Evaluation Metrics Research Articles

Neutral Benchmarking of Survival Models in Health Sciences: Comparative Study of Classical and Machine Learning Techniques

Survival analysis plays a central role in diverse research fields, especially in health sciences. As an analytical tool, it can be used to help improve patients’ survival time, or at least, reduce the prospects of recurrence in cancer studies. However, approaches to the predictive performance of the current survival models mainly center on clinical data along with the classical survival methods. For censored “omics” data, the performance of survival models has not been thoroughly studied, either often due to their high dimensionality issues or reliance on binarizing the survival time for classification analysis. We aim to present a neural benchmark approach that analyzes and compares a broad range of classical and state-of-the-art machine learning survival models for “omics” and clinical datasets. All the methods considered in our study are evaluated using predictability as a performance measure. The study is systematically designed to make 36 comparisons (9 methods over 4 datasets, i.e., 2 clinical and 2 omics), and shows that, in practice, predictability of survival models does vary across real-world datasets, model choice, as well as the evaluation metric. From our results, we emphasize that performance criteria can play a key role in a balanced assessment of diverse survival models. Moreover, the Multitask Logistic Regression (MTLR) showed remarkable predictability for almost all the datasets. We believe this outstanding performance presents a unique opportunity for a wider use of MTLR for survival risk factors. For translational clinicians and scientists, we hope our findings provide practical guidance for benchmark studies of survival models, as well as highlight potential areas of research interest.

Development of a conditional variational autoencoder for handwritten digit recognition.

This study examines the performance of Conditional Variational Autoencoder (CVAE) in handwritten digit recognition. Using the MNIST dataset, two variants of the CVAE models — convolutional and multilevel architecture — were developed and compared. The research methodology includes comprehensive data preprocessing, architecture design, training, and thorough evaluation processes. The obtained data highlight the better performance of the convolutional model-based CVAE in achieving recognition accuracy compared to its multilayer counterpart. Evaluation metrics include analysis of original and reconstructed images, comparison of hidden layer vector distribution patterns, and visualization of loss function dynamics. In addition, the study highlights the practical implications of CVAEs in various fields, highlighting their performance in digit recognition tasks due to their inherent robustness and extraordinary generalizability.

Implementation of K-Nearest Neighbors Algorithm in Analyzing Public Interest in Shoping at Supermarkets

People's shopping patterns and behaviors continue to develop along with technological advances and lifestyle changes, thus requiring retail business actors, especially supermarkets, to better understand their customers' interests and preferences. In this context, accurate analysis of customer shopping interests is very important to improve customer satisfaction and optimize marketing strategies. One solution that can be implemented to analyze people's shopping interests is the application of the K-Nearest Neighbors algorithm, a simple yet effective nearest neighbor-based classification method for recognizing patterns from existing data. This study aims to apply the K-Nearest Neighbors algorithm to classify people's interest in shopping at supermarkets. This study also evaluates the effectiveness and performance of the algorithm in the context of business decision-making in the retail sector. The research methodology includes collecting data on people's shopping interests, data pre-processing, implementing the K-Nearest Neighbors algorithm, and evaluating model performance using evaluation metrics such as accuracy, precision, recall, and F1-score. The results of this study indicate that the K-Nearest Neighbors algorithm is able to achieve an accuracy of 88%, with precision, recall, and F1-score all reaching 92.86%. These results indicate that the K-Nearest Neighbors model is very effective in classifying people's shopping interests, with a low error rate. The resulting confusion matrix also shows the model's ability to identify customers who are interested in shopping with little prediction error. This study concludes that we can rely on the K-Nearest Neighbors algorithm to analyze people's shopping interests in supermarkets. This model not only shows good performance in classification but also has great potential to be implemented in recommendation systems and customer segmentation in the real world. This study contributes to the development of consumer behavior analysis methods in the retail sector, as well as providing a basis for further research to explore other algorithms or combinations of techniques to improve the accuracy and effectiveness of classification models.

