abebual
I'm a researcher working on uncertainty quantification and reliability analysis for aerospace applications. AI and Statistics form the core of my work.
RTXhttps://www.rtx.com/
Pinned Repositories
Breast-Cancer-Classification-with-PyTorch-and-Deep-Learning
In this project we will build a classifier CNN model to detect ICD breast cancer in histopathological images.
Clinical-Deterioration-Prediction-Model---KNN
Clinical Deterioration Prediction Model - KNN
Clinical-Deterioration-Prediction-Model--Ensemble-Algorithms-
Clinical Deterioration Prediction Model - Bayesian Linear Regression
Clinical-Deterioration-Prediction-Model--Inferential-Statistics
Clinical Deterioration Prediction Model: Inferential Statistics
Clinical-Deterioration-Prediction-Model--Logistic-Regression
Clinical-Deterioration-Prediction-Model--Logistic-Regression
Computer-Vision-Project-Monreader
Computer-Vision-Project-Monreader
Predicting-ICU-Patient-Clinical-Deterioration---Report
For this project, I used publicly available Electronic Health Records (EHRs) datasets. The MIT Media Lab for Computational Physiology has developed MIMIC-IIIv1.4 dataset based on 46,520 patients who stayed in critical care units of the Beth Israel Deaconess Medical Center of Boston between 2001 and 2012. MIMIC-IIIv1.4 dataset is freely available to researchers across the world. A formal request should be made directly to www.mimic.physionet.org, to gain access to the data. There is a required course on human research ‘Data or Specimens Only Research’ prior to data access request. I have secured one here -www.citiprogram.org/verify/?kb6607b78-5821-4de5-8cad-daf929f7fbbf-33486907. We built flexible and better performing model using the same 17 variables used in the SAPS II severity prediction model. The question ‘Can we improve the prediction performance of widely used severity scores using a more flexible model?’ is the central question of our project. I used the exact 17 variables used to develop the SAPS II severity prediction algorithm. These are 13 physiological variables, three underlying (chronic) disease variables and one admission variable. The physiological variables includes demographic (age), vital (Glasgow Comma Scale, systolic blood pressure, Oxygenation, Renal, White blood cells count, serum bicarbonate level, blood sodium level, blood potassium level, and blood bilirubin level). The three underlying disease variables includes Acquired Immunodeficiency Syndrome (AIDS), metastatic cancer, and hematologic malignancy. Finally, whether admission was scheduled surgical or unscheduled surgical was included in the model. The dataset has 26 relational tables including patient’s hospital admission, callout information when patient was ready for discharge, caregiver information, electronic charted events including vital signs and any additional information relevant to patient care, patient demographic data, list of services the patient was admitted or transferred under, ICU stay types, diagnoses types, laboratory measurments, microbiology tests and sensitivity, prescription data and billing information. Although I have full access to the MIMIC-IIIv1.4 datasets, I can not share any part of the data publicly. If you are interested to learn more about the data, there is a MIMIC III Demo dataset based on 100 patients https://mimic.physionet.org/gettingstarted/demo/. If you are interested to requesting access to the data - https://mimic.physionet.org/gettingstarted/access/. Linked repositories: Exploratory-Data-Analysis-Clinical-Deterioration, Data-Wrangling-MIMICIII-Database, Clinical-Deterioration-Prediction-Model--Inferential-Statistics, Clinical-Deterioration-Prediction-Model--Ensemble-Algorithms-, Clinical-Deterioration-Prediction-Model--Logistic-Regression, Clinical-Deterioration-Prediction-Model---KNN © 2020 GitHub, Inc.
Term-Deposit-Marketing-Prediction-5MSttBBp1mlzn2hA
The data comes from direct marketing efforts of a European banking institution. The marketing campaign involves making a phone call to a customer, often multiple times to ensure a product subscription, in this case a term deposit. Term deposits are usually short-term deposits with maturities ranging from one month to a few years. The customer must understand when buying a term deposit that they can withdraw their funds only after the term ends. All customer information that might reveal personal information is removed due to privacy concerns.
