Boosting predictions of treatment success

LeBlanc, Michael; Kooperberg, Charles

doi:10.1073/pnas.1008052107

Cited by 14 publications

(7 citation statements)

References 11 publications

(6 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SNP, chr2_171708059_C_T at 2q31.1, is within the GAD1 gene and the expression level of GAD1 is a significant prognostic factor in lung adenocarcinoma [47]. Thus, the interpretation of DNN models may identify novel SNPs with nonlinear association with the breast cancer [48][49][50][51][52].…”

Section: Interpretation Of the Dnn Modelmentioning

confidence: 99%

Deep neural network improves the estimation of polygenic risk scores for breast cancer

et al. 2020

View full text Add to dashboard Cite

Polygenic risk scores (PRS) estimate the genetic risk of an individual for a complex disease based on many genetic variants across the whole genome. In this study, we compared a series of computational models for estimation of breast cancer PRS. A deep neural network (DNN) was found to outperform alternative machine learning techniques and established statistical algorithms, including BLUP, BayesA, and LDpred. In the test cohort with 50% prevalence, the Area Under the receiver operating characteristic Curve (AUC) were 67.4% for DNN, 64.2% for BLUP, 64.5% for BayesA, and 62.4% for LDpred. BLUP, BayesA, and LPpred all generated PRS that followed a normal distribution in the case population. However, the PRS generated by DNN in the case population followed a bimodal distribution composed of two normal distributions with distinctly different means. This suggests that DNN was able to separate the case population into a high-genetic-risk case subpopulation with an average PRS significantly higher than the control population and a normal-genetic-risk case subpopulation with an average PRS similar to the control population. This allowed DNN to achieve 18.8% recall at 90% precision in the test cohort with 50% prevalence, which can be extrapolated to 65.4% recall at 20% precision in a general population with 12% prevalence. Interpretation of the DNN model identified salient variants that were assigned insignificant p values by association studies, but were important for DNN prediction. These variants may be associated with the phenotype through nonlinear relationships.

show abstract

Section: Interpretation Of the Dnn Modelmentioning

confidence: 99%

Deep neural network improves the estimation of polygenic risk scores for breast cancer

et al. 2020

View full text Add to dashboard Cite

show abstract

“…When the derivation of an optimal treatment regime rather than the description of treatment effect heterogeneity is of interest, we can replace the GCV statistic with the average effect size of the optimal treatment rule [Imai and Strauss (2011), Qian and Murphy (2011)]. The proposed methodology with the GCV statistic does not require cross-validation and hence is more computationally efficient than the commonly used methods for estimation of treatment effect heterogeneity such as Boosting [Freund and Schapire (1999), LeBlanc and Kooperberg (2010)], Bayesian additive regression trees (BART) [Chipman, George and McCulloch (2010), Green and Kern (2010a)], and other tree-based approaches [e.g., Su et al (2009), Imai and Strauss (2011), Lipkovich et al (2011), Loh et al (2012), Kang et al (2012)]. While most similar to a Bayesian logistic regression with noninformative prior [Gelman et al (2008)], the proposed method uses LASSO constraints to produce a parsimonious model.…”

mentioning

confidence: 99%

Estimating treatment effect heterogeneity in randomized program evaluation

Imai¹,

Ratkovic²

2013

Ann. Appl. Stat.

448

357

View full text Add to dashboard Cite

When evaluating the efficacy of social programs and medical treatments using randomized experiments, the estimated overall average causal effect alone is often of limited value and the researchers must investigate when the treatments do and do not work. Indeed, the estimation of treatment effect heterogeneity plays an essential role in (1) selecting the most effective treatment from a large number of available treatments, (2) ascertaining subpopulations for which a treatment is effective or harmful, (3) designing individualized optimal treatment regimes, (4) testing for the existence or lack of heterogeneous treatment effects, and (5) generalizing causal effect estimates obtained from an experimental sample to a target population. In this paper, we formulate the estimation of heterogeneous treatment effects as a variable selection problem. We propose a method that adapts the Support Vector Machine classifier by placing separate sparsity constraints over the pre-treatment parameters and causal heterogeneity parameters of interest. The proposed method is motivated by and applied to two well-known randomized evaluation studies in the social sciences. Our method selects the most effective voter mobilization strategies from a large number of alternative strategies, and it also identifies the characteristics of workers who greatly benefit from (or are negatively affected by) a job training program. In our simulation studies, we find that the proposed method often outperforms some commonly used alternatives. . The proposed methods can be implemented via open-source software FindIt [Ratkovic and Imai (2012)], which is freely available at the Comprehensive R Archive Network (http://cran.r-project.org/package=FindIt). This software also contains the results of our empirical analysis.

show abstract

“…A strength of this study is the novel use of a tree‐based model well suited to exploring interactions without a priori specification to empirically identify the most influential subgroups with potential differences in treatment effect. GBM is an innovative method that has been a part of statistical literature for many years but has not been used as a method for identifying subgroups with different treatment effects. The methods described in this paper could be readily applied in other clinical scenarios.…”

Section: Discussionmentioning

confidence: 99%