Houshang Darabi’s research while affiliated with University of Illinois Chicago and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (110)


Figure 1. Architecture of Neural Network Model with Input Layer Variables.
Figure 2. Summary of the results for the baseline model and proposed models.
Figure 3. List of important variables in descending order for model 5, model 18, and model 27 of the training set.
Figure 4. SHAP value impact for the model 1 HOUR ALL-Model 5.
Figure 5. SHAP value impact for the model which includes the base model and biomarkers at 1 h, 6 h, and 24 h-Model 18.

+3

Using a Neural Network Architecture for the Prediction of Neurologic Outcome for Out-of-Hospital Cardiac Arrests Using Hospital Level Variables and Novel Physiologic Markers
  • Article
  • Full-text available

January 2025

·

1 Read

Bioengineering

Martha Razo

·

Pavitra Kotini

·

·

[...]

·

Houshang Darabi

Out-of-hospital cardiac arrest (OHCA) is a major public health burden due to its high mortality rate, sudden nature, and long-term impact on survivors. Consequently, there is a crucial need to create prediction models to better understand patient trajectories and assist clinicians and families in making informed decisions. We studied 107 adult OHCA patients admitted at an academic Emergency Department (ED) from 2018–2023. Blood samples and ocular ultrasounds were acquired at 1, 6, and 24 h after return of spontaneous circulation (ROSC). Six classes of clinical and novel variables were used: (1) Vital signs after ROSC, (2) pre-hospital and ED data, (3) hospital admission data, (4) ocular ultrasound parameters, (5) plasma protein biomarkers and (6) sex steroid hormones. A base model was built using 1 h variables in classes 1–3, reasoning these are available in most EDs. Extending from the base model, we evaluated 26 distinct neural network models for prediction of neurological outcome by the cerebral performance category (CPC) score. The top-performing model consisted of all variables at 1 h resulting in an AUROC score of 0.946. We determined a parsimonious set of variables that optimally predicts CPC score. Our research emphasizes the added value of incorporating ocular ultrasound, plasma biomarkers, sex hormones in the development of more robust predictive models for neurological outcome after OHCA.

Download

Deep-Learning Model for Mortality Prediction of ICU Patients with Paralytic Ileus

November 2024

·

7 Reads

Bioengineering

Paralytic Ileus (PI) patients in the Intensive Care Unit (ICU) face a significant risk of death. Current predictive models for PI are often complex and rely on many variables, resulting in unreliable outcomes for such a serious health condition. Predicting mortality in ICU patients with PI is particularly challenging due to the vast amount of data and numerous features involved. To address this issue, a deep-learning predictive framework was developed using the Medical Information Mart for Intensive Care IV (MIMIC-IV) dataset, which includes data from 1017 ICU patients with PI. By employing SHAP (SHapley Additive exPlanations) analysis, we were able to narrow down the features to six distinct clinical lab items. The proposed framework, called DLMP (Deep Learning Model for Mortality Prediction of ICU Patients with PI), utilizes these six unique clinical lab items: Anion gap, Platelet, PTT, BUN, Total Bilirubin, and Bicarbonate, along with one demographic variable as inputs to a neural network consisting of only two neuron layers. DLMP achieved an outstanding prediction performance with an AUC score of 0.887, outperforming existing predictive models for ICU patients with PI. The DLMP framework significantly enhances the prediction of mortality for PI patients compared to traditional process mining and machine learning models. This model holds considerable potential for prognosis, enabling families to be better informed about the severity of a patient’s condition and to prepare accordingly. Furthermore, the model is valuable for research purposes and clinical trials.



Abstract 346: Deep Learning Approach for Predicting Good Neurological Outcome for Out-of-Hospital Cardiac Arrest Patients

November 2023

·

2 Reads

Circulation

Background: Cardiac arrest continues to pose a significant public health burden, with over 350,000 cases of out-of-hospital cardiac arrests (OHCA) occurring each year in the United States, and nearly 90% of them being fatal. The objective of this work was to develop a deep learning model that can accurately predict the Cerebral Performance Category (CPC) in adult OHCA cases. Methods: Adult OHCA cases at an urban academic emergency department were enrolled between 2018-2023. We evaluated data collected post return of spontaneous circulation (ROSC) at the first hour. Six different groups of data were tested for the prediction of CPC score: (1) Post-Rosc Vitals, (2) Pre-hospital & ED data points, and (3) Hospital Admission data points. The second group of exploratory research variables include: (4) ultrasound variables, (5) biomarkers and (6) Sex steroid hormones. The figure below illustrates the prediction framework for cardiac arrest patients, post ROSC, admitted to the hospital. Results: Of the total of 109 cases that were enrolled, 45% were female and 48% were Black. More than one-third (35%) were discharged alive but only 20% had a CPC of 1-2. While the base model with clinical and demographic variables had an AUC of 0.54, addition of subsequent variables improved the AUC substantially. The AUC improved to 0.59, 0.67 and 0.78 when hormones, ultrasound and biomarker variables were added, respectively, to the base model. However, the optimal model included all variables, resulting in a noteworthy AUC score of 0.861. Conclusion: Our findings emphasize the significance of incorporating novel variables to comprehensively evaluate the outcomes of cardiac arrest patients so that better prediction models can be developed, potentially aiding in modifying procedures and any necessary measures to shift the outcome in favor of preserving patients' lives after OHCA.


