IAES International Journal of Artificial Intelligence (IJ-AI)
Not a member yet
1769 research outputs found
Sort by
A review on long short-term memory combination development
Get PDFLong short-term memory (LSTM) has continued to develop since it was proposed in 1997. LSTM has optimized solutions to various problems. The LSTM cell, architecture, and memory model have been reviewed. A review of LSTM implementation has been carried out in various problem domains. There are combinations of LSTM with other methods to optimize solutions. However, there is no review on the development of LSTM combination (LC). This research reviews the development of the LC model on nine research questions, namely: development framework, data, preprocessing, learning process, tasks, optimization and evaluation, domain problems, trends, and challenges. The results show that the LC model is increasingly widespread in solving problems. The LC model has completed 26 types of tasks. Prediction, detection, forecasting, classification, and recognition are the most frequently performed tasks. LC model development trends show that LSTM is increasingly collaborative with other methods on a wider scope. The challenges identified include research areas, data, model developments, the area of implementation, performance, and efficiency
Deep learning-based feature selection for lung adenocarcinoma classification and biomarker discovery
Get PDFLung adenocarcinoma, a leading cause of cancer-related mortality, underscores the need for reliable diagnostic tools. This study proposes a robust multi-stage feature selection and classification framework for biomarker discovery, using the cancer genome atlas lung adenocarcinoma (TCGA-LUAD) as the primary dataset and GSE19188 for independent validation. The framework combines differential expression analysis (Wilcoxon rank-sum test), joint mutual information maximization (JMIM), and sparse autoencoder-based refinement to identify a compact and predictive set of five genes. These genes are involved in key lung cancer pathways, including epidermal growth factor receptor (EGFR) signaling, cell cycle regulation, and immune response, and include biomarkers such as surfactant protein A2 (SFTPA2), napsin an aspartic peptidase (NAPSA), and T-box transcription factor 4 (TBX4). The hybrid deep learning classifier achieved high accuracy (98.4%) and area under the receiver operating characteristic curve (AUC-ROC) (0.996) on TCGA-LUAD, with strong generalization on GSE19188 (accuracy: 96.7%, AUC-ROC: 0.993%). Overall, the framework offers an interpretable and effective solution for LUAD classification and biomarker identification
Humans’ psychological traits classification from their spending categories using artificial intelligence algorithms
Get PDFThe analysis of human behavior data generated by digital technologies has gained increasing attention in recent years. Spending categories form a significant part of this digital footprint. In this study, we investigate the degree to which human expenditure records can be used to infer psychological traits from transaction data. A broad feature space was constructed, consisting of overall spending behavior, category-related spending behavior, and customer category profiles. These features were examined to identify their correlations with the Big Five personality traits. A dataset containing over 1,200 users’ transaction histories over three months was obtained from Kaggle. Personality trait labels were derived using a percentile-based classification method. Multiple AI algorithms: decision tree (DT), random forest (RF), logistic regression (LR), and support vector machine (SVM) were employed, along with a convolutional neural network (CNN) to classify personality traits. The CNN model, incorporating multi-dimensional convolutional layers and the full feature space, achieved a high accuracy of 99.03%. The outcomes of the experiment indicate the efficiency of combining behavioral features and AI models in psychological trait classification. The study also highlights ethical considerations, including privacy risks and misuse of inferred personality details
Effective modelling of human expressive states from voice by adaptively tuning the neuro-fuzzy inference system
Get PDFThis paper aims to develop efficient speech-expressive models using the adaptively tuning neuro-fuzzy inference system (ANFIS). The developed models differentiate a high-arousal happiness state from a low-arousal sadness state from the benchmark Berlin (EMODB) database. The proposed low-cost flexible developed algorithms are self-tunable and can address several vivid real-world issues such as home tutoring, banking, and finance sectors, criminal investigations, psychological studies, call centers, cognitive and biomedical sciences. The work develops the proposed structures by formulating several novel feature vectors comprising both time and frequency information. The features considered are pitch (F0), the standard deviation of pitch (SDF0), autocorrelation coefficient (AC), log-energy (E), jitter, shimmer, harmonic to noise ratio (HNR), spectral centroid (SC), spectral roll-off (SR), spectral flux (SF), and zero-crossing rate (ZCR). to alleviate the issues of the curse of dimensionality associated with the frame-level extraction, the features are extracted at the utterance level. Several performance parameters have been computed to validate the individual time and frequency models. Further, the ANFIS models are tested for their efficacy in a combinational platform. The chosen features are complementary and the augmented vectors have indeed shown improved performance with more available information as revealed by our results
Artificial intelligence in the United Arab Emirates public sector: a systematic literature review
Get PDFThis systematic literature review examines United Arab Emirates (UAE) public sector artificial intelligence (AI) use, impact, and challenges. Using the preferred reporting items for systematic reviews and meta-analyses (PRISMA) protocol, 20 relevant Scopus articles were selected for the study. Data from selected articles were used to analyse AI's use, benefits, and drawbacks in the UAE's public sector. Quality assessment was done throughout the review process. The results showed that AI is being used more in the UAE's public sector to improve efficiency, cost savings, decision-making, and service delivery. The review also found data, privacy, security, technical, infrastructure, AI, and user challenges. Publication bias and the lack of AI studies in the UAE's public sector limit the study. The findings have major implications for policy and practice, emphasising the need for AI strategies and UAE-specific solutions
