Evolutionary Computing Assisted K-Means Clustering based MapReduce Distributed Computing Environment for IoT-Driven Smart City

Srinivas, Kunal G.; Hosahalli, Doreswamy

doi:10.1109/icccis51004.2021.9397217

Cited by 4 publications

(1 citation statement)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This study identified the similarities in the characteristics of each object in the Aceh regions. In addition, Srinivas and Hosahalli studied a MapReduce distributed computing environment based on clustering using K-means evolutionary computation for an IoT-based smart city [ 21 ].The use of cluster analysis in this work evokes associations with innovation clusters, since the concept of smart city largely relies on innovation implementation. Brakman and van Marrewijk believe that the effect of clusters in cities lies in their impact on the national and regional economies of cities, as their effect has a favorable impact on the existing mechanism in agglomerations [ 22 ] Van Klink and de Langen believe that the city cluster must go through a mandatory life cycle of innovative product development.…”

Section: Literature Reviewmentioning

confidence: 99%

Discovering smart cities’ potential in Kazakhstan: A cluster analysis

Urdabayev,

Kireyeva,

Vasa

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

The potential for developing smart cities in Kazakhstan is evaluated using cluster analysis. Built on previous research focused on clustering the regions of Kazakhstan, this study applies the same method to the cities of the country. The analysis uses indicators related to human capital, infrastructure, education, information technology, production, and other factors to assess the potential of each city. The clustering is performed using Single Linkage, Complete Linkage, and Ward’s methods. The results show that Almaty and Astana are the cities with the highest potential for becoming smart cities. Aktobe is identified as a city with distinctive features that may help or hinder its development as a smart city. The remaining cities are clustered into two groups, with one group having the potential to catch up and maintain the trend of developing smart cities, while the other group is less suitable for starting smart city projects and may require more investment per capita. The study highlights the deep regional inequality affecting the potential to successfully develop and manage smart cities in Kazakhstan. The analysis also reveals some limitations and challenges in the data and variables used, including the lack of data for some variables and the difficulties in "translating" some factors and indicators into quantitative variables for clustering. The study concludes that future research should address these challenges and consider clustering inside certain regions to focus on their unique features. The study recommends launching pilot projects in small cities, with the most successful practices then scaled and implemented in the core smart cities and possibly Aktobe, if it manages to use its advantages to compensate for risks. Overall, this study provides insights into the potential of smart city development in Kazakhstan and can inform policymakers in their efforts to support smart city projects in the country.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Discovering smart cities’ potential in Kazakhstan: A cluster analysis

Urdabayev,

Kireyeva,

Vasa

et al. 2024

PLoS ONE

View full text Add to dashboard Cite

show abstract

Reflecting on a Decade of Evolution: MapReduce‐Based Advances in Partitioning‐Based, Hierarchical‐Based, and Density‐Based Clustering (2013–2023)

Sardar

2024

WIREs Data Min & Knowl

View full text Add to dashboard Cite

The traditional clustering algorithms are not appropriate for large real‐world datasets or big data, which is attributable to computational expensiveness and scalability issues. As a solution, the last decade's research headed towards distributed clustering using the MapReduce framework. This study conducts a bibliometric review to assess, establish, and measure the patterns and trends of the MapReduce‐based partitioning, hierarchical, and density clustering algorithms over the past decade (2013–2023). A digital text‐mining‐based comprehensive search technique with multiple field‐specific keywords, inclusion measures, and exclusion criteria is employed to obtain the research landscape from the Scopus database. The Scopus‐obtained data is analyzed using the VOSViewer software tool and coded using the R statistical analysis tool. The analysis identifies the numbers of scholarly articles, diversities of article sources, their impact and growth patterns, details of most influential authors and co‐authors, most cited articles, most contributing affiliations and countries, and their collaborations, use of different keywords and their impact, and so forth. The study further explores the articles and reports the methodologies employed for designing MapReduce‐based counterparts of traditional partitioning, hierarchical, and density clustering algorithms and their optimizations and hybridizations. Finally, the study lists the main research challenges encountered in the past decade for MapReduce‐based partitioning, hierarchical, and density clustering. It suggests possible areas for future research to contribute further in this field.

show abstract