Figure 4 - uploaded by Carsten Walther
Content may be subject to copyright.
Example of hierarchical clustering: clusters are consecutively merged with the most nearby clusters. The length of the vertical dendogram-lines reflect the nearness. 

Example of hierarchical clustering: clusters are consecutively merged with the most nearby clusters. The length of the vertical dendogram-lines reflect the nearness. 

Source publication
Book
Full-text available
In coupled human-environment systems where well established and proven general theories are often lacking cluster analysis provides the possibility to discover regularities – a first step in empirically based theory building. The aim of this report is to share the experiences and knowledge on cluster analysis we gained in several applications in th...

Context in source publication

Context 1
... are helpful in suggesting the (dis)similar properties and characteristics of the various clusters. In case that the data have a spatial dimension, showing maps can give a clue on how the clusters are geographically distributed, serving to identify and connect features with similar characteristics at different geographical locations (see Figure 14). ...

Similar publications

Article
Full-text available
In this paper, a thermal periodic theory of building integrated semi-transparent photovoltaic thermal (BiSPVT) system has been developed by using the basic energy balance equation for a semi-transparent photovoltaic roof; room air, brick/glazed walls and reinforced concrete cement (RCC) floor of the first room. Since solar irradiation and ambient a...
Article
Full-text available
Purpose – The purpose of this paper is to explore the concept of supply chain quality management (SCQM) and to propose a research model that considers the impact of SCQM practices on firm performance. Design/methodology/approach – Based on a comprehensive literature review, SCQM most relevant practices were analyzed and selected. These practices w...
Article
Full-text available
Purpose – This study applies Oviatt and McDougall’s (2005) model of forces influencing the speed of internationalisation to small, export oriented enterprises. The purpose of this paper is to explore the significance of the forces enabling, motivating, mediating and moderating internationalisation in small and medium-sized enterprises (SMEs), and t...
Article
Full-text available
Background While medical students and residents may be utilizing websites as online learning resources, medical trainees and educators now have the opportunity to create such educational websites and digital tools on their own. However, the process and theory of building educational websites for medical education have not yet been fully explored....
Article
Full-text available
In today’s education, the distance mode represents more than 15% of enrollment in graduate, forum and chat are the most used online conversations tools. It motived the research presented in this paper where the issue is: How many students will participate in an educational chat to ensure a satisfactory level of participation? The aim of this resear...

Citations

... We can hypothesize that there is a certain bias of social desirability that tends to make the students answer what they imagine to be expected or to an ideal that they make of the answers but that in reality is impacted by a series of human and contextual factors. The internal validity of the clusters seems to be acceptable (thanks to the Calinski index which is a good marker of internal validity) [17,41]. However, we do not have the possibility to assess its external validity since we have not found any comparable study so far. ...
Article
Full-text available
Introduction People with substance use disorder (SUD) deal with stigmatization in various areas of life, including healthcare system. In this study, we investigated the attitudes of final-year medical students towards SUD people and attempted to understand their influence. Methods We conducted a two-stage cluster analysis (hierarchical ascending classification followed by K-means clustering) based on the “beSAAS”. We administrated this 23-item questionnaire to 923 final-year medical students in Belgium (response rate = 71,1%). Sociodemographic characteristics were compared between the clusters. Results Four clusters of students with specific characteristics were identified in this study. The first, “The Inclusives” (including 27,9% of respondents) had the least negative attitudes; they wanted to specialize mainly in psychiatry and gynecology. The second, “The Centrists" (23,6%) consisted mainly of male students. They had many private and professional experiences with substance use and considered themselves less healthy than others did. Most wanted to specialize in pediatrics and general practice. Their attitudes were slightly negative towards people with SUD. The third, “The Moralists” (27,6%), were mainly older, from non-European countries, had the least experience with substance use (or contact mainly in hospitals), had the less high mother’s level of education and reported excellent health. They were heading toward other specialties. They had the most stereotypes and moralism, and less treatment optimism. The fourth, “The Specialist care-oriented” (20,8%), were the most in favor of specialized treatment. This group had a higher proportion of Belgian, females, and students who had specific contact with this population. They especially intended to specialize in internal medicine. Conclusion This study revealed 4 profiles of medical students with different attitudes towards SUD people. “The Moralists”, including more than a quarter of the respondents, were characterized by strong stereotypes and moralism and little treatment optimism. These clusters could contribute to the design of a learner-centered program aimed at addressing stigma within the main curriculum.
