Machine learning systems deployed in dynamic environments frequently operate under non-stationary data distributions, where controlled distribution shift can progressively degrade predictive performance. However, many widely used tabular benchmark datasets lack explicit temporal structure, limiting reproducible evaluation of drift adaptation methods. This work proposes a cluster-induced distribution shift simulation framework that transforms static tabular datasets into controlled evolving data streams through structured perturbations across feature-space partitions. Using this framework, six adaptation strategies are systematically evaluated: static learning, sliding-window retraining, global ADWIN retraining, cluster-local ADWIN retraining, random subspace drift detection, and feature-partitioned drift detection. Experiments are conducted on five benchmark datasets covering both classification and regression tasks using diverse predictive model families, including linear models, k-Nearest Neighbours, tree ensembles, boosting methods, and adaptive online learners. The results show that sliding-window retraining often achieves the strongest predictive performance but incurs substantial computational and labelling cost due to frequent global updates. Global ADWIN retraining reduces update frequency while preserving competitive accuracy. In contrast, the proposed cluster-local adaptation strategy consistently achieves a stronger balance between predictive performance and computational efficiency across both classification and regression settings. Under structured distribution-shift scenarios, the method reduces retraining effort by up to 75% while maintaining competitive predictive performance relative to continuously adaptive baselines. For nonlinear regression models, cluster-local adaptation preserves competitive R2 performance while substantially lowering update frequency and training effort. The proposed framework provides a reproducible benchmark for evaluating adaptation under controlled distribution shift on static tabular datasets and demonstrates that cluster-aware drift/shift detection constitutes an effective and computationally efficient alternative to uniform retraining strategies under heterogeneous distributional shift conditions.

Abstract Android is a widely used operating system, primarily found on mobile phones and tablets. Applications (commonly known as “apps”) for android can be easily installed from Google Play, third-party stores, or manually using android package kit (APK) files. Due to its growing popularity, android has attracted significant attention from malicious actors deploying various forms of malware. To address this challenge, artificial intelligence-based approaches are increasingly used to protect systems from cyber-attacks. This research paper focuses on the application of ChatGPT, a powerful large language model, in cybersecurity, specifically for malware detection. It evaluates ChatGPT’s potential as an innovative tool in fighting cyber threats, exploring the process of fine-tuning ChatGPT, its performance and its limitations in malware detection tasks. The objective is to reduce the effort and time required to generate artificial intelligence-based malware detection systems, simplifying their development process. This research shows how ChatGPT can be utilized to generate code for detecting malware in structured datasets with high accuracy. The focus is not on introducing any new algorithms but on allow individuals without programming expertise to create and apply these models effectively.

ABSTRACT The growing popularity of generative AI and large language models (LLMs) has sparked innovation alongside debate, particularly around issues of plagiarism and intellectual property law. However, a less‐discussed concern is the quality of code generated by these models, which often contains errors and encourages poor programming practices. This paper proposes a novel solution by integrating LLMs with automated machine learning (AutoML). By leveraging AutoML's strengths in hyperparameter tuning and model selection, we present a framework for generating robust and reliable machine learning (ML) algorithms. Our approach incorporates natural language processing (NLP) and natural language understanding (NLU) techniques to interpret chatbot prompts, enabling more accurate and customisable ML model generation through AutoML. To ensure ethical AI practices, we have also introduced a filtering mechanism to address potential biases and enhance accountability. The proposed methodology not only demonstrates practical implementation but also achieves high predictive accuracy, offering a viable solution to current challenges in LLM‐based code generation. In summary, this paper introduces a new application of NLP and NLU to extract features from chatbot prompts, feeding them into an AutoML system to generate ML algorithms. This approach is framed within a rigorous ethical framework, addressing concerns of bias and accountability while enhancing the reliability of code generation.

Abstract Android devices have evolved to offer a diverse array of services, spanning applications related to banking, business, health, and entertainment. The widespread adoption of Android devices, coupled with the open-source architecture of the Android operating system, has rendered them a prime target for malicious actors. Among the most perilous threats are Android botnets, which enable malicious actors, often referred to as botmasters, to exert remote control for the execution of destructive attacks. Android botnets have huge potential to be an emerging threat to mobile device security. In this paper, we focus on detecting evolving Android botnets and introduce a new dataset of 3458 apps, represented by 455 permission-based features. We propose an improved multilayer perceptron neural network for zero-day botnet detection. Our methodology, in this way, achieves an accuracy of 98.5%, thus outperforming traditional classifiers. It has a lot of functionality and is based on the neural network approach, making it able to identify slight botnet behaviours in order to improve Android security.

Abstract Android malware has become an increasingly important threat to individuals, organizations, and society, posing significant risks to data security, privacy, and infrastructure. As malware evolves in sophistication and complexity, the detection and mitigation of these malicious software instances have become more challenging and time consuming since the required number of features to identify potential malware can be very high. To address this issue, we have developed an effective feature selection methodology for malware detection in Android. The critical concern in the field of malware detection is the complexity of algorithms and the use of features that are used to detect malware. The present paper delivers a methodology for pre-processing datasets to select the most optimal features that will allow detecting malware, while maintaining very high accuracy. The proposed methodology has been tested on two real world datasets and the results indicate that the number of features is significantly reduced from 489 to between 19 and 28 for the first dataset and from 9503 to between 9 and 27 for the second dataset, whilst the accuracy is maintained as if all features were used.

