Search

Fission product yields (FPYs) play a crucial role in various aspects of nuclear science and technology, including nuclear structure and reactions. However, the inherent constraints of traditional computational methods used in theoretical models, and lack of experimental access to key observables pose challenges in obtaining accurate and comprehensive fission data. Neural Networks (NNs) present a promising solution to address these challenges by effectively modeling and acquiring energy-dependent fission yields. Mixture Density Networks (MDNs) enable learning from available data, predicting unknowns, and quantifying uncertainties simultaneously. Our study demonstrates the effectiveness of MDNs in evaluating fission product yields, particularly in scenarios where experimental data are incomplete. Machine learning algorithms like Gaussian Process Regression (GPR) can capture the distribution of single-fission yields and generate high-quality samples. These samples serve as valuable inputs for MDN networks. This study introduces an MDN approach for evaluating energy-dependent fission mass yields. The results of MDN evaluations indicate satisfactory accuracy in determining both the distribution positions and energy dependencies of FPYs. [ABSTRACT FROM AUTHOR]

This article, X-ray images of the patient were analyzed in order to reduce mortality by early diagnosis of infectious diseases, processing of X-ray images, clarification of its borders, adjustment of image values through histograms was carried out, especially in the detection of COVID disease. Various neural network models were studied using existing lung X-ray images, and the results were obtained using the W-SCBNet neural network model. A result of 93 percent of covid disease recognition was achieved. The results are illustrated by tables and diagrams. A faster and cheaper way to detect Covid-19 and similar infectious diseases using X-ray images has been considered. [ABSTRACT FROM AUTHOR]

Detecting skin cancer early is crucial, especially given its increasing prevalence, particularly among the aging population. Early identification is essential because skin cancer tends to spread as it progresses, and treatment is most effective in its initial stages. The rising incidence, high mortality rate, and costly medical treatments underscore the urgency of early symptom detection. Scientists have developed multiple techniques for early detection that utilize attributes such as the color, symmetry, shape and size of skin bruise to distinguish between benign and malignant instances. The emergence of deep learning (DL) algorithms has brought about a new era of automated diagnostic healthcare systems. In this research, we suggest employing deep learning as a precise tool for identifying bruise regions. Firstly, we improve image quality using Enhanced Super-Resolution Generative Adversarial Networks (ESRGAN), followed by segmentation to isolate Regions of Interest (ROI) and address data imbalances through augmentation. Subsequently, the image undergoes analysis via a convolutional neural network (CNN), Recurrent Neural Networks (RNNs), modified Resnet-50 and Deep Neural Networks (DNNs) variant for skin lesion categorization. By observing these models, we found that Convolutional Neural Networks (CNN) has outperformed its counterparts in feature extraction and categorization assessments. Our research depicts a technique for predicting Merkel Cell Carcinoma and categorizing it through the utilization of image processing and cutting-edge technologies. The outcomes determine if an individual is afflicted by the disease and, if so, identify the specific type. The primary goal of our research is to evaluate skin images and enhance the efficiency of the classification process for different skin conditions. [ABSTRACT FROM AUTHOR]

RMC waste incorporated concrete (RMCWIC) is substantially a different composition. Predicting the compressive strength becomes very critical due to its complex composition. An Artificial Neural Network (ANN) model was used to estimate the 28-day compressive strength (CS28) of RMC waste-incorporated concrete in this study. The average strength of 87 specimens for 28 different mix proportions was considered to develop a data-driven ANN model. The various ingredients of RMCWIC are washout water, dried sludge, silica fume, recovered fine aggregate, recovered coarse aggregate, ordinary Portland cement (OPC), fine aggregate (FA), and coarse aggregate (CA). The ingredients are input parameters and the output parameter is CS28. The output is predicted based on the training carried out with the input parameters. The data-driven artificial neural network model has a good potential for forecasting the 28-day compressive strength of RMCWIC, according to the test findings. [ABSTRACT FROM AUTHOR]

This paper explores the assessment of Diwaniyas Land Surface Temperature (LST) and the changes, in landcover from 2015 to 2021. Due to the challenges of obtaining information through ground based measurements using satellite measurements in the thermal infrared range is seen as a highly attractive option. The findings indicate that LST has increased as a result of urbanization decreased areas and expansion of land. It was observed that regions with vegetation experienced LST while areas with more vegetation had lower LST. These findings can contribute to the formulation of an inclusive climate resilience policy. Enhance Diwaniyas sustainability in the face of climate change impacts. Artificial neural networks were employed in this research to establish rulesets at the object level. By utilizing a Landsat image, network and Support Vector Machine techniques a detailed map of Diwaniya was generated. The accuracy of the maps classification was determined to be 94% with a kappa index value of 0.90. These outcomes demonstrate the effectiveness of the proposed classification method and its potential to offer a model, for applications. [ABSTRACT FROM AUTHOR]

Knowing the quantity of seepage flux from earthen dams is one of the important parameters in monitoring and surveillance activities that help avoid dam failure. In this study, a proposed equation was developed to find the quantity of flux from a chimney drain located in homogeneous earth dam that constructed on impermeable foundation. For this purpose, (SEEP/W) model was applied to generate 729 values of flux quantities that were results from dams with various geometrical shapes, various designs of chimney drain and various reservoir water depths. A methodology which was based on using the dimensional analysis technique and Statistical Package for the Social Sciences (SPSS) was used in the formulation of the proposed equation. The determination coefficient of (R2) of the equation was 0.997. The comparison between computed values of seepage flux from the software (SEEP/W) with the values obtained from application of the proposed equation were found in agreement. Finally, the Artificial Neural Network (ANN) technique was employed to evaluate the significance of each variable in relation to its effect on the quantities of the flux, and it was found that the upstream water level has the largest influence on the seepage flux with a percentage of 64.7%, while the dam slope at upstream side has the least influence percent which is equal to 0.9%. [ABSTRACT FROM AUTHOR]

