Hamza Toulni

Hamza Toulni

Auto-generated profile, based on publicly available information

Hamza Toulni

EIGSI, 282 Route of the Oasis, Mâarif, Casablanca 20140, Morocco

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
6 July 2022 in Mathematics
Verified
Journal Article
Open access
PDF file
CKCédric Stéphane TEKOUABOU KOUMETIO
Cédric Stéphane TEKOUABOU KOUMETIO
Ştefan Cristian GherghinaŞtefan Cristian Gherghina
Ştefan Cristian Gherghina
HTHamza Toulni
Hamza Toulni
PMPedro Neves Mata
Pedro Neves Mata
JMJosé Moleiro Martins
José Moleiro Martins
The diversity of data collected on both social networks and digital interfaces is extremely increased, raising the problem of heterogeneous variables that are not often favourable to classification algorithms. Despite the significant improvement in machine learning (ML) and predictive analysis efficiency for classification in customer relationship management systems (CRM), their performance remains very limited by heterogeneous data processing, class imbalance, and feature scales. This impact turned out to be more important for simple ML methods which in addition often suffer from over-fitting. This paper proposes a succinct and detailed ML model building process including cross-validation of the combination of SMOTE to balance data and ensemble methods for modelling. From the conducted experiments, the random forest (RF) model yielded the best performance of 0.86 in terms of accuracy and f1-scoreusing balanced data. It confirms the literature summary about this topic which shows that RF was among the most effective algorithms for customer predictive classification issues. The constructed and optimized models were interpreted by Shapley values and feature importance analysis which shows that the “age” feature was the most significant while “HasCrCard” was the less one. This process has proven effective in bridging previously reported research gaps and the resulting model should be used for supporting bank customer loyalty decision-making.

Verified
Journal Article
Open access
PDF file
CKCédric Stéphane TEKOUABOU KOUMETIO
Cédric Stéphane TEKOUABOU KOUMETIO
Ştefan Cristian GherghinaŞtefan Cristian Gherghina
Ştefan Cristian Gherghina
HTHamza Toulni
Hamza Toulni
PMPedro Neves Mata
Pedro Neves Mata
MMMário Nuno Mata
Mário Nuno Mata
JMJosé Moleiro Martins
José Moleiro Martins
The use of machine learning (ML) methods has been widely discussed for over a decade. The search for the optimal model is still a challenge that researchers seek to address. Despite advances in current work that surpass the limitations of previous ones, research still faces new challenges in every field. For the automatic targeting of customers in a banking telemarketing campaign, the use of ML-based approaches in previous work has not been able to show transparency in the processing of heterogeneous data, achieve optimal performance or use minimal resources. In this paper, we introduce a class membership-based (CMB) classifier which is a transparent approach well adapted to heterogeneous data that exploits nominal variables in the decision function. These dummy variables are often either suppressed or coded in an arbitrary way in most works without really evaluating their impact on the final performance of the models. In many cases, their coding either favours or disfavours the learning model performance without necessarily reflecting reality, which leads to over-fitting or decreased performance. In this work, we applied the CMB approach to data from a bank telemarketing campaign to build an optimal model for predicting potential customers before launching a campaign. The results obtained suggest that the CMB approach can predict the success of future prospecting more accurately than previous work. Furthermore, in addition to its better performance in terms of accuracy (97.3%), the model also gives a very close score for the AUC (95.9%), showing its stability, which would be very unfavourable to over-fitting.

28 November 2022 in Data
Verified
Journal Article
Open access
PDF file
CKCédric Stéphane TEKOUABOU KOUMETIO
Cédric Stéphane TEKOUABOU KOUMETIO
Jerome ChenalJerome Chenal
Jerome Chenal
ARAZMI Rida
AZMI Rida
HTHamza Toulni
Hamza Toulni
EDEl Bachir DIOP
El Bachir DIOP
Anastasija NikiforovaAnastasija Nikiforova
Anastasija Nikiforova
With the increase in the amount and variety of data that are constantly produced, collected, and exchanged between systems, the efficiency and accuracy of solutions/services that use data as input may suffer if an inappropriate or inaccurate technique, method, or tool is chosen to deal with them. This paper presents a global overview of urban data sources and structures used to train machine learning (ML) algorithms integrated into urban planning decision support systems (DSS). It contributes to a common understanding of choosing the right urban data for a given urban planning issue, i.e., their type, source and structure, for more efficient use in training ML models. For the purpose of this study, we conduct a systematic literature review (SLR) of all relevant peer-reviewed studies available in the Scopus database. More precisely, 248 papers were found to be relevant with their further analysis using a text-mining approach to determine (a) the main urban data sources used for ML modeling, (b) the most popular approaches used in relevant urban planning and urban problem-solving studies and their relationship to the type of data source used, and (c) the problems commonly encountered in their use. After classifying them, we identified the strengths and weaknesses of data sources depending on several predefined factors. We found that the data mainly come from two main categories of sources, namely (1) sensors and (2) statistical surveys, including social network data. They can be classified as (a) opportunistic or (b) non-opportunistic depending on the process of data acquisition, collection, and storage. Data sources are closely correlated with their structure and potential urban planning issues to be addressed. Almost all urban data have an indexed structure and, in particular, either attribute tables for statistical survey data and data from simple sensors (e.g., climate and pollution sensors) or vectors, mostly obtained from satellite images after large-scale spatio-temporal analysis. The paper also provides a discussion of the potential opportunities, emerging issues, and challenges that urban data sources face and should overcome to better catalyze intelligent/smart planning. This should contribute to the general understanding of the data, their sources and the challenges to be faced and overcome by those seeking data and integrating them into smart applications and urban-planning processes.

1 March 2022 in Expert Systems with Applications
Verified
Journal Article
Closed access
CKCédric Stéphane TEKOUABOU KOUMETIO
Cédric Stéphane TEKOUABOU KOUMETIO
EAEl Arbi Abdellaoui Alaoui
El Arbi Abdellaoui Alaoui
ICImane Chabbar
Imane Chabbar
HTHamza Toulni
Hamza Toulni
WCWalid Cherif
Walid Cherif
SHSilkan Hassan
Silkan Hassan

1 May 2014
Verified
Conference Paper
Closed access
HTHamza Toulni
Hamza Toulni
BNBenayad Nsiri
Benayad Nsiri
MBMohammed Boulmalf
Mohammed Boulmalf
MBMohammed Bakhouya
Mohammed Bakhouya
TSTayeb Sadiki
Tayeb Sadiki
Roads' overload remains a major challenge although they are constantly expanded in order to increase their capacity and improve their efficiency. This tends to cause many problems: environmentally, economically and socially. The advancement of urban traffic management systems, construction and expansion of roads are solutions to address this problem and improve the performance of the road circulation, but these solutions are very expensive and also require more space for the development and permanent maintenance. An alternative would be to use the new technologies in the field of communication for sending traffic information such as treacherous road conditions and accident sites by communicating, for a more efficient use of existing infrastructure. In this paper, we present a new approach based on VANETs, which can help drivers have a better trip, by finding the optimal path to reduce travel time and fuel consumption. This approach aims to avoid congestion and queues in a more efficient and optimal use of existing road infrastructure by analyzing the useful and reliable traffic information collected in real time.