Orgeta Gjermeni

@univlora.edu.al

Department of Mathematics and Physics/Faculty of Technical and Natural Sciences
University "IsmailQemali" Vlore

Orgeta Gjermeni


                     Download Compact PDF  
https://researchid.co/orgetagjermni

RESEARCH, TEACHING, or OTHER INTERESTS

Statistics and Probability, Multidisciplinary, Applied Mathematics, Modeling and Simulation
6

Scopus Publications

33

Scholar Citations

3

Scholar h-index

1

Scholar i10-index

Scopus Publications

  • Overfitting Dynamics in Recurrent Neural Networks: A Statistical and Experimental Approach
    Orgeta Gjermëni
    Interdisciplinary Journal of Research and Development, 2026
    Overfitting remains a key challenge in applying Recurrent Neural Networks (RNNs) to sequential forecasting tasks, including carbon emissions modeling. While Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) architectures are widely used, comparative analyses of their overfitting behavior across architectural variants remain limited. The present analysis addresses this gap by examining how structural variables such as model architecture, layer depth, architecture type, model family, and the number of hidden units, together with their two-way interactions with unit configuration, influence overfitting behaviour in RNNs applied to a univariate time series of carbon dioxide (CO2) emissions from land-use change in Albania. The dataset covers the period from 1850 to 2022. Preprocessing steps included Isolation Forest-based outlier detection, LSTM-based imputation, first differencing, Yeo-Johnson transformation, and Min-Max normalization. Six RNN architectures were evaluated, including single-layer models (GRU and LSTM) and their homogeneous and hybrid two-layer variants. Each architecture was trained across hidden unit values, resulting in 402 model instances under a unified configuration. Overfitting ratios were calculated as the ratio of test to training values for each of the four performance metrics: root mean squared error, symmetric mean absolute percentage error, mean absolute scaled error, and normalized mean absolute error. Their distributional properties were also assessed. A non-parametric multivariate analysis of variance was conducted to examine both main effects and two-way interactions, followed by pairwise comparisons within statistically significant structural factors. The results showed that model architecture, layer depth, architecture type, and model family significantly influence overfitting behaviour. Although the number of hidden units did not have a significant main effect, consistent interaction effects suggest that their impact depends on the architectural configuration. Pairwise comparisons revealed that hybrid and homogeneous architectures differed significantly from simple models. No significant difference was found between hybrid and homogeneous architectures, indicating greater similarity between the latter two. These findings emphasize the importance of aligning architectural design with hyperparameter selection when developing RNN-based forecasting models. Methodologically, the study demonstrates the utility of multivariate non-parametric analysis in characterizing generalization behavior. The research insights provided practical guidance for constructing more robust and generalizable RNNs for environmental forecasting and similar applications. Received: 5 January 2026 / Revised: 20 February 2026 / Accepted: 3 March 2026 / Published: 25 March 2026
  • Assessing Non-Linearity and Stationarity in the Time Series of Albania’s Annual Emissions of CO2 from Land-Use Change
    Science and Technology Asia, 2024
  • Exploring Landline Communication Dynamics in Albania: Insights from a Two Non-consecutive Month Comparative Study
    Orgeta Gjermëni
    Lecture Notes in Networks and Systems, 2024
  • Statistical analyses of impressions related to the customer’s experience while attending restaurants located in Vlore, Albania
    Science and Technology Asia, 2020
  • Tourist’s Satisfaction in Terms of Accommodation: A Case Study in Vlore, Albania
    Elenica Pjero (Beqiraj), Orgeta Gjermëni
    Business Perspectives and Research, 2020
    This study aims to explore tourist’s satisfaction on the accommodation provided during their stay in Vlore (Albania) touristic structures, and if there are possible associations between different characteristics related to this service and tourists. Lack of studies on analyzing customer satisfaction in the industry of accommodation, especially for Vlore, have prompted us to undertake this study. The study results are important for local government, the accommodation industry, and is a source of information for whom is interested to improve their accommodation services, or to invest in accommodation industry located in Vlore, Albania. “Netnography” is used to collect data for our research purpose from the reviews in TripAdvisor website. Using descriptive and inferential statistics, this study concludes that 64.9 percent of the ratings are “very good” or “excellent,” regardless of the accommodation structure chosen. Accommodation structures should have a clear defined idea of what kind of tourist they want to attract in a certain period of the year, in order to offer the quality tourists expect. Furthermore, understanding of tourist satisfaction evaluation is important in implementing successful marketing campaigns.
  • Temporal statistical analysis of degree distributions in an undirected landline phone call network graph series
    Orgeta Gjermëni
    Data, 2017
    This article aims to provide new results about the intraday degree sequence distribution considering phone call network graph evolution in time. More specifically, it tackles the following problem. Given a large amount of landline phone call data records, what is the best way to summarize the distinct number of calling partners per client per day? In order to answer this question, a series of undirected phone call network graphs is constructed based on data from a local telecommunication source in Albania. All network graphs of the series are simplified. Further, a longitudinal temporal study is made on this network graphs series related to the degree distributions. Power law and log-normal distribution fittings on the degree sequence are compared on each of the network graphs of the series. The maximum likelihood method is used to estimate the parameters of the distributions, and a Kolmogorov–Smirnov test associated with a p-value is used to define the plausible models. A direct distribution comparison is made through a Vuong test in the case that both distributions are plausible. Another goal was to describe the parameters’ distributions’ shape. A Shapiro-Wilk test is used to test the normality of the data, and measures of shape are used to define the distributions’ shape. Study findings suggested that log-normal distribution models better the intraday degree sequence data of the network graphs. It is not possible to say that the distributions of log-normal parameters are normal.

