The nature of statistical learning theory
The nature of statistical learning theory
A Neural Network Model for Prognostic Prediction
ICML '98 Proceedings of the Fifteenth International Conference on Machine Learning
10-year CVD risk prediction and minimization via InverseClassification
Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium
| Hi-index | 0.00 |
There are three recognized types of Membranoproliferative glomerulonephritis (MPGN). Type II or Dense Deposit Disease (DDD)has a renal survival of 50% at 10 years. The goal of this study was to better identify patients at high risk of early renal failure,and to understand the factors that lead to fast progression of the disease. We identified six diagnostic features on the 98 DDD patients who responed to a Web-based survey, and examined the prognostic performance of these features in isolation and simple combinations. We then combined the features to build predictive models using both Cox proportional hazards regression (CHR), a standard statistical approach, and support vector machines (SVMs), a classification technique from the data mining literature. While the age and gender features showed some prognostic ability, the combined models -- particularly the SVM -- were superior in identifying cases with fast disease progression.This approach can be applied to disease survival analysis and prognosis, and might be useful to healthcare providers and patients in making healthcare decisions.