Journal of Data Science

Causal inference can estimate causal effects, but unless data are collected experimentally, statistical analyses must rely on pre-specified causal models. Causal discovery algorithms are empirical methods for constructing such causal models from data. Several asymptotically correct discovery methods already exist, but they generally struggle on smaller samples. Moreover, most methods focus on very sparse causal models, which may not always be a realistic representation of real-life data generating mechanisms. Finally, while causal relationships suggested by the methods often hold true, their claims about causal non-relatedness have high error rates. This non-conservative error trade off is not ideal for observational sciences, where the resulting model is directly used to inform causal inference: A causal model with many missing causal relations entails too strong assumptions and may lead to biased effect estimates. We propose a new causal discovery method that addresses these three shortcomings: Supervised learning discovery (SLdisco). SLdisco uses supervised machine learning to obtain a mapping from observational data to equivalence classes of causal models. We evaluate SLdisco in a large simulation study based on Gaussian data and we consider several choices of model size and sample size. We find that SLdisco is more conservative, only moderately less informative and less sensitive towards sample size than existing procedures. We furthermore provide a real epidemiological data application. We use random subsampling to investigate real data performance on small samples and again find that SLdisco is less sensitive towards sample size and hence seems to better utilize the information available in small datasets.

Abstract: For model selection in mixed effects models, Vaida and Blan chard (2005) demonstrated that the marginal Akaike information criterion is appropriate as to the questions regarding the population and the conditional Akaike information criterion is appropriate as to the questions regarding the particular clusters in the data. This article shows that the marginal Akaike information criterion is asymptotically equivalent to the leave-one-cluster-out cross-validation and the conditional Akaike information criterion is asymptotically equivalent to the leave-one-observation-out cross-validation.

One of the key features in regression models consists in selecting appropriate characteristics that explain the behavior of the response variable, in which stepwise-based procedures occupy a prominent position. In this paper we performed several simulation studies to investigate whether a specific stepwise-based approach, namely Strategy A, properly selects authentic variables into the generalized additive models for location, scale and shape framework, considering Gaussian, zero inflated Poisson and Weibull distributions. Continuous (with linear and nonlinear relationships) and categorical explanatory variables are considered and they are selected through some goodness-of-fit statistics. Overall, we conclude that the Strategy A greatly performed.