Performance Evaluation of an Object Detection Model Using Drone Imagery in Urban Areas for Semi-Automatic Artificial Intelligence Dataset Construction.

Modern image processing technologies, such as deep learning techniques, are increasingly used to detect changes in various image media (e.g., CCTV and satellite) and understand their social and scientific significance. Drone-based traffic monitoring involves the detection and classification of moving objects within a city using deep learning-based models, which requires extensive training data. Therefore, the creation of training data consumes a significant portion of the resources required to develop these models, which is a major obstacle in artificial intelligence (AI)-based urban environment management. In this study, a performance evaluation method for semi-moving object detection is proposed using an existing AI-based object detection model, which is used to construct AI training datasets. The tasks to refine the results of AI-model-based object detection are analyzed, and an efficient evaluation method is proposed for the semi-automatic construction of AI training data. Different FBeta scores are tested as metrics for performance evaluation, and it is found that the F2 score could improve the completeness of the dataset with 26.5% less effort compared to the F0.5 score and 7.1% less effort compared to the F1 score. Resource requirements for future AI model development can be reduced, enabling the efficient creation of AI training data.

UAV Anomaly Detection Method Based on Convolutional Autoencoder and Support Vector Data Description with 0/1 Soft-Margin Loss

Unmanned aerial vehicles (UAVs) are becoming more widely used in various industries, raising growing concerns about their safety and reliability. The flight data of UAVs can directly reflect their flight health status; however, the rarity of abnormal flight data and the spatiotemporal characteristics of these data represent a significant challenge for constructing accurate and reliable anomaly detectors. To address this, this study proposes an anomaly detection framework that fully considers the temporal correlations and distribution characteristics of flight data. This framework first combines a one-dimensional convolutional neural network (1DCNN) with an autoencoder (AE) to establish a feature extraction model. This model leverages the feature extraction capabilities of the 1DCNN and the reconstruction capabilities of the AE to thoroughly extract the spatiotemporal features from UAV flight data. Then, to address the challenge of adaptive anomaly detection thresholds, this research proposes a nonlinear model of support vector data description (SVDD) utilizing a 0/1 soft-margin loss, referred to as L0/1-SVDD. This model replaces the traditional hinge loss function in SVDD with a 0/1 loss function, with the goal of enhancing the accuracy and robustness of anomaly detection. Since the 0/1 loss function is a bounded, non-convex, and non-continuous function, this paper proposes the Bregman ADMM algorithm to solve the L0/1-SVDD. Finally, the difference between the reconstructed and the actual value is employed to train the L0/1-SVDD, resulting in a hypersphere classifier that is capable of detecting UAV anomaly data. The experimental results using real flight data show that, compared with methods such as AE, LSTM, and LSTM-AE, the proposed method exhibits superior performance across five evaluation metrics.

A Representation-Learning-Based Graph and Generative Network for Hyperspectral Small Target Detection

Hyperspectral small target detection (HSTD) is a promising pixel-level detection task. However, due to the low contrast and imbalanced number between the target and the background spatially and the high dimensions spectrally, it is a challenging one. To address these issues, this work proposes a representation-learning-based graph and generative network for hyperspectral small target detection. The model builds a fusion network through frequency representation for HSTD, where the novel architecture incorporates irregular topological data and spatial–spectral features to improve its representation ability. Firstly, a Graph Convolutional Network (GCN) module better models the non-local topological relationship between samples to represent the hyperspectral scene’s underlying data structure. The mini-batch-training pattern of the GCN decreases the high computational cost of building an adjacency matrix for high-dimensional data sets. In parallel, the generative model enhances the differentiation reconstruction and the deep feature representation ability with respect to the target spectral signature. Finally, a fusion module compensates for the extracted different types of HS features and integrates their complementary merits for hyperspectral data interpretation while increasing the detection and background suppression capabilities. The performance of the proposed approach is evaluated using the average scores of AUCD,F, AUCF,τ, AUCBS, and AUCSNPR. The corresponding values are 0.99660, 0.00078, 0.99587, and 333.629, respectively. These results demonstrate the accuracy of the model in different evaluation metrics, with AUCD,F achieving the highest score, indicating strong detection performance across varying thresholds. Experiments on different hyperspectral data sets demonstrate the advantages of the proposed architecture.