... We can hypothesize that there is a certain bias of social desirability that tends to make the students answer what they imagine to be expected or to an ideal that they make of the answers but that in reality is impacted by a series of human and contextual factors. The internal validity of the clusters seems to be acceptable (thanks to the Calinski index which is a good marker of internal validity) 22,45 . However, we do not have the possibility to assess its external validity since we have not found any comparable study so far. ...
Preprint
Full-text available
Introduction: People with a substance use disorder (SUD) are dealing with many stigmas including from the health care system. In this study, we investigated the attitudes of final year medical students towards SUD people and tried to understand what influence. Methods: We conducted a two-stage cluster analysis (hierarchical ascending classification followed by K-means clustering) based on the “beSAAS”. We administrated this 23-item questionnaire to 923 final-year medical students in Belgium (response rate = 71,1%). Socio-demographic characteristics were compared between the clusters. Results: 4 clusters of students with specific characteristics emerged. The first, “The Inclusive” (including 27,9% of respondents) had the less negative attitudes, they wanted to specialize mainly into psychiatry and gynecology. The second, “The Centrists" (23,6%) consisted mainly of male students. They had many private and professional experiences with substance use and considered themselves less healthy than the others. Most of them wanted to specialize in pediatrics and general practice. Their attitudes were slightly negative towards people with SUD. The third, “The Moralists” (27,6%), were mainly older, from non-European countries, had the least experience with substance use (or contact mainly in hospitals), had the less high mother’s level of education and reported excellent health. They were heading toward other specialities. They had the most stereotypes and moralism and the less treatment optimism. The fourth, “The Specialists” (20,8%), were the most in favor of specialized treatment. This group had a higher proportion of Belgians, females, and students with specific contacts with this population. They intended especially to specialize into internal medicine. This study revealed 4 profiles of medical students with different attitudes towards SUD people. “The Moralists", including more than a quarter of the respondents, were characterized by strong stereotypes and moralism and very little treatment optimism. Conclusion: These clusters could contribute to designing a learner-centered program aimed at addressing stigma within the main curriculum.
... Caquetá is habited by colonos and Yurimaguas by mestizos, who have arrived at these sites through different historical processes and carry diverse cultural heritage. Firstly, the variables were standardized for Caquetá and Yurimaguas separately, where 0 refers to the dataset's minimum value and 1 to its maximum value [67]. Secondly, the cluster analysis was performed using a hierarchical cluster algorithm using the statistics package R [67,68]. ...
... Firstly, the variables were standardized for Caquetá and Yurimaguas separately, where 0 refers to the dataset's minimum value and 1 to its maximum value [67]. Secondly, the cluster analysis was performed using a hierarchical cluster algorithm using the statistics package R [67,68]. Thirdly, a consistency measure was applied to estimate the optimal number of clusters or typologies required to improve the grouping of households, following the methodological steps detailed by Sietz and colleagues [68,69]. ...
Article
Full-text available
The Andean-Amazon foothills region, one of the richest biodiversity ecoregions on earth, is threatened by climate change combined with unsustainable agricultural and extensive livestock farming. These land-use practices tend to reduce the diversification of rural farming, decreasing households' livelihood alternatives and rendering them more vulnerable to climate change. We studied the relationship between rural livelihood diversification and household-level vulnerability to climate change in a sample of Andean-Amazon foothills households in Colombia and Peru. Firstly, we determined typologies of households based on their rural livelihood diversification, including farming diversification (agrobiodiversity and farming activities) and agroecological management practices. Secondly, we evaluated each household typology's vulnerability to climate change by assessing sensitivity and adaptive capacity based on the 'livelihood assets pentagon', which encompasses the five human capitals: natural, social, human, physical, and financial. We concluded that households with higher rural livelihood diversification are less vulnerable to climate change. However, it is impossible to draw significant conclusions about the relationship between the factors of diversification of management practices and vulnerability to climate change because most households have few agroecological practices. Results may inform future interventions that aim to decrease Andean-Amazon foothills households' sensitivity and strengthen their adaptive capacity to climate change.
