Li, Guo-Hung's Backend Projects - HackMD

<style> html, body, .ui-content { background-color: #333; color: #ddd; } .markdown-body h1, .markdown-body h2, .markdown-body h3, .markdown-body h4, .markdown-body h5, .markdown-body h6 { color: #ddd; } .markdown-body h1, .markdown-body h2 { border-bottom-color: #ffffff69; } .markdown-body h1 .octicon-link, .markdown-body h2 .octicon-link, .markdown-body h3 .octicon-link, .markdown-body h4 .octicon-link, .markdown-body h5 .octicon-link, .markdown-body h6 .octicon-link { color: #fff; } .markdown-body img { background-color: transparent; } .ui-toc-dropdown .nav>.active:focus>a, .ui-toc-dropdown .nav>.active:hover>a, .ui-toc-dropdown .nav>.active>a { color: white; border-left: 2px solid white; } .expand-toggle:hover, .expand-toggle:focus, .back-to-top:hover, .back-to-top:focus, .go-to-bottom:hover, .go-to-bottom:focus { color: white; } .ui-toc-dropdown { background-color: #333; } .ui-toc-label.btn { background-color: #191919; color: white; } .ui-toc-dropdown .nav>li>a:focus, .ui-toc-dropdown .nav>li>a:hover { color: white; border-left: 1px solid white; } .markdown-body blockquote { color: #bcbcbc; } .markdown-body table tr { background-color: #5f5f5f; } .markdown-body table tr:nth-child(2n) { background-color: #4f4f4f; } .markdown-body code, .markdown-body tt { color: #eee; background-color: rgba(230, 230, 230, 0.36); } a, .open-files-container li.selected a { color: #5EB7E0; } </style> # Li, Guo-Hung's Backend Projects --- # Project List --- [TOC]  ## Working Projects ### Mayo Clinic Project #### Advisor Dr. Jim long, AESOP tech #### Project introduction Because medical behaviors are different between countries, we try to use Mayo Clinic data to build statistic models and applied to products #### Target Build association models for medicines, diagnoses, procedures, surgery, lab test, diagnoses extracted by NLP #### Contribution 1. ETL of **billions of prescription data** 2. Build statistic association model 3. Define the rules and parameters --- ### Prescription Error Detection System #### Advisor Dr. Jim long, AESOP tech #### Project introduction Detect wrong prescribed medicines and raise alert for doctors. #### Target Complete and refine detection algorithm #### Contribution 1. ETL of **billions of prescription data** 2. Build statistic association model 3. Define the rules and parameters --- ### Medical Records Data Mining System #### Advisor Dr. Jim long, AESOP tech #### Project introduction Extract diagnoses from free text medical records and recommend doctors to add them into problem list #### Target Build a web-based system and data mining api. #### Contribution 1. Develop **Frontend** and **Backend** program - **UI/UX** - **API** 2. **Data mining** 3. Add information of National Health Insurance like usage freguency, CC, primaryDx... --- ### DeepDRG: Health Insurance Declaration Recommender System #### Advisor Dr. Jim long, AESOP tech #### Project introduction Using **GRU** model with patient's medicine record, procedure record, operation record and problem list to **recommend proper main diagnosis code**. #### Target **Speed up the coding diagnosis from minutes to seconds** and improve the accuracy of coding. #### Contribution 1. Collect 15 years Urology and Chest Medicine inpatients' data. 2. Data modeling and evaluation. 3. Get **over 80% in precision** --- ## Research Projects ### NTU MED GOD, Lifestyle Team #### Advisor Prof. Fei-Pei Lai, principal investigator of Medical Informatics Laboratory Dr. Jung-Yien Chien, pulmonologist in Mational Taiwan University Hospital #### Project introduction Dashboards of wearable devices data for medical caregivers to observe patients' health condition. #### Target Build a complete system and apply to more medical research. #### Contribution 1. Develop and maintain of subject **data monitoring platform** - **UI/UX** - **Database management** - **Bash script writing** - **Data fetching for medical research** 2. Collect lifestyle data, environmental data, and concatenate with a clinical questionnaire 3. Prediction of acute exacerbation of COPD in 7 days 4. Get **over 80% in accuracy** without clinical questionnaire 5. Prove that lifestyle data could be as important as a clinical questionnaire in disease prediction --- ### Establishing A Machine-Learning Model Based On Perioperative Data To Predict And Prevent Anesthesia-Related Adverse Outcomes #### Collaborator Dr. Pei-Fu Chen, anesthesiologist at Far Eastern Memorial Hospital #### Project introduction Establish a risk predicting model to reduce the risks of post-anesthesia mortality and morbidity when assessing patients in a pre-anesthesia clinic. #### Target Reduce the mortality and morbidity of patients. #### Contribution 1. Data cleaning, outlier removing, extreme imbalanced data preprocessing (400:1) 2. Features include demography data, lab data, medical history, vital sign, and departments 3. Modeling with machine learning 4. Performance evaluation 5. Visualization 6. Deployment 7. **Get over 50% in F1-score when predicting the mortality rate** 8. **Get more than 40% F1-score when predicting the risk of 5 common complications** --- ### Enhanced Prediction Of Mortality In Critically Ill Patients By Combining Machine Learning Models: A Three-Step Strategy And Predicted Risk-Stratified Analysis #### Collaborator Dr. Yu-Chang Yeh #### Project introduction Build an AI prediction model of mortality and length of stay. #### Target Reduce the mortality and morbidity of patients by paying more attention on patients in higher risk. #### Contribution 1. Data cleaning 2. Feature engineering 3. Modeling with machine learning 4. Get over 80% in accuracy, 90% in AUROC, and 60% in F1-score 5. Prediction performance is better than APACHE II and APACHE IV --- ### Onset Date Extraction For Stroke Patient #### Collaborator Associate Prof. Pei-Fang Tang #### Project introduction Extract the onset date of stroke patients and their NIH Stroke Scale (NIHSS) from the admission summary. #### Target Reduce more than 30% of the time for clinical caregivers to survey medical records. #### Contribution 1. Data cleaning 2. Use regular expression to filter out the sentences without date-alike strings. 3. Use LSTM to vote if the reserved data contains the onset date or not. 4. Get **over 80% in accuracy** --- ### Physical Activity Level Identification #### Advisor Prof. Chia-Tai Chan, principal investigator of Ubiquitous Computing Laboratory #### Project introduction To monitoring user's daily activity levels, I had developed an Android App to collect user's demography data and acceleration data. Then I modeling these data with machine learning models and compare the performance. #### Contribution 1. **App development** 2. Data collection 3. Feature engineering 4. Modeling with machine learning 5. **Get over 90% in accuracy** 6. Submit Conference paper and it has been adapted --- ### Lateralization Of Women’S Brain Change With Age #### Advisor Prof. Yu-Te Wu, principal investigator of Medical Image and Signal Analysis Laboratory #### Project introduction The lateralization of brain function refers to some neural functions, or cognitive processes tend to be more dominant in one hemisphere than the other. The project is to find out the relation between age and the change of lateralization pattern. #### Contribution 1. Use 100 females' T1 images with cortical surface extraction and curvature_based alignment. 2. Use 5 features of brain structure include: - fractal dimension(FD) - segment thickness - local gyrification index(LGI) - surface area - volume 3. Calculate the lateralization index as (R-L)/(R+L). 4. Calculate Pearson correlation coefficients for their lateralization index and ages.