A Comprehensive Survey of Masked Faces: Recognition, Detection, and Unmasking

Masked face recognition (MFR) has emerged as a critical domain in biometric identification, especially with the global COVID-19 pandemic, which introduced widespread face masks. This survey paper presents a comprehensive analysis of the challenges and advancements in recognizing and detecting individuals with masked faces, which has seen innovative shifts due to the necessity of adapting to new societal norms. Advanced through deep learning techniques, MFR, along with face mask recognition (FMR) and face unmasking (FU), represents significant areas of focus. These methods address unique challenges posed by obscured facial features, from fully to partially covered faces. Our comprehensive review explores the various deep learning-based methodologies developed for MFR, FMR, and FU, highlighting their distinctive challenges and the solutions proposed to overcome them. Additionally, we explore benchmark datasets and evaluation metrics specifically tailored for assessing performance in MFR research. The survey also discusses the substantial obstacles still facing researchers in this field and proposes future directions for the ongoing development of more robust and effective masked face recognition systems. This paper serves as an invaluable resource for researchers and practitioners, offering insights into the evolving landscape of face recognition technologies in the face of global health crises and beyond.

The relationship between accessibility and land prices: A focus on accessibility to transit in the 15-min city

In the context of local residents’ travel activities, access to transportation facilities is crucial and a key factor in achieving the goals of a 15-min city. However, current research on accessibility and land prices seldom fully considers the exploration of the relationship between the two using machine learning models. In this study, we conduct an analysis based on land price and accessibility in Beijing. Firstly, factors such as accessibility, population density, land use types, the number of Points of Interest (POI), and housing prices are selected as independent variables for the model, with land price as the dependent variable. Secondly, the CatBoost model is employed to investigate the complex relationship between accessibility to transit and land prices. Finally, the study introduces the XGBoost model and the ordinary least squares (OLS) method for comparison, validating the effectiveness of the CatBoost model in studying the complex relationship between accessibility and land prices through the comparison of model performance evaluation metrics. A clear nonlinear relationship exists between public transportation accessibility and land prices. Accessibility to bus within a 15-min bicycle is positively associated with land price; its trend shows a sharp increase followed by a gradual increase, and then another sharp increase. Accessibility to metro within the multistage distance threshold walk has a roughly positive association with land price within a certain range. While accessibility to bus within a 15-min walk and accessibility to bus within the multistage distance threshold walk have a negative association with land price. The nonlinear patterns between the four accessibility variables and land prices differ and exhibit significant spatial heterogeneity. Both housing prices and population density positively correlate with land prices. Housing prices sharply increase, then gradually, while population density gradually rises, then sharply. This paper explores the relationship between the two, reinforcing research on the mutual influence of public transportation facilities and surrounding environmental elements. It provides a theoretical basis for promoting transit-oriented development (TOD) and city planning.

Improving emi performance in automotive data transmission systems using spread spectrum modulation

This paper investigates the enhancement of electromagnetic interference (EMI) performance in automotive data transmission systems through the application of spread spectrum modulation techniques. The significance of this research lies in the growing complexity and density of electronic systems within modern vehicles, which increases the potential for EMI-related issues that can compromise the reliability and safety of automotive communication systems. The primary objective of this study is to evaluate the effectiveness of spread spectrum modulation in mitigating EMI and improving the overall performance of data transmission in automotive environments. To achieve this objective, the research utilizes both theoretical analysis and experimental validation. The methods include a detailed review of current EMI mitigation strategies, the design and implementation of spread spectrum modulation techniques in a controlled environment, and subsequent testing to measure the impact on EMI performance. Key metrics for evaluation include signal integrity, data transmission rate, and EMI reduction levels. The results of the study indicate a significant improvement in EMI performance when spread spectrum modulation is applied. The experimental data demonstrate a notable reduction in EMI levels, leading to enhanced signal integrity and more reliable data transmission. These findings suggest that spread spectrum modulation is a viable solution for addressing EMI challenges in automotive data transmission systems

