Dataset Information

Estimating and modelling bias of the hierarchical partitioning public-domain software: implications in environmental management and conservation.

ABSTRACT: BACKGROUND: Hierarchical partitioning (HP) is an analytical method of multiple regression that identifies the most likely causal factors while alleviating multicollinearity problems. Its use is increasing in ecology and conservation by its usefulness for complementing multiple regression analysis. A public-domain software "hier.part package" has been developed for running HP in R software. Its authors highlight a "minor rounding error" for hierarchies constructed from >9 variables, however potential bias by using this module has not yet been examined. Knowing this bias is pivotal because, for example, the ranking obtained in HP is being used as a criterion for establishing priorities of conservation. METHODOLOGY/PRINCIPAL FINDINGS: Using numerical simulations and two real examples, we assessed the robustness of this HP module in relation to the order the variables have in the analysis. Results indicated a considerable effect of the variable order on the amount of independent variance explained by predictors for models with >9 explanatory variables. For these models the nominal ranking of importance of the predictors changed with variable order, i.e. predictors declared important by its contribution in explaining the response variable frequently changed to be either most or less important with other variable orders. The probability of changing position of a variable was best explained by the difference in independent explanatory power between that variable and the previous one in the nominal ranking of importance. The lesser is this difference, the more likely is the change of position. CONCLUSIONS/SIGNIFICANCE: HP should be applied with caution when more than 9 explanatory variables are used to know ranking of covariate importance. The explained variance is not a useful parameter to use in models with more than 9 independent variables. The inconsistency in the results obtained by HP should be considered in future studies as well as in those already published. Some recommendations to improve the analysis with this HP module are given.

SUBMITTER: Olea PP

PROVIDER: S-EPMC2908144 | biostudies-literature | 2010

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Estimating and modelling bias of the hierarchical partitioning public-domain software: implications in environmental management and conservation.

Olea Pedro P PP Mateo-Tomás Patricia P de Frutos Angel A

PloS one 20100721 7

<h4>Background</h4>Hierarchical partitioning (HP) is an analytical method of multiple regression that identifies the most likely causal factors while alleviating multicollinearity problems. Its use is increasing in ecology and conservation by its usefulness for complementing multiple regression analysis. A public-domain software "hier.part package" has been developed for running HP in R software. Its authors highlight a "minor rounding error" for hierarchies constructed from >9 variables, howeve ...[more]

PMID: 20657734

Similar Datasets

Project description:BACKGROUND:Expanding access to contraception and ensuring that need for family planning is satisfied are essential for achieving universal access to reproductive healthcare services, as called for in the 2030 Agenda for Sustainable Development. Monitoring progress towards these outcomes is well established for women of reproductive age (15-49 years) who are married or in a union (MWRA). For those who are not, limited data and variability in data sources and indicator definitions make monitoring challenging. To our knowledge, this study is the first to provide data and harmonised estimates that enable monitoring for all women of reproductive age (15-49 years) (WRA), including unmarried women (UWRA). We seek to quantify the gaps that remain in meeting family-planning needs among all WRA. METHODS AND FINDINGS:In a systematic analysis, we compiled a comprehensive dataset of family-planning indicators among WRA from 1,247 nationally representative surveys. We used a Bayesian hierarchical model with country-specific time trends to estimate these indicators, with 95% uncertainty intervals (UIs), for 185 countries. We produced estimates from 1990 to 2019 and projections from 2019 to 2030 of contraceptive prevalence and unmet need for family planning among MWRA, UWRA, and all WRA, taking into account the changing proportions that were married or in a union. The model accounted for differences in the prevalence of sexual activity among UWRA across countries. Among 1.9 billion WRA in 2019, 1.11 billion (95% UI 1.07-1.16) have need for family planning; of those, 842 million (95% UI 800-893) use modern contraception, and 270 million (95% UI 246-301) have unmet need for modern methods. Globally, UWRA represented 15.7% (95% UI 13.4%-19.4%) of all modern contraceptive users and 16.0% (95% UI 12.9%-22.1%) of women with unmet need for modern methods in 2019. The proportion of the need for family planning satisfied by modern methods, Sustainable Development Goals (SDG) indicator 3.7.1, was 75.7% (95% UI 73.2%-78.0%) globally, yet less than half of the need for family planning was met in Middle and Western Africa. Projections to 2030 indicate an increase in the number of women with need for family planning to 1.19 billion (95% UI 1.13-1.26) and in the number of women using modern contraception to 918 million (95% UI 840-1,001). The main limitations of the study are as follows: (i) the uncertainty surrounding estimates for countries with little or no data is large; and (ii) although some adjustments were made, underreporting of contraceptive use and needs is likely, especially among UWRA. CONCLUSIONS:In this study, we observed that large gaps remain in meeting family-planning needs. The projected increase in the number of women with need for family planning will create challenges to expand family-planning services fast enough to fulfil the growing need. Monitoring of family-planning indicators for all women, not just MWRA, is essential for accurately monitoring progress towards universal access to sexual and reproductive healthcare services-including family planning-by 2030 in the SDG era with its emphasis on 'leaving no one behind.'