... We used cluster analysis to analyze both forms of archetype. Cluster analysis is a statistical method that group objects with similar attributes into exclusive clusters (Janssen et al. 2012). Clustering is a common approach to identify patterns across complex social-ecological systems and is often used in archetype analysis . ...
Article
Full-text available
In African small-scale agriculture, sustainable land and water management (SLWM) is key to improving food production while coping with climate change. However, the rate of SLWM adoption remains low, suggesting a gap between generalized SLWM advantages for rural development across the literature, and the existence of context-dependent barriers to its effective implementation. Uganda is an example of this paradox: the SLWM adoption rate is low despite favorable ecological conditions for agriculture development and a large rural population. A systemic understanding of the barriers hindering the adoption of SLWM is therefore crucial to developing coherent policy interventions and enabling effective funding strategies. Here, we propose a cross-scale archetype approach to identify and link barriers to SLWM adoption in Uganda. We performed 80 interviews across the country to build cognitive archetypes, harvesting stakeholders' perceptions of different types of barriers. We complemented this bottom-up perspective with a spatial archetype analysis to contextualize these results across different social-ecological regions. We found poverty trap, overpopulation, risk aversion, remoteness, and post-conflict patriarchal systems as cognitive archetypes that synthesize the different dynamics of barriers to SLWM adoption in Uganda. Our results reveal both specific and cross-cutting barriers. Ineffective extension services emerges as a ubiquitous barrier, whereas gender inequality is a priority barrier for large supported farms and farms in drier lowlands in northern Uganda. The combination of cognitive and spatial archetypes proposed here can help to overcome ineffective "one-size-fits-all" solutions and support context-specific policy plans to scale up SLWM, rationing resources to support sustainable intensification of agriculture.
... For KIDS data, also three was chosen as the number of centroids for which stability measures scored highest, however not across all ES (e.g., crop farming). Using clustering is a common method to study social-ecological systems (e.g., Janssen et al., 2012). Cluster algorithms come with trade-offs when comparing results to potential future studies in the field. ...
Article
Full-text available
This paper takes a new look on transition processes in social-ecological systems, identified based on household use of direct ecosystem services in a case study in KwaZulu-Natal, South Africa. We build on the assumption that high dependence on local ecosystems for basic needs satisfaction corresponds to a “green loop” type of system, with direct feedbacks between environmental degradation and human well-being. Increasing use of distant ecosystems marks a regime shift and with that, the transition to “red loops” in which feedbacks between environmental degradation and human well-being are only indirect. These systems are characterized by a fundamentally different set of sustainability problems as well as distinct human-nature connections. The analysis of a case study in KwaZulu-Natal, South Africa, shows that social-ecological systems identified as green loops in 1993, the average share of households using a characteristic bundle of direct ecosystem services drops consistently (animal production, crop production, natural building materials, freshwater, wood). Conversely, in systems identified as red loops, mixed tendencies occur which underpins non-linearities in changing human-nature relationships. We propose to apply the green to red loop transition model to other geographical contexts with regards to studying the use of local ecosystem services as integral part of transformative change in the Anthropocene.
... farmer and trader profits) (Holden et al., 2017). Alternatively, simulation outcomes could instead be categorised post-hoc, with the use of statistical clustering (Janssen et al., 2012) or multivariable regression techniques empirically analysing the relationships between scenario trajectories and outcomes from the bottom-up. However, the normative futures here help to distinguish the scenarios and feedbacks leading to aspirational futures defined a priorilaying the foundations for future studies to disaggregate outcome dynamics within the 'win-win-win' spacesuch as between farmers with different assets (e.g. ...
