Correct Prediction Research Articles

Applying Resin Radial Injection for Manufacturing Fiber-Reinforced Polymer Composite: Advanced Mathematical Modeling and Simulation

Recently, the liquid composite molding technique (LCM) has been used for producing fiber-reinforced polymer composites, since it allows the molding of complex parts, presenting good surface finishing and control of the mechanical properties of the product at the end of the process. Studies in this area have been focused on resin transfer molding (RTM), specifically on the resin rectilinear infiltration through the porous preform inserted in the closed cavity neglecting the sorption effect of the polymeric fluid by the reinforcement. Thus, the objective of this work is to predict resin radial flow in porous media (fibrous preform), including the effect of resin sorption by fibers considering a one-dimensional approach. For correct prediction of the flow behavior inside the porous media, an advanced modeling approach composed of the mass conservation equation and Darcy’s law is used, and the solution of the coupled equation is obtained. Transient results of the flow front location, velocity and pressure within the mold during the resin infiltration are shown, the effects of different parameters for resin (viscosity), reinforcement (sorption term, permeability and porosity) and process (injection pressure and injection radius) are analyzed, and an in-depth discussion is performed.

Effects of Turbulence Modeling on the Simulation of Wind Flow over Typical Complex Terrains

The correct prediction of the wind speed and turbulence levels over complex terrain is essential for accurately assessing wind turbine wake recovery, power production, safety, and wind farm design. In this paper, two modified RANS turbulence models are proposed, which are innovative variants of the conventional SST k-ω model and the linear Reynolds stress model (RSM) featuring optimized closure constants. Then, these two modified models and their origin models are applied to compare and analyze wind flows from a 3D hill wind tunnel experiment and two field measurements over typical complex terrain, including Askervein hill and Bolund island, with the aim of analyzing the sensitivity of wind flows to different RANS turbulence models. The study focuses on analyzing the effects of different turbulence models on the self-sustainability of wind speed and turbulent kinetic energy upstream of the computational domain and on the accuracy of wind flow prediction over complex terrain. The results show that our modified RSM model shows better agreement with the available experimental data on the upstream and leeward sides of all simulated hills. The wind speed on the leeward slope is particularly sensitive to the turbulence model, with a maximum difference in the relative root mean square error (RRMSE) that can reach 11% among the four models. The accuracy of the turbulent kinetic energy depends on the self-sustainability of the upstream turbulent kinetic energy and the predictive ability of the turbulence model for separated flows, and the maximum difference in the RRMSE of the four models can reach 47%. In addition, the advantages and disadvantages of the tested models are discussed to provide guidance for model selection during wind flow simulations in complex terrain.

Supervised Automated Kinetic Perimetry (SAKP) Using Simulated Visual Field Data - Presentation of a New Examination Technique.

The aim of this study was to develop, optimise, train, and evaluate an algorithm for performing Supervised Automated Kinetic Perimetry (SAKP) using digitalised perimetric simulation data. The original SAKP algorithm was based on findings from a multicentre study to establish reference values by semi-automated kinetic perimetry (SKP) combined with an automated examination method with moving stimuli ("Program K", developed in Japan). The algorithm evaluated the outer angles of isopter segments and responded to deviations from expected values by placing examination vectors to measure the outer boundaries of the visual field (VF). Specialised interpolation methods were also used to create individual 3D hills of vision and local "probing vectors" to optimise the eccentricity of the vector origins. This algorithm was trained iteratively on seven representative digitalised 3D VF results from five typical classes and optimised in each step: (1) Normal VF, (2) Central scotoma, (3) Concentric VF constriction, (4) Retinal nerve fibre layer defects in the visual field (VFDs), (5) VFDs with respect to the vertical meridian. The optimised SAKP algorithm was then applied to a new set of twenty 3D VF results of varying origin and severity. The primary targets were measured in agreement between actual calculated VF expressed as accuracy (A), that is, the ratio between the area containing correct predictions and total area of predictions measured between 0 = worst and 1 = best, and examination duration (T). The results are given as median (and interquartile range). We also verified the test's robustness by varying individual error rates (ERs) and error magnitudes (EMs). The median and interquartile range (IQR, in brackets) for the total of representative VFs were 0.93 (0.02) for A and 7.0 min (5.2 min) for T, respectively. A gave the best result for altitudinal VFDs and VFDs with hemianopic character and macular sparing (0.98 each) and worst in superior wedge-shaped VFDs (0.78); T was shortest in blind spot displacement (3.9 min) and longest in hemianopic VFDs with hemianopic character and macular sparing with preserved temporal crescent (12.1 min). Error rate and magnitude (up to 30% each) only showed a comparatively low influence on A and T. The SAKP algorithm presented here achieves a comparatively high degree of accuracy and robustness for actual, simulated visual field data within acceptable examination times. This algorithm is currently being prepared for application in real patient examinations under clinical conditions.

