An application of spectral clustering approach to detect communities in data modeled by graphs

Mouden, Zakariyaa Ait El; Jakimi, Abdeslam; Hajar, Moha

doi:10.1145/3320326.3320330

Cited by 9 publications

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work presented in this paper is a part of a research project of knowledge extraction from data modeled by graphs [6,7] where the main objective is the classify a set of individuals using graph approaches such as spectral clustering algorithms [8][9][10][11].…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

A New Algorithm for Storing and Migrating Data Modelled by Graphs

Mouden

Jakimi

2020

Int. J. Onl. Eng.

Self Cite

View full text Add to dashboard Cite

<span>NoSQL databases have moved from theoretical solutions to exceed relational databases limits to a practical and indisputable application for storing and manipulation big data. In term of variety, NoSQL databases store heterogeneous data without being obliged to respect a predefined schema such as the case of relational and object-relational databases. NoSQL solutions surpass the traditional databases in storage capacity; we consider MongoDB for example, which is a document-oriented database capable of storing unlimited number of documents with a maximal size of 32TB depending on the machine that runs the database and also the operating system. Also, in term of velocity, many researches compared the execution time of different transactions and proved that NoSQL databases are the perfect solution for real-time applications. This paper presents an algorithm to store data modeled by graphs as NoSQL documents, the purpose of this study is to exploit the high amount of data stored in SQL databases and to make such data usable by recent clustering algorithms and other data science tools. This study links relational data to document datastores by defining an effective algorithm for reading relational data, modelling those data as graphs and storing those data as NoSQL documents.</span>

show abstract

Section: Related Workmentioning

confidence: 99%

“…Where n is the order of the graph (n = |V|) and each vertex vi (i ϵ [1, n]) is defined as follows (6).…”

Section: T = Directed | Undirected (3)mentioning

confidence: 99%

See 1 more Smart Citation

A New Algorithm for Storing and Migrating Data Modelled by Graphs

Mouden

Jakimi

2020

Int. J. Onl. Eng.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Other works [15,16] propose multi-class classifiers to predict normal, bacteria and Covid-19 from chest X-ray images. Recent works are considering graph analytics as a powerful tool for tracking Covid-19 as a complex network, using graph-based machine learning algorithms such as spectral clustering [17] and Graph neural networks [19].…”

Section: Related Workmentioning

confidence: 99%

Towards Using Recurrent Neural Networks for Predicting Influenza-like Illness: Case Study of Covid-19 in Morocco

Taj¹,

Mouden²,

Jakimi³

et al. 2020

IJATCSE

View full text Add to dashboard Cite

The influenza does not affect people's health only, but it is also an essential topic of governments and health care facilities. Early analysis, prediction and response is the most effective control method for flu epidemics. The Artificial Intelligence (AI) scientists are conducting their efforts to develop supervised and unsupervised models in order to analyse epidemics. In this paper, we present the most used Machine Learning (ML) and Deep Learning (DL) models in order to understand Covid-19'sbehaviour by analysing time series data. Among several algorithms of ML, Recurrent Neural Network (RNN) was chosen for tracking this epidemic and predicting its future outbreak. Since the appearance of the first case of COVID-19 in Morocco, the cumulative number of reported infectious cases continues to increase, however this number varies according to the regions of the Kingdom. Also, in this paper, we propose an analysis and prediction model of influenza-like illness Covid-19 by regional distribution. The proposed model is further used to obtain statistical summaries.

show abstract

“…Classification techniques such as multiclass classifiers are widely used to classify images into a predefined set of classes, also called output layers in Neural networks [5], those techniques give good results for Covid-19 detection such as the application to X-ray images [6,7] or the classification of the infected areas by regions especially using Length Short-Term Memory (LSTM) networks [4,8]. While, clustering techniques are widely used for grouping data into unknown set of clusters according to their similar behaviors such as the behavior of the Covid-19's outbreak in function of time and other factors [9,10].…”

Section: Introductionmentioning

confidence: 99%

A Graph-based Methodology for Tracking Covid-19 in Time Series Datasets

Mouden

Jakimi

Taj

et al. 2020

2020 IEEE 2nd International Conference on Electronics, Control, Optimization and Computer Science (ICECOCS)

Self Cite

View full text Add to dashboard Cite

Covid-19 has become a wide field of scientific research. Starting from biology to bioinformatic solutions, Artificial Intelligence has contributed in turn as a powerful tool for tracking and predicting the outbreak of Covid-19 using different types of datasets. Chest X-ray images are widely used in computer vision applications and Time series datasets are used for predicting the spread of the novel coronavirus. Graph analytics is a recent field of study that links the mathematical definition and operations of a graph to its application in computer science as a complex data structure, this combination has played a critical role in making graph-based applications present in different fields. One of the most powerful graph analytics is community detection which is an intelligent and unsupervised grouping of a set of graph structured data using the similarity between them. The aim of this work is to highlight the importance of graph-based algorithms in tracking Covid-19 using time series datasets, our work will also focus on Spectral Clustering (SC) as a community detection approach to extract clusters from the input datasets. Further applications are needed in order to validate the proposed theoretical approach.

show abstract

An application of spectral clustering approach to detect communities in data modeled by graphs

Cited by 9 publications

References 13 publications

A New Algorithm for Storing and Migrating Data Modelled by Graphs

A New Algorithm for Storing and Migrating Data Modelled by Graphs

Towards Using Recurrent Neural Networks for Predicting Influenza-like Illness: Case Study of Covid-19 in Morocco

A Graph-based Methodology for Tracking Covid-19 in Time Series Datasets

Contact Info

Product

Resources

About