Comparative analysis of machine learning algorithms to improve the diagnosis of African swine fever

Abstract

This study focused on improving the diagnosis of African swine fever (ASF) by improving accuracy, reliability and precision using various machine learning algorithms. Data from the European Union Reference Laboratory for ASF and the EU Animal Disease Information System (ADIS) containing clinical information on wild and domestic pigs were processed and converted into numerical and categorical formats for analysis. Various machine learning models were tested as part of the study, including linear regression, Bayesian regression, support vector machines (SVM), decision trees, random forest classifiers, artificial neural networks (ANNs) and logistic regression. The models were evaluated using metrics such as accuracy, precision, recall, F1-score and root mean square error (RMSE). The results showed that the SVM achieved the lowest performance with 44.4% accuracy, followed by the logistic regression model with 61.5%. The neural network model achieved 69% accuracy, closely followed by the decision tree model with 70%. The linear regression model performed slightly better with 74% accuracy. Bayesian regression achieved a higher accuracy of 80%, while the Random Forest model outperformed all others, achieving the highest accuracy of 88.9%. These results make an important contribution to economic efficiency and innovative applications in veterinary medicine and improve disease management through advanced machine learning techniques.

Keywords: {'name': 'African swine fever'}, {'name': 'Causal Inference'}, {'name': 'Ecological context of ASF'}, {'name': 'Machine Learning'}

Authors & affiliations

Steven Lububu
Affiliation not supplied

Michael Twum-Darko
Affiliation not supplied

Deposited references

Supplied references, not a list of publications citing this article.

Agbo, S. O., Samuel, S. E., & Idoko, D. J. (2022). Expert system-based machine learning analysis for diagnosis and treatment of swine diseases using decision tree, logistic regression and support vector machine algorithms. Retrieved from https://www.ftstjournal.com/uploads/docs/73%20Article%2049%20pp%20364-374.pdf.

DOI: 10.1371/journal.pone.0192565

DOI: 10.1111/tbed.12695

DOI: 10.3390/ani11092692

DOI: 10.3390/pathogens10010084

DOI: 10.1038/s42256-022-00445-z

DOI: 10.1186/s40813-018-0109-2

DOI: 10.1111/tbed.12129

DOI: 10.1111/tbed.12381

DOI: 10.1080/22221751.2022.2146537

DOI: 10.1111/j.1865-1682.2012.01317.x

DOI: 10.1016/j.virusres.2019.197676

DOI: 10.1186/s40813-015-0013-y

DOI: 10.1136/vr.103593

DOI: 10.3390/v11090864

DOI: 10.3390/v13112285

DOI: 10.1038/s41467-020-17419-7

DOI: 10.1016/j.jcpa.2014.09.003

DOI: 10.1098/rsos.220638

DOI: 10.1016/j.vetmic.2012.11.030

DOI: 10.15623/ijret.2015.0406007

DOI: 10.1016/j.tvjl.2017.12.025

DOI: 10.3389/fvets.2019.00486

DOI: 10.1128/JCM.43.1.112-119.2005