Statistical analysis of absolute photopeak efficiency of clover HPGe detectors

The variation of absolute photopeak efficiency of an array of clover HPGe detectors as a function of γ-ray energy has been obtained using statistical model analysis. Machine learning technique has been implemented for determining the energy-dependent polynomial of photopeak efficiency. The distribution of parameters of the best-fit model has been deduced by incorporating Markov chain Monte Carlo method working with the basic framework of Bayesian analysis which also provides the correlation between the parameters. This method has the advantage of calculating the posterior probability distribution by maximizing the likelihood function of the experimental data over standard least-squares method. Excellent agreement among the results procured from machine learning, Markov chain Monte Carlo, and least-squares method has been achieved and the correct prediction of the nature of the variation extrapolation region as well as interpolation region has been discussed.

Analysis of Public Sentiment Regarding the Issue of Cancelling the Revision of the 2024 Regional Election Law with NLP

The latest changes regarding the age requirements for regional head candidates in the 2024 Election by the Supreme Court have given rise to various responses from the public. This decision has become an important issue that is widely discussed on social media and mass media, where this decision has caused various reactions, both positive, negative and neutral. Sentiment analysis is important to determine public opinion on this decision. This study aims to analyze public sentiment on the Issue of Cancellation of the Revision of the 2024 Pilkada Law. This study uses the Natural Language Processing (NLP) method with the Naive Bayes algorithm. The data used is text data in the form of public opinion on the Issue of Cancellation of the Revision of the 2024 Pilkada Law collected via Twitter. This data is then processed using NLP techniques such as cleaning, tokenization, normalization, filtering and stemming. After that, public opinion on the Issue of Cancellation of the Revision of the 2024 Pilkada Law is classified into positive, negative and neutral sentiment categories. The results of the study showed that the results of the model performance evaluation using data testing produced an accuracy of 92%, so it can be concluded that this model is good enough for sentiment analysis. Accuracy shows that the model makes correct predictions on 92% of the total data, which is dominated by the neutral class.

Loss calculation and temperature prediction method for High-Speed amorphous alloy PMSM considering the coupling effect of Multi-Physical field factors

The correct prediction of the loss of amorphous alloy stator cores (AASC) is challenging due to factors such as temperature fluctuations, interference fit stress, and skin effect in high frequency environments. In the complicated operating conditions of high-speed amorphous alloy permanent magnet synchronous motor (HSAAPMSM), an accurate prediction of AASC loss is accomplished by introducing a refined modeling approach of iron loss that considers the interaction of several physical field components. This approach combines with the 3D lumped-parameter thermal network model to generate bidirectional iterative coupling between the temperature distribution of the entire machine and different forms of losses in HSAAPMSM. A steady-state temperature rise result from Maxwell-Fluent bidirectional magnetothermal coupling simulation and test validates the effectiveness and computational superiority of the temperature model that considers the coupling effect of multi-physical field variables.

Multicenter evaluation of Fourier transform infrared (FTIR) spectroscopy as a first-line typing tool for carbapenemase-producing Klebsiella pneumoniae in clinical settings.

