A novel quantum-inspired binary bat algorithm for leukocytes classification in blood smear

In the field of medical sciences, day-to-day procedure is followed for identification of bone marrow and immune system related diseases, which is most of the time carried out manually. The notion is to perform differential and qualitative analysis of leukocytes for the timely diagnosis of these diseases. In this article, a systematized solution is offered for the classification of leukocytes in blood smear. The proposed model incorporates the optimistic aspects of nature-inspired and quantum-inspired algorithms; this model tends to be perfect blend of both the techniques. For reducing the dimensionality, that is, irrelevant features; the quantum-inspired binary bat algorithm (QBBA) has been used in the proposed model. The optimality of features selected has been computed with the help of accuracy measure using various machine learning classifiers like Logistic Regression, KNN, Random Forest, Decision Tree. The performance of QBBA and its customary algorithms has been compared and the results depict that QBBA outperforms binary bat algorithm for the same set of population. QBBA comes out as an influential algorithm with an average accuracy of 98.31% and also possess enhanced noise invulnerability. The proposed QBBA can also find its usage in thorough haematological analysis.

Automated summary

This article proposes a systematic solution for classifying leukocytes in blood smears, combining the advantages of nature-inspired and quantum-inspired algorithms, with the quantum-inspired binary bat algorithm showing significant effectiveness in feature selection. The research findings demonstrate that QBBA outperforms traditional algorithms in the same population, achieving an average accuracy of 98.31% and enhanced noise resistance capabilities.