Multiple Algorithms Research Articles

USING PREDICTIVE ANALYTICS TO IMPROVE SURVEILLANCE OF HEAT-RELATED ILLNESS DURING MILITARY TRAINING

Background: Heat-related illnesses are a critical concern for military personnel, especially those unfamiliar with hot climate regions. The Royal Thai Army (RTA) implements a 10-week military training program encompassing four phases: (1) Heat acclimatization training, (2) Combat fundamentals and unarmed combat training, (3) Armed combat training and tactical training, and (4) Field training exercise and evaluations. Objective: This study aimed to conduct a predictive analysis of heat-related illnesses to enhance prevention programs. Methods: The study utilized secondary data from the RTA Medical Department, incorporating variables such as age, occupation, education, underlying diseases, smoking and alcohol consumption, sleep duration, exercise, medication history, body mass index (BMI), weight loss, body temperature, dark urine, heat rash, and environmental humidity. Multiple machine learning algorithms were employed to develop predictive models. Results: The samples comprised 809 male recruits (103,051 encounters) with an average age of 22. Approximately 12% of the recruits had a BMI ≥30 kg/m2, while nearly 70% and 90% reported tobacco use and alcohol consumption in the past 12 months, respectively. Among the recruits, 16% reported substance use within the preceding 30 days. The eXtreme Gradient Boosting (XGB) model achieved 91% accuracy in predicting heat-related illnesses before Phase 2. The top five predictive variables were Lopburi Province (central region), Songkhla Province (southern region), and Bangkok (capital city), sleep duration before joining military training (hours), and age (years). Conclusion: This study, which applied machine learning techniques to predict heat-related illnesses among Thai recruits, can potentially impact the health and training of military personnel. The comparative analysis of various algorithms identified the XGB model as the optimal performer in predicting heat-related illnesses during the combat fundamentals and unarmed combat training phase. However, it is essential to note that further study is needed to enhance the applicability of our predictive model, which includes expanding its use to new cohorts of Thai conscripts, underscoring our research’s ongoing nature.

RSD-based high-performance radix-4 Montgomery Modular Multiplication for Elliptic Curve Cryptography

This paper proposes a high-performance radix-4 Montgomery Modular Multiplication (MMM) algorithm and its corresponding hardware architecture for Elliptic Curve Cryptography (ECC), in which the quotient and the partial product accumulation are computed in parallel in each iteration. Additionally, in this MMM, the Redundant Signed Digit (RSD) representation and the Signed Digit Adder (SDA) are used to eliminate the long carry chain and achieve parallel computation, as well as remove pre-computation and integrate modular reduction operations. Our MMM algorithm is implemented in 256-bit and 1024-bit versions on Xilinx Virtex-6 and Virtex-7 FPGAs, respectively. It consumes only 1.55k/10.18k Look-Up Tables (LUTs), takes 133/517 clock cycles, and runs at maximum frequencies of 558.8/641.7 MHz. According to the comparison in terms of Area Time Product (ATP), our design can achieve the ATP of 0.369 over the 256-bit NIST prime domain, which is approximately half of that of the state-of-the-art works. The Scalar Point Multiplication (SPM) scheme using this MMM algorithm consumes 14.19k LUTs and completes a single Scalar Point Multiplication (SPM) operation in 0.217 ms, and it also has a lower ATP than most other SPM algorithms currently in existence.

Forecasting financial distress in listed Chinese construction firms: leveraging ensemble learning and non-financial variables

The construction industry, characterized by high business failure rates due to complex, capital-intensive projects, is crucial for China’s economy. Predicting financial distress and early warnings is crucial to prevent crises. While financial ratios are commonly used, non-financial information remains underexplored, with limited empirical evidence supporting its effectiveness in enhancing predictive accuracy. Additionally, most studies on financial distress prediction focus on constructing single classifiers without utilizing ensemble models that integrate multiple algorithms, resulting in suboptimal prediction accuracy. To address these problems, this study presents a model that uses accounting variables and firm characteristics, construction market dynamics, and macroeconomic indicators to predict the probability of failure one to two years in advance. The model uses a soft voting-based ensemble algorithm, financial ratios refined through the recursive feature elimination with cross-validation (RFECV) algorithm, and dataset balancing via Synthetic Minority Over-Sampling Technique + Tomek Link (SMOTETomek). The comparison results indicate that the predictive performance of the soft voting ensemble model outperforms all single classifiers across all combinations of input variables and prediction years. Additionally, incorporating non-financial variables, such as firm characteristics, construction market dynamics, and macroeconomic indicators, further enhances the model’s accuracy. The proposed model can be effectively employed to help stakeholders mitigate the risks associated with the financial distress of construction companies before the project implementation phase.