RECENT SCHOLAR PUBLICATIONS

  • Overfitting Dynamics in Recurrent Neural Networks: A Statistical and Experimental Approach
    O Gjermëni
    Interdisciplinary Journal of Research and Development 13 (1), 142-152 , 2026
    2026
  • Exploring the Latent Dimensions of STEM Orientation among Albanian Upper-Secondary Students
    O Gjermëni, M Ramosaçaj, E Kushta, A Denaj
    UniVlora Scientific Journal 1 (2), 122-136 , 2025
    2025
  • Assessing Non-Linearity and Stationarity in the Time Series of Albania’s Annual Emissions of CO2 from Land-Use Change
    O Gjermëni
    Science & Technology Asia 29 (4), 39-50 , 2024
    2024
    Citations: 1
  • Exploring Landline Communication Dynamics in Albania: Insights from a Two Non-consecutive Month Comparative Study
    O Gjermëni
    Innovative Computing and Communications. ICICC 2024. Lecture Notes in … , 2024
    2024
  • Likelihood of AI Tools Adoption and Interest in Professional Development Opportunities in Higher Education: An Ordinal Logistic Regression Analysis
    O Gjermëni
    The Eurasia Proceedings of Educational and Social Sciences 35, 217-229 , 2024
    2024
    Citations: 2
  • Unraveling the Impact of Service Quality on Hotel Ratings: A Comprehensive Study of Hotels and Aparthotels in Vlore County, Albania
    O Gjermëni
    The Albanian Journal of Economy & Business (ALJEB) 39 (5), 176-186 , 2024
    2024
    Citations: 1
  • ARTIFICIAL INTELLIGENCE PERCEPTIONS IN HIGHER EDUCATION: A COMPREHENSIVE ANALYSIS IN THE ALBANIAN CONTEXT
    O Gjermëni
    International Conference on New Research and Advances on Computer Science … , 2023
    2023
    Citations: 2
  • Statistical Analyses of Impressions Related to the Customer’s Experience while Attending Restaurants Located in Vlore, Albania
    O Gjermëni, E Pjero
    Science & Technology Asia 25 (1), 46-61 , 2020
    2020
    Citations: 2
  • Tourist’s satisfaction in terms of accommodation: A case study in Vlore, Albania
    E Pjero, O Gjermëni
    Business Perspectives and Research 8 (1), 67-80 , 2020
    2020
    Citations: 11
  • ASSORTATIVITY AND RECIPROCITY IN TEMPORAL LANDLINE PHONE CALL NETWORK GRAPH SERIES
    O Gjermëni
    Scientific Challenges for Sustainable Development (SCfSD18) 5 , 2018
    2018
  • ANALIZË STATISTIKORE MBI DINAMIKËN E GRAF RRJETAVE TË TELEFONISË FIKSE
    O Gjermëni
    https://upt.edu.al/wp-content/uploads/2018/05/Orgeta-Gjermeni-ANALIZE … , 2018
    2018
  • Temporal Statistical Analysis of Degree Distributions in an Undirected Landline Phone Call Network Graph Series
    O Gjermëni
    Data 2 (4), 33 , 2017
    2017
    Citations: 4
  • From Small World Phenomenon to Correlation Analysis in a Temporal Landline Phone Call Network Graph Series
    O Gjermëni, M Ramosaço
    International Journal of Applied Physics and Mathematics 7 (4), 275-283 , 2017
    2017
  • Survival Analysis on a Landline Phone Call Duration Data
    O Gjermëni
    International Bulletin of Mathematical Research (IBMR) 4 (3), 35-44 , 2017
    2017
  • SMALL WORLD PHENOMENON IN A LANDLINE PHONE CALL NETWORK GRAPH
    O Gjermëni
    Scientific Challenges for Sustainable Development 2017-SCfSD17 4, 283-297 , 2017
    2017
  • Descriptive analysis of characteristics: A case study of a phone Call network graph
    O Gjermëni, M Ramosaco
    2016
    Citations: 8
  • POWER LAW DISTRIBUTION AS A COMPONENT OF THE VERTEX DEGREE DISTRIBUTION ON A SOCIAL UNIVERSITY NETWORK COURSE
    O Gjermeni
    European Scientific Journal 11 (20) , 2015
    2015
  • Assessing Clustering in a Social University Network Course
    O Gjermëni
    2015
  • Power-Law versus Lognormal Distribution in a Phone Call Network Graph
    O Gjermëni, M Ramosaco, D Zotaj
    Proceedings of International Conference on Application of Information and … , 2015
    2015
    Citations: 2

