Multivariate Regression Trees for Analysis of Abundance Data

Larsen, David R.; Speckman, Paul L.

doi:10.1111/j.0006-341x.2004.00202.x

Cited by 106 publications

(53 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also collated a range of meteorological parameters to investigate whether local weather conditions could be correlated with the observed fluctuations in aerosol bacterial populations. Using multivariate regression tree analysis (22,23), we examined such correlations, with tree topology and splitting parameters suggesting that sample location (in this case two geographically proximate cities) was less of a factor in explaining the variability of aerosol bacterial composition than temporal or meteorological influences (Fig. 3).…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Urban aerosols harbor diverse and dynamic bacterial populations

Brodie

DeSantis

Parker

et al. 2007

Proc. Natl. Acad. Sci. U.S.A.

584

529

View full text Add to dashboard Cite

Considering the importance of its potential implications for human health, agricultural productivity, and ecosystem stability, surprisingly little is known regarding the composition or dynamics of the atmosphere's microbial inhabitants. Using a custom high-density DNA microarray, we detected and monitored bacterial populations in two U.S. cities over 17 weeks. These urban aerosols contained at least 1,800 diverse bacterial types, a richness approaching that of some soil bacterial communities. We also reveal the consistent presence of bacterial families with pathogenic members including environmental relatives of select agents of bioterrorism significance. Finally, using multivariate regression techniques, we demonstrate that temporal and meteorological influences can be stronger factors than location in shaping the biological composition of the air we breathe.16S rRNA ͉ biosurveillance ͉ aerobiology ͉ microarray ͉ climate change

show abstract

Section: Resultsmentioning

confidence: 99%

“…Multivariate regression tree analysis (22,23) was carried out by using the package ''mvpart'' within the ''R'' statistical programming environment. A Bray-Curtis-based distance matrix was created by using the function ''gdist.''…”

Section: Real-time Quantitative Pcr Confirmation Of Phylochip-observedmentioning

confidence: 99%

Urban aerosols harbor diverse and dynamic bacterial populations

Brodie

DeSantis

Parker

et al. 2007

Proc. Natl. Acad. Sci. U.S.A.

584

529

View full text Add to dashboard Cite

show abstract

“…For multivariate responses, the influence function is a combination of influence functions appropriate for any of the univariate response variables discussed in the previous paragraphs: i.e., indicators for multiple binary responses (Zhang 1998;Noh, Song, and Park 2004), Logrank or Savage scores for multiple failure times (for example tooth loss times Su and Fan 2004) and the original observations or a rank transformation for multivariate regression (De'ath 2002;Larsen and Speckman 2004).…”

Section: J-class Classificationmentioning

confidence: 99%

Unbiased Recursive Partitioning: A Conditional Inference Framework

Hothorn

Hornik

Zeileis

2006

Journal of Computational and Graphical Statistics

3,123

2,667

View full text Add to dashboard Cite

Recursive binary partitioning is a popular tool for regression analysis. Two fundamental problems of exhaustive search procedures usually applied to fit such models have been known for a long time: Overfitting and a selection bias towards covariates with many possible splits or missing values. While pruning procedures are able to solve the overfitting problem, the variable selection bias still seriously effects the interpretability of tree-structured regression models. For some special cases unbiased procedures have been suggested, however lacking a common theoretical foundation. We propose a unified framework for recursive partitioning which embeds tree-structured regression models into a well defined theory of conditional inference procedures. Stopping criteria based on multiple test procedures are implemented and it is shown that the predictive performance of the resulting trees is as good as the performance of established exhaustive search procedures. It turns out that the partitions and therefore the models induced by both approaches are structurally different, indicating the need for an unbiased variable selection. The methodology presented here is applicable to all kinds of regression problems, including nominal, ordinal, numeric, censored as well as multivariate response variables and arbitrary measurement scales of the covariates. Data from studies on animal abundance, glaucoma classification, node positive breast cancer and mammography experience are re-analyzed.

show abstract

“…De'Ath's version of the tree is available as the R package mvpart (De'Ath 2006). Various authors, such as Larsen and Speckman (2004) and Hsiao and Shih (2007), have proposed alternative versions of this estimation method.…”

Section: Introductionmentioning

confidence: 99%

RE-EM trees: a data mining approach for longitudinal and clustered data

2011

View full text Add to dashboard Cite

Longitudinal data refer to the situation where repeated observations are available for each sampled object. Clustered data, where observations are nested in a hierarchical structure within objects (without time necessarily being involved) represent a similar type of situation. Methodologies that take this structure into account allow for the possibilities of systematic differences between objects that are not related to attributes and autocorrelation within objects across time periods. A standard methodology in the statistics literature for this type of data is the mixed effects model, where these differences between objects are represented by so-called "random effects" that are estimated from the data (population-level relationships are termed "fixed effects," together resulting in a mixed effects model). This paper presents a methodology that combines the structure of mixed effects models for longitudinal and clustered data with the flexibility of tree-based estimation methods. We apply the resulting estimation method, called the RE-EM tree, to pricing in online transactions, showing that the RE-EM tree is less sensitive to parametric assumptions and provides improved predictive power compared to linear models with random effects and regression trees without random effects. We also apply it to a smaller data set examining accident fatalities, and show that the RE-EM tree strongly outperforms a tree without random effects while performing comparably to a linear model with random effects. We also perform extensive simulation experiments to show that the estimator improves predictive performance relative to regression trees without random effects and is comparable or superior to using linear models with random effects in more general situations.

show abstract

Multivariate Regression Trees for Analysis of Abundance Data

Cited by 106 publications

References 12 publications

Urban aerosols harbor diverse and dynamic bacterial populations

Urban aerosols harbor diverse and dynamic bacterial populations

Unbiased Recursive Partitioning: A Conditional Inference Framework

RE-EM trees: a data mining approach for longitudinal and clustered data

Contact Info

Product

Resources

About