Text-Processing-Naive-Bayes-in-Python
In the mini-project, you'll learn the basics of text analysis using a subset of movie reviews from the rotten tomatoes database. You'll also use a fundamental technique in Bayesian inference, called Naive Bayes. This mini-project is based on Lab 10 of Harvard's CS109 class. Please free to go to the original lab for additional exercises and solutions.
ValueInvestor-Predicting-TimeSeries-Data
Value Investor
abebual's Repositories
abebual/Predicting-ICU-Patient-Clinical-Deterioration---Report
For this project, I used publicly available Electronic Health Records (EHRs) datasets. The MIT Media Lab for Computational Physiology has developed MIMIC-IIIv1.4 dataset based on 46,520 patients who stayed in critical care units of the Beth Israel Deaconess Medical Center of Boston between 2001 and 2012. MIMIC-IIIv1.4 dataset is freely available to researchers across the world. A formal request should be made directly to www.mimic.physionet.org, to gain access to the data. There is a required course on human research ‘Data or Specimens Only Research’ prior to data access request. I have secured one here -www.citiprogram.org/verify/?kb6607b78-5821-4de5-8cad-daf929f7fbbf-33486907. We built flexible and better performing model using the same 17 variables used in the SAPS II severity prediction model. The question ‘Can we improve the prediction performance of widely used severity scores using a more flexible model?’ is the central question of our project. I used the exact 17 variables used to develop the SAPS II severity prediction algorithm. These are 13 physiological variables, three underlying (chronic) disease variables and one admission variable. The physiological variables includes demographic (age), vital (Glasgow Comma Scale, systolic blood pressure, Oxygenation, Renal, White blood cells count, serum bicarbonate level, blood sodium level, blood potassium level, and blood bilirubin level). The three underlying disease variables includes Acquired Immunodeficiency Syndrome (AIDS), metastatic cancer, and hematologic malignancy. Finally, whether admission was scheduled surgical or unscheduled surgical was included in the model. The dataset has 26 relational tables including patient’s hospital admission, callout information when patient was ready for discharge, caregiver information, electronic charted events including vital signs and any additional information relevant to patient care, patient demographic data, list of services the patient was admitted or transferred under, ICU stay types, diagnoses types, laboratory measurments, microbiology tests and sensitivity, prescription data and billing information. Although I have full access to the MIMIC-IIIv1.4 datasets, I can not share any part of the data publicly. If you are interested to learn more about the data, there is a MIMIC III Demo dataset based on 100 patients https://mimic.physionet.org/gettingstarted/demo/. If you are interested to requesting access to the data - https://mimic.physionet.org/gettingstarted/access/. Linked repositories: Exploratory-Data-Analysis-Clinical-Deterioration, Data-Wrangling-MIMICIII-Database, Clinical-Deterioration-Prediction-Model--Inferential-Statistics, Clinical-Deterioration-Prediction-Model--Ensemble-Algorithms-, Clinical-Deterioration-Prediction-Model--Logistic-Regression, Clinical-Deterioration-Prediction-Model---KNN © 2020 GitHub, Inc.
abebual/Clinical-Deterioration-Prediction-Model---KNN
Clinical Deterioration Prediction Model - KNN
abebual/Clinical-Deterioration-Prediction-Model--Ensemble-Algorithms-
Clinical Deterioration Prediction Model - Bayesian Linear Regression
abebual/Clinical-Deterioration-Prediction-Model--Inferential-Statistics
Clinical Deterioration Prediction Model: Inferential Statistics
abebual/Clinical-Deterioration-Prediction-Model--Logistic-Regression
Clinical-Deterioration-Prediction-Model--Logistic-Regression
abebual/LLMs-for-Flight-Safety
The project aims to enhance the handling of Federal Aviation Regulation (FAR) Title 14, Part 33 and Advisory Circular (AC), related to aircraft engine airworthiness standards by employing Deep Learning Models.
