Optimal Weight Parameters Research Articles

CAE-ResVGG FusionNet: A Feature Extraction Framework Integrating Convolutional Autoencoders and Transfer Learning for Immature White Blood Cells in Acute Myeloid Leukemia

Acute myeloid leukemia (AML) is a highly aggressive cancer form that affects myeloid cells, leading to the excessive growth of immature white blood cells (WBCs) in both bone marrow and peripheral blood. Timely AML detection is crucial for effective treatment and patient well-being. Currently, AML diagnosis relies on the manual recognition of immature WBCs through peripheral blood smear analysis, which is time-consuming, prone to errors, and subject to inter-observers' variation. This study aimed to develop a computer-aided diagnostic framework for AML, called "CAE-ResVGG FusionNet", that precisely identifies and classifies immature WBCs into their respective subtypes. The proposed framework leverages an integrated approach, by combining a convolutional autoencoder (CAE) with finely tuned adaptations of the VGG19 and ResNet50 architectures to extract features from CAE-derived embeddings. The process begins with a binary classification model distinguishing between mature and immature WBCs followed by a multiclassifier further classifying immature cells into four subtypes: myeloblasts, monoblasts, erythroblasts, and promyelocytes. The CAE-ResVGG FusionNet workflow comprises four primary stages, including data preprocessing, feature extraction, classification, and validation. The preprocessing phase involves applying data augmentation methods using geometric transformations and synthetic image generation using the CAE to address imbalance in the WBC distribution. Feature extraction involves image embedding and transfer learning, where CAE-derived image representations are used by a custom integrated model of VGG19 and ResNet50 pretrained models. The classification phase employs a weighted ensemble approach that leverages VGG19 and ResNet50, where the optimal weighting parameters are selected using a grid search. The model performance was assessed during the validation phase using the overall accuracy, precision, and sensitivity, while the area under the receiver characteristic curve (AUC) was used to evaluate the model's discriminatory capability. The proposed framework exhibited notable results, achieving an average accuracy of 99.9%, sensitivity of 91.7%, and precision of 98.8%. The model demonstrated exceptional discriminatory ability, as evidenced by an AUC of 99.6%. Significantly, the proposed system outperformed previous methods, indicating its superior diagnostic ability.

Application of The Double Exponential Smoothing Brown Method to Consumer Price Index Forecasting in Sorong City before and during The Covid-19 Pandemic

Application of The Double Exponential Smoothing Brown Method to Consumer Price Index Forecasting in Sorong City before and during The Covid-19 Pandemic

An Improved and Optimized Gated Recurrent Unit and Long Short-Term Memory Model for Fake News Detection

This study presents a novel approach for detecting counterfeit news, employing an advanced hybrid model that integrates Enhanced Gated Recurrent Unit and Long Short-Term Memory networks, termed as IGRU-LSTM. Initially, the database is assembled from the Information Security and Object Technology (ISOT) database and Wikipedia databases. From the database, the real and fake news is detected by considering news reviews. The dataset may contain unwanted information line URLs and symbols, which should be corrected to achieve efficient fake news detection. So, the pre-processing technique should be considered such as special symbol removal, URLS removal, upper to lower case conversion and replace contractions. After that, the pre-processed data is sent to the embedding procedure for word embedding. Finally, the IGRU-LSTM classifier is utilized for classifying real and fake news detection. In the combined GRU-LSTM framework, we incorporate the Enhanced Wild Horse Optimisation (EWHO) algorithm to optimize the selection of optimal weighting parameters. We utilize MATLAB for implementing this method. To evaluate the effectiveness of our approach, we analyze key performance metrics like precision, recall, and accuracy, and compare them with established methods including CNN-PSO, CNN-FO, and standard CNN.

Estimation of weight window parameters based on recursive Monte Carlo approach for reactor shielding problems

Weight window, a variance reduction tool, is often used to improve the performance of radiation shielding calculations. One common issue with the weight window is determining the optimal set of weight window parameters for solving deep penetration shielding problems. To address this issue, the recursive Monte-Carlo methodology has been used with the in-house TRAWEI code. The program is responsible for generating the optimal weight parameters in a single run with minimum computational time. This paper presents the results of a numerical test conducted using a simple reactor model to evaluate the performance of the developed weight generator program. The findings reveal that MCNP simulations utilizing TRAWEI-generated weight values exhibit significantly higher calculation efficiency compared to both analog simulation and MCNP simulation using weights generated by the existing MCNP weight window generator. Overall, the utilization of the RMC methodology has shown its potential to significantly contribute to deep penetration shielding calculations.

