Olivia Roxana Alecsoiu

Dr. Olivia Roxana Alecsoiu

Universitatea Constantin Brancusi, Târgu Jiu, Gorj, Romania
orcid

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
7 July 2025 in IEEE Access
Verified
Journal Article
Open access
NRNidadavolu Venkat Durga Sai Siva Vara Prasad Raju
Nidadavolu Venkat Durga Sai Siva Vara Prasad Raju
NFNuruzzaman Faruqui
Nuruzzaman Faruqui
NPNikhil Patel
Nikhil Patel
OAOlivia Roxana Alecsoiu
Olivia Roxana Alecsoiu
PTPriyabrata Thatoi
Priyabrata Thatoi
AAA. K. M. Azad
A. K. M. Azad
The legal industry struggles with inefficiencies, high costs, and manual-intensive workflows. Traditional AI lacks adaptability in optimizing legal operations. To address this, we propose LegalMind, an agentic AI-driven framework leveraging DeepSeek R1 for intelligent legal process automation and cost reduction. LegalMind integrates a structured legal dataset and fine-tunes DeepSeek R1 to enhance decision-making and workflow efficiency. Experimental results show a 42.6% cost reduction and a 60.8% improvement in document processing speed over baseline AI models. Scalability tests confirm the system’s ability to handle 100,000 queries efficiently. Real-world case studies validate LegalMind’s effectiveness in law firms, corporate legal departments, and government agencies, demonstrating significant reductions in case preparation time and operational costs. These findings highlight the transformative potential of agentic AI in legal automation, optimizing workflows and improving decision support.

14 April 2025 in IEEE Access
Verified
Journal Article
Open access
OAOlivia Roxana Alecsoiu
Olivia Roxana Alecsoiu
NFNuruzzaman Faruqui
Nuruzzaman Faruqui
APAndreea Anamaria Panagoret
Andreea Anamaria Panagoret
CICeauşescu Aurelian Ionuţ
Ceauşescu Aurelian Ionuţ
DPDragoş Mihai Panagoret
Dragoş Mihai Panagoret
MMMiruna Angela Mutu
Miruna Angela Mutu
The rapid expansion of e-commerce has increased the complexity of operational processes, making it challenging to manage tasks such as product cataloging, customer responses, delivery updates, and feedback collection efficiently. These challenges often result in elevated operational costs and decreased customer satisfaction. This paper introduces EcoptiAI, an agentic AI-powered framework leveraging a transformer-based model as an intelligent agent to automate essential e-commerce processes. EcoptiAI minimizes manual effort, streamlines workflows, and optimizes procedural costs by handling tasks such as product description generation, catalog updates, personalized customer communication, and empathetic delivery status updates. The system employs an empathetic tone for delay notifications, distinguishing itself from standard cold responses, and automates customer feedback analysis, ensuring an enhanced customer experience. The transformer model underlying EcoptiAI has been trained using a uniquely structured dataset created from diverse e-commerce-related sources, including product catalogs, customer reviews, and operational logs. The experimental analysis demonstrates that EcoptiAI reduces procedural costs by 52.7% on average and achieves high-performance metrics, with an accuracy of 92.42%, precision of 92.44%, recall of 92.40%, and an F1-score of 92.41%. The findings indicate the transformative potential of agentic AI in driving cost-effective, automated e-commerce operations while enhancing customer satisfaction. This paper provides a comprehensive evaluation of EcoptiAI’s design, implementation, and impact, paving the way for scalable and intelligent e-commerce automation solutions.