Project description:Backgroundnationally determined contributions (NDCs) serve to meet the goals of the Paris Agreement of staying "well below 2°C", which could also yield substantial health co-benefits in the process. However, existing NDC commitments are inadequate to achieve this goal. Placing health as a key focus of the NDCs could present an opportunity to increase ambition and realise health co-benefits. We modelled scenarios to analyse the health co-benefits of NDCs for the year 2040 for nine representative countries (ie, Brazil, China, Germany, India, Indonesia, Nigeria, South Africa, the UK, and the USA) that were selected for their contribution to global greenhouse gas emissions and their global or regional influence.MethodsModelling the energy, food and agriculture, and transport sectors, and mortality related to risk factors of air pollution, diet, and physical activity, we analysed the health co-benefits of existing NDCs and related policies (ie, the current pathways scenario) for 2040 in nine countries around the world. We compared these health co-benefits with two alternative scenarios, one consistent with the goal of the Paris Agreement and the Sustainable Development Goals (ie, the sustainable pathways scenario), and one in line with the sustainable pathways scenario, but also placing health as a central focus of the policies (ie, the health in all climate policies scenario).FindingsCompared with the current pathways scenario, the sustainable pathways scenario resulted in an annual reduction of 1·18 million air pollution-related deaths, 5·86 million diet-related deaths, and 1·15 million deaths due to physical inactivity, across the nine countries, by 2040. Adopting the more ambitious health in all climate policies scenario would result in a further reduction of 462 000 annual deaths attributable to air pollution, 572 000 annual deaths attributable to diet, and 943 000 annual deaths attributable to physical inactivity. These benefits were attributable to the mitigation of direct greenhouse gas emissions and the commensurate actions that reduce exposure to harmful pollutants, as well as improved diets and safe physical activity.InterpretationA greater consideration of health in the NDCs and climate change mitigation policies has the potential to yield considerable health benefits as well as achieve the "well below 2°C" commitment across a range of regional and economic contexts.FundingThis work was in part funded through an unrestricted grant from the Wellcome Trust (award number 209734/Z/17/Z) and supported by an Engineering and Physical Sciences Research Council grant (grant number EP/R035288/1).

Project description:BackgroundWhile Expressed Sequence Tags (ESTs) have proven a viable and efficient way to sample genomes, particularly those for which whole-genome sequencing is impractical, phylogenetic analysis using ESTs remains difficult. Sequencing errors and orthology determination are the major problems when using ESTs as a source of characters for systematics. Here we develop methods to incorporate EST sequence information in a simultaneous analysis framework to address controversial phylogenetic questions regarding the relationships among the major groups of seed plants. We use an automated, phylogenetically derived approach to orthology determination called OrthologID generate a phylogeny based on 43 process partitions, many of which are derived from ESTs, and examine several measures of support to assess the utility of EST data for phylogenies.ResultsA maximum parsimony (MP) analysis resulted in a single tree with relatively high support at all nodes in the tree despite rampant conflict among trees generated from the separate analysis of individual partitions. In a comparison of broader-scale groupings based on cellular compartment (ie: chloroplast, mitochondrial or nuclear) or function, only the nuclear partition tree (based largely on EST data) was found to be topologically identical to the tree based on the simultaneous analysis of all data. Despite topological conflict among the broader-scale groupings examined, only the tree based on morphological data showed statistically significant differences.ConclusionBased on the amount of character support contributed by EST data which make up a majority of the nuclear data set, and the lack of conflict of the nuclear data set with the simultaneous analysis tree, we conclude that the inclusion of EST data does provide a viable and efficient approach to address phylogenetic questions within a parsimony framework on a genomic scale, if problems of orthology determination and potential sequencing errors can be overcome. In addition, approaches that examine conflict and support in a simultaneous analysis framework allow for a more precise understanding of the evolutionary history of individual process partitions and may be a novel way to understand functional aspects of different kinds of cellular classes of gene products.

Dataset Information

Estimating and modelling bias of the hierarchical partitioning public-domain software: implications in environmental management and conservation.

Publications

Estimating and modelling bias of the hierarchical partitioning public-domain software: implications in environmental management and conservation.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets