Predictive Modelling for determining patients’ length of stay
Predictive Modelling for determining patients’ length of stay
Tools Used: Python, MariaDB, SQL, Docker
Period: Aug 2022 - Dec 2022
The relevance of optimizing resources at hospitals saw new heights during the covid pandemic as hospitals were faced with shortages of hospital beds
A model was generated to predict the patient’s length of stay in the hospital based on their health condition
Performed skew analysis on the predictors to eliminate outliers by assuming a cutoff with minimal data loss
Trained and tested various machine learning algorithms by hyperparameter tuning on data to identify the best-performing model
The image shown is the ROC curves of the best-performing model, to know more about the project refer to the project report in the git link provided.