ClustGeo: an R package for hierarchical clustering with spatial constraints

Chavent, Marie; Kuentz-Simonet, Vanessa; Labenne, Amaury; Saracco, Jérôme

doi:10.1007/s00180-018-0791-1

Cited by 111 publications

(116 citation statements)

References 14 publications

(14 reference statements)

Supporting

Mentioning

104

Contrasting

Unclassified

Order By: Relevance

“…We chose the partition P α in K clusters to minimize within‐cluster inertia, I α . For our study, this simplified (for the k th cluster) to minimizing a sum of the pairwise squared distances with respect to explanatory variables

false(d_{0}^{2} false)

and spatial variables

false(d_{1}^{2} false)

I_{α} false(C_{k}^{α} false) = false(1 - normalα false) \sum_{i \in C_{k}^{normalα}} \sum_{j \in C_{k}^{α}} \frac{d_{0, i j}^{2}}{2 n_{k}} + normalα \sum_{i \in C_{k}^{normalα}} \sum_{j \in C_{k}^{α}} \frac{d_{1, i j}^{2}}{2 n_{k}},

where n k is the number of counties in cluster k and α is the mixing parameter that weights the degree to which nonspatial or spatial information contributes to inertia, α ∈ [0, 1] (Chavent et al ). Low inertia signifies high homogeneity (similarity) within cluster k , and a measure of total within‐cluster inertia is simply the sum of within‐cluster inertia for the K clusters.…”

Section: Methodsmentioning

confidence: 99%

“…where n k is the number of counties in cluster k and α is the mixing parameter that weights the degree to which nonspatial or spatial information contributes to inertia, α ∈ [0, 1] (Chavent et al 2018). Low inertia signifies high homogeneity (similarity) within cluster k, and a measure of total within-cluster inertia is simply the sum of withincluster inertia for the K clusters.…”

Section: Clustering To Form Multicounty Management Unitsmentioning

confidence: 99%

See 1 more Smart Citation

A Flexible Model‐based Approach to Delineate Wildlife Management Units

Caudell

Brooke

2020

Wildl. Soc. Bull.

View full text Add to dashboard Cite

Delineation of management units for harvested wildlife should strive to maximize homogeneity within each unit subject to constraints imposed by geography and heterogeneity in factors related to human‐caused changes in vital rates. Prior efforts to delineate management units for white‐tailed deer (Odocoileus virginianus) used regression to identify environmental and sociological factors related to female harvest density, followed by spatially constrained cluster analysis to select multicounty units for management purposes. We used county‐level data in Indiana, USA, from 2012 to 2017 to present a modified approach that 1) simultaneously models variation in 4 types of human‐mediated deer mortality with spatial, environmental, and socio‐cultural explanatory variables; 2) clusters counties using a partial contiguity constraint that explicitly considers nonspatial variables; and 3) relies on expert opinion to assess model adequacy and suggest improvements. Multivariate negative binomial regression of mortality from 4 sources (legal antlerless and antlered harvest, reported deer–vehicle collisions, crop depredation permits) yielded additive models with adequate fit and sizable effects for all univariate models due to latitude (standardized coefficient, β range = −0.18–0.24), and the nonspatial variables deer habitat (β range = 0.26–1.02) and hunter density (β range = 0.25–1.02). Fraction of cropland enrolled in the Conservation Reserve Program was positively related to harvest (β range = 0.10–0.13), and general human development was positively related to deer collision deaths (β range = 0.25–0.28) and take with depredation permits (β range = 0.07–0.24). Clustering counties with partial contiguity increased within‐unit nonspatial homogeneity 1.7–1.9 times more than clustering with strict contiguity. Responses to online surveys by 10 agency biologists and conservation officers provided congruence (mean ± SD = 85.2 ± 3.4%) of expert and model‐derived maps and useful suggestions for incorporating a separate urban management unit. Combining mortality‐based statistical models and expert assessment offers a rigorous yet flexible approach to delineating spatial units for wildlife management that is easily adaptable for various harvested species in multiple jurisdictions. © 2020 The Wildlife Society.

show abstract

false(d_{0}^{2} false)

and spatial variables

false(d_{1}^{2} false)

I_{α} false(C_{k}^{α} false) = false(1 - normalα false) \sum_{i \in C_{k}^{normalα}} \sum_{j \in C_{k}^{α}} \frac{d_{0, i j}^{2}}{2 n_{k}} + normalα \sum_{i \in C_{k}^{normalα}} \sum_{j \in C_{k}^{α}} \frac{d_{1, i j}^{2}}{2 n_{k}},

Section: Methodsmentioning

confidence: 99%

Section: Clustering To Form Multicounty Management Unitsmentioning

confidence: 99%

A Flexible Model‐based Approach to Delineate Wildlife Management Units

Caudell

Brooke

2020

Wildl. Soc. Bull.