A flexible and efficient algorithm for high dimensional support vector regression

In high dimensional statistical learning, variable selection and handling highly correlated phenomena are two crucial topics. Elastic-net regularization can automatically perform variable selection and tends to either simultaneously select or remove highly correlated variables. Consequently, it has been widely applied in machine learning. In this paper, we incorporate elastic-net regularization into the support vector regression model, introducing the Elastic-net Support Vector Regression (En-SVR) model. Due to the inclusion of elastic-net regularization, the En-SVR model possesses the capability of variable selection, addressing high dimensional and highly correlated statistical learning problems. However, the optimization problem for the En-SVR model is rather complex, and common methods for solving the En-SVR model are challenging. Nevertheless, we observe that the optimization problem for the En-SVR model can be reformulated as a convex optimization problem where the objective function is separable into multiple blocks and connected by an inequality constraint. Therefore, we employ a novel and efficient Alternating Direction Method of Multipliers (ADMM) algorithm to solve the En-SVR model, and provide a complexity analysis as well as convergence analysis for the algorithm. Furthermore, extensive numerical experiments validate the outstanding performance of the En-SVR model in high dimensional statistical learning and the efficiency of this novel ADMM algorithm.

Analyzing the dynamics between crude oil spot prices and futures prices by maturity terms: Deep learning approaches to futures-based forecasting

Predicting crude oil prices is critical due to their significant impact on the global economy. While numerous studies have focused on time series forecasting and machine learning methods to predict crude oil spot prices, the use of futures prices segmented by the maturity terms as predictors is comparatively underexplored. This study explored a deep learning approach to investigate using futures prices across different maturities to predict spot prices. Specifically, this paper analyzes the predictive power of one-, two-, three-, six-, and 12-month crude oil futures contracts. This study employs multiple deep learning algorithms, including Multilayer Perceptron (MLP), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Convolutional Neural Network (CNN), and Temporal Convolutional Neural Network (TCN), to forecast crude oil spot prices. This research investigates the performance of these machine learning models when exploring the relationship between futures and spot prices of crude oil. In addition, this research incorporates extensive hyperparameter tuning to enhance the interpretability of the machine learning models when forecasting spot prices using futures prices, thereby contributing to the field of Explainable Artificial Intelligence (XAI) with an optimal set of hyperparameters. In summary, this research systematically shows results demonstrating predictive power in terms of XAI between spot prices and futures prices with different maturities, and machine learning algorithms.

Transient pressure prediction in large-scale underground natural gas storage: A deep learning approach and case study

Large-scale underground natural gas storage (UNGS) facilities typically allocate only one month annually for well shut-in to facilitate pressure build-up. This limited well shut-in period affects the subsequent pressure measurements and evaluation of gas storage capacity. This study proposes a novel workflow to directly predict the transient pressure behavior of UNGS during pressure build-up without requiring additional well shut-in operations. The proposed workflow uses the gas injection-withdrawal rate as a dynamic input feature for a deep learning model, with reservoir pressure as the output feature. An enhanced WA-BiLSTM model integrates multiple physical mechanisms and advanced optimization algorithms. The model achieves mean squared error (MSE) of 2 × 10-3, which is less than 5 % of traditional model's results. Field data from the largest Hutubi UNGS exhibit a prediction accuracy of 99.0 % during the well shut-in phase. High-precision prediction results with low MSE ensure the reliability of pressure derivative data. By analyzing the predicted pressure data, the actual gas storage volume is 104.75 × 108 m3, accounting for 97.9 % of the facility's designated storage capacity.

Determinants of Visual Impairment Among Chinese Middle-Aged and Older Adults: Risk Prediction Model Using Machine Learning Algorithms.

