A Comparative Analysis of Machine Learning Algorithms for Allergy Scale Prediction

In this comprehensive machine learning study, we investigated the efficacy of various algorithms in predicting dust mite allergy severity using data collected from the Alex-X® detection devices. There were a total of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 1 0}$</tex> participants. The research examined seven distinct machine learning algorithms to determine their predictive capabilities in dust mite allergy assessment. In comparative algorithmic analysis, the experimental results demonstrated hierarchical performance variations across multiple machine learning models. The Random Forest algorithm exhibited superior predictive accuracy at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 0. 1 5 \%}$</tex>, establishing its prominence in the classification task. Neural Network implementations achieved the second-highest accuracy at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 8. 9 3 \%}$</tex>, followed by K-Nearest Neighbors (KNN) with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 5. 6 8 \%}$</tex> accuracy. Linear Regression demonstrated moderate performance at 84.52 %, while the Decision Tree algorithm achieved 83.15% accuracy. Stochastic Gradient Descent (SGD) and Naïve Bayes algorithms showed relatively lower performance metrics at <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{8 2. 4 3 \%}$</tex> and 79.84% accuracy, respectively.