An Optimal Nonlinear Model Predictive Control- Based Motion Cueing Algorithm Using Cascade Optimization and Human Interaction

Nonlinear model predictive control has been used in motion cueing algorithms recently to consider the nonlinear dynamics model of the system. The entire motion cueing algorithm indexes, including the physical and dynamical constraints of the actuators and physical constraints of passive joints, can be controlled with precision using nonlinear model predictive control. However, several weighting parameters in the nonlinear model predictive control-based motion cueing algorithm (including driving sensation, motion description of the actuators, and passive joints) require proper and laborious tuning to attain an optimal design structure. In this work, the optimal weighting parameters of a nonlinear predictive control-based motion cueing algorithm model are calculated using cascade optimisation and human interaction. A cascade optimisation method consisting of a particle swarm optimisation and genetic algorithm is designed to identify the best weighting parameters compared to those from one optimiser. In addition, the human decision-making units are added to the two-level cascade optimiser to determine the best solution from a Pareto front. The proposed cascade optimiser decreases the run-time with better extraction of the optimal weighting parameters to increase the motion fidelity compared to a single optimiser. It should be noted that the proposed methodology is applied along longitudinal channel. While the same methodology can be applied along lateral, heave and yaw channels for further evaluation of the proposed method. The proposed model is simulated utilising the MATLAB software and the results prove the efficiency of the newly proposed model compared to those from the previous single optimiser in reproducing more accurate motion signals with better usage of the driving motion platform workspace.

Optimal Weighting of Preclinical Alzheimer's Cognitive Composite (PACC) Scales to Improve their Performance as Outcome Measures for Alzheimer's Disease Clinical Trials.

Cognitive composite scales constructed by combining existing neuropsychometric tests are seeing wide application as endpoints for clinical trials and cohort studies of Alzheimer's disease (AD) predementia conditions. Preclinical Alzheimer's Cognitive Composite (PACC) scales are composite scores calculated as the sum of the component test scores weighted by the reciprocal of their standard deviations at the baseline visit. Reciprocal standard deviation is an arbitrary weighting in this context, and may be an inefficient utilization of the data contained in the component measures. Mathematically derived optimal composite weighting is a promising alternative. Sample size projections using standard power calculation formulas were used to describe the relative performance of component measures and their composites when used as endpoints for clinical trials. Power calculations were informed by (n=1,333) amnestic mild cognitive impaired participants in the National Alzheimer's Coordinating Center (NACC) Uniform Data Set. A composite constructed using PACC reciprocal standard deviation weighting was both less sensitive to change than one of its component measures and less sensitive to change than its optimally weighted counterpart. In standard sample size calculations informed by NACC data, a clinical trial using the PACC weighting would require 38% more subjects than a composite calculated using optimal weighting. These findings illustrate how reciprocal standard deviation weighting can result in inefficient cognitive composites, and underscore the importance of component weights to the performance of composite scales. In the future, optimal weighting parameters informed by accumulating clinical trial data may improve the efficiency of clinical trials in AD.

Generation of Optimal Weight Values Based on the Recursive Monte Carlo Method for Use in Monte Carlo Deep Penetration Calculations

In nuclear engineering, Monte Carlo (MC) methods are commonly used for reactor analysis and radiation shielding problems. These methods are capable of dealing with both simple and complex system models with accuracy. The application of MC methods experiences challenges when dealing with the deep penetration problems that are typically encountered in radiation shielding cases. It is difficult to produce statistically reliable results due to poor particle sampling in the region of interest. Therefore, such calculations are performed by the Monte Carlo N-Particle Transport (MCNP) code in association with the weight window (WW) variance reduction technique, which increases the particle statistics in the desired tally region. However, for large problems, MCNP’s built-in weight window generator (WWG) produces zero WW parameters for tally regions located far from the source. To address this issue, the recursive Monte Carlo (RMC) method was proposed. This paper focuses on the RMC methodology and its implementation in the Helmholtz-Zentrum Dresden-Rossendorf’s (HZDR’s) in-house code TRAWEI, which is responsible for producing optimal zone weight parameters used for optimizing deep penetration MC calculations. In addition, this paper discusses the verification of the TRAWEI weight generator program to that of an existing MCNP WWG. The performance of TRAWEI-generated weight values is assessed using a handful of test cases involving two shield materials. Globally, the TRAWEI-generated weight values improved not only the statistical variance and computational efficiency of the MC run compared to the analog MCNP simulation but also those of the simulation with WW values generated by the standard MCNP WWG.