Visual impairment (VI) is a prevalent global health issue, affecting over 2.2 billion people worldwide, with nearly half of the Chinese population aged 60 years and older being affected. Early detection of high-risk VI is essential for preventing irreversible vision loss among Chinese middle-aged and older adults. While machine learning (ML) algorithms exhibit significant predictive advantages, their application in predicting VI risk among the general middle-aged and older adult population in China remains limited. This study aimed to predict VI and identify its determinants using ML algorithms. We used 19,047 participants from 4 waves of the China Health and Retirement Longitudinal Study (CHARLS) that were conducted between 2011 and 2018. To envisage the prevalence of VI, we generated a geographical distribution map. Additionally, we constructed a model using indicators of a self-reported questionnaire, a physical examination, and blood biomarkers as predictors. Multiple ML algorithms, including gradient boosting machine, distributed random forest, the generalized linear model, deep learning, and stacked ensemble, were used for prediction. We plotted receiver operating characteristic and calibration curves to assess the predictive performance. Variable importance analysis was used to identify key predictors. Among all participants, 33.9% (6449/19,047) had VI. Qinghai, Chongqing, Anhui, and Sichuan showed the highest VI rates, while Beijing and Xinjiang had the lowest. The generalized linear model, gradient boosting machine, and stacked ensemble achieved acceptable area under curve values of 0.706, 0.710, and 0.715, respectively, with the stacked ensemble performing best. Key predictors included hearing impairment, self-expectation of health status, pain, age, hand grip strength, depression, night sleep duration, high-density lipoprotein cholesterol, and arthritis or rheumatism. Nearly one-third of middle-aged and older adults in China had VI. The prevalence of VI shows regional variations, but there are no distinct east-west or north-south distribution differences. ML algorithms demonstrate accurate predictive capabilities for VI. The combination of prediction models and variable importance analysis provides valuable insights for the early identification and intervention of VI among Chinese middle-aged and older adults.

Key platelet genes play important roles in predicting the prognosis of sepsis

Sepsis is a life-threatening organ malfunction induced by an imbalanced immunological reaction to infection in the host. Many studies have utilized traditional RNA sequencing (RNA-seq) data to identify important biological targets to predict sepsis prognosis. However, alterations in core cells and functional status cannot be effectively detected in sepsis patients. The goal of this study was to identify key cells through single-cell RNA-seq (scRNA-seq), and combine bulk RNA-seq data and multiple algorithm analysis to construct a stable prognostic model for sepsis. The scRNA-seq and bulk RNA-seq data from sepsis patients were collected from the Gene Expression Omnibus (GEO) database. The R package “Seurat” was used to process the scRNA-seq data. Cell communication was investigated using the R package “CellChat”. The pseudo-time of the cells was calculated using the R package “monocle”. The R package “limma” was used to identify differentially expressed genes (DEGs) between the sepsis group and the control group. Weighted gene correlation network analysis (WGCNA) was used to identify critical modules. Eight kinds of machine learning and 90 algorithm combinations were used to construct the prognostic model for sepsis. Quantitative real-time PCR (qRT‒PCR) was performed to determine the expression of key genes in the cecal ligation and puncture (CLP)-induced sepsis mouse model. The immunological status and related properties of DEGs were then investigated in the high- and low-risk groups delineated by the model. By combining the scRNA-seq data from nine samples, 13 clusters and 9 cell types were identified. CellChat analysis revealed that the number and strength of interactions between platelets and a variety of cells increased. We identified key platelet genes from the scRNA-seq data and combined these genes and the results of differential analysis and WGCNA of the bulk RNA-seq data. After univariate Cox regression analysis, we calculated the Cindex of the model constructed by the combination of 90 algorithms, and we finally determined the “CoxBoost + Lasso” combination. Multivariate Cox regression was used to construct the final prognostic model. The qRT-PCR results revealed significant differences in five key prognostic genes between the CLP and sham groups. The data was classified into high- and low-risk groups based on the model score. The high-risk group had a poorer survival rate and less immune infiltration. We identified the importance of platelets in sepsis patients through scRNA-seq, and established prognostic models with key genes that were identified via scRNA-seq combined with bulk RNA-seq analysis. The results of this model were closely associated with patient survival rates and immunological status and this model is useful for the prognostic management of sepsis.

A Dual-Focus Workflow for Simultaneously Engineering High Activity and Thermal Stability in Methyl Parathion Hydrolase.

