Machine learning applications for COVID-19 outbreak management.

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Epidemiology

Background:

• The COVID-19 pandemic necessitated rapid advancements in healthcare, with machine learning (ML) emerging as a key technology.
• Medical imaging (CT, X-ray) and ML techniques provide powerful tools for pandemic response and patient management.

Purpose of the Study:

• To conduct a systematic literature review (SLR) on the diverse applications of ML in combating the COVID-19 pandemic.
• To identify and categorize the primary uses, methodologies, and challenges of ML in pandemic-related medical research.

Main Methods:

• Systematic literature review (SLR) methodology was employed to analyze existing research.
• Categorization of ML applications into seven key areas: imaging, survival analysis, forecasting, economic/geographical issues, monitoring, drug development, and hybrid applications.
• Identification of frequently used ML algorithms (CNNs, LSTMs, RNNs, GANs, etc.) and software libraries (Keras).

Main Results:

• ML applications in COVID-19 span diagnostics, patient monitoring, drug discovery, and forecasting.
• Conventional neural networks (CNNs) and Long Short-Term Memory (LSTM) networks are prominent ML techniques.
• Medical imaging is utilized in 20.4% of applications, with Keras being the most common library (24.4%).
• Evaluation often focuses on flexibility and accuracy, with safety aspects frequently overlooked.

Conclusions:

• ML offers significant potential for improving COVID-19 detection, management, and future pandemic preparedness.
• Further research is needed to address overlooked aspects like safety and explore novel hybrid ML applications.
• Standardization of evaluation metrics and a focus on safety are crucial for advancing ML in pandemic response.