Dataset Information

Evaluation of Causal Structure Learning Methods on Mixed Data Types.

ABSTRACT: Causal structure learning algorithms are very important in many fields, including biomedical sciences, because they can uncover the underlying causal network structure from observational data. Several such algorithms have been developed over the years, but they usually operate on datasets of a single data type: continuous or discrete variables only. More recently, we and others have proposed new causal structure learning algorithms for mixed data types. However, to-date there is no study that critically evaluates these methods' performance. In this paper, we provide the first extensive empirical evaluation of several popular causal structure learning methods on mixed data types and in a variety of parameter settings and sample sizes. Our results serve as a guide as to which method performs the best in a given context, and as such they are a first step towards a "method selection guide" for those running causal modeling methods on real life datasets.

SUBMITTER: Raghu VK

PROVIDER: S-EPMC6510516 | biostudies-literature | 2018 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Evaluation of Causal Structure Learning Methods on Mixed Data Types.

Raghu Vineet K VK Poon Allen A Benos Panayiotis V PV

Proceedings of machine learning research 20180801

Causal structure learning algorithms are very important in many fields, including biomedical sciences, because they can uncover the underlying causal network structure from observational data. Several such algorithms have been developed over the years, but they usually operate on datasets of a single data type: continuous or discrete variables only. More recently, we and others have proposed new causal structure learning algorithms for mixed data types. However, to-date there is no study that cr ...[more]

PMID: 31080946

Similar Datasets

Project description:BackgroundDespite the central role of mixed methods in health research, studies evaluating online methods training in the health sciences are nonexistent. The focused goal was to evaluate online training by comparing the self-rated skills of scholars who experienced an in-person retreat to scholars in an online retreat in specific domains of mixed methods research for the health sciences from 2015-2023.MethodsThe authors administered a scholar Mixed Methods Skills Self-Assessment instrument based on an educational competency scale that included domains on: "research questions," "design/approach," "sampling," "analysis," and "dissemination" to participants of the Mixed Methods Research Training Program for the Health Sciences (MMRTP). Self-ratings on confidence on domains were compared before and after retreat participation within cohorts who attended in person (n = 73) or online (n = 57) as well as comparing across in-person to online cohorts. Responses to open-ended questions about experiences with the retreat were analyzed.ResultsScholars in an interactive program to improve mixed methods skills reported significantly increased confidence in ability to define or explain concepts and in ability to apply the concepts to practical problems, whether the program was attended in-person or synchronously online. Scholars in the online retreat had self-rated skill improvements as good or better than scholars who participated in person. With the possible exception of networking, scholars found the online format was associated with advantages such as accessibility and reduced burden of travel and finding childcare. No differences in difficulty of learning concepts was described.ConclusionsKeeping in mind that the retreat is only one component of the MMRTP, this study provides evidence that mixed methods training online was associated with the same increases in self-rated skills as persons attending online and can be a key component to increasing the capacity for mixed methods research in the health sciences.

Dataset Information

Evaluation of Causal Structure Learning Methods on Mixed Data Types.

Publications

Evaluation of Causal Structure Learning Methods on Mixed Data Types.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets