Vilnius Tech DSpace Repository
Not a member yet
52562 research outputs found
Sort by
BiLSTM-CNN with Bayesian Optimization for Accurate Long-Term Load Forecasting: Cross-Regional Insights from Morocco and Spain
No full textAccurate long-term electricity demand forecasting (LTLF) is critical for strategic planning, particularly in the context of escalating climate issues and the intricacies of energy systems. This research proffers a pioneering hybrid deep learning model that integrates Bidirectional Long Short-Term Memory (BiLSTM) networks and Convolutional Neural Networks (CNN), which is optimised using Bayesian Optimisation (BO), with a view to enhancing forecast reliability across diverse energy landscapes. Utilising hourly data from Morocco and Spain in 2017, the model captures significant seasonal, meteorological, and socioeconomic factors that influence power usage. The integration of advanced feature engineering techniques, including lag features and rolling statistical windows, enhances temporal representation, while CNN layers facilitate the extraction of spatial relationships. The model provides reliable 30-day forecasts, validated with MAE, RMSE, and R2 metrics. The model demonstrates higher accuracy in Spain (RMSE: 798.03 kW, R2: 0.9693) and performs well in Morocco (RMSE: 1836.91 kW, R2: 0.9324), thus demonstrating its versatility. This methodology provides a scalable solution for utilities and regulators looking to address long-term demand uncertainties and promote renewable integration.Taip / Ye
AI-Based Advancements for Comprehensive Mangrove Analysis Suitability Mapping
No full textHabitat loss remains a critical global issue linked to climate change. The Philippines, recognized as a global hotspot for marine biodiversity, is home to extensive mangrove forests. Despite their ecological significance, these forests are increasingly threatened by deforestation and various anthropogenic pressures. To support conservation efforts and mitigate habitat loss, this study employs habitat suitability modeling to identify optimal afforestation zones for mangrove species. Machine learning algorithms, including Maximum Entropy (MaxEnt), Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Network (ANN), were utilized to analyze mangrove distribution. A total of 17 environmental factors, encompassing bioclimatic and marine variables, along with mangrove sapling presence points, were integrated into the models. Experimental results demonstrated that the RF model outperformed others, achieving an area under the curve (AUC) value of 0.97 in predicting potential mangrove habitats. Based on RF model predictions, the model identified high-suitability zones for mangrove restoration and afforestation. The generated suitability maps provide valuable insights for developing science-based adaptation policies, strategies, and measures aimed at enhancing the resilience of mangrove ecosystems against present and future climate challenges.Taip / Ye
New Trends in Contemporary Economics, Business and Management. Selected Proceedings of the 15th International Scientific Conference “Business and Management 2025”
No full textThe Proceedings contain the selected papers from five sections: Advanced Economic Development,
Green Economy and Sustainable Development, Business Technologies and Sustainable
Entrepreneurship, Finance and Investment: New Challenges and Opportunities, New Perspectives
on Management and Resilience of Business Organisations.Taip / Ye
The role of AI-powered chatbots in enhancing customer experience: systematic literature review
No full textThe engagement of chatbots in customer experience management is increasing as a disruptive innovation in
the digital ecosystem. The goal of the research conducted by the authors was to explore the application of AI chatbots
to be used in e-commerce to improve customer satisfaction levels. The systematic literature review and data-analysis
presented in the paper involve the peer-reviewed literature published in the last three years. This study includes peerreviewed
research articles on AI-powered chatbots and customer experience published between 2021 and 2024. After
searching Scopus and PubMed and applying exclusion criteria, five relevant full-text articles from the last three years
were selected. Scopus and PubMed databases were used for record searching; finally, twenty-one study-relevant articles
were chosen for the research. After applying some exclusion criteria, which were duplicates, other than the English
language, non-AI-powered chatbots, and not accessed for full-text articles, five peer-reviewed articles were finally used
for analysis. The PRISMA approach was utilized to synthesize the findings. Cross-study compatibility, validity, and reliability
of sources were validated and assessed to derive more accurate data-analysis results. Consequent themes include
the impact of anthropomorphism on perceived personalization qualities, behavioral trust and intention, and the part
played by chatbots in reducing perceived risks in electronic commerce. The results of the research support chatbots as
stewards of business and customer service operations to achieve global sustainability and citizenship objectives, including
inclusive interactions for all users.Taip / Ye
Scaling sustainability: the economic and environmental impact of circular economy clusters
No full textThis study examines established circular economy clusters in Europe, focusing on their structure, strategies,
and impact. The principal objectives are to categorize clusters based on circular models, assess their economic
and environmental contributions, and identify key success factors. The investigation covers industrial symbiosis, biobased