Early use of infection control methods is critical for preventing the spread of antimicrobial resistance. Whole-genome sequencing (WGS) is considered the gold standard for investigating outbreaks; however, the turnaround time is usually too long for clinical decision-making and the method is also costly. The aim of this study was to evaluate the performance of Fourier transform infrared (FTIR) and artificial intelligence tools as a first-line typing tool for typing carbapenemase-producing Klebsiella pneumoniae (CPK) in the hospital setting. For this purpose, we analyzed 365 CPK isolates from two tertiary hospitals in Spain in parallel by applying unsupervised principal component analysis (PCA) and supervised algorithms (artificial neural network [ANN], support vector machine [SVM] linear, SVM radial basis function in the IR Biotyper software, and random forest in the Clover MSDAS software). Concordance with FTIR clustering considering the sequence type (ST) and the clonal cluster, obtained by cgMLST for reference purposes, was measured using the adjusted Wallace index (AWI), yielding values of 0.611 and 0.652, respectively. Different regions of the spectra were studied in relation to repeatability and reproducibility, and the polysaccharides region proved the best for FTIR differential analysis. The best results for accuracy were obtained using the ANN algorithm in the IR Biotyper software, with 80.5% of correct prediction. Regarding accuracy, the poorest results were obtained for isolates belonging to ST392 (55.5%) and the best results for ST307 (94.4%). The findings demonstrate the utility of the FTIR method as a rapid, inexpensive, first-line typing tool for detecting CPK, preserving WGS for confirmation and further characterization.

Naturally constrained reject option classification

Existing Reject Option Classification formulations typically learn a function (e.g., softmax threshold) to select/reject uncertain classifier predictions. Such approaches leverage a user-defined cost of rejection or constraints on the accuracy or coverage of the selected predictions. We formulate a new objective for applications that have no such costs/constraints by using a natural constraint on the rejected predictions. Our proposed Reject Option Classification formulation eliminates regions of random chance classification in the decision space of any neural classifier and dataset. The goal is to maximize accuracy in the selected region while permitting a reasonable degree of prediction randomness in the rejected region. Optimally, the hope would be to reject more incorrect than correct predictions. We employ a novel selection/rejection function and learn per-class softmax thresholds using a validation set. Results demonstrate the advantages of our proposed method compared to naïvely thresholding calibrated/uncalibrated softmax scores. We evaluate 2-D points, imagery, and text classification datasets using state-of-the-art pretrained and learned models. Source code is available at https://github.com/osu-cvl/learning-idk.

Spatial-temporal characteristics and the importance of environmental factors in relation to algal blooms in coastal seas

In the prediction of algal blooms with the artificial intelligence (AI) technology, the proper selection of the indices is crucial to the quality of the results. Although eutrophication and climate change have been identified as the main contributors to algal blooms, limited knowledge of the relationships among environmental factors hinders the accurate and correct AI prediction. To explore the importance of environmental factors in the occurrence of algal blooms in coastal seas and facilitate the selection of inputs for the prediction model, this study examines observed water quality data and analyzes their relationship using machine learning methods. It was found that nutrients, water temperature, and salinity play important roles in controlling algal blooms in coastal seas, and their influence varies depending on locations. The high-risk cases of algal bloom mainly occur in environments with medium salinity (22–28 psu), particularly in open estuaries with high nutrient levels. According to the feature importance analysis, water temperature presented the most significant impact on algal bloom risk compared to other environmental factors. The variations of water temperature and Chl-a concentration are synchronous when the water temperature is below 32 °C. Additionally, the relationships between inorganic nutrients (NO3−-NO2− and Ortho-P) and Chl-a exhibit asynchronous across multi-temporal scales. The uptake of inorganic nutrients occurs in large quantities within the 24 h prior to the boom in Chl-a concentration. This demonstrates that variations in concentration of inorganic nutrients can be considered as crucial factors in algal bloom predictions, and their significant decrease can be served as an important signal of an impending algal bloom. This study enhances our understanding of the mechanisms behind algal blooms and facilitates their management and prediction strategies in coastal seas.