Enhanced conditional random field‐long short‐term memory for name entity recognition in English texts

SummaryNamed entity recognition (NER) is an essential topic in the real world during the advanced development of technologies. Hence, in this article, develop enhanced conditional random field‐long short‐term memory (ECRF‐LSTM) for NER in the English language. The proposed ECRF‐LSTM is the combination of conditional random field‐long short‐term memory (ECRF‐LSTM) and chaotic arithmetic optimization algorithm (CAOA). The proposed research concentrated to perform NER for Indian names from the given input database for Indian digital database management and processing. The Chaotic AOA leads to fast convergence and helps to avoid the local optima. The proposed method is working with three phases preprocessing phase, the feature extraction phase, and the NER phase. In the initial stage, the datasets are collected from the online system. In the preprocessing phase, the removal of the URL, removal of special symbol, username removal, tokenization, and stop word removal are done. After that, the essential features such as domain weight, event weight, textual similarity, spatial similarity, temporal similarity, and relative document‐term frequency difference are extracted and then applied to train the proposed model. To empower the training phase of the CRF‐LSTM method, CAOA is utilized to select optimal weight parameter coefficients of CRF‐LSTM for training the model parameters. The proposed method is validated by statistical measurements and compared with the conventional methods such as convolutional neural network‐particle swarm optimization and convolutional neural network respectively. The proposed method achieves 98.91 accuracy, 97.36 sensitivity, 97.19 specificity, 97.54 precision, and 97.63 recall which is better than the existing methodologies.

Filter-Based Ensemble Feature Selection and Deep Learning Model for Intrusion Detection in Cloud Computing

In recent years, the high improvement in communication, Internet of Things (IoT) and cloud computing have begun complex questioning in security. Based on the development, cyberattacks can be increased since the present security techniques do not give optimal solutions. As a result, the authors of this paper created filter-based ensemble feature selection (FEFS) and employed a deep learning model (DLM) for cloud computing intrusion detection. Initially, the intrusion data were collected from the global datasets of KDDCup-99 and NSL-KDD. The data were utilized for validation of the proposed methodology. The collected database was utilized for feature selection to empower the intrusion prediction. The FEFS is a combination of three feature extraction processes: filter, wrapper and embedded algorithms. Based on the above feature extraction process, the essential features were selected for enabling the training process in the DLM. Finally, the classifier received the chosen features. The DLM is a combination of a recurrent neural network (RNN) and Tasmanian devil optimization (TDO). In the RNN, the optimal weighting parameter is selected with the assistance of the TDO. The proposed technique was implemented in MATLAB, and its effectiveness was assessed using performance metrics including sensitivity, F measure, precision, sensitivity, recall and accuracy. The proposed method was compared with the conventional techniques such as an RNN and deep neural network (DNN) and RNN–genetic algorithm (RNN-GA), respectively.

Гібридний накопичувач електроенергії типу «акумуляторна батарея — суперконденсатор» для мікромобільного електротранспорту

The study proves the importance of hybridising the power supply system of micromobile electric vehicles, taking into account the strict limitations on the final size and weight of the storage device, using the example of an electric bicycle. The authors offer a brief overview of the main hybridization topologies and analyze the advantages and disadvantages of each of them. The paper describes the possibilities of using supercapacitors as an additional source of energy for a hybrid system and proposes a way to deal with the disadvantages of using standard batteries by combining them with an additional energy source. A mathematical basis is given to such a solution being able to unload the main power source by drawing excessive current at peak consumption times, thereby providing a wider range of operating voltages compared to using only a standard power source. To ensure optimal weight and size parameters of the drive, an active system with a controlled auxiliary power source is chosen. In this topology, the second converter is installed between the additional energy source, the supercapacitor storage device, and the battery. It is proposed to implement the schematic of this converter using a Chuck converter with an automated control system containing a feedback loop and based on a microcontroller, which ensures a relatively quick change of the control algorithm depending on the configuration and requirements for the system. The paper also presents the algorithm of the storage system operation for possible modes of electric bicycle operation. The transfer characteristic is calculated using the averaged state method. To confirm the operability of the auxiliary power source, a converter circuit is developed and simulation modelling is carried out in MATLAB Simulink. The system simulation allowed assessing the nature of the transient processes of the open-loop system and the accuracy of the transfer characteristic calculation.