Reporting the Progress and Performance Evaluation of an Ongoing Integrated Program for Recruitment, Retention, and Graduation of High-Achieving, Low-income Engineering Students Reporting the Progress and Performance Evaluation of an Ongoing Integrated Program for Recruitment, Retention, and Graduation of High- Achieving, Low-income Engineering Students

Abstract The present paper reports an update on an NSF-funded S-STEM program currently in its last year at the University of Illinois Chicago. Lessons learned during the project implementation are also listed in the paper. A summary of the paper materials will be presented at the ASEE 2023 Annual Conference and Exposition as part of the NSF Grantees Poster Session. The project's objectives are 1) enhancing students' learning by providing access to extra and co-curricular experiences, 2) creating a positive student experience through mentorship, and 3) ensuring successful student placement in the STEM workforce or graduate/professional degree program. As part of this project, students are provided with financial assistance. A total of three Cohorts of students are supported by the project: Engineering students who started as freshmen, including 18 students of Cohort I and 13 students of Cohort II, and 19 students who transferred from various community colleges to Cohort III. More than 60% of the students are classified as minorities. This project has resulted in the creation of several support and intervention programs, including a Summer Bridge Program, an Engineering Success Initiative course, a Service Learning Project course, and an integrated mentoring program that matches each student with an academic mentor (a faculty) and an industry mentor. The paper will summarize the lessons learned from the support programs. Out of the 18 students recruited by this program as Cohort I, all have already graduated, and 16 have started a job. Cohort II students will graduate in Spring 2023, and Cohort III students will graduate in Spring or Fall of 2023. Two students dropped out of the University in their first year, and three dropped out of the University in their second year. More information is provided in this paper regarding student retention and performance (Grade Point Average).




Module presentation details after preprocessing.
Mapping activity types to learning style features based on FSLSM.
Features categories in each quarter.
Results of machine learning models-State I.
Statistical comparisons of assessment grades between Supported and Not Supported categories.
Personalized Learning in Virtual Learning Environments Using Students’ Behavior Analysis

April 2023

·

256 Reads

·

15 Citations

Education Sciences

In recent years, many research studies have focused on personalized e-learning. One of the most crucial parts of any learning environment is having a learning style that focuses on individual learning. In this paper, we propose an approach to personalizing learning resources based on students’ learning styles in a virtual learning environment to enhance their academic performance. Students’ interactions with the learning management system are utilized to analyze learners’ behaviors. The Felder–Silverman Learning Style Model (FSLSM) is used to map students’ interactions with online learning resources to learning style (LS) features. The learning style and demographic features are then utilized for training machine learning models to predict students’ academic performance in each quarter of courses. The most accurate prediction model for each quarter is then used to find learning style features that maximize students’ pass rates. We statistically prove that students whose actual learning style features were close enough to the ones calculated by the approach achieved better grades. To improve students’ academic performance each quarter, we suggest two strategies based on the learning style features calculated by the process.


AXDP Performance compared to baseline models
NN Architecture for Datasets.
Adjacency Matrix Deep Learning Prediction Model for Prognosis of the Next Event in a Process

January 2023

·

103 Reads

IEEE Access

Prediction of the next event is important for organizations to improve and optimize their system process to achieve organizational goals. Existing predictive models are limited since they use discovery algorithms that might not be able to conserve the sequences of events as reported in the event log. Discovery algorithms alter the sequence of events in two ways, either the algorithms generate additional sequence of events not found in the event logs or remove the order of events. Since prediction relies on these process algorithms, the prediction model can suffer and produce underperforming results. Models that do not use discovery algorithms, such as deep learning models, ignore completely the sequence of events. In order to overcome these limitations, we propose a new algorithm called AXDP (Adjacency Matrix Deep Learning Prediction Model). AXDP predicts the next event of a process using graph theory techniques, specifically adjacency matrices and predicts using the power of deep learning models. AXDP has a major advantage, that sequence of events is conserved, resulting in better prediction of the next event. When testing AXDP on eight publicly available datasets, AXDP outperforms what we believe to be the most recent and best predictive models that exist for prediction of the next event for six of the eight datasets.