The forecasting tools presented in this paper were applied in the flood-prone city of Marikina, a highly urbanized city located in the National Capital Region (NCR) of the Philippines. The city is traversed by Marikina River, which inundates nearby communities during heavy downpours. Marikina City has a long history of flooding. Perhaps the most devastating among these not only in Marikina but in the entire NCR is the Typhoon Ondoy which resulted in hundreds of deaths and huge amount of damage to properties. Despite the constant threat of flooding, residents stay in the city and have learned to adapt to the realities of the environmental hazards. For its part, the local government of Marikina has developed an alert system to warn residents of impending flood during heavy rains. Depending on the water level recorded at Sto. Niño station, which is located at a strategic point along Marikina River, residents are advised to either prepare or go to evacuation centers. The disaster response efforts would be best supplemented by real-time forecasts of water level to give both the local government and the residents ample time to prepare. This is the motivation for this study. In this paper, machine learning algorithms such as random forest, artificial neural network and support vector machine were applied on historical data of rainfall and water level recorded at different stations to develop tools to forecast water level at Sto. Niño at least one hour ahead. Results indicated that the random forest model performed better overall than other forecasting models since it has the lowest error metric and the highest correlation coefficient. To supplement the disaster management program and for budget planning purposes, a probabilistic approach was used to obtain long-term forecast in terms of annual frequency and duration of water level reaching at least a specified height. [ABSTRACT FROM AUTHOR]

Rainfall forecasting plays a crucial role in managing water resources, agricultural planning, and mitigating the risks of disasters. In this study, we propose an approach using multi-output neural network to predict rainfall rates specifically in East Java Province, Indonesia. The neural network model is trained using historical rainfall data obtained from eleven meteorological stations across East Java Province. To optimize the model's performance, we used resilient propagation algorithm during the training process. VAR (Vector Autoregressive) modeling was utilized to determine the significant lags to be used as input variables in the neural network model. This enabled us to capture the relationships and dependencies between different time steps. The model's performance was evaluated using two metrics, Root Mean Squared Error (RMSE) and Akaike's Information Criterion (AIC). The best-performing model was the Feed Forward Neural Network Model, which outperformed the VAR models, with a lower score of RMSE and AIC value. Furthermore, the model consistently delivered accurate predictions across all sampled meteorological stations, indicating its suitability and reliability for rainfall forecasting in East Java Province. The proposed multi-output neural network approach offers an effective solution for rainfall forecasting in East Java Province. Its implementation can significantly contribute for decision-making in water resource management and disaster mitigation. Providing accurate rainfall forecasts can aid in planning and preparing for potential weather-related challenges and minimizing the adverse impacts of extreme weather events in the region. [ABSTRACT FROM AUTHOR]

Interest rates are among the most important macroeconomic factors considered by financial market participants, both government and private entities, when making investment and policy decisions. Pricing and valuation of financial securities rely heavily on both current and future interest rates, among others. A reliable forecast of interest rates is a requisite to sound management of exposure to different types of risk including market risk, liquidity risk, and credit risk. Determining an optimal debt strategy for the government also requires reliable forecast of interest rates. These are among the reasons why the area of forecasting the term structure of interest rates has created a huge literature. Several theoretical models, such as factor models and stochastic models, have long been the forecasting tools used by big financial institutions. The advent of machine learning has recently provided good alternatives to theoretical models. This paper aims to add to the literature on the use of machine learning to forecast short-term interest rates by applying Artificial Neural Networks (ANNs). In particular, we propose two deep learning models – Multilayer Perceptrons (MLP) and Vanilla Generative Adversarial Network (VGAN) – to forecast the 1-month, 3-month, 6-month, and 1-year Philippine (PH) BVAL rates. As a base case, 16 economic variables were considered as inputs. Considering the principle of parsimony, we used the statistical methods of Backward Stepwise Regression and Variance Inflation Factor (VIF) Testing to also create models with smaller number of input variables. Thus, we applied MLP and VGAN to each of these three sets of input variables. We determined the best model for each interest rate tenor based on the graphical simulation and comparison of the actual and predicted values alongside the validation scores of the model. In general, both MLP and VGAN produced reliable forecasts of the interest rates. [ABSTRACT FROM AUTHOR]

This study article emphasises the Long Short-Term Memory (LSTM) model while utilising cutting-edge machine learning techniques to provide a thorough investigation into the subject of predicting the development of construction and demolition trash. For urban planners and environmentalists, the efficient management of building waste is a critical issue given global trends towards growing urbanisation. We carefully preprocessed the sizeable historical data sets from Shanghai and Hong Kong used in this study to ensure quality and representativeness. The study compares the LSTM model to machine learning techniques like Artificial Neural Networks (ANN), Random Forests (RF), and Support Vector Machines (SVM). When the model's ability to predict is thoroughly examined, it is clear that the LSTM routinely outperforms its competitors in terms of precision, F1 score, recall, and accuracy. Our findings also show how cutting-edge machine learning has the revolutionary potential to solve waste management and urban planning issues. LSTM and similar models provide a practical way to lessen environmental impact and lower waste disposal costs in urban settings by maximising resource allocation and promoting sustainable waste management practises. The ongoing discussion about garbage disposal in structures and sustainable urban development is furthered by this effort. It demonstrates the significance of LSTM and modern machine learning algorithms in transforming waste management paradigms and ultimately pointing us in the direction of a more environmentally friendly and sustainable future. [ABSTRACT FROM AUTHOR]

In-depth research on hybrid fiber-reinforced concrete (HyFRC), a substance with potential applications in structural engineering, is presented in this article. We examine the flexural characteristics of HyFRC, a composite made of steel, glass, and polypropylene fibres, with an eye towards how different fibre combinations affect the mechanical properties of concrete. To assess HyFRC's efficacy, a comprehensive experimental programme was run. At a total fibre volume fraction of 0.75%, the results demonstrated a noticeably higher compressive strength, splitting tensile strength, and flexural strength when steel and polypropylene fibres were combined. Effective fracture control allows the concrete to survive higher compressive loads, increasing compressive strength by up to 27% compared to non-fibrous control concrete. A crucial component of concrete performance, the splitting tensile strength in HyFRC was significantly increased thanks to the interaction between steel and polypropylene fibres. Along with better ductility, controlled fracture propagation, and greater strain capacity, HyFRC also shown these properties. The HyFRC also showed a significant increase in flexural strength, making it a suitable material for structural uses, particularly when combined with steel and polypropylene fibres. The features of the load-deflection demonstrated ductile behaviour, in stark contrast to the brittle failure seen in the control concrete. Concurrent with experimental study, an Artificial Neural Network (ANN) model with a remarkable total efficiency of 99.84% was developed, allowing precise predictions of flexural strength based on concrete composition. [ABSTRACT FROM AUTHOR]

