Prefacedoi: 10.1088/1742-6596/1743/1/011001pmid: N/A
The International Conference on Mathematics & Data Science (ICMDS’20), was held on June 29-30, 2020 as a virtual conference due to the growing concerns over the coronavirus outbreak (COVID-19), and in order to protect the well-being of our attendees, partners, and staff as our number one priority. the ICMDS'20 event aims in its first edition to gather international and national researchers, interdisciplinary scientists, professors, students and professionals seeking to exchange and discuss their ideas and inspirations on Applied Mathematics and Data Science. Besides proclaiming knowledge in the fields of applied mathematics and data science, ICMDS'20 strives to promote scientific research and innovation among junior and senior researchers.This scientific event brings together more than 300 national and international researchers in mathematics and data sciences. On top of the local participants coming from different national universities, international participants are also registered from different countries, namely France, India, the United States, the Comoros, Senegal, Algeria and Iraq. The event has four plenary conferences and around 140 communications that span over four parallel sessions which are well balanced in content. At the end of the conference, 42 papers were selected to be published in the journal of physics conference series (JPCS).List of committees are available in this Pdf.
Peer review declarationdoi: 10.1088/1742-6596/1743/1/011002pmid: N/A
All papers published in this volume of Journal of Physics: Conference Series have been peer reviewed through processes administered by the Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.• Type of peer review: Double-blindConference submission management system: Through our own website www.icmds.info• Number of submissions received: 210• Number of submissions sent for review: 180• Number of submissions accepted: 42• Acceptance Rate (Number of Submissions Accepted / Number of Submissions Received X 100): 20%• Average number of reviews per paper: 2• Total number of reviewers involved:25Any additional info on review process: Each selected paper passed through a screening process and plagiarism check. Afterwards, selected papers went through a double-blind review, two reports were made for each paper. Two positive reports mean that the paper is accepted.Contact person for queries: Amine Laghrib, FST Beni-Mellal University Sultan Moulay Slimane, [email protected]
IGAEM: Improved Genetic Algorithm based Entity MatchingAassem, Y.; Hafidi, I.; Aboutabit, N.
doi: 10.1088/1742-6596/1743/1/012001pmid: N/A
The presence of duplicate records is a major data quality concern in huge datasets. To detect duplicates, entity matching is used as an essential step of the data cleaning process to map records that refer to the same real-world entity. Most of proposed algorithms require labeled data in order to train a classifier. However, we cannot always obtain labeled data. In our paper we propose an unsupervised approach for entity matching problem using an improved version of genetic algorithm. We explain the main improvements added to genetic algorithm and the encoding strategy to encode partitions in the form of a chromosome. Different similarity functions are used to compute similarities between records. The obtained results prove that our proposition stands as a powerful approach in the entity matching field where it outperforms the traditional genetic algorithm based approach.
Comparative study of optimization techniques in deep learning: Application in the ophthalmology fieldMustapha, Aatila; Mohamed, Lachgar; Ali, Kartit
doi: 10.1088/1742-6596/1743/1/012002pmid: N/A
The optimization is a discipline which is part of mathematics and which aims to model, analyse and solve analytically or numerically problems of minimization or maximization of a function on a specific dataset. Several optimization algorithms are used in systems based on deep learning (DL) such as gradient descent (GD) algorithm. Considering the importance and the efficiency of the GD algorithm, several research works made it possible to improve it and to produce several other variants which also knew great success in DL. This paper presents a comparative study of stochastic, momentum, Nesterov, AdaGrad, RMSProp, AdaDelta, Adam, AdaMax and Nadam gradient descent algorithms based on the speed of convergence of these different algorithms, as well as the mean absolute error of each algorithm in the generation of an optimization solution. The obtained results show that AdaGrad algorithm represents the best performances than the other algorithms with a mean absolute error (MAE) of 0.3858 in 53 iterations and AdaDelta one represents the lowest performances with a MAE of 0.6035 in 6000 iterations. The case study treated in this work is based on an extract of data from the keratoconus dataset of Harvard Dataverse and the results are obtained using Python.
Identification of Complex Network Influencer using the Technology for Order Preference by Similarity to an Ideal SolutionAit Rai, K.; Agouti, T.; Machkour, M.; Antari, J
doi: 10.1088/1742-6596/1743/1/012004pmid: N/A
Marketing through social networks is a recent approach which consists in using these networks to convince potential consumers with the quality of products or services offered by a company. Marketing is developing very quickly, particularly on Facebook, Twitter, LinkedIn, Instagram, YouTube, etc. The major advantage of social networks is the possibility of influencing a panel of people according to their interests but without having the feeling of being guided. Identifying influencers is an interesting topic in social networks, and centrality measures are among the methods used to address this topic. Each measure has some shortcomings. In this paper, we gather centrality measures by using Technology for Order Preference by Similarity to an Ideal Solution (TOPSIS) method, which is a Multi-Criteria Decision Making (MCDM) to identify potential influences in a social network. A case study is presented to explain carefully TOPSIS and to illustrate the effectiveness of the proposed method, three real datasets are used for the experiments. The results show that TOPSIS can rank spreaders more accurately than centrality criteria.