29 August 2025 in Behavioral Sciences
Verified
Journal Article
Open access
PDF file
CPCătălin Peptan
Cătălin Peptan
FMFlavius Cristian MARCAU
Flavius Cristian MARCAU
OAOlivia Roxana Alecsoiu
Olivia Roxana Alecsoiu
DPDragos Mihai Panagoret
Dragos Mihai Panagoret
MCMarian Emanuel Cojoaca
Marian Emanuel Cojoaca
GVGheorman Victor
Gheorman Victor
This study examines the evolution of willingness to accept the monkeypox (Mpox) vaccine in Romania between 2022 and 2025. It explores key sociodemographic and behavioral predictors of vaccine acceptance and investigates how public perceptions—particularly concerning disease severity and conspiracy beliefs—have shifted across two independent cross-sectional samples. Two nationally distributed surveys were conducted in July 2022 (n = 820) and January–February 2025 (n = 1029), targeting Romanian residents aged 18 and above. Data were analyzed using descriptive statistics, Chi-square tests, Kolmogorov–Smirnov tests, and a Random Forest classification model to assess the relative importance of predictors of vaccine acceptance. Between 2022 and 2025, vaccine acceptance increased modestly, particularly among individuals aged 36–65 and those with prior experience of voluntary or COVID-19 vaccination. Random Forest analysis identified behavioral factors as the strongest predictors of acceptance in both years, while the influence of education and gender varied over time. Belief in conspiracy theories slightly declined and lost predictive relevance by 2025. Perceptions of pandemic potential and fear of infection also decreased, suggesting reduced risk salience and possible pandemic fatigue. Despite a slight upward trend, overall Mpox vaccine acceptance in Romania remains among the lowest in Europe. These findings highlight the need for targeted public health communication, particularly toward skeptical or demographically vulnerable groups. Prior vaccination behavior emerged as a key driver of acceptance, indicating that trust-building strategies should capitalize on existing pro-vaccination habits. Future research should adopt qualitative and longitudinal approaches to better capture the evolving psychosocial dynamics of vaccine hesitancy.

26 October 2022 in Sustainability
Verified
Journal Article
Open access
PDF file
OAOlivia Roxana Alecsoiu
Olivia Roxana Alecsoiu
CMChirculescu Maria-Felicia
Chirculescu Maria-Felicia
APAndreea Panagoret
Andreea Panagoret
The continued interest in this research topic can be justified by the fact that a direct correlation between financial performance and managerial performance has not yet been reached at the level of an entity in general and implicitly at the level of an energy entity. The assumed objective of the research was to make a theoretical, methodological, and empirical evaluation of the cause-effect relationships between financial performance and managerial performance. Regarding the working instruments for data collection and interpretation, the most commonly used research method was the comparative method used both in the approach of theoretical aspects and during the empirical research carried out. In addition, longitudinal methods were used both by presenting the evolution in time of the concepts debated in the work, as well as by the evolution in time of the performance indicators. At the end of the work, the observation method was used to draw partial conclusions. The results of the research confirm the hypothesis according to which the decision to implement the organizational change at the level of the entity has a significant contribution to the efficiency of the activity, as can be seen from the evolution of the net result.

26 June 2026 in Computational Economics
Verified
Journal Article
Open access
OAOlivia Roxana Alecsoiu
Olivia Roxana Alecsoiu
Dumitrescu CatalinDumitrescu Catalin
Dumitrescu Catalin
ATAlina Iuliana Tabirca
Alina Iuliana Tabirca
Florin RADUFlorin RADU
Florin RADU
EVEmilia Vasile
Emilia Vasile
Valentin RaduValentin Radu
Valentin Radu
The system's automated trading decisions in the presented multi-agent system are based on predictions regarding the evolution of a product's trend, generated by intelligent agents that use artificial intelligence methods in their analysis. The primary objective of the research presented in this article is to conduct the necessary study to design, implement, and realize a multi-agent transaction system in a real-world environment that automates electronic commerce activities. The benefits of automating e-commerce activities are defined by a well-trained system that reacts much faster to changes in market trends than the human factor, and its decisions are not affected by the inconsistency that often characterizes human decisions. The results of implementing the automatic multi-agent trading system demonstrate that an architecture based on intelligent agents can serve as a basis for creating e-commerce-specific applications for financial transactions.