Industrial fermentation applications typically require enzymes that exhibit high stability and activity at high temperatures. However, efforts to simultaneously improve these properties are usually limited by a trade-off between stability and activity. This report describes a computational strategy to enhance both activity and thermal stability of the mesophilic organophosphate-degrading enzyme, methyl parathion hydrolase (MPH). To predict hotspot mutation sites, we assembled a library of features associated with the target properties for each residue and then prioritized candidate sites by hierarchical clustering. Subsequent in silico screening with multiple algorithms to simulate selective pressures yielded a subset of 23 candidate mutations. Iterative parallel screening of mutations that improved thermal stability and activity yielded, MPHase-m5b, which exhibited 13.3 °C higher Tm and 4.2 times higher catalytic activity than wild-type (WT) MPH over a wide temperature range. Systematic analysis of crystal structures, molecular dynamics (MD) simulations, and quantum mechanics/molecular mechanics (QM/MM) calculations revealed a wider entrance to the active site that increased substrate access with an extensive network of interactions outside the active site that reinforced αβ/βα sandwich architecture to improve thermal stability. This study thus provides an advanced, rational design framework to improve efficiency in engineering highly active, thermostable biocatalysts for industrial applications.

A Dual‐Focus Workflow for Simultaneously Engineering High Activity and Thermal Stability in Methyl Parathion Hydrolase

AbstractIndustrial fermentation applications typically require enzymes that exhibit high stability and activity at high temperatures. However, efforts to simultaneously improve these properties are usually limited by a trade‐off between stability and activity. This report describes a computational strategy to enhance both activity and thermal stability of the mesophilic organophosphate‐degrading enzyme, methyl parathion hydrolase (MPH). To predict hotspot mutation sites, we assembled a library of features associated with the target properties for each residue and then prioritized candidate sites by hierarchical clustering. Subsequent in silico screening with multiple algorithms to simulate selective pressures yielded a subset of 23 candidate mutations. Iterative parallel screening of mutations that improved thermal stability and activity yielded, MPHase‐m5b, which exhibited 13.3 °C higher Tm and 4.2 times higher catalytic activity than wild‐type (WT) MPH over a wide temperature range. Systematic analysis of crystal structures, molecular dynamics (MD) simulations, and quantum mechanics/molecular mechanics (QM/MM) calculations revealed a wider entrance to the active site that increased substrate access with an extensive network of interactions outside the active site that reinforced αβ/βα sandwich architecture to improve thermal stability. This study thus provides an advanced, rational design framework to improve efficiency in engineering highly active, thermostable biocatalysts for industrial applications.

Area-Time-Efficient Secure Comb Scalar Multiplication Architecture Based on Recoding

With the development of mobile communication, digital signatures with low latency, low area, and high security are in increasing demand. Elliptic curve cryptography (ECC) is widely used because of its security and lightweight. Elliptic curve scalar multiplication (ECSM) is the basic arithmetic in ECC. Based on this background information, we propose our own research objectives. In this paper, a low-latency and low-area ECSM architecture based on the comb algorithm is proposed. The detailed methodology is as follows. The recoding-k algorithm and randomization-Z algorithm are used to improve security, which can resist sample power analysis (SPA) and differential power analysis (DPA). A low-area multi-functional architecture for comb is proposed, which takes into account different stages of the comb algorithm. Based on this, the data dependency is considered and the comb architecture is optimized to achieve a uniform and efficient execution pattern. The interleaved modular multiplication algorithm and modified binary inverse algorithm are used to achieve short clock cycle delay and high frequency while taking into account the need for a low area. The proposed architecture has been implemented on Xilinx Virtex-7 series FPGA to perform ECSM on 256-bits prime field GF(p). In the hardware architecture with only 7351 slices of resource usage, a single ECSM only takes 0.74 ms, resulting in an area-time product (ATP) of 5.41. The implementation results show that our design can compete with the existing state-of-the-art engineering in terms of performance and has higher security. Our design is suitable for computing scenarios where security and computing speed are required. The implementation of the overall architecture is of great significance and inspiration to the research community.