Similarity based deep learning model for movie recommendation system

“Movie Recommendation Systems” helps user get relative & relevant items within millions of items. “Movie recommendation system’s” main task is to offer personalized content through information filtering. Here through this paper, we want to develop Similarity Based Deep Learning Model (SDLM) for automatic movie recommendation system. The projected technique is developed to identify the best rated movies and automatic movie recommendation system. This SDLM is a combination of “Spiking Neural Network (SNN)” and “Ebola Optimization Search Algorithm (EOSA)”. In the SNN, the EOSA is utilized to select optimal weighting parameters. The User Profile Correlation-Based Similarity (UPCS) is utilized along with proposed techniques to enable efficient movie recommendation system. To validate the proposed methodology, the movie databases is obtained from the online solutions. The proposed methodology is executed in MATLAB in addition performances can be assessed by “performance measures like recall, precision, accuracy, recall, specificity, sensitivity and F_Measure”. The projected methodology can be compared with the conventional methods such as “ODLM, Recurrent Neural Network (RNN) and Artificial Neural Network (ANN)” respectively.

CDIPEM: Cardiovascular disease incidence probability estimation CNN model

Heart disease affects the majority of the world’s population, with a high rate of morbidity and mortality. The prediction of heart disease at an early stage is viewed as one of the imperative issues in clinical trials. Since there is a huge amount of medical data and it keeps on increasing, it becomes very difficult to analyze and process such data. Hence, machine learning algorithms become an obvious choice to handle such data. This paper presents a classification framework for clinical decision-making that uses an optimized convolutional neural network (CNN). The pre-processed clinical dataset is used for the training and testing of the classifier. A publicly available UCI dataset is used to examine how the system performs. The statistical features are extracted at the beginning and then subsequent data minimization is carried out. Therein-after, the prediction of heart diseases is done with the help of CNN. The identification of optimal weights is done to enhance the performance of CNN which in turn gives accurate disease prediction results. In this paper, we propose a hybrid Particle Swarm Optimization (PSO) and Gray wolf Optimization (GWO) technique namely PS-GW technique to identify the optimal weight parameters. Finally, performance analysis is done and the experimental results are compared with the existing approaches which show improved classification accuracy over conventional optimization techniques and some well-known classifiers such as k-nearest neighbour, decision tree, logistic regression, naive bayes and random forest.

Structural Prediction Analysis of Cross-Sea Cable-Stayed Bridges during Operation Based on Existing Traffic Volumes

In order to study the safety state of the structure of a cross-sea cable-stayed bridge during its operation period, this paper proposes a combined long-term traffic prediction model based on the XGBoost (eXtreme Gradient Boosting) model and LSTM (Long Short Term Memory) model in the context of a cross-sea cable-stayed bridge in Qingdao. XGBoost is an optimized distributed gradient enhancement library. LSTM is a neural network for processing long sequence data. The LSTM model and the XGBoost model were first built separately, and then a genetic algorithm was used to select the optimal weight parameters to combine the two models for prediction. Based on the traffic prediction results of the combined LSTM-XGBoost model, a finite element model was established using numerical analysis. The effect of different traffic volumes on the deflection and stresses in the span of the main beam and the stresses in the diagonal cables was analyzed using the time course analysis method. From the point of view of structural safety, the maximum of future traffic limits and more reasonable average traffic speeds are given to provide a basis for the later management of the bridge.

