ENBIS-17 in Naples

9 – 14 September 2017; Naples (Italy) Abstract submission: 21 November 2016 – 10 May 2017

Identification of Protein Based Biomarkers for Gliomas

12 September 2017, 16:20 – 16:40


Submitted by
Sanjeev Sabnis
Sanjeev Sabnis (Indian Institute of Technology Bombay), Brijesh Singhal (Walmart)
This article presents a case study involving a Big Data. One of the recent statistical methodologies for variable selection, namely, Sure Independence Screening (SIS) and regularised statistical methodology, namely, Elastic Net Regularisation (a convex combination of the L_{1} and L_{2}, respective, penalties of the lasso and ridge regression methods) have been used to statistically analyse Glioma serum samples in which n < < p. The tuning parameter of Elastic Net Regularisation can be chosen using variety of loss functions such as Binomial Deviance, Mean Squared Error, Mean Absolute Error, Area Under the Curve and the Misclassification Error. The models obtained using two-step screening procedure involving SIS and LASSO (Elastic Net Regularisation) turn out to be better than those obtained only on LASSO (Elastic Net Regularisation).

Return to programme