production, and policy-led frameworks. The study employs a comparative analysis of key performance indicators
(KPIs) and policy frameworks. The findings indicate that collaboration, innovation, and government support are
critical factors in driving cluster success while regulatory and market barriers remain challenges. The conclusion emphasizes
the need for policy incentives, cross-sector cooperation, and investment in circular innovation to accelerate
sustainability.Taip / Ye
Biblioteka informuoja, 2025 Nr. 2 (699)
No full textNaujai į Web of Science ir Scopus įtrauktų Vilnius Gedimino technikos darbuotojų publikacijų sąrašai ir kitos bibliotekos aktualijos.2 (699)202
Biblioteka informuoja, 2025 Nr. 20 (717)
No full textNaujai į Web of Science ir Scopus įtrauktų Vilnius Gedimino technikos darbuotojų publikacijų sąrašai ir kitos bibliotekos aktualijos.20 (717)202
Senti Guide: A Machine Learning-Based Sentiment Analysis System for Student Feedback Evaluation
No full textIn academic settings, student feedback plays a crucial role in evaluating events and improving future planning. However, traditional manual feedback processing is time-consuming, error-prone, and inconsistent, often leading to delays in decision-making. To address this, the Senti Guide web application was developed, incorporating sentiment analysis to automatically classify student feedback into positive, neutral, or negative sentiments. The system utilizes Support Vector Machines (SVM) with a Radial Basis Function (RBF) kernel for sentiment classification, ensuring efficient handling of nonlinear relationships in the data. This method allows the system to effectively differentiate between complex patterns in student feedback. Preprocessing steps such as normalization, tokenization, and stop-word removal are applied to clean and structure the feedback data, preparing it for optimal analysis. Furthermore, hyperparameter tuning is employed to optimize the model’s performance, improving its accuracy, precision, recall, and F1 score. The model demonstrates significant performance improvements, achieving high classification accuracy of 90.05%, precision of 90.56%, recall of 90.05%, and F1-score of 90.04%. By automating sentiment analysis, Senti Guide provides the SSG with a scalable solution to modernize evaluation methods and enhance the future quality of school events. This research underscored the significance of employing machine learning to improve feedback processing, offering valuable insights for the Supreme Student Government to enhance event planning and student engagement.Bohol Island State University Clarin Campus administrators and the Supreme Student GovernmentTaip / Ye
Financial inclusion, institutional quality, and capital market development: evidence from Africa
No full textThis paper examines the development of capital markets in the African context, incorporating financial
inclusion and institutional quality. The study utilizes a unique data set for the period between 2004 and 2021 for
countries in the African region. Our findings, based on the PMG-Panel ARDL approach, indicate that while financial
inclusion and institutional quality significantly promote capital market development in the long run, their effect is
not statistically significant in the short run. Conversely, the interaction between financial inclusion and institutional
quality has a significant negative effect on capital markets in the long run. These findings underscore the centrality
of financial inclusion and institutional quality in capital market development. Similarly, the study suggests that poorquality
institutions result in an exclusive financial sector, thus adversely affecting the development of capital markets.
The study recommends that policymakers and researchers consider the importance of quality institutions as they seek
to strengthen their financial sectors.Taip / Ye
A Hybrid Approach in Developing a Recommendation System for Personalized Selection of Locations for a Visit
No full textPersonalized recommendation systems play a crucial role in enhancing user experiences by providing tailored suggestions based on individual preferences and contextual factors. This paper presents a hybrid approach in developing a recommendation system for selecting locations to visit, integrating user-defined filters, contextual data, and collective user feedback. The proposed system leverages a deep neural network to analyze various inputs, including explicit user preferences (e.g., desired atmosphere, type of location, etc.), dynamic contextual factors (e.g., weather conditions, temperature, etc.), and historical user data (e.g., ratings, recommendation trends for similar preferences, etc.). By combining content-based filtering with collaborative filtering techniques, the model aims to improve the accuracy and relevance of recommendations. The system classifies locations as suitable or unsuitable based on the given criteria, providing users with adaptive and context-aware suggestions. The hybrid nature of the approach allows for a more comprehensive understanding of user needs while incorporating real-time environmental conditions. Experimental validation is conducted to assess the effectiveness of the model in generating accurate recommendations. The results highlight the advantages of integrating multiple data sources and deep learning techniques to enhance accuracy and achieve high-quality recommendations. This research contributes to the development of intelligent recommendation systems by proposing a scalable and adaptable framework for personalized location selection.Taip / Ye