Feature Selection Methods for ICU Mortality Prediction Model
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
The goal of this research is to offer insightful information that can improve Ethiopia's intensive care unit (ICU) services. There is an increased risk of patients' death in Intensive Care Units (ICUs). This is because of several variables, including preexisting medical issues, lack of resources, and delayed decisions. Healthcare professionals can better prioritize their patients in need of intensive care, distribute resources more efficiently, and enhance patient outcomes by using predictive models to estimate ICU mortality. ICU data is collected from five Ethiopian public hospitals to develop a machine learning method for predicting ICU mortality. The data includes demographic features, vital signs, lab results, and discharge status of 10,798 ICU dataset records. We employed a range of feature selection techniques, such as filters, wrappers, and embedding methods, to identify the most crucial features for mortality prediction. We also compared the performance of two machine learning algorithms, Random Forest and Decision Tree. These models are trained using ICU data with features encompassing age, length of stay, temperature, neutrophil, Diagnosis (DX) condition, PH, and Lymphocite. These features are selected using Recursive Feature Elimination (RFE). Using a number of different evaluation methods, including accuracy (99.7%), precision (99.4%), recall (98.8%), F1 score (99.1%), and area under the curve (AUC) (99.3%), Random Forest performed better than Decision Tree. Based on our findings, we made recommendations for healthcare practitioners and policy makers. We also suggest key future research directions for researchers in the area.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.003 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it