In this study, we evaluate the accuracy with which two artificial intelligence models—Adaptive Neuro Fuzzy Inference System (ANFIS) and Artificial Neural Networks (ANN)—can estimate the output of solar energy in the context of a photovoltaic system. A solar power plant's actual data is employed, and input variables used include various sun radiation and meteorological characteristics. The primary objective is to forecast solar energy output since doing so will help solar energy generation systems be optimised. Performance is assessed using Mean Square Error (MSE) and the Coefficient of Determination (R2). ANFIS's accuracy of 70.42% illustrates how effective it is at locating particular data associations. In terms of prediction accuracy, the ANN model surpasses ANFIS, showing promising accuracy with an R2 of 77.77 and an MSE of 4.8517e+03. The study's results demonstrate that the Artificial Neural Network (ANN) model forecasts solar energy output significantly more precisely than ANFIS, underscoring its potential to improve the dependability and efficiency of solar energy generation systems and deepen our understanding of forecasting renewable energy sources. [ABSTRACT FROM AUTHOR]

In this investigation, the performance of powder-metallurgically manufactured cutting tools made of Fe, 2% C, 2% W, and 1.5 Ti will be assessed. The making of the cutting tool involves the processes of hot forging, sintering, machining, and induction hardening. The induction hardening procedure contributes to the better endurance of the cutting tool. The synthetic cutting tool is evaluated using the L9 orthogonal array. Tool wear and surface roughness are reaction parameters, whereas cutting speed, feed rate, and depth of cut are machining parameters. When turning AISI 1018 low carbon steel, it was discovered that the tool made of powder metallurgy had a superior surface polish than the traditional HSS tool. The experiment is carried out using the Taguchi technique, which is based on machine learning. The Signal to Noise ratio and the ANOVA statistical techniques were used to examine the impact and contribution of the machining parameter on the response outcome. An ANOVA revealed that feed rate has a greater effect on surface quality than cutting speed does on tool wear. Findings from confirmation tests show that the P/M cutting tool performed satisfactorily under the intended ideal machining conditions. The neural network that has been created predicts results more precisely. [ABSTRACT FROM AUTHOR]

The well-timed recognition of cancer and understanding the genetic paths that fuel its development are crucial for enhancing treatment efficacy and survival rates, with significant social and financial insinuations. Over the past two decades, researchers have explored the use of machine learning (ML) and deep learning (DL) methods to classify different types of cancer. These techniques hold promise for improving cancer diagnosis and treatment outcomes, ultimately contributing to better health outcomes for individuals artificial by cancer. In this paper is about the use of machine learning and deep learning techniques to classify DNA cancer data. We have two systems for classifying cancerous or normal DNA by K Nearest Neighbors (KNN), Probabilistic Neural Network (PNN), Long Short Term Memory (LSTM), and Convolution Neural Network (CNN) methods. These ways are improved using Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) in order to select the best features from the dataset towards an increase in diagnosis accuracy. The study assesses and appraises answers of the most correct algorithm used during classification by utilizing confusion matrices. Additionally, the paper considers processes of handling data such as deleting alike columns, normalizing, filling in nulls, etc., using missing value techniques. The research offers valuable insights into evolutionary optimized deep learning approaches for classifying DNA cancer data that may help improve outcomes of diagnosis and treatment for patients with cancer. [ABSTRACT FROM AUTHOR]

In this study, finger vein recognition technology has been focused to enhance security of biometric authentication. A new feature extraction technique has been proposed and applied on a dataset containing 3,816 finger-vein images from 106 individuals. This technique preserves the necessary data of images while decreasing the size and improve classification accuracy. It has multiple layers of neural networks that includes Conv1D, MaxPooling1D, LeakyReLU, and Dense layer to form a model. The results indicated that this model was very effective in detecting the venous structure having 27 layers with accuracy of 99.99%, a recall rate of 99.99%, an F1-score of 99.99% and Precision rate of 99.99%. The study indicated the potentialities and opportunities in developing new biometric and secure systems through the use of finger-vein recognition technology. [ABSTRACT FROM AUTHOR]

The primary objective of this study is to identify Distributed Denial of Service (DDoS) assaults by utilising Novel Deep Neural Networks that are coupled to recurrent neural networks in order to divert server traffic. Instructions & Materials Required: ClincalcGpower was utilised in order to ascertain the sample size, with the pre-test power being set at 0.8 and the alpha being set at 0.05. The identification of Distributed Denial of Service assaults was accomplished through the training of a Novel Deep Neural Network and a Recurrent Neural Network, each of which underwent ten iterations. As a whole, the collection has 20,000 different kinds of data. With everything taken into consideration, there are a total of 804 columns and 65 rows when it comes to the classification of DDoS attacks. In light of this, the mean accuracy of a Novel Deep Neural Network is a remarkable 92.76 percent, which is significantly higher than the accuracy of recurrent neural network technology because of this. which is 87.53 percent. The significance threshold of the accuracy is 0.001 when using an independent sample T-test with a p-value of less than 0.05. Given this information, it is possible to draw the conclusion that the two approaches are very different from one another. Based on the comparison of the two methods, it can be concluded that the Novel Deep Neural Network technique is superior than the recurrent neural network in terms of the average accuracy it achieves in recognising Distributed Denial of Service attacks. [ABSTRACT FROM AUTHOR]

The primary purpose of this research is to accurately diagnose lung nodules by contrasting and comparing three-dimensional spatial transform networks with innovative two-dimensional convolutional neural network models. The Instruments and Methods Used in Research: It is possible for the spatial transformer network to go through end-to-end training with the rest of the network due to the fact that it is integrated with the other network components. In this particular instance, they do accuracy calculations by utilising cutting-edge Matlab software and establish the sample size by utilising clinicalc.com. It is a difficult challenge for radiologists to identify nodules that are present in the lungs. One percent of the students are enrolled, the G power is eighty percent, and the threshold alpha is five percent. Automatic identification and categorization of pulmonary nodules in medical imaging has been accomplished through the development of a number of techniques, particularly those that make use of deep learning in conjunction with convolutional neural networks (CNN). In the next section, we will discuss the preliminary information that is required to make use of CNNs in lung nodule research, and we will also take a brief look at a few medical imaging datasets. When compared to the 2.5D CNN, which achieves an accuracy rate of 92.3 percent with a significance value of p=0.001 and a confidence interval of 95 percent, the findings of the research study indicate that the 3D spatial transform is superior with an accuracy rate of 89.29 percent for the 3D spatial transform. In terms of statistics, there is a considerable disparity between the two sets of data. In comparison to the 2.5D CNN Model, it would appear that the Novel 2.5D CNN, which has a mean value of 92.3 percent and a 3D spatial transform of 89.29 percent, comes out on top in terms of accuracy. [ABSTRACT FROM AUTHOR]