abebual/Timeliness-Analysis-Productive-Safety-Net-Programme-PSNP-in-Ethiopia
The Productive Safety Net Programme (PSNP) is Ethiopia’s rural safety net designed to support poor food insecure rural households through the provision of timely and predictable benefits. The PSNP was launched in 2005 and currently provides support to approximately 2 million eligible households (8 million beneficiaries) in: Afar, Amhara, Dire Dawa, Harari, Oromia, Southern Nations, Nationalities and Peoples (SNNP), Somali and Tigray. Households that have able-bodied adult labour engage in Public Works (PW) and receive transfers for six months of the year. Households without labour capacity, Permanent Direct Support (PDS) clients, receive 12 months of unconditional transfers. The timing of PW and associated transfers vary from region to region. While many woredas are scheduled to pay benefits to PW clients during the months February to July (for January-June entitlements) all woredas in Somali region and some woredas in Oromia have different transfer schedules. One of the major agreements and changes made during the PSNP 4 Mid Term Review (MTR) was not to finance the federal contingency budget from the existing resources and scaling up safety net support in response to drought shocks primarily due to funding gap for the core program which will in turn help (i) strengthen the linkage between the PSNP and Humanitarian Food Assistance (HFA) and (ii) support the application of a common set of operational procedures to the provision of the PSNP transfers and transfers to the non PSNP households in response to drought. Linked to this for 2017 ad-hoc Federal contingency resources were mobilized through the World Bank Group, USAID through its PSNP implementing NGO Partners, DFID, WFP and UNICEF financial contributions to finance payments to be transferred to PSNP and non PSNP beneficiaries affected by the on-going drought. In the PSNP 4 logical framework, there are indicators that track timeliness of Federal contingency budget utilization. Specifically, these performance targets indicate that percentage of clients receiving contingency resources within 60 days of identification of needs .
abebual/Computer-Vision-Project-Monreader
Computer-Vision-Project-Monreader
abebual/Term-Deposit-Marketing-Prediction-5MSttBBp1mlzn2hA
The data comes from direct marketing efforts of a European banking institution. The marketing campaign involves making a phone call to a customer, often multiple times to ensure a product subscription, in this case a term deposit. Term deposits are usually short-term deposits with maturities ranging from one month to a few years. The customer must understand when buying a term deposit that they can withdraw their funds only after the term ends. All customer information that might reveal personal information is removed due to privacy concerns.
abebual/ValueInvestor-Predicting-TimeSeries-Data
Value Investor
abebual/AutoEM
abebual/cgan-based-ids
abebual/cGAN_GBM_IDS_app
cGAN based generated IoT intrusion detected by GBM multiclass classifier model
abebual/Clinical-Deterioration-Prediction-Model---Clustering
Clinical Deterioration Prediction Model - Clustering
abebual/Conda-Start_Guide
This guide only instructs you on how to set up Conda.
abebual/dais-2022-AE-demo
Build auto-encoder for anomaly detection, trained and inference distributedly
abebual/Deployment-flask
abebual/Exploratory-Data-Analysis-Clinical-Deterioration
This notebook includes an EDA Presentation based on Clinical Deterioration (cd) Final Dataset.
abebual/generative_inpainting
DeepFill v1/v2 with Contextual Attention and Gated Convolution, CVPR 2018, and ICCV 2019 Oral
abebual/Happy-Customer-Bpojxqbxe0zbwgEV
Bpojxqbxe0zbwgEV
abebual/lime
Lime: Explaining the predictions of any machine learning classifier
abebual/monte-carlo-reinforcement-learning-model
Solving the UNO Card Game using Monte-Carlo Reinforcement Learning
abebual/nbeatsx
abebual/neuromancer
Pytorch-based framework for solving parametric constrained optimization problems, physics-informed system identification, and parametric model predictive control.
abebual/pretrained-microscopy-models
abebual/Probabilistic-Programming-and-Bayesian-Methods-for-Hackers
aka "Bayesian Methods for Hackers": An introduction to Bayesian methods + probabilistic programming with a computation/understanding-first, mathematics-second point of view. All in pure Python ;)
abebual/pymc-experimental
abebual/RTX-coe_training
Training examples
abebual/sdr-hazards-classification
Collaboration work between FAA and Boeing on identifying safety hazards in Service Difficulty Reports (SDR)
abebual/shapiq
SHAP Interaction Quantification (short SHAP-IQ) is an XAI framework extending on the well-known shap explanations by introducing interactions i.e. synergy scores.