Enhancing microgrid production through particle swarm optimization and genetic algorithm
Get PDFThe growing demand for sustainable and efficient energy solutions has led to research on optimizing renewable energy sources within microgrid systems. This study presents a comparative analysis of two prominent optimization techniques, particle swarm optimization (PSO) and genetic algorithm (GA), to enhance solar photovoltaic PV and wind production in microgrids. The aim is to achieve a balanced and efficient energy generation that closely matches the load demand, thereby minimizing energy wastage and ensuring a reliable energy supply. The two algorithms are employed using data representing PV and wind production, as well as load consumption, over a 24-hour period. The results are evaluated based on their ability to reduce the gap between energy production and load demand. Our findings reveal compelling insights into the performance of GA and PSO in the context of microgrid optimization. To validate the results obtained from the simulation, the PSO algorithm is implemented on an actual cart Digital Signal Processor DSP platform, using a processor-in-the-loop (PIL). This successful real-world application highlights the practical viability of utilizing PSO to improve solar PV and wind energy generation within microgrids
Dealing imbalance dataset problem in sentiment analysis of recession in Indonesia
Get PDFGlobal recession news dominates social media, particularly in Indonesia, with social news platforms on Twitter generating public responses and re-tweetings on the issue. Mining these opinions from Twitter using a sentiment analysis approach yields invaluable insights. The research stages included data collection, pre-processing, data labeling using the lexical-based method like valence aware dictionary and sentiment reasoner (VADER) and TextBlob, sampling techniques using synthetic minority oversampling technique (SMOTE) and random over sampling (ROS) before and after splitting data, and modeling using machine learning such as support vector machines (SVM), k-nearest neighbour (KNN), naive Bayes, and model evaluation. The problem is that almost 300,000 data collected from NodeXL are unbalanced. The findings show that models with balanced datasets show better model evaluation results. The sampling technique was carried out before and after splitting the data. The model evaluation results show that the Bernoulli-naive Bayes algorithm, with the VADER labeling technique, and the SMOTE sampling technique after splitting data, obtains the best accuracy of 84%, and using the ROS technique obtains an accuracy of 81%. On the other hand, with the SMOTE and ROS technique before splitting data on the SVM algorithm, it gets the best accuracy of 93% from before if only using SVM only reached 84%
Automated diagnosis of brain tumor classification and segmentation of magnetic resonance imaging images
Get PDFBrain tumors are one of the most prevalent disorders of the central nervous system and are dangerous. For patients to receive the best treatment, early diagnosis is crucial. For radiologists to correctly detect brain tumor images, an automated approach is required. The identification procedure can be time-consuming and prone to mistakes. In this work, the issue of fully automated brain tumor classification and segmentation of magnetic resonance imaging (MRI) including meningioma, glioma, pituitary, and no tumor is taken into consideration. In this study, convolutional neural network (CNN) and mask region-based convolutional neural network (R-CNN) are proposed for classification and segmentation problems respectively. This study employed 3,200 images as a training set and the system achieved an accuracy of 96% for classifying the tumors and 94% accuracy in segmentation of tumors
Federated inception-multi-head attention models for cyber-attacks detection
Get PDFWith the proliferation of internet of things (IoT) devices, ensuring the security of these interconnected systems has become a critical concern. Cyberattacks targeting IoT devices pose significant threats to individuals and organizations due to the generation of vast amounts of data across many connected devices, which traditional centralized methods cannot solve. Federated learning (FL) could be a promising solution to mitigate privacy concerns associated with centralized approaches and address cybersecurity concerns. This paper uses FL and deep learning (DL) approaches to cybersecurity in IoT applications. The goal of cyber security is achieved by forming a federation of acquired and shared models at the head of the various participants. We use inception time and multi-head attention (CNN) algorithm based on FL to detect cyber-attacks and avoid data privacy leaks under two distribution modes, namely IID and Non-IID. In contrast, the FedAvg and FedMA algorithms aggregate local model updates. A global model is produced after several communication rounds between the IoT devices and the model parameter server. Cyber threats are simulated using edge-IIoT datasets. Experiment results show that the federated inception model's best global accuracy was 93, 91%, and 93, 49% using multi-head attention
Adaptive radio propagation model for maximizing performance efficiency in smart city disaster management application
Get PDFClimate change poses several environmental threats like floods to urban environment; thus, effective and reliable communication of emergency information is needed during massive breakdown of network infrastructure. This paper presents a mobile adhoc network (MANETs) based effective information such as calls, image, and videos communication system that is compatible with current 3GPP and 5G communication network. Here in maintaining connectivity the information is communicated between different MANET nodes in a multi-hop manner. However, designing radio propagation is challenging considering higher local emergency request congestion at different terrain with varying speed of users. The current radio propagation model is designed without considering the effect of line-of-sight between communicating device and are not adaptive to different environment considering urban disaster management environment. This paper develops an adaptive radio propagation (ARP) model namely expressway, city and semiurban. Then, in reducing congestion and improving network performance efficiency the work introduced an adaptive medium access control (AMAC) protocol. The MAC incorporates a dynamic network controller (DNC) to optimize the contention window size in dynamic manner according to current traffic demands. The AMAC protocol achieves much improved throughput with lesser packet loss in comparison with existing MAC (EMAC) model considering different radio propagation model introduced in this work