Spring 2021, Machine Learning STOR 565
Final Project (Heart failure clinical records)
Data link: https://archive.ics.uci.edu/ml/datasets/Heart+failure+clinical+records#
Final project link: https://cymichael.github.io/files/Final_report_group_Yi%20Cui.pdf
Abstract:
The goal of this research is to use a variety of machine learning methods to predict the survival of patients with heart failure based on their clinical, body, and lifestyle information. We also would like to find the best statistical models for our data set and identify the most important predictors in these models. Specifically, we use both unsupervised learaning (i.e., PCA and clustering) and supervised learning methods (i.e., kNN, logistic regressions, LASSO, and decision trees) on our data set.
Consistent with the existing literature, we find that age, the level of serum creatinine in the blood, and the speed at which the blood running through the heart (i.e., ejection fraction) are the three most important predictors for the survival of patients with heart failure. In specific, younger patients, a lower level of serum creatine, and a higher level of ejection fraction are more likely to survive with heart failure.
Presentation record link:
https://www.bilibili.com/video/BV1YV411J7vz/?vd_source=86bbdd0218dd3a7a94ae36979723c137