Modified mutual information feature selection algorithm to predict COVID-19 using clinical data

Clinical Snapshot

Key Findings

• P Value: Not reported

• Effect Size: 95% overall accuracy

• Primary Outcome: COVID-19 prediction accuracy using MMI-SVM algorithm

• Nnt Or Sensitivity: Sensitivity and specificity not reported - only overall accuracy of 95%

• Confidence Interval: Not reported

Clinical Application

Unclear - depends on specific blood parameters required, computational resources, and integration with existing laboratory systems Would need TGA approval as diagnostic software. Integration with existing pathology services and Medicare rebate structure unclear. RACGP guidelines would need updating to incorporate such tools into diagnostic algorithms. Patients requiring COVID-19 diagnostic testing with available blood test results

Abstract

The COVID-19 pandemic has profoundly impacted health, emphasizing the need for timely disease detection. Blood tests have become key diagnostic tools due to the virus's effects on blood composition. Accurate COVID-19 prediction through machine learning requires selecting relevant features, as irrelevant features can lower classification accuracy. This study proposes Modified Mutual Information (MMI) for feature selection, ranking features by relevance and using backtracking to find the optimal subset. Support Vector Machines (SVM) are then used for classification. Results show that MMI with SVM achieves 95% accuracy, outperforming other methods, and demonstrates strong generalizability on various benchmark datasets.