Stabilizing Sparse Cox Model Using Statistic and Semantic Structures in Electronic Medical Records

Stability in clinical prediction models is crucial for transferability between studies, yet has received little attention. The problem is paramount in high dimensional data, which invites sparse models with feature selection capability. We introduce an effective method to stabilize sparse Cox model of time-to-events using statistical and semantic structures inherent in Electronic Medical Records (EMR). Model estimation is stabilized using three feature graphs built from (i) Jaccard similarity among features (ii) aggregation of Jaccard similarity graph and a recently introduced semantic EMR graph (iii) Jaccard similarity among features transferred from a related cohort. Our experiments are conducted on two real world hospital datasets: a heart failure cohort and a diabetes cohort. On two stability measures – the Consistency index and signal-to-noise ratio (SNR) – the use of our proposed methods significantly increased feature stability when compared with the baselines.

[1]  Erhan Bilal,et al.  Improving Breast Cancer Survival Analysis through Competition-Based Multidimensional Modeling , 2013, PLoS Comput. Biol..

[2]  Shie Mannor,et al.  Sparse algorithms are not stable: A no-free-lunch theorem , 2008, Allerton 2008.

[3]  Bhanukiran Vinzamuri,et al.  Cox Regression with Correlation Based Regularization for Electronic Health Records , 2013, 2013 IEEE 13th International Conference on Data Mining.

[4]  Holger Fröhlich,et al.  Including network knowledge into Cox regression models for biomarker signature discovery , 2014, Biometrical journal. Biometrische Zeitschrift.

[5]  R. Tibshirani.  The lasso method for variable selection in the Cox model. , 1997, Statistics in medicine.

[6]  A. Barabasi,et al.  Network medicine : a network-based approach to human disease , 2010 .

[7]  Wei Luo,et al.  An integrated framework for suicide risk prediction , 2013, KDD.

[8]  Bhanukiran Vinzamuri,et al.  Active Learning based Survival Regression for Censored Data , 2014, CIKM.

[9]  Qiang Yang,et al.  A Survey on Transfer Learning , 2010, IEEE Transactions on Knowledge and Data Engineering.

[10]  John Blitzer,et al.  Regularized Learning with Networks of Features , 2008, NIPS.

[11]  Rui Feng,et al.  NETWORK-REGULARIZED HIGH-DIMENSIONAL COX REGRESSION FOR ANALYSIS OF GENOMIC DATA. , 2014, Statistica Sinica.

[12]  Shie Mannor,et al.  Sparse Algorithms Are Not Stable: A No-Free-Lunch Theorem , 2008, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[13]  Peng Zhao,et al.  On Model Selection Consistency of Lasso , 2006, J. Mach. Learn. Res..

[14]  Ludmila I. Kuncheva,et al.  A stability index for feature selection , 2007, Artificial Intelligence and Applications.

[15]  Martin Ester,et al.  Optimally discriminative subnetwork markers predict response to chemotherapy , 2011, Bioinform..

[16]  Holger Fröhlich,et al.  Review Biomarker Gene Signature Discovery Integrating Network Knowledge , 2012 .

[17]  Jieping Ye,et al.  Sparse methods for biomedical data , 2012, SKDD.

[18]  Jorge Nocedal,et al.  On the limited memory BFGS method for large scale optimization , 1989, Math. Program..

[19]  Wei Luo,et al.  Stabilized sparse ordinal regression for medical risk stratification , 2014, Knowledge and Information Systems.

[20]  Jinchi Lv,et al.  High-Dimensional Sparse Additive Hazards Regression , 2012, 1212.6232.

[21]  Peter C Austin,et al.  Automated variable selection methods for logistic regression produced unstable models for predicting acute myocardial infarction mortality. , 2004, Journal of clinical epidemiology.

[22]  Trevor Hastie,et al.  Regularization Paths for Cox's Proportional Hazards Model via Coordinate Descent. , 2011, Journal of statistical software.