This research primarily aims to identify human action recognition (HAR) using a Conventional Neural Network (CNN) instead of an Artificial Neural Network (ANN) (ANN). The Parts and Methods: It is in this research piece that the CNN and ANN classifiers are put to use. The CNN classifier achieves a rate of 99.20% accuracy, whereas the ANN classifier only manages 92.55% accuracy. The publicly available Kaggle database was used as the raw material for this analysis. For the purpose of determining human action recognition, a sample size of twenty persons was utilised. Throughout the research, ten people participated in Groups 1 and 2. For this computation, we employed the G-power formula with alpha and beta coefficients of 0.05 and 0.2, respectively. The use of a 95% confidence interval was also adopted. The results show that Traditional Neural Network (CNN) classifiers have an accuracy rate that is 99.20 times greater than the Artificial Neural Network (ANN) classifier, which stands at 92.55%. With a rate of accuracy 99.20% greater than the present level, the CNN classifier outperforms the ANN classifier. The independent sample T test, which produced a p-value lower than 0.05, demonstrated that the difference between the two algorithms was statistically significant, with a value of 0.041. Finally, in terms of human action recognition prediction accuracy, the CNN classifier outperforms the ANN classifier. [ABSTRACT FROM AUTHOR]

Recently, numerous normalization layers for the purpose of stabilizing the training of deep neural networks (DNN) have been developed. Group normalization is one such technique that expands upon instance normalization and layer normalization by permitting some flexibility about the quantity of groups it employs. However, in order to ascertain the most effective number of groups, it is necessary to conduct time-consuming studies involving trial-and-error hyperparameter adjustment. For this study, we lay out a method that is both practical and effective for deciding how many groups to use. The initial observation is that the group normalization layer's gradient behavior is affected by the number of groups. By deducing the optimal group size, we may calibrate the gradient scale for optimization using gradient descent. For the first time, this research suggests a maximum group size that accounts for theoretical underpinnings, architectural concerns, and the capacity to independently produce adequate value for each layer. All sorts of neural network topologies, tasks, and datasets showed that the proposal method outperformed the state-of-the-art procedures. [ABSTRACT FROM AUTHOR]

Disasters are huge challenges to mankind. With the advent of modern and digital technologies most of the disasters can be predicted. But they cannot be a hundred percent prevented. However, the loss of life can be prevented or reduced by being disaster ready. The research work in this paper introduces a novel optimized framework of Disaster Response Supply Chain Management (DR-SCM) for effective management of relief operations by integrating AI, context based IOT, cloud and fog computing with remote sensed data. Framework proposes the collection of real time data obtained from remote sensing which is passed to the IOT framework. Fog computing connected to IOT framework reduces the burden on centralized cloud center by filtering the context-based data pertaining to the disaster area and builds the disaster dataset. Cloud centers run Artificial Intelligence (AI) models like Artificial Neural Network, Convolution Neural Network used to classify and predict disaster data for storing in knowledge base (KB). KB is used by DR-SCM for planning, allocation, and management of relief operations with optimized processes and logistics. [ABSTRACT FROM AUTHOR]

This article introduces the utilization of Artificial Neural Network (ANN), an algorithm within the domain of machine learning, to establish a forecasting framework for the damping efficacy of Magnetorheological Fluid (MRF) foam dampers. MRF foam stands as an intelligent material that allows for the modulation of its rheological characteristics through the influence of an external magnetic field. Predicting the damping performance is crucial for the development of effective MRF foam dampers for real-time applications. The proposed ANN-based model considers the damping force as the response of interest, which is influenced by the damper parameters such as foam density, viscosity of oil, and size of iron particles. Anticipated as a dependable instrument, the MATLAB simulation is poised to serve as a trustworthy resource for forecasting and approximating the damping force exerted by MRF foam dampers across a spectrum of pragmatic damper parameter configurations. [ABSTRACT FROM AUTHOR]

One essential component of human-computer interaction is gesture recognition, plays a pivotal role in our study on a Deep Learning-Driven Gesture Controlled Air Canvas. Analogous to emotions in the realm of human experiences,gestures serve as a rich non-verbal communication medium, shaping interactions with digital canvases. A Digital canvas is a white board that turns our hand gestures into required text. Our system introduces a groundbreaking flexible air canvas model that capitalizes on gesture recognition to enable users to manipulate a standard screen as their painting canvas. Gesture-Controlled air canvas empowers users to effortlessly generate artistic expressions directly through their gestureson the screen. The study aims to leverage deep learning techniques for enhanced gesture recognition. By integrating both vocal and linguistic features and employing deep neural networks for classification tasks, our research aspires to elevate the performance of gesture recognition systems. This study critically examines existing methodologies, addresses drawbacks in current techniques, and seeks to advance our understanding of these characteristics to foster the development of a more intuitive and emotionally resonant interaction with digital canvases. [ABSTRACT FROM AUTHOR]

In recent times, abstract text summarization has made great strides by moving away from linear models based on sparse and manually-crafted features and towards nonlinear neural network models that take use of rich inputs. Deep learning models have proven effective in NLP applications because they can model complex data patterns without the need for human-created features. Learning sequentially has acknowledged a lot of consideration in the last several years. Because entire training of encoder-decoder neural systems in tasks like machine translation has proven successful, research using similar architectures in other transduction tasks, such as paraphrase creation or abstractive summarization, has developed. In this study, we provide a neural network based abstractive text summarizer. The attention method was implemented to fix the issue of processing lengthy sequences of input text. A dataset including two news summaries was used to train the model. The goal of the approach is to make use of a mechanism for directing focus at the sentence level to help with focus at the word level. The results are superior than those of competing models in the research. The experimental findings on both datasets demonstrate that the provided model efficiently increases ROUGE scores and provides a more concise summary of the original document without losing any key details. [ABSTRACT FROM AUTHOR]

One application of mathematical modelling in finance is stock price modelling, which moves like a random walk. The movement of stock prices, moves up or down, does not follow a certain pattern and therefore requires proper modelling and analysis to predict the stock prices. This research uses two approaches, Geometric Brownian Motion (GBM) and Artificial Neural Network (ANN) to predict the daily close price. GBM is a stochastic model while ANN is one of the machine learning classification methods. These two methods are used to predict the daily closed price of four Indonesia BUMN Bank. Predicted price is compared with the actual price and used to measure the performance of each the method. [ABSTRACT FROM AUTHOR]

