Abstract
The cryptocurrency market presents immense opportunities and significant risks due to its high volatility. Accurate price forecasting is crucial for informed investment decisions, enabling investors to optimize portfolio allocation, mitigate risk, and potentially maximize returns. Existing forecasting methods often struggle with the inherent non-stationarity and complexity of cryptocurrency price dynamics. This study addresses this challenge by developing a system for high-frequency forecasting of the closing prices of ten leading cryptocurrencies. We compare various machine learning models, including recurrent neural networks (RNNs), time series analysis (ARIMA), and conventional regression algorithms, using minute-step Bitcoin price data over a 30-day period to predict prices 60 min ahead. Our findings demonstrate that the GRU neural network exhibits superior predictive accuracy (MAPE = 0.09%, MSE = 5954.89, RMSE = 77.17, MAE = 60.20), outperforming other models considered. This improved forecasting accuracy contributes to the existing literature by providing empirical evidence for GRU's effectiveness in the volatile cryptocurrency market and offers practical insights for investment strategies. A web application integrating the best-performing model further facilitates real-time price prediction for multiple cryptocurrencies.

Abstract
Stress in the workplace is a major problem that affects people of all ages, backgrounds, and occupations. It can contribute to various health problems, from anxiety to insomnia, among others. Workplace stress significantly impacts employee well-being and productivity. Current stress-management approaches, while valuable, primarily address stress after it has occurred. This highlights the critical need for proactive systems capable of anticipating individual stress and preventing negative health consequences. This research presents the design and initial implementation of a novel microservice-based recommendation system for proactively mitigating workplace stress among computer users. The system leverages predicted stress levels to deliver timely, personalized, and easily implemented interventions. This study focuses on evaluating the system's architecture, core functionalities, and initial performance using a content-based filtering approach. A pilot study demonstrated the system's feasibility, highlighting areas for future development.

Abstract
This study presents a machine learning-based approach to predicting the outcosmes of NBA games, with the aim of enhancing decision-making in sports betting and performance analysis. Using a dataset spanning 20 NBA seasons (2003-2023), we incorporated key features such as team statistics, player performance metrics, and external factors like team fatigue and rankings. The methodology followed the CRISP-DM process, involving data preprocessing, feature selection, and model evaluation. We experimented with multiple classification algorithms, including Logistic Regression, Random Forest, Gradient Boosting, and ensemble methods, to identify the best-performing models. Feature selection techniques such as LASSO and decision tree-based methods were employed to optimize model performance. Our best model, combining team rankings, statistics, and fatigue factors, achieved an accuracy rate of 64.1% and an F1 score of 72.4%, reflecting the complexity of NBA game outcome prediction. The study highlights the importance of key features like team rankings and the challenges posed by the dynamic nature of the NBA. Future research will explore additional qualitative factors, such as emotional states and team dynamics, and employ more advanced machine learning techniques like deep learning to further improve prediction accuracy.

Abstract
Fighting the propagation of misinformation within a social media group or community by focusing on identifying dishonest members who deliberately try to quash any constructive social movement is very challenging because such people use advanced tactics to create division and doubt by manipulating information. The present research aims to develop a hybrid heuristic model to identify those who intentionally spread misleading information on social media to jeopardize a social movement. We frame this issue under the heading of Graph Semi-supervised Learning (GSSL), and we propose a hybrid model that falls under the heuristic approach, called Label Propagation-Gated Recurrent Unit (LP-GRU). LP-GRU can effectively identify perpetrators of disinformation within social communities by fusing community structure from the Label Propagation algorithm with behavioral patterns identified by GRU. Compared to previous heuristic approaches, we achieve up to 76% accuracy when using the LP-GRU model on augmented semi-synthetic social network data. © 2025 IEEE.

Abstract
Digital Transformation Models (DTM) and Digital Maturity Models (DMM) are two artefacts that guide the planning and implementation of Digital Transformation (DT) initiatives. When used in a combined approach, a DTM-DMM pairing could support DT managers and practitioners, as DTs are holistic and complex initiatives with high failure rates. However, no study has yet systematically addressed the compatibility amongst artefacts. This paper, therefore, aims to analyse the compatibility between academic DTMs and DMMs. Based on architectural compatibility and conceptual similarity, we provide a structured and replicable mixed methods approach to assessing artefact compatibility. To achieve this, we start with a systematic literature review to identify existing academic DTMs and DMMs, analyse the models and group them according to their scope. After, we employ quantitative similarity analysis techniques (Term Frequency-Inverse Document Frequency and Bidirectional Encoder Representations from Transformers combined with Cosine Similarity) and perform a qualitative compatibility analysis to establish ground truth. Based on this analysis, we apply the Receiver Operating Characteristic Curve technique to define threshold values for compatibility assessment. The threshold values were used to suggest compatible DTM-DMM pairings, resulting in nine DTM-DMM binomials for Small and Medium-sized Enterprises. The findings support managers and practitioners in selecting DTM-DMM pairs to guide DT initiatives while offering academics a mixed-methods approach based on the similarity analysis field to evaluate artefact compatibility systematically. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

Abstract
Exploring a group of influential spreaders to acquire maximum influence has become an emerging area of research in complex network analysis. The main challenge of this research is to identify the group of important nodes that are scattered broadly, such that the propagation ability of information is maximum to a network. Researchers proposed many centrality-based approaches with certain limitations to identify the influential nodes (spreaders) considering different properties of the networks. To find a group of spreaders, the VoteRank (a voting mechanism) based method produces effective results with low time complexity, wherein each iteration, the node votes for its neighbors by its voting capability, and the node obtaining the maximum vote score is identified as an influential spreader. The major loophole of existing VoteRank methods is measuring the voting capability based on the degree, k-shell index, or contribution of neighbors methods, which does not efficiently identify the spreaders from the diverse regions based on their spreading ability. In this paper, we propose a novel Community-based VoteRank method (CVoteRank) to identify a group of influential spreaders from diverse network regions by which the diffusion process is enhanced. Firstly, we measure every node's spreading ability based on intra- and inter-connectivity structure in a community, which signifies the local and global importance of the node. To identify the seed nodes, we assign the spreading ability to that node's voting capability and iteratively calculate the voting score of anode based on its neighboring voting capability and its spreading ability. Then, the node acquiring the maximum voting score is identified as the influential spreader in each iteration. Finally, to solve the problem of influence overlapping, CVoteRank reduces the voting capability of the neighboring nodes of the identified spreader. The efficiency of CVoteRank is evaluated and compared with the different state-of-the-art methods on twelve real networks. Utilizing the stochastic susceptible-infected-recovered epidemic method, we calculate the infected scale, final infected scale, and the average shortest path length among the identified spreaders. The experimental results show that CVoteRank identifies the most efficient spreaders with the highest spreading ability within a short period and the maximum reachability, and the identified spreaders are situated at diverse portions of the networks.