Effect of Northeast Pacific wind on the improvement of El Niño prediction in a climate model

Abstract The role of the Northeast Pacific wind on the dynamics of El Niño has been proposed in previous theoretical studies. However, the actual effect of Northeast Pacific wind on improving El Niño prediction remains unclear, and the mechanisms through which Northeast Pacific improves El Niño predictions are not fully understood. In this study, we investigate the role of Northeast Pacific wind in improving El Niño prediction. By utilizing CESM that assimilates the Northeast Pacific wind, the model improves prediction skills of ENSO indices at a lead time of 6-12 months, with a three-time correlation higher than predictions without nudged Northeast Pacific wind. Nudging Northeast Pacific wind in the model enhances prediction of Central-Pacific (CP) El Niño, thus leading to an approximate 30% increase in the hit ratio of correct prediction El Niño. However, the model overestimates CP El Niño predictions when the nudged Northeast Pacific wind is incorporated. The mechanisms for the improvement of Northeast Pacific wind on El Niño predictions are investigated. In spring, El Niño development is attributed to the Wind-Evaporation-SST (WES) feedback and oceanic downwelling induced by subtropical westerly wind stress. In contrast, during summer, El Niño development is dominated by the Summer Deep Convection (SDC) response. During subsequent summer-autumn, anomalous surface latent heat flux related to wind speed, and net shortwave radiation associated with the low cloud-SST feedback contribute to El Niño development. These processes are more favorable for CP El Niño development. This study implies that accurately simulating air-sea coupling associated with Northeast Pacific wind may be an effective approach to improving El Niño prediction.

Which is More Accurate? Urologist or STONE Nomogram: An Original Study

Background: Comparing and determining the accuracy of -urolo gists and STONE nomogram prediction about the Percutaneous Nephrolithotomy (PCNL) outcomes in urolithiasis patients. Methods: This retrospective cohort study was performed on 70 patients who underwent PCNL in Firoozgar Hospital (Tehran, Iran) from March 2019 to April 2020. Two expert urologists, who were not informed about the actual result of the procedure and did not have access to the post-operative data, were asked to predict the success rate of PCNL based on pre-operative imaging. The remaining stone size of less than 4 mm was considered as stone-free in PCNL. On the other hand, the STONE nomogram calculated the stone-free rate, and if the predicted success rate was over 60%, PCNL was considered successful based on the STONE nomogram. Results: The accuracy of the urologist’s prediction was 76.42%. The STONE nomogram correctly predicted the PCNL result in 59 cases (84.28%). There was a significant correlation between the actual success rate of the PCNL and the nomogram’s prediction (p-value= 1.013×10-5). There was no significant difference between the surgeons and the nomogram’s prediction (p-value=0.336). The number of stones did not significantly affect the correct prediction of surgical results (p-value<0.05). Upper calyx stones caused a significant difference between correct prediction of surgical results by surgeons and nomograms (p-value<0.05). Conclusion: No significant difference was found between surgeons’ prediction and STONE nomogram, but only in patients with kidney stones in the upper calyx; the nomogram’s prediction accuracy was higher.

Short-Term Passenger Flow Prediction in Urban Rail Transit based on Hybrid Deep Learning Models

Urban rail transit systems are essential for efficient, reliable, and environmentally friendly transportation in modern cities. Correct and effective short-term passenger flow prediction is crucial for optimizing operational efficiency, enhancing service quality, and ensuring passenger safety. Traditional prediction methods often fail to capture the complex, non-linear, and dynamic patterns in urban rail transit systems. This study uses a hybrid deep learning model combining Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks to predict short-term passenger flow in urban rail transit. By integrating the strengths of CNN in capturing spatial features and LSTM in modeling temporal dependencies, the hybrid model aims to improve prediction accuracy. Using historical data from Hangzhou Metro, the study demonstrates the model's effectiveness in predicting passenger flow, which reasonably has a good fit. Additionally, models with different convolutional layers show different performances. These improved predictions offer valuable insights for transit authorities, enabling them to make more informed decisions regarding train scheduling, resource allocation, and emergency response planning. By anticipating passenger demand more accurately, authorities can optimize the deployment of trains, reduce waiting times, enhance passenger comfort, and improve overall service reliability.