In the last few years, Brushless DC Motor (BLDC) was performed in different implementations, essentially sensitive applications; motors often run in hard loading conditions, especially after increasing production with complicated operating conditions. Therefore, overheating, overload, and other complex operating states are responsible for motor bearing faults. Bearing faults are essentially popular because of the details of drive operation. The monitoring and personifying bearing conditions are comparatively rare in many kinds of motor applications. Early diagnosis of bearing faults is meaningful in ensuring the seamless operation of BLDC motors in the industrial environment. The detection and classification of faults help to execute maintenance operations in an effective manner. This work shows an approach for identifying rolling elements imperfections in the BLDC motor in unsteady operating states. By using a Discrete Wavelet Transform (DWT), vibration and Stator current signal measurements represent the fault indexes to extract significant features; these features are highly reduced through the applications of Principle Components Analysis (PCA) to remove the redundant, irrelevant, or noisy data. A Recurrent Neural Network (RNN) and Feed Forward Neural Network (FFN) are used to expose and classify the existence of bearing faults and compare them from training time and accuracy; these networks were trained on the selected signal and can effectually recognize between healthy and fault signals, the accuracy of the faults diagnosis system equal to 99. 44% to recognize the kind of fault. The contribution of current work can be summarized as follows: due to the dynamic behavior of electrical machines, recurrent Neural Networks have been implemented instead of static Neural Networks that have been used in almost all previous work. Furthermore, the proposed technique can diagnose faults under stationary and non-stationary operating conditions. [ABSTRACT FROM AUTHOR]

An artificial neural network approach is presented in this article to simulate a fault detection scheme and classify the fault type of distribution lines. A new approach that combines neural network modules with digital logic provides the protection technology's fault detection and classification abilities. Inputs used for the proposed training design were samples from 3-phase voltage and current, voltage with current during the zero sequences, and angle during the negative sequence. The network has been trained using the Levenberg-Marquardt algorithm. Digital logic with artificial intelligence was used to make an accurate final classification decision. The data for the distribution line was generated with simulation MATLAB (11-kilo volt). The work with an integrated system has been tested. The findings demonstrated that the artificial neural network performance is outstanding and accurate, given that it can classify various faults. This can be seen from the training results, where the confusion matrix showed the results of 91.7%, and the training regression level was one or very close to it, while the mean square error was very low. [ABSTRACT FROM AUTHOR]

This review paper examines the application of back propagation and expert system algorithms in artificial neural networks for the medical diagnosis of diabetes. With the high prevalence of diabetes worldwide, timely diagnosis of addiction is essential for effective control. This paper will discuss issues related to data availability, interpretation, and scalability and explore potential future directions. This review paper serves as a valuable resource for researchers and healthcare professionals interested in leveraging artificial neural networks for the diagnosis of diabetes, facilitating advancements in this critical field of study. [ABSTRACT FROM AUTHOR]

Fault detection is a critical stage in power systems, requiring swift and accurate identification. This paper introduces an Artificial Neural Network (ANN) tailored for fault identification, utilizing a deep learning approach. The proposed model marks a pioneering effort, employing ANN for fault identification using raw and sampled data from three-phase current and voltage signals collected through simulation from a 14-bus system across various fault and no-fault classes. Performance evaluation metrics such as accuracy, sensitivity, specificity, precision, and F1 score are employed to assess the model's efficacy. Results demonstrate that the ANN model achieves exceptional accuracy and efficiency in fault detection, showcasing its potential as a pivotal tool for enhancing the reliability and performance of electrical power systems. Notably, the suggested model attains a remarkable 99% accuracy across all tested fault scenarios, F1Score 0.99 for fault and non-fault detection with closely matched testing and training loss curves. [ABSTRACT FROM AUTHOR]

This study presents the development and evaluation of a medical diagnostic system for diabetes utilizing Artificial Neural Networks (ANNs) with Back Propagation and an Expert System Algorithm. The dataset, sourced from Kaggle, encompasses a comprehensive array of medical parameters related to diabetic patients. After preprocessing, the dataset was divided into training, validation, and testing sets. A Multi-Layer Perceptron (MLP) classifier model was trained using the training data and assessed on the validation and testing sets, achieving a validation accuracy of 97.72% and a testing accuracy of 71.43%. Examination of the confusion matrix provided insights into the model's performance strengths and areas for improvement. Integration of an expert system algorithm enhanced the diagnostic system's robustness. The study highlights the potential of ANNs in medical diagnostics, albeit with acknowledged limitations, and suggests avenues for future research to optimize model performance and expand dataset generalization. Overall, this study marks an important step forward in the use of artificial intelligence for diabetes diagnosis and improved patient care. [ABSTRACT FROM AUTHOR]

This review paper finely analyzes eye-to-text communication and explores the development and evaluation of an advanced eye-controlled communication system. The study focuses on the algorithms for eye tracking and gaze estimation, the Artificial Neural Networks (ANNs), and the training process for converting eye movements into textual output. The article highlights Computer Vision (CV) algorithms utilized for real-time eye tracking and gaze estimation, emphasizing the robustness and accuracy of the algorithms. The authors detail the techniques employed, including pupil detection, iris segmentation, and gaze estimation, highlighting their effectiveness in capturing and analyzing eye movements. Moreover, the article discusses challenges faced during training that provide insights into potential improvements for future work. The review paper presents comprehensive experimental results, including ANN comparisons with existing methods such as gaze estimation error and user satisfaction, thoroughly assessing the system's capabilities. The system enables users to express themselves and interact with digital devices more independently, enhancing the Quality of Life (QoL) for individuals with limited motor abilities. This review aims to pinpoint the deficiencies left unaddressed by researchers, including issues such as head motion, low illumination, low-resolution cameras, and user fatigue. These identified shortcomings emphasize the pressing demand for eye-to-text communication systems. [ABSTRACT FROM AUTHOR]

Pneumonia is a prevalent lung disease worldwide, and timely recognition is crucial. Machine learning has been increasingly applied in various fields, including healthcare, but it has limitations when dealing with a small dataset for training models. In this study, we focused on diagnosing pneumonia using X-ray images and developed an approach that employs the ANN algorithm. We used a small dataset of RGB images to test, train, and validate the ANN, and we tested the algorithm with 15 neurons. This paper examined 500 patient X-ray pictures from the Kaggle dataset for normal and abnormal cases. Ten thousand pixels were selected from these images, with 7,000 pixels allocated for training the ANN and 1,500 pixels each for testing and validation. The outcomes showed that the ANN with 15 neurons achieved remarkable performance on all performance indices, including specificity, F1-score, accuracy, sensitivity, and precision, with a score of 99.9%. Furthermore, the suggested approach outperformed the current solution for pneumonia diagnosis utilizing the ANN model. [ABSTRACT FROM AUTHOR]