MOST CITED SCHOLAR PUBLICATIONS

  • Tourist’s satisfaction in terms of accommodation: A case study in Vlore, Albania
    E Pjero, O Gjermëni
    Business Perspectives and Research 8 (1), 67-80 , 2020
    2020
    Citations: 11
  • Descriptive analysis of characteristics: A case study of a phone Call network graph
    O Gjermëni, M Ramosaco
    2016
    Citations: 8
  • Temporal Statistical Analysis of Degree Distributions in an Undirected Landline Phone Call Network Graph Series
    O Gjermëni
    Data 2 (4), 33 , 2017
    2017
    Citations: 4
  • Likelihood of AI Tools Adoption and Interest in Professional Development Opportunities in Higher Education: An Ordinal Logistic Regression Analysis
    O Gjermëni
    The Eurasia Proceedings of Educational and Social Sciences 35, 217-229 , 2024
    2024
    Citations: 2
  • ARTIFICIAL INTELLIGENCE PERCEPTIONS IN HIGHER EDUCATION: A COMPREHENSIVE ANALYSIS IN THE ALBANIAN CONTEXT
    O Gjermëni
    International Conference on New Research and Advances on Computer Science … , 2023
    2023
    Citations: 2
  • Statistical Analyses of Impressions Related to the Customer’s Experience while Attending Restaurants Located in Vlore, Albania
    O Gjermëni, E Pjero
    Science & Technology Asia 25 (1), 46-61 , 2020
    2020
    Citations: 2
  • Power-Law versus Lognormal Distribution in a Phone Call Network Graph
    O Gjermëni, M Ramosaco, D Zotaj
    Proceedings of International Conference on Application of Information and … , 2015
    2015
    Citations: 2
  • Assessing Non-Linearity and Stationarity in the Time Series of Albania’s Annual Emissions of CO2 from Land-Use Change
    O Gjermëni
    Science & Technology Asia 29 (4), 39-50 , 2024
    2024
    Citations: 1
  • Unraveling the Impact of Service Quality on Hotel Ratings: A Comprehensive Study of Hotels and Aparthotels in Vlore County, Albania
    O Gjermëni
    The Albanian Journal of Economy & Business (ALJEB) 39 (5), 176-186 , 2024
    2024
    Citations: 1
  • Overfitting Dynamics in Recurrent Neural Networks: A Statistical and Experimental Approach
    O Gjermëni
    Interdisciplinary Journal of Research and Development 13 (1), 142-152 , 2026
    2026
  • Exploring the Latent Dimensions of STEM Orientation among Albanian Upper-Secondary Students
    O Gjermëni, M Ramosaçaj, E Kushta, A Denaj
    UniVlora Scientific Journal 1 (2), 122-136 , 2025
    2025
  • Exploring Landline Communication Dynamics in Albania: Insights from a Two Non-consecutive Month Comparative Study
    O Gjermëni
    Innovative Computing and Communications. ICICC 2024. Lecture Notes in … , 2024
    2024
  • ASSORTATIVITY AND RECIPROCITY IN TEMPORAL LANDLINE PHONE CALL NETWORK GRAPH SERIES
    O Gjermëni
    Scientific Challenges for Sustainable Development (SCfSD18) 5 , 2018
    2018
  • ANALIZË STATISTIKORE MBI DINAMIKËN E GRAF RRJETAVE TË TELEFONISË FIKSE
    O Gjermëni
    https://upt.edu.al/wp-content/uploads/2018/05/Orgeta-Gjermeni-ANALIZE … , 2018
    2018
  • From Small World Phenomenon to Correlation Analysis in a Temporal Landline Phone Call Network Graph Series
    O Gjermëni, M Ramosaço
    International Journal of Applied Physics and Mathematics 7 (4), 275-283 , 2017
    2017
  • Survival Analysis on a Landline Phone Call Duration Data
    O Gjermëni
    International Bulletin of Mathematical Research (IBMR) 4 (3), 35-44 , 2017
    2017
  • SMALL WORLD PHENOMENON IN A LANDLINE PHONE CALL NETWORK GRAPH
    O Gjermëni
    Scientific Challenges for Sustainable Development 2017-SCfSD17 4, 283-297 , 2017
    2017
  • POWER LAW DISTRIBUTION AS A COMPONENT OF THE VERTEX DEGREE DISTRIBUTION ON A SOCIAL UNIVERSITY NETWORK COURSE
    O Gjermeni
    European Scientific Journal 11 (20) , 2015
    2015
  • Assessing Clustering in a Social University Network Course
    O Gjermëni
    2015