Ismail El Bazi

Ismail El Bazi

Auto-generated profile, based on publicly available information

Ismail El Bazi

Sultan Moulay Slimane University, Beni Mellal, Morocco

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

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

Representative Publications

View All Publications
1 August 2022 in International Journal of Electrical and Computer Engineering (IJECE)
Verified
Journal Article
Open access
PDF file
SMSoukaina Mihi
Soukaina Mihi
BABrahim Ait Ben Ali
Brahim Ait Ben Ali
IBIsmail El Bazi
Ismail El Bazi
SASara Arezki
Sara Arezki
NLNabil Laachfoubi
Nabil Laachfoubi
The heavy involvement of the Arabic internet users resulted in spreading data written in the Arabic language and creating a vast research area regarding natural language processing (NLP). Sentiment analysis is a growing field of research that is of great importance to everyone considering the high added potential for decision-making and predicting upcoming actions using the texts produced in social networks. Arabic used in microblogging websites, especially Twitter, is highly informal. It is not compliant with neither standards nor spelling regulations making it quite challenging for automatic machine-learning techniques. In this paper’s scope, we propose a new approach based on AutoML methods to improve the efficiency of the sentiment classification process for dialectal Arabic. This approach was validated through benchmarks testing on three different datasets that represent three vernacular forms of Arabic. The obtained results show that the presented framework has significantly increased accuracy than similar works in the literature.

1 September 2022 in Indonesian Journal of Electrical Engineering and Computer Science
Verified
Journal Article
Open access
PDF file
BABrahim Ait Ben Ali
Brahim Ait Ben Ali
SMSoukaina Mihi
Soukaina Mihi
IBIsmail El Bazi
Ismail El Bazi
NLNabil Laachfoubi
Nabil Laachfoubi
Named entity recognition is an essential task for various applications related to natural language processing (NLP). It aims to retrieve a variety of named entities (NEs) from text and categorize them according to predetermined target categories. In many cases, using the entire feature set can be time-consuming and negatively impact the performance. Moreover, it is challenging to find the relevant subsets of features for a particular task due to the high number. The feature selection technique is an unsupervised process for selecting informative features by creating a new subset of informative features. This technique is used to enhance the underlying algorithm's performance. This article implements an effective feature selection algorithm using particle swarm optimization (PSO) to identify and classify the Arabic NEs in the text from social media. PSO is a search algorithm that utilizes a population of particles in a multidimensional space. The proposed method is evaluated using two publicly available Arabic Dialect social media datasets. It is demonstrated through comparisons with both baselines and previous models that the new approach achieves significant accuracy with considerably reduced feature sets in all parameters.

1 February 2026 in Bulletin of Electrical Engineering and Informatics
Verified
Journal Article
Open access
BGBrahim Ghazoui
Brahim Ghazoui
IBIsmail El Bazi
Ismail El Bazi
IEIbtissam Essadik
Ibtissam Essadik
BBBrahim Ait Benali
Brahim Ait Benali
HMHicham Moussa
Hicham Moussa
Named entity recognition (NER) is vital for turning unstructured social media text into structured information. However, Arabic tweets pose distinct challenges; informality, brevity, dialectal variation, and inconsistent orthography. This study targets those challenges by coupling targeted data augmentation with a transformer model, bert-base-arabertv2. We design a lightweight augmentation pipeline—synonym replacement, name and location replacement, and deletion of third-person Arabic names—to expand linguistic variety and reduce overfitting under limited annotation. The approach is simple, but deliberate: preserve labels when substituting entities with type-consistent alternatives; remove corresponding tags when deleting names; and keep tweet semantics intact where possible. We then fine-tune bert-base-arabertv2 on the combined original and augmented data and evaluate on a held-out set of tweets. The result is a substantial gain in overall performance: F1=0.93 with augmentation versus 0.72 without. These findings indicate that controlled, label-aware augmentation can improve robustness and generalization for Arabic tweet NER, where data scarcity and linguistic variability otherwise degrade accuracy. Beyond empirical gains, our work offers a practical recipe—clear augmentation heuristics and a standard transformer backbone—that can be replicated and adapted to similar low-resource, noisy domains. This contributes to more reliable Arabic social media analysis and downstream information extraction.

17 October 2026
Verified
Book Chapter
Closed access
AOAbdallah Ousri
Abdallah Ousri
YJYoussef Jouali
Youssef Jouali
IBIsmail El Bazi
Ismail El Bazi
SGSoufiane Ghyati
Soufiane Ghyati
As states adopt platform-based models of public governance, understanding how institutional capabilities and governance arrangements translate into societal value has become a central concern for public administration research. No quantitative model had jointly tested how public digital capability, data governance, and platform governance interact to shape digital platform use and digital public value creation in emerging economies such as Morocco. This chapter develops and empirically validates an integrative tripartite structural model grounded in Public Value Theory, New Public Management, and Sociotechnical Theory, using PLS-SEM on a sample of 350 Moroccan public employees, civil servants, and citizens. All seven hypotheses were confirmed at p < 0.001: public digital capability and platform governance emerged as the strongest direct determinants of platform use, which fully mediates digital public value creation. The chapter offers theoretical and managerial insights for decision-makers engaged in digital transformation in emerging platform states.

15 October 2026
Verified
Book Chapter
Closed access
ITIkram Talha
Ikram Talha
ZRZakaria Rafae
Zakaria Rafae
IBIsmail El Bazi
Ismail El Bazi
ACAbdelmajid Charrass
Abdelmajid Charrass
Virtual influencers have created new opportunities for sustainability communication. However, existing research remains fragmented and lacks an integrated theoretical explanation of how virtual influencers shape sustainability-related consumer responses. This study addresses this gap through a theory-building literature review integrating the Theory–Context–Methodology (TCM) and Antecedents–Decisions–Outcomes (ADO) frameworks. Following a PRISMA-informed review process, five conceptually rich studies were synthesized to identify dominant theories, research contexts, methodological approaches, and key explanatory mechanisms. Building on this synthesis, the study proposes the Sustainable Influence Legitimacy Framework (SILF), which explains how virtual influencer characteristics and sustainability message design shape consumer evaluations and sustainability outcomes. By incorporating moral legitimacy, AI transparency, resistance mechanisms, and the intention–behavior gap, SILF provides an integrated theoretical foundation and a future research agenda.

29 July 2026
Verified
Book Chapter
Closed access
IBIsmail El Bazi
Ismail El Bazi
JEJamal Ezzine
Jamal Ezzine
BGBrahim Ghazoui
Brahim Ghazoui
This chapter presents an agentic intelligence framework for autonomous phishing detection and response in the context of evolving cyber warfare strategies. It introduces the Autonomous Phishing Detection and Action (APDA) architecture, which integrates large language models with perception, reasoning, action, and feedback layers to enable adaptive and explainable decision-making. This chapter examines how adversarial techniques, such as prompt injection, jailbreaking, hallucination, and data poisoning, threaten LLM-based cybersecurity systems, and outlines the corresponding mitigation strategies embedded within each architectural layer. A structured evaluation framework is proposed, including technical and operational metrics, alongside an illustrative enterprise email threat scenario. This chapter contributes a novel layered approach that bridges advances in natural language processing with autonomous cyber defense, offering a scalable and governance-aware solution for next-generation security operations.