KF-Swoosh: An Efficient Spark-Based Entity Resolution Algorithm for BigDataAlami, L.; Aassem, Y.; Hafidi, I.
doi: 10.1088/1742-6596/1743/1/012005pmid: N/A
Entity matching (EM), which is, the task of identifying records that refer to the same entity, is a critical task when constructing data warehouses. This task is often very expensive at the running time because data must be compared in pairs. This problem becomes more important when dealing with large-scale data. We propose a new parallel algorithm that divides the data using K-Medoid algorithm implemented with Spark framework. The computational experiments are done and show that we can improve the solution of a set of instances in a reduced execution time.
The Classification of the Environmental Indicators using ELECTRE TRI Method for Loukkos Basin in MoroccoAziz, Layla; Achki, Samira; Chalh, Ridouane.
doi: 10.1088/1742-6596/1743/1/012006pmid: N/A
Multi-Criteria Analysis (MCA) has found many applications in both technical and research sectors. MCA is a way to break the problem into more practicable elements, to permit data and decisions to be judgements to support the elements, and make the right decision. The aim of this paper is to analyze, compare, and make decisions of various current, and future scenarios of different quantifiable indicators for different considerations and various socio-economic aspects. Furthermore, this analysis is used to improve or at least to preserve the environment and natural resources in the basin. In this study an application by a real data set is made, these data are evaluated and extracted from classified satellite images of Loukkos basin, the classification of this satellite images regroups several classes of the data set such as agglomeration, dams, watercourses, croplands, bare soils and forests … etc. In reality, these data come from different sources like watershed information system (drinking water supply, irrigation system), transportation infrastructures (roads, dams), natural resources (Water, soils, and vegetation), human activities (agriculture, urbanization, and industry) and different socio-economic factors (demography). The main objective of this work is sorting the environmental indicators using the ELECTRE TRI tool, where ten alternatives are considered. We focus the classification of the real data set into the altitude, and the combined surface area factors. The obtained results prove that the classification is stable and the multi-criteria approach ELECTRE-TRI is suitable to a better sorting of the environmental indicators for the Loukkos Basin located in Morocco.
Transfer Learning for classifying front and rear views of vehiclesBaghdadi, Sara; Aboutabit, Noureddine
doi: 10.1088/1742-6596/1743/1/012007pmid: N/A
Various computer systems have been proposed to classify vehicles according to several criteria (category, brand, model). Unfortunately, there is not much research on the classification of views, especially front and rear views. Several factors make this classification very difficult including similarity in shape, size, and color. This work aims to classify front and rear views of vehicles using the Transfer Learning (TL) approach. Here, we used a pre-trained CNN (AlexNet) that has been trained on more than a million images and can classify images into 1000 object categories. Thus, we transferred its learned knowledge and applied it to our new task (Classifying vehicle views). We conducted then two experiments. The first experiment has two scenarios: the first scenario is devoted to Transfer Learning using the AlexNet model, and the second scenario aims to build a network from scratch inspired from AlexNet. Experimental results reveal that the Transfer Learning approach gives high results. On the other hand, in the second experiment, we decided to use TL-AlexNet to extract features and train them with an SVM classifier instead of fully connected layers. And also, we combined the SVM with the fully connected layers. The accuracy rates have been improved after this experiment.
Multi-Agent Reinforcement learning Approach to IoT CoordinationBelkeziz, Radia; Jarir, Zahi; El Kassmi, Ilyass
doi: 10.1088/1742-6596/1743/1/012008pmid: N/A
Nowadays, the IoT is evolving at a very fast pace and has proven its usefulness in several areas by creating better applications and services. However, more flexible approaches proving a well-defined architecture meeting the general requirements and building blocks of IoT are still needed despite the results obtained in the literature. In this paper, we focus on the IoT coordination challenge which represents a fundamental property allowing things to collaborate and make a decision when an appropriate change is detected in its environment. This contribution proposes an agent-based approach coupled with Q-learning which is a reinforcement learning technique, to compensate for coordination in its entirety, namely objective coordination and subjective coordination. To illustrate this approach, an evacuation use case is presented.