Article
Full-text available
Context There is growing recognition that food systems must adapt to become more sustainable and equitable. Consequently, in developing country contexts, there is increasing momentum away from traditional producer-facing value chain upgrades towards efforts to increase both the availability and affordability of nutritious foods at the consumer level. However, such goals must navigate the inherent complexities of agricultural value chains, which involve multiple interactions, feedbacks and unintended consequences, including important but often surprising trade-offs between producers and consumers. Objective and methods Based around the 'Loop' horticultural aggregation scheme of Digital Green in Bihar, India, we develop a system dynamics modelling framework to survey the value chain trade-offs emerging from upgrades that aim to improve the availability of fruits and vegetables in small retail-oriented markets. We model the processes of horticultural production, aggregation, marketing, and retailing – searching for futures that are ‘win-win-win’ for: (i) the availability of fruits and vegetables in small retail markets, (ii) the profits of farmers participating in aggregation, and (iii) the sustainability of the initial scheme for Digital Green as an organisation. We simulate two internal upgrades to aggregation and two upgrades to the wider enabling environment through a series of 5000 Monte Carlo trajectories – designed to explore the plausible future dynamics of the three outcome dimensions relative to the baseline. Results We find that ‘win-win-win’ futures cannot be achieved by internal changes to the aggregation scheme alone, emerging under a narrow range of scenarios that boost supplies to the small retail market whilst simultaneously supporting the financial takeaways of farmers. In contrast, undesirable producer versus consumer trade-offs emerge as unintended consequences of scaling-up aggregation and the introduction of market-based cold storage. Significance This approach furthers ongoing efforts to capture complex value chain processes, outcomes and upgrades within system dynamics modelling frameworks, before scanning the horizon of plausible external scenarios, internal dynamics and unintended trade-offs to identify ‘win-win-win’ futures for all.
... We applied a clustering procedure, as described by Janssen et al. (2012) and Kok et al. (2010Kok et al. ( , 2016, to identify archetypical patterns or profiles of risk, based on recurrent indicator combinations. This cluster-based approach was previously applied by Sietz et al. (2011Sietz et al. ( , 2012Sietz et al. ( , 2017, Kok et al. (2010) and Vidal Merino et al. (2019. ...
Article
Full-text available
More frequent and intense climate hazards, a predicted outcome of climate change, are likely to threaten existing livelihoods in rural communities, undermining households’ adaptive capacity. To support households’ efforts to manage and reduce this risk, there is the need to better understand the heterogeneity of risk within and between communities. The Intergovernmental Panel on Climate Change revised their climate vulnerability framework to incorporate the concept of risk. This study contributes towards the operationalization of this updated framework by applying a recognized methodology to the analysis of the climate-related risk of rural households. Using a mixed-method approach, including a cluster analysis, it determined and assessed archetypical patterns of household risk. The approach was applied to 170 households in two villages, in different agro-ecological zones, in the Vhembe district municipality of South Africa’s Limpopo province. Six archetypical climate-risk profiles were identified based on differences in the core components of risk, namely the experience of climate hazards, the degree of exposure and vulnerability, and the associated impacts. The method’s application is illustrated by interpreting the six profiles, with possible adaptation pathways suggested for each. The archetypes show how climate-related risk varies according to households’ livelihood strategies and capital endowments. There are clear site-related distinctions between the risk profiles however the age of the household and the gender of the household head also differentiate the profiles. These different profiles suggest the need for adaptation responses that account for these site-related differences, while still recognizing the heterogeneity of risk at the village level.
... Data processing [25] of GED of variant set of genes are also one of the factors to consider to obtain functionally [26] similar groups of genes. Thus different types of clustering principles (like Hierarchical, Simple K-means and DBSCAN etc.) are used to obtain similar set of genes as given in Fig. 1 [27]. The DB-Index and Dunn-index are used for the validation of obtained clusters. ...
