In this study, we introduce a pioneering framework, DroneSSL, that integrates the concept of spatial crowdsourcing with TinyML to enhance anomaly detection in Internet Drone Things (IoDT). This innovative approach leverages drones and unmanned ground vehicles (UGVs) for expansive data collection environments are typically inaccessible or hazardous, such as during Australian bushfire incidents. By employing lightweight machine learning models alongside advanced communication technologies,...

In the evolving landscape of consumer electronics, Generative Adversarial Learning-based Trust Management (GALTrust) framework emerges as a novel solution, uniquely combining Networks (GANs) and type-2 fuzzy logic to tackle trust management challenges within Internet Drone Things (IoDT). Addressing pivotal needs spatial crowdsourcing scenarios like bushfire management, GALTrust significantly overcomes limitations posed by traditional machine learning methods in detecting emergent types...

The wealth of gene expression values being generated by high throughput microarray technologies leads to complex dimensional datasets. Moreover, many cohorts have the problem imbalanced classes where number patients belonging each class is not same. With this kind dataset, biologists need identify a small informative genes that can be used as biomarkers for disease.This paper introduces Balanced Iterative Random Forest (BIRF) algorithm select most relevant disease from high-throughput data....

We introduce "TMIoDT," a pioneering framework aimed at bolstering communication security in the Internet of Drone Things (IoDT) integrated with Open Radio Access Networks (Open RAN), specific focus on bushfire monitoring applications. Our novel contributions include seamless integration digital twin technology blockchain to establish robust trust management system IoDT context. This approach addresses critical vulnerabilities associated unsecured wireless networks IoDT, such as data...

The identification of a subset genes having the ability to capture necessary information distinguish classes patients is crucial in bioinformatics applications. Ensemble and bagging methods have been shown work effectively process gene selection classification. Testament that random forest which combines decision trees with improve overall feature classification accuracy. Surprisingly, adoption these support vector machines has only recently received attention but mostly on not selection....

We introduce the AI-Generated Optimal Decision (AIGOD) algorithm and Deep Diffusion Soft Actor-Critic (DDSAC) framework, marking a significant advancement in integrating Human Digital Twins (HDTs) with Content (AIGC) within IoMT-based smart homes. Our innovative Content-as-a-Service (AIGCaaS) architecture, optimized for IoMT environments, leverages network edge servers to enhance selection of Service Providers (AISPs) tailored unique characteristics individual HDTs. Extensive experiments...

Early damage detection is critical for a large set of global ageing infrastructure. Structural Health Monitoring systems provide sensor-based quantitative and objective approach to continuously monitor these structures, as opposed traditional engineering visual inspection. Analysing sensed data one the major (SHM) challenges. This paper presents novel algorithm detect assess in structures such bridges. method applies tensor analysis fusion feature extraction, further uses one-class support...

We introduce an ensemble model approach for multimodal sentiment analysis, focusing on the fusion of textual and video data to enhance accuracy depth emotion interpretation. By integrating three foundational models-IFFSA, BFSA, TBJE-using advanced techniques, we achieve a significant improvement in analysis performance across diverse datasets, including MOSI MOSEI. Specifically, propose two novel models-IFFSA which utilise large language models BERT GPT-2 extract features from text modality...

Abstract Identifying cancer risk groups by multi-omics has attracted researchers in their quest to find biomarkers from diverse risk-related omics. Stratifying the patients into using genomics is essential for clinicians pre-prevention treatment improve survival time and identify appropriate therapy strategies. This study proposes a framework that can extract features various omics simultaneously. The employs autoencoders learn non-linear representation of data applies tensor analysis...

A new heuristic optimization algorithm is presented to solve the nonlinear problems. The proposed utilizes a stochastic method achieve optimal point based on simplex techniques. dual distributed stochastically in search space find best point. Simplexes share and worst vertices of one another move better through space. applied 25 well-known benchmarks, its performance compared with grey wolf optimizer (GWO), particle swarm (PSO), Nelder-Mead algorithm, hybrid GWO combined pattern (hGWO-PS),...

In this paper, we focused on the development and verification of a solid robust framework for structural condition assessment real-life structures using measured vibration responses, with presence multiple progressive damages occurring within inspected structures. A self-tuning learning method was proposed. Damage sensitive features were extracted frequency domain decomposition (FDD) approach to fuse all followed by random projection algorithm dimensionality reduction. An automatic parameter...

Multi-label classification is defined as the problem of identifying multiple labels or categories new observations based on labeled training data. Multi-labeled data has several challenges, including class imbalance, label correlation, incomplete multi-label matrices, and noisy irrelevant features. In this article, we propose an integrated approach with space imbalance (ML-CIB) for simultaneously model addressing aforementioned challenges. The learns a matrix captures correlations, because...

The one-class support vector machines with Gaussian kernel function is a promising machine learning method which have been employed extensively in the area of anomaly detection. However, generalization performance OCSVM profoundly influenced by its model parameter σ. This paper proposes new algorithm named Edged Support Vector (ESV) for tuning parameter. semantic idea this based on inspecting spatial locations selected samples. selects optimal value σ leads to decision boundary that has all...

Abstract Data-driven machine learning models, compared to numerical demonstrated promising improvements in detecting damage structural health monitoring (SHM) applications. In such approaches, sensors’ data are used train a model either centralized (server) or locally inside each sensor unit node (client). The often leads computing and privacy issues as wireless transmission costs data-sensitive vulnerability, especially real-time settings. decentralized also poses different challenges...

Recommender systems have been successfully used in many domains with the help of machine learning algorithms. However, such applications tend to use multi-dimensional user data, which has raised widespread concerns about breach users’ privacy. Meanwhile, wearable technologies enabled users collect fitness-related data through embedded sensors monitor their conditions or achieve personalized fitness goals. In this article, we propose a novel privacy-aware recommender system. We introduce...