Machine Learning in Predicting Hemoglobin Variants

Abstract

Disease diagnosis is of the utmost importance in providing appropriate medical treatment. Genetic diseases, such as hemoglobinopathies and thalassemia, need to be diagnosed accurately and on time. Though Hb variants are diagnosed using a HPLC-based hemoglobin typing machine. appropriate interpretation of the data obtained is still necessary and this requires trained professionals. Machine learning helps to interpret the obtained data and in predicting the type of Hb variants, thus reducing the workload of health professionals. In this study, the obtained data are classified using the following classifiers, namely logistic regression, support vector classifier (SVC), k-nearest neighbor (KNN), Gaussian naï ve bayes, perceptron classifier, linear SVC, stochastic gradient descent, decision tree, random forest, and multi-layer perceptron. The pre-processing, visualization and the classification steps were implemented using Python 2.7 on an Intel Core i5 computer. The performance of each classifier was then tested by initially creating a confusion matrix. Indices including "precision," "recall," and "f1-score" were used to quantify the quality of each model. KNN, decision tree, and random forest show better classification results in comparison to the other classifiers. With a precision of 93.89%, recall of 92.78%, and f1-score of 93.33%, the decision tree and random forest classifiers prove to be better classifiers in predicting the Hb variants with a higher accuracy rate.

Comparing different supervised machine learning algorithms for disease prediction

“…For these 49 diseases, 50 algorithms were found to show the superior accuracy. One disease has two algorithms (out of 5) that showed the same higher-level accuracies [36]. To sum up, 49 diseases were predicted in 48 articles considered in this study and 50 supervised machine learning algorithms were found to show the superior accuracy.…”

“…For each of the remaining 48 articles, the performance outcomes of the supervised machine learning algorithms that were used for disease prediction were gathered. Two diseases were predicted in one article [17] and two algorithms were found showing the best accuracy outcomes for a disease in one article [36]. In that article, five different algorithms were used for prediction analysis.…”

A comparative analysis of haemoglobin variants using machine learning algorithms

“…However, these results give an accurate and deep understanding of the performance of the models. It can be noted that the overall performance of the models depends on the size of the dataset and the distinction of features [13].…”

“…As surmised by Borah et al [13], a large dataset will help in improving the accuracy of a model. Therefore, the use of machine learning approach in blood laboratory-based diagnosis could lead to a fundamental change in differential diagnosis and result in the modification of the currently accepted guidelines.…”

Applications of Artificial Intelligence in Thalassemia: A Comprehensive Review

“…For instance, the decision tree and random forest classifiers were proven to be the main methods in terms of classifying the Hb variants. Moreover, studying a large number of samples may help in generating greater precision and F1-score values, which imply greater accuracy [49]. Moreover, two predictive methods were established, one method for the detection of Bthalassemia trait (BTT), and the other for the identification of hemoglobin E (HbE) trait and BTT using the decision tree, Naïve Bayesian classifier, and artificial neural network.…”