Semi-analytical peridynamic method for modal analysis of acoustoelastic Lamb waves

Lamb wave has been widely used as a non-destructive testing tool for inspecting the defects or damage in the plate system. A comprehensive understanding and correct prediction of the modal characteristics of Lamb waves are of high importance for ensuring successful practical applications. In this paper, a new method called the semi-analytical peridynamic (SAPD) method for analyzing wave propagation is developed. This method, within the framework of the general acoustoelasticity theory, uses the peridynamic differential operator to transform the equations of motion for guided waves in prestressed anisotropic media and the boundary conditions from local differential forms to nonlocal integral forms. By introducing meshfree discretization and Lagrange multipliers, these governing equations can be reorganized into a standard generalized eigenvalue formalism and solved. The effectiveness and accuracy of the SAPD method are first verified through comparison with the exact solutions. Phase and group velocity dispersion curves and displacement distributions of Lamb waves in three typical cases are then calculated to study the effects of material heterogeneity, applied stress and residual stress on the propagation of Lamb waves. Since complex grid generation algorithms are avoided, the SAPD method exhibits the advantages in terms of simplicity and implementation.

The Space–Time Representation of Extraordinary Rainfall Events

ABSTRACTExtraordinary events are rarely observable in a single rainfall gauge, and this make extremely challenging the correct prediction of their arrivals. However, it may be possible to develop a more robust approach by employing a space–time modelling scheme that is able to capture the spatial dynamics of such phenomena. Therefore, a space–time Poisson model of rainfall cells with circular shape and random depth has been exploited for the first time to interpret the behaviour of this family of extraordinary events. This category of events that may be connected to larger meteorological phenomena not necessarily connected with local heterogeneity of the landscape. Following the identification of the observed extraordinary event across southern Italy, six zones with significantly different dynamics in terms of the frequency of such extremes were identified. Subsequently, a simple mathematical representation was adopted to calibrate the model parameters, leading to an estimate of regional probability distributions defined on the space–time occurrences of extraordinary events over homogeneous zones. The approach allows to overcome the limitations posed by point observations allowed the definition of a probability distribution that pertains to an entire area rather than just a point. The obtained quantiles of rainfall estimated seems to align well with the upper bound of the probability distribution of the annual maxima observed over the areas of interests.

Enhancing urban flow prediction via mutual reinforcement with multi-scale regional information

Intelligent Transportation Systems (ITS) are essential for modern urban development, with urban flow prediction being a key component. Accurate flow prediction optimizes routes and resource allocation, benefiting residents, businesses, and the environment. However, few methods address the spatial–temporal heterogeneity of urban flows. Existing methods typically capture spatial features solely from urban flows, but spatial feature sensitivity becomes a bottleneck when dealing with small or noisy datasets. To address this issue, we propose a method for urban flow prediction via mutual reinforcement with multi-scale regional information (MR-UFP). Firstly, we employ spatial–temporal random masking and spatial–temporal contrastive learning pre-training to directly mine spatial–temporal heterogeneity from historical flow data. Secondly, we transform the task of spatial feature extraction and embedding for urban flow prediction into a mutual reinforcement task by multi-scale region classification auxiliary task. The adaptive environment fusion module and real-time graph processing dynamically correlate regional and environmental features during flow prediction. To balance the mutual reinforcement of the two tasks, we design a joint loss function to optimize feature embedding and feedback correction, ensuring robust and accurate urban flow prediction. Extensive experiments on two real-world datasets demonstrate that MR-UFP outperforms baseline models, showcasing its robustness and effectiveness even with minimal data.