Looking Ahead: Structure of an Industry Mentorship Program for Undergraduate Engineering Students

September 2022

·

60 Reads

·

3 Citations

AEE Journal

The matriculation, retention, and graduation of students is a critical and ongoing effort in undergraduate engineering education. Implementing measures is vital in increasing student motivation to continue pursuing engineering and is especially important for individuals from low-income and academically talented students. At the University of Illinois at Chicago (UIC), our S-STEM program includes students from all six departments in the College of Engineering and includes multiple implementations such as a summer bridge program, faculty mentoring, and a service-learning project. Here, we define and implement a longitudinal Industry Mentorship (IM) program structure. While the implementation of this program is still ongoing, we present here the basic structure of the IM program, the rationale for the structure, and some preliminary results of implementation.


Citations (61)


... LSTM (Long Short-Term Memory) based neural networks are widely used for time series forecasting [17,18]. To increase the potential for identifying hidden representations, it was decided to use two LSTM cells instead of one. ...

Reference:

Methodology of Data Popularity Forecasting in High-Energy Physics Experiments on Unbalanced and Irregular Time-series Data
Multivariate LSTM-FCNs for Time Series Classification
  • Citing Preprint
  • January 2018

... The progress of the project has been disseminated through three poster presentations [3] [6] [7]. In addition, engineering identity focused interviews with Cohort I Scholars have been conducted and the results have been published [8]. ...

Reporting the Progress and Latest Status of an Ongoing S-STEM Project

... The progress of the project has been disseminated through four poster presentations [3] [5] [6] [7]. In addition, engineering identity-focused interviews with cohort I scholars have been conducted, and the results have been published [8]. The execution details and assessment results of the Summer Bridge Program were published at an educational conference [9]. ...

Low-Income, High-Achieving Students and Their Engineering Identity Development After One Year of Engineering School
  • Citing Conference Paper
  • July 2021

... These models not only predict academic outcomes but also help in identifying factors that influence students' learning achievements in Massive Open Online Courses (MOOCs) [25]. Moreover, the FSLSM has been utilized to personalize learning in virtual learning environments by analyzing students' behaviors and mapping them to learning style features [26]. Such personalized approaches enhance the effectiveness of online learning by tailoring educational experiences to individual students' needs. ...

Personalized Learning in Virtual Learning Environments Using Students’ Behavior Analysis

Education Sciences

... These partnerships aim to provide meaningful career guidance and connections that complement students' academic learning. Research shows professional mentorship enables crucial real-world perspective and experiential learning opportunities for mentees [8][9][10][11]. ...

Looking Ahead: Structure of an Industry Mentorship Program for Undergraduate Engineering Students

AEE Journal

... Decay replay mining has been successfully applied to various problems [18], including healthcare [10], [19], [20], [21]. While this method utilizes outcomes from process discovery, the neural network is yet not fully connected with the Petri net to its full extent [22]. This article is an extended work that has been originally presented at the 2021 International Conference on Cyber-physical Social Intelligence [22]. ...

Masking Neural Networks Using Reachability Graphs to Predict Process Events
  • Citing Conference Paper
  • December 2021

... Deep learning with Neural Networks (NNs) have shown an extraordinary performance in prediction across research in many industries including the medical field. Several significant disorders such as diabetes [1], coronavirus disease 2019 (COVID-19) [2], paralytic ileus (PI) [3] and heart failure (HF) [4] have been analyzed using deep learning models to predict outcomes. Yet, there remains a gap for one of the most lethal public health problems in the United States, out-of-hospital cardiac arrest (OHCA). ...

A process mining- deep learning approach to predict survival in a cohort of hospitalized COVID‐19 patients

BMC Medical Informatics and Decision Making

... While several CVD-related risk factor analyses are based on modelspecific variable importance, these metrics can make it difficult to understand positive/negative trait contributions to the target disease and individual sample contributions of input traits to the target disease (Dong et al. 2023;Hu, Liu, and Li 2020;Ji et al. 2020). Few studies identified risk factors for spatiotemporal CVD patterns using the XAI, but none addressed important categories of factors such as sociodemographic factors, air pollutants, and meteorological variables (Harford et al. 2022;Nakashima et al. 2021;Shimada-Sammori et al. 2023). Therefore, we applied the proposed XAI-based risk factor framework to understand the spatiotemporal association between CVD and various risk factors in South Korea from 2010 to 2019. ...

Utilizing Community Level Factors to Improve Prediction of Out of Hospital Cardiac Arrest Outcome using Machine Learning
  • Citing Article
  • July 2022

Resuscitation

... Pishgar et al. 28 presented a process mining and DL method to forecast unscheduled 30-day readmissions from heart failure patients in the intensive care unit. They used event logs and the Decay Replay Mining (DREAM) algorithm to extract temporal data. ...

Prediction of unplanned 30-day readmission for ICU patients with heart failure

BMC Medical Informatics and Decision Making