Transformer Fault Diagnosis Based on an Improved Sine Cosine Algorithm and BP Neural Network

Background: The operation state evaluation and fault location of the transformer is one of the technical bottlenecks restricting the safe power grid operation. Methods: A hybrid intelligent method based on the Improved Sine Cosine Algorithm and BP neural network (ISCA-BP) is developed to improve the accuracy of transformer fault diagnosis. First, the cloud model is introduced into the Sine Cosine Algorithm (SCA) to determine the conversion parameter of each individual to balance the global search and local exploitation capabilities. After that, six popular benchmark functions are used to evaluate the effectiveness of the proposed algorithm. Finally, based on the dissolved gas analysis technology, the improved SCA algorithm is employed to find the optimal weight and threshold parameters of the BP neural network, and the transformer fault classification model is established. Results: Simulation results indicate that the improved SCA algorithm exhibits strong competitiveness. Furthermore, compared with the BP neural network optimized by the Sine Cosine Algorithm (SCA-BP) and BP neural network, the ISCA-BP method can significantly improve the diagnostic accuracy of transformer faults. Conclusion: The proposed intelligent method can provide a valuable reference idea for transformer fault classification.

Improved heat coefficients for joint-space metabolic energy expenditure model during level, uphill, and downhill walking.

A previously developed joint-space metabolic energy expenditure (MEE) model includes subject-specific parameters and was validated using level walking gait data. In this work, we determine how well this joint-space model performs during various walking grades (-8%, 0%, and 8%) at 0.8 m·s ⁻1 and 1.3 m·s ⁻1 using published gait data in the literature. In response to those results, we formulate an optimization problem and solve it through the particle swam method plus fmincon function in MATLAB to identify a new optimal weighting parameter set for each grade that produces more accurate predicted MEE and we compare our new findings with seven other MEE models in the literature. The current study matched the measured MEE the best with the lowest RMSE values for level (0.45 J·kg ⁻1·m ⁻1) and downhill (0.82 J·kg ⁻1·m ⁻1) walking and the third lowest RMSE value for uphill (1.56 J·kg ⁻1·m ⁻1) walking, where another MEE model, Looney et al., had the lowest RMSE for uphill (1.27 J·kg ⁻1·m ⁻1) walking. Bland-Altman plots and three independent-samples t-tests show that there was no statistical significant difference between experimentally measured MEE and estimated MEE during the three walking conditions, meaning that the three new optimal weighting parameter sets can be used with 6 degree of freedom (DOF) lower extremity motion data to better estimate whole body MEE in those scenarios. We believe that this work is a step towards identifying a single robust parameter set that allows for accurate estimation of MEE during any task, with the potential to mitigate a limitation of indirect calorimetry requiring lengthy steady state motion.

A homogenization-planning-tracking method to solve cooperative autonomous motion control for heterogeneous carrier dispatch systems

Taxiing aircraft and towed aircraft with drawbar are two typical dispatch modes on the flight deck of aircraft carriers. In this paper, a novel hierarchical solution strategy, named as the Homogenization-Planning-Tracking (HPT) method, to solve cooperative autonomous motion control for heterogeneous carrier dispatch systems is developed. In the homogenization layer, any towed aircraft system involved in the sortie task is abstracted into a virtual taxiing aircraft. This layer transforms the heterogeneous systems into a homogeneous configuration. Then in the planning layer, a centralized optimal control problem is formulated for the homogeneous system. Compared with conducting the path planning directly with the original heterogeneous system, the homogenization layer contributes to reduce the dimension and nonlinearity of the formulated optimal control problem in the planning layer and consequently improves the robustness and efficiency of the solution process. Finally, in the tracking layer, a receding horizon controller is developed to track the reference trajectory obtained in the planning layer. To improve the tracking performance, multi-objective optimization techniques are implemented offline in advance to determine optimal weight parameters used in the tracking layer. Simulations demonstrate that smooth and collision-free cooperative trajectory can be generated efficiently in the planning phase. And robust trajectory tracking can be realized in the presence of external disturbances in the tracking phase. The developed HPT method provides a promising solution to the autonomous deck dispatch for unmanned carrier aircraft in the future.

Neuroadaptive Robust Speed Control for PMSM Servo Drives with Rotor Failure

In this paper, a neuroadaptive robust trajectory tracking controller is utilized to reduce speed ripples of permanent magnet synchronous machine (PMSM) servo drive under the presence of a fracture or fissure in the rotor and external disturbances. The dynamics equations of PMSM servo drive with the presence of a fracture and unknown frictions are described in detail. Due to inherent nonlinearities in PMSM dynamic model, in addition to internal and external disturbances; a traditional PI controller with fixed parameters cannot correctly regulate the PMSM performance under these scenarios. Hence, a neuroadaptive robust controller (NRC) based on a category of on-line trained artificial neural network is used for this purpose to enhance the robustness and adaptive abilities of traditional PI controller. In this paper, the moth-flame optimization algorithm provides the optimal weight parameters of NRC and three PI controllers (off-line) for a PMSM servo drive. The performance of the NRC is evaluated in the presence of a fracture, unknown frictions, and load disturbances, likewise the result outcomes are contrasted with a traditional optimized PID controller and an optimal linear state feedback method.

