Abstract:
Inflammatory mediators play an important function in a variety of disorders. IL-13 tightly
regulates immune responses, notably those associated with allergies and inflammatory
reactions. Predicting interleukin-13 (IL-13) activity is critical since it can be used to
identify those who are more likely to develop IL-13-driven illnesses such as asthma and
atopic dermatitis. Because the major goal of this study is to construct an accurate
Interleukin-13 prediction model utilizing ensemble machine learning methods, we
present an enhanced prediction of IL-13-inducing peptides here. The positive and
negative datasets were collected from a recent study (IL13Pred), and feature extraction
was performed using the ILearnplus package. We used the Best Peptide Sequence
Extractor and reported the results of various techniques individually. The data collection
was unbalanced, therefore we used the Adasyn Algorithm to balance it. For feature
selection, modern feature engineering approaches (Recursive Shapley Value) were
used.The results show that when utilizing our StackingClassifier, specific feature sets as
CKSAAP, DPC, CTDC, and CTraid may be used to accurately classify data.Our StackingClassifier has improved in terms of accuracy, AUC, and MCC value. Machine
learning techniques for interleukin-13 prediction contribute to a better knowledge of IL-13
and its possible consequences in healthcare. It improves the accuracy and reliability of
Interleukin-13 prediction, allowing for more informed medical decisions to be made for
better patient care and treatment outcomes. This study's successful conclusion
increases our understanding of IL-13 prediction while also highlighting the potential of
machine learning technologies in addressing complicated biological challenges.
