Kutta Optimizer (RKO) to create MSRUN, improving search performance.
• MSRUN was employed for synchronized parameter optimization and feature selection within the MSRUN-KELM framework for MM diagnosis.
• The MSRUN algorithm was validated using IEEE CEC2014 benchmark functions.

Main Results:

• The MSRUN algorithm demonstrated significantly improved search performance compared to standard RKO.
• The MSRUN-KELM framework achieved a diagnostic accuracy of 93.88% for multiple myeloma.
• Key performance metrics included a Matthews correlation coefficient of 0.922677 and sensitivities of 93.41% and 93.19%.

Conclusions:

• The MSRUN-KELM framework is an effective tool for analyzing multi-characteristic indexes for multiple myeloma diagnosis.
• This approach shows promise as a potential diagnostic tool for multiple myeloma.