Utilization of ResNet Architecture and Transfer Learning Method in the Classification of Faces of Individuals with Down Syndrome

Classifying the faces of individuals with Down Syndrome poses a significant challenge in image processing and genetic anomaly detection. This study leverages the ResNet34 architecture and transfer learning methods to improve classification accuracy for Down Syndrome facial recognition. Three experiments were conducted, varying the batch size, learning rate, and number of epochs. In the first experiment, the model achieved an accuracy of 82.83%, precision of 0.8362, recall of 0.8350, and an F1 score of 0.8348, showing promising performance but falling short of the target accuracy of 85%. The second experiment yielded the best results, with an accuracy of 87.88%, precision of 0.8956, recall of 0.8956, and an F1 score of 0.8956, indicating an optimal balance between correct predictions and errors. The third experiment resulted in the lowest accuracy, at 80.47%, with a precision of 0.8272, recall of 0.8249, and an F1 score of 0.8247, signifying a decline in performance compared to the other trials. Among the three experiments, the best configuration was achieved in the second trial, as the high recall value is crucial in medical contexts to ensure that as many individuals with Down Syndrome are correctly detected as possible, minimizing the risk of serious consequences due to false negatives.

Abstract A062: Insightful analysis: Harnessing aqueous humor ctDNA for retinoblastoma stratification

Abstract Retinoblastoma (RB) the most common intraocular malignancy in children poses a significant challenge as traditional tissue biopsies are not feasible. Due to its critical location within the eye and the high risk of tumor seeding, clinicians are limited in performing molecular analysis-based risk stratification prior to enucleation. In this study, we explored the potential of aqueous humor (AH) as a source for minimally invasive liquid biopsies, aiming to establish a molecular profiling methodology for circulating tumor DNA (ctDNA) for risk stratification. We analyzed AH samples and corresponding tumor tissues from 19 enucleated eyes of patients diagnosed with RB. The extracted ctDNA from AH underwent low-coverage whole-genome sequencing (lcWGS) and methylation profiling. These data were then compared with the genomic and methylation landscapes of the matched tumor DNA. Additionally, we assessed the ability of AH- derived ctDNA to predict retinoblastoma methylation cluster. Our findings demonstrate a high concordance between ctDNA from AH and the primary tumor DNA in both methylation patterns and genomic alterations detected by lcWGS allowing the correct prediction of the retinoblastoma molecular group. The shared aberrations underscore the potential of AH as a reliable surrogate for tumor tissue in retinoblastoma. Additionally, the methylation profiles revealed consistent epigenetic modifications, further supporting the accuracy of AH-derived ctDNA in reflecting the tumor’s molecular characteristics. This study highlights the feasibility and effectiveness of using AH for molecular ctDNA genetic analysis in retinoblastoma. The overall high concordance between ctDNA profiles with primary tumor DNA and the ability to accurately predict RB methylation classes, highlights the superiority of the liquid biopsy approach overcoming the limitations of traditional biopsies and transforming the potential for early molecular analysis and risk stratification. This revolutionary technique could redefine pre-enucleation diagnostics, offering unprecedented insights and paving the way for tailored, precision medicine in retinoblastoma care. Citation Format: Sophia H Montigel, Tatsiana Ryl, Stefanie Volz, Elena Afanasyeva, Tatjana Wedig, Nathalie Schwarz, Stefan M Pfister, Kristian W Pajtler, Petra Ketteler, Kendra K Maaß. Insightful analysis: Harnessing aqueous humor ctDNA for retinoblastoma stratification [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr A062.