Immune-related glycosylation genes based classification predicts prognosis and therapy options of osteosarcoma

Osteosarcoma is the most common primary bone malignancy, with a very poor prognosis. Aberrant glycosylation is close involvement in osteosarcoma. Accordingly, this study aimed at investigating the role of glycosylation genes in the prognosis and therapy options of osteosarcoma. The microenvironment of osteosarcoma was assessed using estimate algorithm. A total of 20 immune-related glycosylation genes (IRGGs) was identified using Pearson correlation analysis. Accordingly, osteosarcoma patients were divided into C1 and C2 type using consensus clustering. Multiple algorithms (Xcell, MCP-counter, ssGSEA, epic, quantiseq), cancer immune cycle analysis, and GSVA were applied to estimate the immune, molecule and metabolism characteristics of osteosarcoma, indicating that C1 type was featured with high immune infiltration, high glycosylation, enriched MEK signaling, and good prognosis, while C2 type was characterized by more metastasis, enriched immunotherapy-positive gene signatures, high tumor mutation burden, and poor prognosis. Results from TIDE algorithm and immunotherapy datasets suggested the C2 type’s preference of immune checkpoint inhibitors (ICIs), while data of GDSC, CMap analysis and cell experiments indicated that C1 type was sensitivity to MEK inhibitor PD0325901. In addition, univariate Cox and Lasso analysis was combined to establish an IRGGs’ risk score containing 6 genes (B3GNT8, FUT7, GAL3ST4, GALNT14, HS3ST2, and MFNG). The data of DCA and ROC indicated its well prediction of prognosis in osteosarcoma. Finally, cellular location analysis showed that the 6 genes not only distributed in tumor cells but also in immune cells. In summary, the classification and risk score based on IRGGs effectively predicted the prognosis and therapy options of osteosarcoma. Further studies on IRGGs may contribute to the understanding of cancer immunity in osteosarcoma.

Lightweight algorithm based on you only look once version 5 for multiple class defect detection on wind turbine blade surfaces

A lightweight wind turbine blade surface multiple class defect detection algorithm based on You Only Look Once version 5 (YOLOv5) is proposed to address the issues of low accuracy, poor efficiency, and high power consumption associated with manual and traditional automation algorithms used for detecting surface defects on wind turbine blades. Firstly, by replacing the Concentrated-Comprehensive Convolution Block module in YOLOv5 with the GhostBottleneckv2 module, the overall parameter count of the model is significantly reduced. Additionally, advocating the inclusion of the Convolutional Block Attention Module (CBAM) into the YOLOv5 backbone network to address the issue of small targets and significant scale variations in surface defects found on wind turbine blades. The YOLOv5 feature fusion module replaces the Feature Pyramid Network (FPN) structure with the Bidirectional Feature Pyramid Network (BiFPN) structure to enhance the fusion capability of multi-scale weighted features. Finally, the optimized model files underwent structured pruning through the application of pruning strategies, retraining fine-tuning, and knowledge distillation techniques. The experimental results on the self-made dataset of surface defects on wind turbine blades indicate that the optimized model has a parameter count that is only 48.79% of the small version of the YOLOv5 model. However, its mean Average Precision (mAP) and frame rate have increased by 5.05% and 8.1 frames/s, respectively, reaching 94.56% and 43.7 frames/s. The lightweight algorithm proposed in this study meets the accuracy and real-time requirements for surface defect detection on wind turbine blades.

A Prospective Study on Risk Prediction of Preeclampsia Using Bi-Platform Calibration and Machine Learning.

Preeclampsia is a pregnancy syndrome characterized by complex symptoms which cause maternal and fetal problems and deaths. The aim of this study is to achieve preeclampsia risk prediction and early risk prediction in Xinjiang, China, based on the placental growth factor measured using the SiMoA or Elecsys platform. A novel reliable calibration modeling method and missing data imputing method are proposed, in which different strategies are used to adapt to small samples, training data, test data, independent features, and dependent feature pairs. Multiple machine learning algorithms were applied to train models using various datasets, such as single-platform versus bi-platform data, early pregnancy versus early plus non-early pregnancy data, and real versus real plus augmented data. It was found that a combination of two types of mono-platform data could improve risk prediction performance, and non-early pregnancy data could enhance early risk prediction performance when limited early pregnancy data were available. Additionally, the inclusion of augmented data resulted in achieving a high but unstable performance. The models in this study significantly reduced the incidence of preeclampsia in the region from 7.2% to 2.0%, and the mortality rate was reduced to 0%.

