Bell-Shaped Fuzzy Least Square Twin SVM With Biomedical Applications

In practical applications, datasets frequently encompass noise, outliers, and imbalanced classes, which can markedly affect a model's generalization performance. Support vector machine (SVM) and its twin variant i.e., TWSVM tend to be biased towards the majority class samples, leading to misclassification of the minority class samples. TWSVM suffers from this biasness as it generates hyperplanes without considering preceding data information. To address the aforementioned issues, we propose bell-shaped fuzzy least square twin support vector machine for imbalance data (BSFLSTSVM-ID). The proposed BSFLSTSVM-ID allocates weight to the majority class samples through a novel membership function, namely, “class probability and bell-shaped” (CPBS). The CPBS function is amalgamation of class probability, bell-shaped function, and imbalance ratio of the dataset. The bell-shaped function's value diminishes as data points move farther away from the class center, reducing the influence of noise or outliers in constructing hyperplanes. To underscore the importance of samples from minority class, a weight of one is assigned to them. Furthermore, the proposed BSFLSTSVM-ID utilizes the conjugate gradient method to handle the challenge of matrix inversion. To demonstrate its effectiveness, we conducted experiments on 59 UCI and KEEL datasets with imbalance ratios from 1 to 72.69. Additionally, we tested the proposed BSFLSTSVM-ID model's scalability on NDC datasets and applied it to diagnosing breast cancer and Alzheimer's disease using the BreakHis and Alzheimer's Disease Neuroimaging Initiative (ADNI) datasets, respectively. The results show that the proposed BSFLSTSVM-ID outperforms baseline models, highlighting its potential for tackling classification challenges in the biomedical domain.