In the landscape of predictive analytics, commendable success has been garnered by Machine Learning models in the accurate estimation of age. The critical ability to discern deviations from typical age progression, which may signify accelerated aging or variations from the norm, underscores the demand for diagnostic techniques that are both efficient and accurate. Over time, contributions have been presented for the task that makes use of data-driven techniques for modeling the system, with recent attention being centered on the popularity of deep neural networks, commonly known as deep learning. This review aims to serve as an in-depth analysis of the existing literature on deep learning in age estimation, including a thorough overview of the main deep learning architectures implemented for this end. A quantitative assessment of the collective body of publication research in the field to date is considered. Moreover, the various approaches to age assessment are criticized in the evaluation, indicating the few but significant advantages and disadvantages of each used framework. Unexamined routes as well as really innovative ways of dealing with age estimation in the rapidly changing field are researched in this article. Finally, as a result of such thorough examination, the paper aims to become a very useful and solid resource for the audience of both novice and expert students equally. [ABSTRACT FROM AUTHOR]

Blind or visually impaired people face several challenges in safely navigating their environments due to their lack of inclusive real-time obstacle detection and surrounding information aid. In this paper, a smart stick for visually impaired people was designed and practically implemented. The smart stick system contains heart rate, ultrasound, and moisture sensors. The system's core utilizes an Arduino Nano microcontroller programmed in C++ and GPS, GSM/GPRS, a vibration motor, a buzzer, and a battery. Ultrasound and soil moisture sensors are used to detect the obstacles and the nature of the floor faced by the visually impaired person during navigation in various environments. The heart rate was employed to monitor the health condition of the visually impaired. Moreover, the position of the visually impaired individual was tracked in real-time based on GPS, and the geolocations were transmitted to the family members through the GSM/GPRS network to facilitate the monitoring of the user. The vibration motor and buzzer alert the user to unwanted environments, like obstacles, wet surfaces, or elevated heart rates. The performance of the alert decision was enhanced by utilizing an Artificial Neural Network (ANN) implemented in Matlab software. The analysis results revealed exceptionally high correlation coefficients (R²) across all datasets: 0.997 for training, 0.988 for testing, 0.982 for validation, and 0.994 for all combined datasets, as determined by the applied ANN. Furthermore, the alert decision error was enhanced to 4.096×10-5. This error revealed that applying ANN resulted in a remarkable reduction in errors relative to those obtained from statistical analysis of differences among the smart stick system and benchmark systems. [ABSTRACT FROM AUTHOR]

The progress in technology and the swift development of brain imaging techniques have had a lasting and crucial impact on transforming our comprehension of brain structure and operation. The utilisation of image-processing techniques in the realm of medical science has proven to be highly advantageous in improving the early diagnosis and treatment of various conditions. Deep neural networks (DNNs) have repeatedly exhibited exceptional performance in various tasks, such as classification and segmentation. Expanding upon the aforementioned groundwork, this study presents an innovative methodology for Convolutional Neural Network (CNN) implementation, employing diverse optimisation algorithms. This method combines different pre-processing techniques, resulting in a substantial reduction in feature set size for subsequent classification tasks using CNN. In order to assess the efficacy of this methodology, a dataset consisting of brain images was employed, and the performance of the suggested classifier was juxtaposed with that of other well-established classifiers, including a pretrained model. The results clearly indicate that the proposed CNN model with Adam optimizer achieved 89.7 accuracy method surpasses the performance of existing approaches. [ABSTRACT FROM AUTHOR]

In telecommunications, the spectrum encompasses a range of allocated frequencies crucial for data transmission, constituting a finite resource. Historically, manual allocation processes have led to suboptimal utilization, leaving a substantial portion of this spectrum untapped. This research delves into the integration of Artificial Intelligence (AI) into the realm of telecommunications, with a primary focus on optimizing spectrum management and tackling prevalent industry challenges. The core objective of this study is to leverage the potential of existing AI and Machine Learning algorithms, particularly Artificial Neural Networks (ANNs), to enhance various facets of the telecommunications sector. These enhancements encompass Predictive Maintenance, Virtual Assistance, Network Optimization, Fraud Prevention, and Revenue Growth. To facilitate the seamless integration of AI, a proposed architecture includes vital components like an information observation base, a learning engine, and an application programming interface (API) designed to facilitate efficient communication with the telecommunications network. Several AI techniques, including anomaly identification, prediction algorithms, recommendation systems, and detection/classification algorithms, are put forth to effectively address diverse aspects of telecom spectrum management. While AI applications in telecommunication services are still in their nascent stages, this research envisions promising outcomes that will alleviate spectrum shortages and enhance overall industry efficiency and competitiveness. The proposed architecture exhibits adaptability, making it versatile enough to address a wide range of applications, spanning from predicates to activities and objective definitions. The research project's potential impact on the telecommunications sector is substantial, as AI holds the capability to streamline intricate network management, optimize resource allocation, and elevate network monitoring and control. As data science tools and AI applications continue to advance, the telecom industry can anticipate accelerated progress and heightened competitiveness. The amalgamation of AI, Machine Learning, and robust big data tools offers a potent solution to address critical industry challenges, ultimately benefitting both service providers and consumers. [ABSTRACT FROM AUTHOR]

Expansion of land cover (LC) is significantly impacted by man-made activities, particularly in areas that are expanding quickly. A CA-ANN model, which stands for cellular automata and artificial neural network, the combination is one of the many models that are commonly employed to evaluate potential changes in land cover (LC) using satellite images using Landsat 5 TM, Landsat 7 ETM+ and Landsat 8 OLI with spatial resolution 30 m. Examining shifts in LC patterns was the driving force behind this article over the previous thirty years (1993 to 2023) in Kirkuk City, Iraq. The available satellite pictures were analyzed using the Maximum Likelihood Supervised Classification (MLSC) approach in this article. The output of this study shows that from 1993 to 2023, the urban area is projected to rise by almost 215%, while the water area increased by nearly 150%, but the vegetation areas decreased from 1993 to 2023 by about 150%, however, the bare land slightly increased by about 2%. The CA-ANN model, whose kappa value was 0.97 and accuracy percentage was 93.06%, was made to make future LC expectations. The findings from the LC prediction model for the future indicate that by 2033 and 2043, the urban area will grow by 28% and 28.3% respectively, even though there will be less vegetation and less barren ground, factors that significantly influence urban planning and their implications for environmental management such as Environmental quality, City scale, Economic level, Industrial structure, Population structure and Government management and services. This study's results may inform future efforts to improve the advancement of long-term sustainability in urban design and administration, as well as inform decision-making processes by governing bodies focused on improving ecological and environmental conditions within the designated study region. [ABSTRACT FROM AUTHOR]