Acoustic process monitoring during the laser beam welding of stainless-steel foils using an adjustable ring mode laser beam source

Electrification of the mobility sector is vital to meet the targets for reducing greenhouse gas emissions. Besides battery-based mobility solutions, polymer electrolyte membrane fuel cells (PEMFCs) are a promising technology for electrifying drive trains, especially in heavy-duty applications, such as maritime or logistics. Bipolar plates, a key component of PEMFCs, can consist of two stainless-steel foils that must be welded to be gas-tight. In order to join the two metal foils, laser beam welding is the state-of-the-art technology. Current challenges include process instabilities at higher welding speeds, such as the humping effect, which can cause weld seam imperfections. Therefore, applying sensors for laser beam welding is a promising approach to monitor the welding process. AISI 316L foils were welded within the scope of this work with various process parameters using an adjustable ring mode laser beam source. Additionally, an optical microphone was used as a process monitoring system. By applying different parameter settings and due to the introduction of artificial faults, weld seam defects, such as a burn-through or a gap, were induced. After utilizing a noise reduction algorithm for the acoustic signals, numerous features in the time and frequency domains were extracted, with which multiple machine learning algorithms were trained and compared concerning their performance. A light gradient boosting machine was identified as a suitable machine learning model for weld seam classification. Finally, hyperparameter tuning was conducted, which resulted in a cross-validation accuracy of 94.78%, depending on the quality categories considered.

Proposal of a facet-to-facet multiple flow direction algorithm based on geometrical and mathematical analysis of physical dispersion over triangle facet

Flow direction algorithms have been widely used to extract crucial terrain attributes of great hydrological and geomorphological significance. However, essential distinctions between the empirically-designed strategies of typical algorithms and the natural rules of physical dispersions bring various problems (e.g. parallel channel, artificial dispersion), leading to the low size and extent precisions of estimated results. In this work, geometrical and mathematical analysis is conducted on the inherent characteristics of physical dispersions along slope lines on local terrains. On each 3 × 3 window of digital elevation model (DEM), center pixel is divided into eight non-overlapping sub-facets. Necessary and sufficient (NS) conditions of size relationships between the elevations of adjacent pixels are summarized to directly identify the receiving facets of a sub-facet. Then, strict mathematical relations are derived between slope direction of a sub-facet and flow proportions allocated to receiving facets. A strategy is designed to re-adjust receiving facets and flow proportions for the boundary flow of adjacent facets. Lastly, a multiple-flow-direction algorithm called TFGA is proposed with the NS condition of size relationships, mathematical relation of slope direction with flow proportion, and adjustment strategy of boundary flow. Case studies are conducted for investigating the total contributing areas (TCA) and specific contributing areas (SCA) estimated by TFGA. Results reveal all-side superiorities of TFGA to typical algorithms in spatial patterns, error indicators and statistic characteristics of estimated TCAs and SCAs. Particularly, TFGA improves size precision of estimated results by approximately one order of magnitude. In a conclusion, we highly recommend TFGA for digital elevation analysis.

Growth-adaptive distillation compressed fusion model for network traffic identification based on IoT cloud-edge collaboration

The development of the Internet of Things (IoT) has led to the rapid growth of the types and number of connected devices and has generated large amounts of complex and diverse traffic data. Traffic identification on edge servers solves the real-time and privacy requirements of IoT management and has attracted much attention, but still faces several problems: (1) traditional machine learning (ML) models rely on artificially constructed features, and the existing deep learning (DL) traffic identification models have reached their performance limit; and (2) insufficient computing resources of edge servers limit the possible improvement in the performance of deep learning models by increasing the number of parameters and structural complexity. To address these issues, we propose a lightweight fusion model. First, the Network-in-Network (NiN) model and Random Forest (RF) model are used on the cloud server to construct a traffic identification fusion model. The excellent representation extraction capability of the NiN compensates for the RF’s dependence on manual feature extraction, and its modular structure is suitable for the subsequent model compression operations. Then, the NiN was distilled. We propose Growth-Adaptive Distillation to lightweight the NiN model, which can reduce the operation of manually adjusting the structure of the student model and ensure the efficiency and low power consumption of the fusion model deployment. In addition, both the RF in the cloud and the distilled NiN are deployed on the edge server. Comparisons with multiple algorithms on two network traffic datasets show that the proposed model achieves state-of-the-art performance while ensuring the use of minimal computational resources.

