POST-PANDEMIC ANALYSIS OF HOUSE PRICE PREDICTION IN SURABAYA: A MACHINE LEARNING APPROACH

The COVID-19 outbreak caused a slowdown in the Indonesian economy, as it did in many other impacted nations. Consequently, the housing market in Indonesia, along with other industries, deteriorated. Other post-pandemic issues displace the property industry's priorities in Indonesia. Determining a fair property price is a problem occurring because of the economic slowdown. Property sellers expected their property selling prices to be the same before the pandemic or even increase, but property agents hoped the properties would be selling fast, creating a sense of distrust between the seller and the property agents. This work aims to develop a machine learning-based prediction model for real estate agents to use in determining property prices, with the expectation that the resulting predictions will be more accurate and supported by the data, increasing seller and buyer confidence. Following the suggestion from previous studies, several supervised algorithms such as Linear Regression, Decision Tree, and Random Forest were used to develop the model. Training data were collected from five property agents in Surabaya and as well as web scraping from the online home sales portals. Findings from the study show that Random Forest performs best in predicting with the highest coefficient of determination and lowest error. Using evaluation measures such as Mean Absolute Percent Error (MAPE), the error was calculated to be 23%, which is acceptable for prediction. © 2022 Science Press. All rights reserved.

Modernizing Pressure Injury Risk Assessment in the ICU in the COVID Era: Ensemble Super-Learning and Explainable AI

Objectives: Hospital-acquired pressure injuries (HAPrI) are areas of injury to the skin and/or underlying tissues. Risk stratification is essential for guiding prevention in the ICU, but current risk assessment tools require labor-intensive input. This motivates a tactical, parsimonious, and automatic risk profiling algorithm, that can be based on readily available clinical measures (e.g., COVID status, race, Medicare/Medicaid status). Additionally, International Pressure Injury Prevention guidelines call for the development of machine learning-based risk assessment algorithms that are clinician-interpretable and context-informed. Method(s): Adult patients admitted to one of two ICUs between April 2020, and April 2021 were eligible for inclusion. Discrete and ensemble super-learning models, adjusting for class imbalance, were created from a rich library of candidate base learners. For explainability, SHAP (SHapley Additive exPlanations) global and local values were derived to help explain variable average marginal contributions (across all permutations) to the model. An iteration of clinical expert review was performed with the SHAP values, and simulations of patient profiles and results were used to reformat and re-weight predictor variables. All analysis was run in open Python (version 3.7), and code/results will be made available via a GitHub page. Result(s): The final sample consisted of 1,911 patients (removing 9 with missing pressure injury status). Hospital-acquired pressure injuries (defined as stage 2, or worse) occurred in 18.5% of the sample (n=354). We achieved the best overall performance on the testing data with a stacked ensemble using three base models: random forest (rf), gradient boosted machine (gbm), and neural network (NN) (Performance on 20% holdout: Accuracy: 81%;AUC: 0.77;AUCPR: 0.53). Conclusion(s): Prediction engineering should be done in collaboration with clinical experts to optimize tactical implementation to both optimize performance, with minimal interruption to workflow. XAI enhanced adoption of the experts' advice based on the selected model features. Copyright © 2022

Changes in Practice/Outcomes of Pediatric/Congenital Catheterization in Response to the First Wave of COVID.

Background: The Coronavirus disease 2019 (COVID-19) pandemic has posed tremendous stress on the health care system. Its effects on pediatric/congenital catheterization program practice and performance have not been described. Objectives: To evaluate how case volumes, risk-profile, and outcomes of pediatric/congenital catheterization procedures changed in response to the first wave of COVID-19 and after that wave. Methods: A multicenter retrospective observational study was performed using Congenital Cardiac Catheterization Project on Outcomes Registry (C3PO) data to study changes in volume, case mix, and outcomes (high-severity adverse events [HSAEs]) during the first wave of COVID (March 1, 2020, to May 31, 2020) in comparison to the period prior to (January 1, 2019, to February 28, 2020) and after (June 1, 2020, to December 31, 2020) the first wave. Multivariable analyses adjusting for case type, hemodynamic vulnerability, and age group were performed. Hospital responses to the first wave were captured with an electronic study instrument. Results: During the study period, 12,557 cases were performed at 14 C3PO hospitals (with 8% performed during the first wave of COVID and 32% in the postperiod). Center case volumes decreased from a median 32.1 cases/mo (interquartile range: 20.7-49.0) before COVID to 22 cases/mo (interquartile range: 13-31) during the first wave (P = 0.001). The proportion of cases with risk factors for HSAE increased during the first wave, specifically proportions of infants and neonates (P < 0.001) and subjects with renal insufficiency (P = 0.02), recent cardiac surgery (P < 0.001), and a higher hemodynamic vulnerability score (P = 0.02). The observed HSAE risk did not change significantly (P = 0.13). In multivariable analyses, odds of HSAE during the first wave of COVID (odds ratio: 0.75) appeared to be lower than that before COVID, but the difference was not significant (P = 0.09). Conclusion: Despite increased case-mix complexity, C3PO programs maintained, if not improved, their performance in terms of HSAE. Exploratory analyses of practice changes may inform future harm-reduction efforts.