Magnetorheological (MR) material is a smart material composed of one of the magnetic particles of carbonyl iron (CIP) in a non-magnetic medium that can be controlled via magnetic fields. One of the groups of MR materials is MR plastomer (MRP) which consists of CIPs in the low-crosslinking polymer network such as plastomer, and it is introduced to compromise the limitations in MR fluid and MR elastomer. Somehow, the new modern approach particularly machine learning needs to be highlighted to acquire accuracy in the prediction trend of the MRP's properties, including viscosity. In fact, the approach may reduce time and work consuming in the development process of MRP. Therefore, using an artificial neural network (ANN) as one of the machine learning approaches is an effective medium for forecasting the behavior of the material and resolving constitutive model concerns. [ABSTRACT FROM AUTHOR]

In the rapidly evolving landscape of digital communication, ensuring the reliability and security of networks is paramount, with the escalating volume and complexity of network traffic necessitating robust anomaly detection mechanisms to safeguard against potential threats and disruptions. This paper introduces a novel hybrid deep learning system designed for effective network anomaly detection, employing a two-stage approach. The first stage utilizes a Convolutional Neural Network (CNN) as a binary classifier to identify broad anomalous patterns in network traffic. Subsequently, the second stage employs a Deep Neural Network (DNN) as a multi-class classifier to further categorize the detected anomalies into specific classes. Notably, a unique feedback mechanism is incorporated into the system, leveraging false results, including false negatives and false positives, to iteratively refine the model through retraining and reclassification, thereby mitigating false results. An innovative aspect of the proposed approach lies in the incorporation of a feedback loop that addresses false results. False negatives and false positives identified during the classification process are used to augment the training dataset, allowing the system to learn from its own mistakes. The model is then retrained, and the classification process is repeated, resulting in a more resilient and adaptive anomaly detection system. To evaluate the performance of the proposed hybrid system, extensive experiments are conducted using the CICIDS2017 dataset, a widely recognized benchmark in the field of cybersecurity. Comparative analyses with existing methods showcase the superior capabilities of the hybrid system in terms of accuracy, false positive reduction, and adaptability to evolving network threats. The results indicate that the hybrid deep learning system not only outperforms traditional approaches but also exhibits an enhanced ability to learn and adapt from misclassifications, thereby continually improving its performance over time. In conclusion, the presented two-stage hybrid deep learning system, coupled with the adaptive feedback mechanism, represents a significant advancement in network anomaly detection, promising heightened accuracy and robustness in the ever-evolving landscape of cybersecurity. [ABSTRACT FROM AUTHOR]

Target strategy is an important part of low-carbon city construction, and reasonable formulation of target strategy can effectively improve resource utilization, reduce environmental pollution, and promote sustainable urban development. Therefore, this paper proposes an artificial neural network algorithm, which comprehensively analyzes the connotation of artificial intelligence, and discusses the low-carbon city research model under the background of artificial intelligence to determine the corresponding target strategy analysis. Firstly, artificial intelligence theory is used to analyze urban construction. The artificial intelligence the low-carbon city construction strategy, forms a strategy evaluation scheme, the target strategy results. the evaluation accuracy, energy efficiency and carbon emission level of low-carbon city construction strategy by artificial neural network algorithm are better than those of previous urban planning. [ABSTRACT FROM AUTHOR]

This paper delves into the training and testing of two pre-trained models of a Convolutional Neural Network (CNN) to classify images of clothing from the Fashion MNIST dataset and determine the classification accuracy and performance of both the models. The Fashion MNIST dataset is a collection of 60,000 28x28 grayscale images of 10 different types of clothing. The images are well-labeled and comparatively easy to classify, making them a good starting point for learning about CNNs. The article starts with a summary of the Fashion MNIST dataset. There are 60,000 training photos and ten thousand test ones in the dataset. T-shirt/top, Trouser, Pullover, Dress, Coat, Sandal, Shirt, Sneaker, Bag and Ankle boot are the pictures' ten constituent classes. The CNN architecture used in the tests is described in the next part of the study. CNN architecture is made up of two convolutional layers, two max-pooling layers, and two fully linked layers. The convolutional layers employ 32 3x3 filters, whereas the max pooling layers employ a pool size of 2x2. There are 128 neurons in the fully linked layer and 10 neurons in the unconnected layer. The Adam optimizer is chosen to train the CNN framework utilizing the learning rate of 1e-5. The machine learning algorithm underwent training for 30 epochs with a batch size of 64. The experiments' shortcomings are highlighted in the conclusion section. It is certainly possible that the CNN model would not perform as well on other datasets. Another limitation is that the experiments were conducted only for a set of hyperparameters. The outcomes of the experiments are presented in the results section. The training accuracy of the Convolutional Neural Network models VGG19 and Resnet50 are 94.58 and 99.41 respectively. The testing accuracies for VGG19 and Resnet50 models are 91.92 and 90.25 respectively. The inference latencies for both cases were found to be 5.439 and 3.522 respectively. [ABSTRACT FROM AUTHOR]

In recent years there has been a growing interest in stock market investment as it helps to accumulate wealth for investors, which tends to beat the returns obtained by traditional investment instruments (Fixed deposits, PPF etc.). But investors are not aware of the stock market behavior i.e., they do not know which stocks to buy or to sell to gain profit. Considering this problem and the opportunity to use Neural Networks for prediction, in this project we have developed a neural network model using LSTM and trained the model with different stock market indicators data as input (for various stocks like TCS, INFOSYS, HDFC, and DIVISLAB) for an accurate prediction of the stock prices. The indicators used are RSI and Moving Averages. This model now predicts the closing price of the next day which helps the investors to select the stocks to invest in, that bring profit. Also, a comparison is made to identify the best indicator, giving more accurate results. [ABSTRACT FROM AUTHOR]

The main objective of the research is to evaluate the effectiveness of two methods for human behavior recognition forecasting: one uses an artificial neural network technology, and the other uses support vector machines. To determine the sample amount for the expected result of human behavior recognition using the artificial neural network approach, we utilize G-power 0. 8 using an alpha of 0.8 and a confidence limit of 95%. Group 1 has 25 participants and Group 2 has 25. While Support Vector Machine achieves a precision of 91.8469, New Artificial Neural Network achieves 97.3989. A newly developed artificial neural network significantly surpasses support vector machines from incorporating the top aspects of both approaches. [ABSTRACT FROM AUTHOR]