Blockchain has been viewed as a breakthrough and an innovative technology due to its privacy, security, immutability, and data integrity characteristics. The consensus layer of the blockchain is the backbone and the most important layer of the blockchain architecture because it acts as the performance and security manager of the blockchain. The detection of Long-Range Attacks (LRA) on the Proof-of-Stake (PoS) blockchain is a complex task. Earlier studies have shown various challenges in detecting long-range attacks and monitoring the activities of validator nodes on the blockchain network. Thus, this paper proposes a novel dataset for node classification on a proof-of-stake permissionless blockchain and proposes a Deep Learning method that can be used to classify nodes into malicious or non-malicious nodes to mitigate long-range attacks with high accuracy. The performance metrics for the model are compared and measured which suggest the developed performance of the proposed model. The proposed solution can serve as a guide on how future researchers and blockchain developers can simulate and curate proof-of-stake datasets and goes further to demonstrate that artificial intelligence models can be used as a mitigating checkpoint for long-range attacks. The dataset in the paper is publicly available and can be used by other researchers to detect other activities and behaviors on a permissionless blockchain. These techniques can further enhance security, performance and create fairness on the proof-of-stake consensus.

Transport infrastructure relies heavily on extended multi sensor networks and data streams to support its advanced real time monitoring and decision making. All relevant stakeholders are highly concerned on how travel patterns, infrastructure capacity and other internal / external factors (such as weather) affect, deteriorate or improve performance. Usually new network infrastructure can be remarkably expensive to build thus the focus is constantly in improving existing workflows, reduce overheads and enforce lean processes. We propose suitable graph-based workflow monitoring met­hods for developing efficient performance measures for the rail industry using extensive business process workflow pattern analysis based on Case-based Reasoning (CBR) combined with standard Data Mining methods. The approach focuses on both data preparation, cleaning and workflow integration of real network data. Preliminary results of this work are promising since workflow integration seems efficient against data complexity and domain peculiarities as well as scale on demand whilst demonstrating efficient accuracy. A number of modelling experiments are presented, that show that the approach proposed here can provide a sound basis for the effective and useful analysis of operational sensor data from train Journeys.

With widespread modernization, digitization and transformations of most of industries, Artificial Intelligence (AI) has become the key enabler in that modernization journey. AI offers substantial capabilities to solve new problems and optimise existing solutions specialising on specific problems and learning from different domains. AI solutions can be either explainable or black box ones with the latter being urged to improve since they cannot trust. Case-based Reasoning (CBR) is an explainable AI approach where solutions are provided along with relevant explanations in terms of why a solution was selected. However, CBR, like most other explainable approaches, has several limitations in terms of scalability, large data volumes, domain complexity, that reduce its ability to scale any CBR system in industrial applications. In this paper, we provide a heterogeneous CBR framework - DeepKAF where we combine CBR paradigm with Deep Learning architectures to solve complicated Natural Language Processing (NLP) problems (eg. mixed language and grammatically incorrect text).DeepKAF is built based on continuous research in the area of Deep Learning and CBR. DeepKAF has been implemented and used across different domains, test use cases and research models as an ensemble deep learning and CBR Architecture.

Social engineering attacks are well known attacks in the cyberspace and relatively easy to try and implement because no technical knowledge is required. In various online environments such as business domains where customers talk through a chat service with employees or in social networks potential hackers can try to manipulate other people by employing social attacks against them to gain information that will benefit them in future attacks. Thus, we have used a number of natural language processing steps and a machine learning algorithm to identify potential attacks. The proposed method has been tested on a semi-synthetic dataset and it is shown to be both practical and effective.

Purpose This paper aims to address privacy concerns that arise from the use of mobile recommender systems when processing contextual information relating to the user. Mobile recommender systems aim to solve the information overload problem by recommending products or services to users of Web services on mobile devices, such as smartphones or tablets, at any given point in time and in any possible location. They use recommendation methods, such as collaborative filtering or content-based filtering and use a considerable amount of contextual information to provide relevant recommendations. However, because of privacy concerns, users are not willing to provide the required personal information that would allow their views to be recorded and make these systems usable. Design/methodology/approach This work is focused on user privacy by providing a method for context privacy-preservation and privacy protection at user interface level. Thus, a set of algorithms that are part of the method has been designed with privacy protection in mind, which is done by using realistic dummy parameter creation. To demonstrate the applicability of the method, a relevant context-aware data set has been used to run performance and usability tests. Findings The proposed method has been experimentally evaluated using performance and usability evaluation tests and is shown that with a small decrease in terms of performance, user privacy can be protected. Originality/value This is a novel research paper that proposed a method for protecting the privacy of mobile recommender systems users when context parameters are used.