Researches

✅ Predict Mortality of ICU Patients

I am working on 37 clinical variables of 8000 ICU patients to predict their mortality results. I develop the programs to perform different machine learning algorithm, and try to find smarter ways to deal with missing values. I learned to use R studio to better visualize the data as well as the prediction results and present the analysis to research team members. It is a challenging process for me, but I always enjoy it. I am very interested in applying machine learning models in the field of healthcare, and I hope my research will help ICU patients in some day.

✅ Analyze Effects of Different Features on House Prices

House price is always important for us. For this project, I visualized the cross-sectional data using the software Stata and made a presentation of data and predictions. After spending lots of time in data preprocessing, I created a hedonic regression model to analyze a package of attributes and their effects on house sales prices. What is more, I performed time series analysis on house prices and the reported the trend and cyclic patterns. Some simple regularized linear regression model as well as SVR model are used as baseline comparision.

✅ Summarize Wiki Pages Using Natural Language Processing Techniques

Besides healthcare, natural language processing is another area that attracts me the most. I used Python to crawl text files from Wikipedia. After doing some basic tokenization and lemmatization in the preprocessing step, I used language models extract the underlying patterns, and tried to summarize the important topics related to certain keywords. This helped us better curate and analyze key information, which I believe, will become more and more important in the era of big data.

✅ Additional information

Currently I'm working with Professor Joyce Ho on machine learning algorithms with healthcare problems. To learn more, please feel free to contact with me! My email address is emorywangzining@gmail.com.