Mode Analysis for the Rotating Aerodynamic Disturbance of a Transonic Fan

Abstract The non-synchronous blade vibration induced by a rotating aerodynamic disturbance is encountered for the fan of a turbofan engine working at 65% speedline. To analyze the unsteady characteristics, two methods based on the phase analysis and the compressive sensing (CS) for spatially-undersampled data are proposed and applied to obtain the circumferential mode order of the aerodynamic disturbance, and their performance has been carefully examined at various parameters. It was shown that when there are sufficient probes, i.e., 7–8 in the present study, both the two methods can give a correct prediction to the dominant mode order (the highest is 26 in this work) for all non-engine-order (EO) frequencies. However, if less probes are used, the deterioration of accuracy can be observed for both the two methods and a better performance can be observed for the CS method. In order to suppress the blade vibration, the fan blade is trimmed to alter its solid eigenfrequencies and avoid the resonance, which is proved to be an effective anti-vibration design. The existence of the aerodynamic disturbance after the blade is trimmed also suggests that the vibration is not self-excited but induced by the resonance of the flow excitation and the structure. For the cases with and without vibration, the origin of these non-EO frequencies can be well explained by the spinning mode theory and the interaction between the blade vibration and a high-order blade passing frequency has been observed in this study.

Loan Approval Prediction Improved by XGBoost Model Based on Four-Vector Optimization Algorithm

Abstract. This paper discusses an improved XGBoost model based on four-vector optimization algorithm to improve the accuracy of loan approval prediction. Through the analysis of correlation heat maps, we found that there were significant positive and negative correlations among some variables, which laid the foundation for subsequent machine learning analysis. Based on this, we compare the traditional machine learning algorithm with the improved model to evaluate its performance in loan approval forecasting. In the confusion matrix analysis of the training set, the improved XGBoost model demonstrated excellent performance, with all loan approval predictions being correct with 100% accuracy. However, the performance in the test set was slightly different, with 1,182 projects receiving correct loan approval predictions and 99 projects forecasting errors. Of these, 26 projects that should have been predicted to be "unapproved" were incorrectly labeled as "approved," while 73 projects that should have been predicted to be "approved" were incorrectly labeled as "unapproved." These results suggest that we still need to pay attention to the misjudgment of the model in practical application. By synthesizing all model evaluation indicators, we found that the improved XGBoost model based on the four-vector optimization algorithm has a higher accuracy in loan approval prediction than the traditional XGBoost model, with an increase of 1.5%. In addition, the other evaluation indicators also show a trend of significantly better than the traditional model. This study shows that the four-vector optimization algorithm can effectively improve the application effect of XGBoost model in the field of loan approval, and provide more accurate data support and decision-making basis for the financial industry. In the future, we will continue to explore the potential and application prospects of this algorithm in other fields.

Wastewater treatment monitoring: Fault detection in sensors using transductive learning and improved reinforcement learning

Wastewater treatment plants (WWTPs) increasingly utilize sensors to optimize operations and ensure treated water quality. These sensors’ rich datasets are well-suited for automated monitoring and fault detection. This study introduces a deep learning method for fault detection in sensors designed to tackle significant challenges, including a class imbalance in datasets where normal operational data significantly outnumber anomalies and sensitivity to hyperparameters. We employ a novel spatial attention-based transductive long short-term memory (TLSTM) network designed to detect subtle temporal variations in time-series data, facilitating the binary classification of faults in key processes like oxidation and nitrification. To address the challenge of data imbalance prevalent in WWTP monitoring, our model integrates the off-policy proximal policy optimization (Off-Policy PPO) framework. This adaptation enhances the traditional PPO algorithm for off-policy learning environments, improving data utilization and algorithm stability. In this system, data points are treated as a sequence of decisions, with the neural network functioning as an intelligent agent. The Off-Policy PPO approach employs a reward mechanism that prioritizes the correct prediction of minority-class instances over majority-class ones by assigning higher rewards. Moreover, the model incorporates the differential evolution (DE) algorithm for autonomous hyperparameter optimization, thereby minimizing reliance on manual tuning. Our rigorous testing on the Valdobbiadene dataset shows that our approach outperforms existing methods. Additionally, we apply transfer learning (TL) to the BSM1 dataset to further validate the model’s effectiveness. Achieving F-measures of 87.24% on the Valdobbiadene dataset and 82.48% on the BSM1 dataset demonstrates the model’s capability to promptly identify faults, significantly enhancing the reliability and efficiency of WWTP monitoring systems.