2013/12/07 by Md Hasinur Rahaman Khan, Khan, Md Hasinur Rahaman, J.E. Shaw +1
Biochemistry, Genetics and Molecular Biology · Decision Sciences · Mathematics · #FOS: Computer and information sciences #Gene expression and cancer classification #Methodology (stat.ME) #Optimal Experimental Design Methods #Statistical Methods and Inference
paper · pdf · doi:10.48550/arxiv.1312.2079
openalex publication_date 2013/12/07 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
The accelerated failure time (AFT) models have proved useful in many\ncontexts, though heavy censoring (as for example in cancer survival) and high\ndimensionality (as for example in microarray data) cause difficulties for model\nfitting and model selection. We propose new approaches to variable selection\nfor censored data, based on AFT models optimized using regularized weighted\nleast squares. The regularized technique uses a mixture of L1 and L2 norm\npenalties under two proposed elastic net type approaches. One is the the\nadaptive elastic net and the other is weighted elastic net. The approaches\nextend the original approaches proposed by Ghosh (2007), and Hong and Zhang\n(2010) respectively. We also extend the two proposed approaches by adding\ncensoring observations as constraints into their model optimization frameworks.\nThe approaches are evaluated on microarray and by simulation. We compare the\nperformance of these approaches with six other variable selection\ntechniques--three are generally used for censored data and the other three are\ncorrelation-based greedy methods used for high-dimensional data.\n