Article
Full-text available
Accurate brain tumor segmentation plays a significant role in the area of radiotherapy diagnosis and in the proper treatment for brain tumor detection. Typically, the brain tumor has poor boundary and low contrast between normal and lesion soft tissues that makes segmentation of brain tumor in the CT images a challenging task. This paper presents a novel approach to brain image segmentation using pulse-coupled neural network (PCNN) and zero level set (ZL) with Sobolev gradient (SG) method. In this article, PCNN is designed to use as an edge mapper to provide a regional description for the ZL to segregate the CT images based on contour maps. The PCNN is used to estimate the exact threshold to obtain the prominent edges of the images. Resulting edges are utilized in the ZL to extract image contour from the source image. Due to the over-sensitivity of the ZL method on the initial contour, a level set with the SG has been equipped to overcome the limitation of the ZL method. The experimental results show satisfactory segmentation outcomes with excellent accuracy and acceleration in comparison with the state-of-the-art methods.
... For the cluster analysis a two-step cluster procedure, previously used to identify archetypes of farmers' vulnerability in the southern Peruvian Andes (Sietz et al. 2012), was used (as described by Janssen et al. 2012) to identify archetypes of FHS vulnerability to climate-related hazards. Despite potential variations on the relative importance of indicators, all indicators were given an equal weighting as a first approximation in the present analysis. ...
Chapter
Agriculture in the Andes is subject to multiple climate-related risks, typical of complex mountain ecosystems. Most of the strategies used to confront or reduce these risks are based on the adaptive capital of the farm households, such as the availability of labor, extension, and distribution of agricultural land, access to markets, among others. In order to increase the adaptive capacity of farm households, it is first necessary to understand the heterogeneity of the factors that explain their vulnerability. This article presents an analysis of archetypes (patterns) of climate vulnerability based on empirical data of farm household systems in the central Andes of Peru. The study uses mixed methods, combining qualitative tools and quantitative techniques, including cluster analysis. The results demonstrate the suitability of the methodology for explaining the vulnerability of farm household systems to climate-related hazards. For the case study, seven factors explain differences in vulnerability between five archetypes of agricultural households, including agricultural area, availability of irrigation, use of different agro-ecological zones, and access to non-agricultural employment.
... For this purpose we applied a formalized method based on clustering [25,26]. This method has been successfully applied before to identify and interpret socio-ecological problems which similarly generate vulnerability in global drylands on regional and local scales [18,[27][28][29]]. ...
... 3. Identification of vulnerability profiles and their spatial distribution We submit these indicators to a cluster analysis [26,30] to determine in how far typical indicator value combinations, which resemble vulnerability-creating mechanisms occur. Thereby the center of each resulting cluster signifies an urban vulnerability profile. ...
... We applied the established cluster analysis method k-means to integrate the 11 datasets and to identify typical combinations in the data structure [26,30]. This method is suitable for large datasets and generates rather compact clusters [78] which fits our objective to identify typical vulnerability profiles, represented by the cluster center. ...
Article
Full-text available
Coastal areas are urbanizing at unprecedented rates, particularly in low- and middle-income countries. Combinations of long-standing and emerging problems in these urban areas generate vulnerability for human well-being and ecosystems alike. This baseline study provides a spatially explicit global systematization of these problems into typical urban vulnerability profiles for the year 2000 using largely sub-national data. Using 11 indicator datasets for urban expansion, urban population growth, marginalization of poor populations, government effectiveness, exposures and damages to climate-related extreme events, low-lying settlement, and wetlands prevalence, a cluster analysis reveals a global typology of seven clearly distinguishable clusters, or urban profiles of vulnerability. Each profile is characterized by a specific data-value combination of indicators representing mechanisms that generate vulnerability. Using 21 studies for testing the plausibility, we identify seven key profile-based vulnerabilities for urban populations, which are relevant in the context of global urbanization, expansion, and climate change. We show which urban coasts are similar in this regard. Sensitivity and exposure to extreme climate-related events, and government effectiveness, are the most important factors for the huge asymmetries of vulnerability between profiles. Against the background of underlying global trends we propose entry points for profile-based vulnerability reduction. The study provides a baseline for further pattern analysis in the rapidly urbanizing coastal fringe as data availability increases. We propose to explore socio-ecologically similar coastal urban areas as a basis for sharing experience and vulnerability-reducing measures among them.