View full text Add to dashboard Cite

show abstract

“…The R package "ClustGeo" (Chavent, Kuentz-Simonet, Labenne, & Saracco, 2018) was implemented to create a hierarchy based on spatial location to create the intermediate groupings between 33 pairs and 1 pair. At the broadest spatial level, or the "global" level, the data contain 1 pair of pooled localities sampled in 1974 and 2014.…”

Section: Analyzing Spatial Effects On Assemblage Changementioning

confidence: 99%

Temporal dynamics of stream fish assemblages and the role of spatial scale in quantifying change

Zbinden

2020

Ecology and Evolution

View full text Add to dashboard Cite

Spatial grain of studies of communities is often based on arbitrary convention.Few studies have examined how spatial scaling of grain size affects estimates of compositional change over time, despite its broad implications.2. Fish assemblage structure was compared between 1974 and 2014 at 33 sampling locations in the Muddy Boggy River drainage, USA. The two main objectives for this comparison were to quantify change in assemblage structure and to test for a relationship between compositional change and spatial scale. Spatial scale was manipulated by pooling assemblage data into a continuous range of groups, which increased in size from K = 33 pairs (i.e., local scale) to K = 1 pair (i.e., global scale), via clustering algorithm based on pair-wise fluvial distance.3. Local assemblages (stream reaches) varied in the degree of assemblage change over time (range = 0.10-0.99 dissimilarity; mean = 0.66). The global assemblage (drainage), however, remained relatively similar. A discontinuity in the relationship between compositional change and spatial scale occurred at K = 15 (mean dissimilarity = 0.56; p = .062), and this grouping is roughly the size of the headwater/ tributary drainages (i.e., stream order ≤ 3) in the study system. 4. Spatial scale can impact estimates of biodiversity change over time. These results suggest assemblages are more dynamic at individual stream reaches than at the scale of the entire drainage. The decline in assemblage change at the spatial scale of K = 15 deserves further attention given the marginal significance, despite a small sample size (n = 15). This pattern could suggest regional and meta-community processes become more important in shaping assemblage dynamics at the scale of headwater drainages, whereas the factors responsible for driving individual stream reach dynamics (e.g., stochasticity) become less important. Defining assemblages at a larger scale will result in different estimates of species persistence. Biodiversity monitoring efforts must take the effect of spatial scaling into consideration. K E Y W O R D Sbiodiversity monitoring, community stability, domains of scale, headwater streams, persistence, spatial clustering, spatial grain, temporal beta diversity

show abstract

“…FAMD is implemented in R packages FactoMineR (Lê, Josse, & Husson, ), PCAmixdata (Chavent, Kuentz‐Simonet, Labenne, & Saracco, ) and ade4 (Hill & Smith, ).…”

Section: Tandem Approachmentioning

confidence: 99%

Distance‐based clustering of mixed data

Velden

D’Enza

Markos

2018

WIREs Computational Stats

View full text Add to dashboard Cite

Cluster analysis comprises of several unsupervised techniques aiming to identify a subgroup (cluster) structure underlying the observations of a data set. The desired cluster allocation is such that it assigns similar observations to the same subgroup. Depending on the field of application and on domain‐specific requirements, different approaches exist that tackle the clustering problem. In distance‐based clustering, a distance metric is used to determine the similarity between data objects. The distance metric can be used to cluster observations by considering the distances between objects directly or by considering distances between objects and cluster centroids (or some other cluster representative points). Most distance metrics, and hence the distance‐based clustering methods, work either with continuous‐only or categorical‐only data. In applications, however, observations are often described by a combination of both continuous and categorical variables. Such data sets can be referred to as mixed or mixed‐type data. In this review, we consider different methods for distance‐based cluster analysis of mixed data. In particular, we distinguish three different streams that range from basic data preprocessing (where all variables are converted to the same scale), to the use of specific distance measures for mixed data, and finally to so‐called joint data reduction (a combination of dimension reduction and clustering) methods specifically designed for mixed data. This article is categorized under: Statistical Learning and Exploratory Methods of the Data Sciences > Clustering and Classification Statistical Learning and Exploratory Methods of the Data Sciences > Exploratory Data Analysis Statistical and Graphical Methods of Data Analysis > Dimension Reduction

show abstract

ClustGeo: an R package for hierarchical clustering with spatial constraints

Cited by 111 publications

References 14 publications

A Flexible Model‐based Approach to Delineate Wildlife Management Units

A Flexible Model‐based Approach to Delineate Wildlife Management Units

Temporal dynamics of stream fish assemblages and the role of spatial scale in quantifying change

Distance‐based clustering of mixed data

Contact Info

Product

Resources

About