This study investigates a new method for COVID-19 outbreak prediction by means of an Artificial neural network (ANN) using convolutional layers (CNN). Compared to with an Adaptive boosting classifier. In this method, two groups are used. With the help of CNNs, one group and two more groups with Ada boost method employed a 20-person sample for each research to evaluate the precision of COVID-19 predictions. Applying G power and a pretest power of 80%, we determined the sample size. The CNN and Ada boost algorithms were trained using chest X-rays of both healthy and COVID-19 afflicted people. The results show that CNN achieved 98.5% accuracy and Ada boosted 79% accuracy, with a p-value of less than 0.05 indicating statistical significance. To review: Deep Learning Network algorithm has achieved better accuracy compared to the Ada boost classifier. [ABSTRACT FROM AUTHOR]

This research tests two neural network types—artificial neural networks and convolutional neural networks—to see whether one is better at predicting the spread of COVID-19. We evaluated the performance of Convolutional A Team Approach to COVID-19 Pandemic Forecasting with the Use of ML and CNNs using two sets of 20 samples. G power, which had an 80% pretest power, was used to compute the sample size. Data models were developed using Chest X-rays from healthy individuals and COVID-19 patients were used in conjunction with artificial neural network (ANN) and convolutional neural network (CNN) methods. The results show that CNN has a 98.5% accuracy rate and ANN 95.33% accuracy rate, with a p-value less than 0.05 indicating statistical significance. When compared with the Artificial Neural Network method, the Convolutional Neural Network method yields better results in terms of accuracy. [ABSTRACT FROM AUTHOR]

To be more specific, we are interested in comparing the accuracy of Support Vector Machine with a novel artificial neural network classifier when it comes to fingerprint verification. We compare the new method of artificial neural networks to an SVM. Following evaluating the suggested procedure with a combined total of fifteen samples in both groups, the G power value stands at eighty percent. When it comes to identifying biometric identity, the Support Vector Machine technique only manages 83% accuracy; in contrast, the Novel Artificial Neural Network technique approach achieves an impressive 87% accuracy. With a protest power analysis completed at 85% and test sizes of 15 for one group and 30 for the overall set, the Novel Artificial Neural Network approach outperformed SVM marginally in terms of mean accuracy and standard deviation (1.25369 vs. 1.30700). The Novel Artificial Neural Network Method (1.25369) outperforms the SVM method (1.30700) with a significant value of 0.01 (p<0.05) in terms of mean exactness and standard deviation, although the SVM algorithm (83.10%) ranks somewhat lower. The results demonstrated that when it came to biometric expectancies, the Novel Artificial Neural Network Method Algorithm outperformed the SVM algorithm. [ABSTRACT FROM AUTHOR]

In order to maintain good health and avoid illness, people need access to potable water. People often become sick from water-related causes because they drink polluted water. This is why monitoring the water quality is essential. The Novel Tensor model is built using time-, location-, and water-related variables. A support vector regression approach is used to forecast the novel water quality by the suggested system, which takes water samples from the DWS system. This research employed a dataset consisting of about 1993 samples, with 127 samples analyzed statistically using SPSS. The results suggest that this new tensor has the potential to produce the best results with the most accuracy (p=.676). The goal of this research is to improve water quality monitoring by using a novel approach to existing image processing classifiers. When compared to older, less sophisticated algorithms, modern artificial neural network methods perform far better. [ABSTRACT FROM AUTHOR]

This research intends to evaluate two classifiers, an ANN and an NB, to achieve a higher degree of precision in tumor identification using bone x-ray images. Things You Must Have and Perform: The NTHU Computer Vision Lab has made their data available in a publicly accessible dataset, which is utilized in this work. Our capacity to categorize tumor cells from bone x-ray photographs was proved with a beta of 0.2 and an alpha range of 95% using G-power 0.8. The research made use of 280 specimen's total, 140 of which were split evenly between Group 1 and Group 2. Utilizing the Naïve Bayes (NB) classifier with a sample size of 10, the Artificial Neural Network (ANN) is able to detect and group cancer cells in bone x-ray pictures. With a score of 92.0988%, the ANN classifier achieves better results than the NB classifier. With a p-value of only 0.031, the results of this investigation are considered statistically significant. Compared to NB classifiers, ANN classifiers perform better when it comes to identifying and categorizing cancer cells in x-ray pictures of the bone. [ABSTRACT FROM AUTHOR]

The main objective of the research is to evaluate the efficacy, specificity, and sensitivity of two approaches—a support vector machine (SVM) and an artificial neural network (ANN)—in differentiating between benign and malignant lung tumours. For this investigation, data sets were acquired from Github. For the ANN and SVM, a total of 40 lung CT scans were employed. With a review course power of 0.8 and a maximum acceptability error of 0.5, the sample size in each group is examined. When used to identify cancers in lung CT images, Artificial Neural Networks perform with an accuracy of 92%, whereas Support Vector Machines do better with an accuracy of 84% and a significant level of p 0.05. When it comes to distinguishing between benign and malignant lung tumours, the artificial neural network technology outperforms the Support Vector Machine method in terms of overall accuracy, sensitivity, and specificity. [ABSTRACT FROM AUTHOR]

This project aims to compare coalition game theory with state-of-the-art artificial neural networks in order to optimize the data transmission rate for secondary users of cognitive radio networks. Actions to Take and Items to Bring: The SIMATS system that conducted the desired operation. This experiment made use of two groups. Twenty people were divided up between the two groups at random. Technologies for spectrum identification that employ innovative artificial neural networks were included in Group 1. The cognitive radio network's major actors prepared the second group's samples. To evaluate the Coalition Game Theory–based algorithm, we used 80% of G power. The difference between the two data sets is 0.018 (p 0.05). The finished item: While coalition game theory yielded an accuracy of 89.0815, the simulation results showed that the artificial neural network achieved an average accuracy of 96. 7347. In the end Artificial neural networks optimize throughput in secondary user communication more accurately and precisely than Coalition game theory. [ABSTRACT FROM AUTHOR]

The primary goal of this work is to improve brain tumor new photo classification accuracy by using ant colony optimization instead of artificial neural networks. The ANN and ant colony were optimized using Matlab software. The classification accuracy of brain tumor fresh pictures is investigated in both the ANN and the ANT colonies. With a G-power value of 80%, brain malignancies may now be classified by a revolutionary image method using ANN with a sample size of 23 and ant colony optimization. After the study was completed, it was found that the accuracy of Ant colony optimization is 74.68%, which is much greater than the accuracy of ANN, which is 60.51%. The significance threshold, or p<0.05, is p=0.03. Because of its great accuracy, ant colony optimization outperforms the ANN algorithm, according to the results of the examination of the two algorithms' results. [ABSTRACT FROM AUTHOR]