Plasma cell-free DNA methylome-based liquid biopsy for accurate gastric cancer detection.

Early detection plays a critical role in mitigating mortality rates linked to gastric cancer. However, current clinical screening methods exhibit suboptimal efficacy. Methylation alterations identified from cell-free DNA (cfDNA) present a promising biomarker for early cancer detection. Our study focused on identifying gastric cancer-specific markers from cfDNA methylation to facilitate early detection. We enrolled 150 gastric cancer patients and 100 healthy controls in this study, and undertook genome-wide methylation profiling of cfDNA using cell-free methylated DNA immunoprecipitation and high-throughput sequencing. We identified 21 differentially methylated regions (DMRs) between the gastric tumor and nontumor groups using multiple algorithms. Subsequently, using the 21 DMRs, we developed a gastric cancer detection model by random forest algorithm in the discovery set, and validated the model in an independent set. The model was able to accurately discriminate gastric cancer with a sensitivity and specificity of 93.90% and 95.15% in the discovery set, respectively, and 88.38% and 94.23% in the validation set, respectively. These results underscore the efficacy and accuracy of cfDNA-derived methylation markers in distinguishing early stage gastric cancer. This study highlighted the significance of cfDNA methylation alterations in early gastric cancer detection.

Improvement of biodegradability prediction QSAR system through optimal combination of algorithms

Abstract This paper introduces a novel quantitative structure-activity relationship (QSAR) system for predicting the biodegradability of chemicals. The system focuses on evaluating the ratio of biochemical oxygen demand (BOD) to theoretical oxygen demand (ThOD) (%BOD/ThOD) and the presence of degradation products to categorize chemicals as “readily biodegradable” (RB) or “not readily biodegradable” (NB). The conventional laboratory testing method for biodegradability assessment is time-consuming and expensive, thus highlighting the need for a quantitative structure-activity relationship (QSAR) system that can predict biodegradability without the need for extensive testing. The authors acquired a dataset of 3948 chemicals from the National Institute of Technology and Evaluation (NITE) database, which included information on molecular structure, %BOD/ThOD values, and the presence of degradation products. The QSAR system consists of two prediction models: a %BOD/ThOD prediction model and a discrimination prediction model. Multiple algorithms were employed in each model to enhance prediction accuracy. The %BOD/ThOD prediction model utilizes eight algorithms, including linear, nonlinear, and tree-structured models. The discrimination prediction model employs ten algorithms, and the outcomes are combined using majority voting. Explanatory variables for both models include molecular descriptors and spatial structure derived from the chemical’s molecular representation. The performance of the QSAR system was evaluated based on training and validation datasets. Results showed that the combination of three %BOD/ThOD prediction algorithms improved the average correction rate compared to using all algorithms. Similarly, selecting the discrimination prediction algorithm further enhanced the average correction rate.

Maximizing chitin and chitosan recovery yields from Fusarium verticillioides using a many-factors-at-a-time approach

The extraction of chitin and chitosan presents challenges due to the complexity of the process and the influence of many variables. This study aimed to optimize chitin and chitosan extraction from Fusarium verticillioides by analyzing many additives and processing variables and modeling their yields using multiple linear regression (MLR) and evolutionary algorithms. FT-IR analysis confirmed the presence of characteristic bands in the extracted samples, and SEM analysis further revealed the microfibrillar appearance of the chitin and the dense, non-porous structure of the chitosan. The Ant Lion Optimizer (ALO) was employed to select significant factors and optimize model parameters. A transformation was applied to capture nonlinear relationships, and the fine-tuned models showed improved predictive power, with p-values of 0.00203 for chitin and 0.00884 for chitosan. Multi-objective optimization (MOO) using the Adaptive Geometry Estimation-based Multi-Objective Evolutionary Algorithm (AGE-MOEA) further identified significant factors for optimal yields, achieving 3 g of Arginine, 100 ml of culture medium volume, 7 to 11 days of incubation time, 0.2 to 1.76 ml of Oligochitin, 1.4 g of FeSO4, 1.5 g of K2HPO4, and 1 g of NaCl. Therefore, the integration of ALO and AGE-MOEA algorithms effectively modeled and optimized chitin and chitosan yields by maximizing biopolymer recovery, enabling significant industrial exploitation.