Using machine learning techniques to develop forecasting algorithms for postoperative complications: Protocol for a retrospective study

Keyphrases

• Postoperative Complications 100%
• Machine Learning Techniques 100%
• Forecasting Algorithm 100%
• Postoperative Outcomes 66%
• Tertiary Care Hospital 33%
• Adult Patients 33%
• Patient Outcomes 33%
• Density-based 33%
• Logistic Regression 33%
• Randomized Controlled Trial 33%
• In-hospital Mortality 33%
• Laboratory Values 33%
• University of Washington 33%
• Public Health Risk 33%
• Patient Risk 33%
• St. Louis 33%
• Population Level 33%
• Risk Profile 33%
• Healthcare 33%
• Postoperative Respiratory Failure 33%
• Patient History 33%
• Time Series Data 33%
• Ethics 33%
• Electronic Health Records 33%
• Care Plan 33%
• Kernel Density Estimation 33%
• Successful Development 33%
• Patient Journey 33%
• Vital Signs 33%
• Human Research Protections 33%
• Convolutional Neural Network 33%
• Feature Extraction 33%
• Postoperative Acute Kidney Injury 33%
• Outcome Trajectories 33%
• Perioperative Time 33%
• Second-order Features 33%
• Dynamic Risk 33%
• Nadaraya-Watson 33%
• First-order Features 33%
• Traditional Prediction Models 33%
• Postoperative Adverse Events 33%
• Watson Kernel 33%
• Shapelet Method 33%

Medicine and Dentistry

• Clinician 100%
• Retrospective Study 100%
• Postoperative Complication 100%
• Time Series Analysis 100%
• Electronic Patient Record 50%
• Tertiary Care 50%
• Adverse Outcome 50%
• Randomized Controlled Trial 50%
• Logistic Regression Analysis 50%
• Hospital Mortality 50%
• Respiratory Failure 50%
• Acute Kidney Injury 50%
• Public Health 50%
• Morbidity 50%
• Vital Sign 50%
• Feature Extraction 50%
• Prediction Model 50%
• Medical History 50%
• Convolutional Neural Network 50%