State of Health Prediction of Lithium-Ion Batteries Based on the Discharge Voltage and Temperature

Accurate state of health (SOH) prediction of lithium-ion batteries is essential for battery health management. In this paper, a novel method of predicting the SOH of lithium-ion batteries based on the voltage and temperature in the discharging process is proposed to achieve the accurate prediction. Both the equal voltage discharge time and the temperature change during the discharge process are regarded as health indicators (HIs), and then, the Pearson and Spearman relational analysis methods are applied to evaluate the relevance between HIs and SOH. On this basis, we modify the relevance vector machine (RVM) to a multiple kernel relevance vector machine (MKRVM) by combining Gaussian with sigmoid function to improve the accuracy of SOH prediction. The particle swarm optimization (PSO) is used to find the optimal weight and kernel function parameters of MKRVM. The aging data from NASA Ames Prognostics Center of Excellence are used to verify the effectiveness and accuracy of the proposed method in numerical simulations, whose results show that the MKRVM method has higher SOH prediction accuracy of lithium-ion batteries than the relevant methods.

Classification Based on Brain Storm Optimization With Feature Selection

Classification is one of the most classic problems in machine learning. Due to the global optimization ability, evolutionary computation (EC) techniques have been successfully applied to solve many problems and the evolutionary classification model is one of the methods used to solve classification problems. Recently, some evolutionary algorithms (EAs) such as the fireworks algorithm (FWA) and brain storm optimization (BSO) algorithm have been employed to implement the evolutionary classification model and achieved the desired results. This means that it is feasible to use EC techniques to solve the classification problem directly. However, the existing evolutionary classification model still has some disadvantages. The limited datasets used in the experiment make the experimental results not convincing enough, and more importantly, the structure of the evolutionary classification model is closely related to the dimension of datasets, which may lead to poor classification performance, especially on large-scale datasets. Therefore, this paper aims at improving the structure of the evolutionary classification model to improve classification performance. Feature selection is an effective method to deal with large datasets, firstly, we introduce the concept of feature selection and use the different feature subsets to construct the structure of the evolutionary classification model. Then, the BSO algorithm is employed to implement the evolutionary classification model to search the optimal structure by search for the optimal feature subset. Moreover, the optimal weight parameters corresponding to the different structures are also searched by the BSO algorithm while searching the optimal feature subset. For verification of the classification effectiveness of the proposed method, 11 different datasets are selected for experiments. The results show that it is feasible to optimize the structure of the evolutionary classification model by introducing feature selection. Moreover, the new method has better classification performance than the original method, especially on large-scale or high-dimensional datasets.

A new finite difference scheme for the 3D Helmholtz equation with a preconditioned iterative solver

In this paper, we propose a new finite difference scheme for the 3D Helmholtz problem, which is compact and fourth-order in accuracy. Different from a standard compact fourth-order one, the new scheme is specially established based on minimizing the numerical dispersion, by approximating the zeroth-order term of the equation with a weighted-average for the values at 27 points. To determine optimal weight parameters, an optimization problem is formulated and then dealt with the singular value decomposition method based on the dispersion equation. For the proposed scheme, by skillfully splitting the 3D error equation into several 1D difference problems, the solution's uniqueness and convergence are derived with an effort. To solve the resulting linear system stemming from difference discretization, which is sparse and large-sized, we develop a Bi-CGSTAB iterative solver based on the preconditioning of shifted-laplacian and 3D full-coarsening multigrid. The shifted-laplacian is used to generate the preconditioner with a discretization by the proposed compact fourth-order scheme, while the full-coarsening multigrid with matrix-based prolongation operators is built to approximate the inverse of the preconditioner. Finally, numerical examples are presented to demonstrate the efficiency of the new difference scheme and the preconditioned solver.