An advanced dual-layered framework for sustainable supply chain performance

Purpose This study aims to evaluate the sustainable supply chain performance indicators. At a macro level, the identification of the sustainable supply chain management (SSCM) performance indicators is done through exhaustive literature survey and interviews with experts. Furthermore, these indicators are evaluated through a hybrid approach, i.e. total weighted interpretive structural modelling (TWISM) followed by analytic hierarchical process (AHP). Design/methodology/approach Micro small and medium enterprises (MSMEs) in India are a major contributor to nation’s GDP. However, this sector struggles to comprehend benefits from implementation of SSCM due to a lack of appropriate performance evaluation metrics. The purpose of this paper is to contribute to the body of knowledge in SSCM by proposing and evaluating a set of SSCM performance indicators. Findings The paper highlights the SSCM performance indicators and concludes that business strategies, implementation planning and impact of stakeholders are the top SSCM performance indicators (SPIs). Therefore, the decision-makers must initially focus on strategic requirements which foster the implementation of SSCM, thereby ensuring profitability for all stakeholders. Research limitations/implications Although the proposed framework was validated through a case study on Indian automobile component manufacturing MSMEs, future research would explore the extension of the framework to other industries. Originality/value The originality of this study lies in the application of the novel TWISM-AHP tool. Furthermore, the SPIs identified in the study, consider the integration of the triple bottom line from the MSME perspective. The TWISM-AHP analysis will be beneficial for SC decision-makers to enhance the SSCM performance based on the identified indicators and their criticality.

Enhancing Indoor Localization Accuracy in Wireless Sensor Networks: A Hybrid Approach Integrating Hierarchical Structure Poly Particle Swarm Optimization and Teaching–Learning-Based Optimization (HSPPSO-TLBO)

Background and Objective: Enhancing localization accuracy while minimizing development costs poses significant challenges in deploying and managing wireless sensor networks (WSNs). This paper presents an advanced algorithm for node localization in indoor environments, integrating sophisticated optimization techniques. The hybrid algorithm, HSPPSO-TLBO, combines Hierarchical Structure Poly Particle Swarm Optimization (HSPPSO) with Teaching– Learning-Based Optimization (TLBO). Methods: The proposed algorithm HSPPSO-TLBO aims to minimize the mean squared range error (MSRE) resulted by calculating internal distances between nodes using Received Signal Strength Indicator (RSSI). TLBO, with its robust global search capabilities, complements HSPPSO’s local search, preventing convergence to inappropriate local optima. HSPPSO-TLBO offers easy implementation and leverages the cost-free feature of RSSI, making it an attractive choice for enhancing localization precision. Results: Simulation results demonstrate the superior performance of HSPPSO-TLBO compared to other algorithms using different meta-heuristic optimization techniques. The outstanding performance of HSPPSO-TLBO is evident across various evaluation metrics, including localization error, localization rate, and simulation runtime. Conclusion: The proposed algorithm utilizing HSPPSO and TLBO is exceptionally effective in improving localization precision in indoor WSNs due to several key characteristics. These include the seamless integration and easy implementation of both HSPPSO and TLBO, along with the costfree advantage of using the RSSI technique. This combination makes the algorithm a highly functional solution for improving localization accuracy. other: N/A

Analysis of methods and algorithms for image filtering and quality enhancement

This paper addresses the prevalent issue of image noise and presents methods for its mitigation. The paper describes, analyses and tests a variety of image filtering techniques, with specific reference to their use in different contexts. The filtering methods can be classified into two principal categories: linear filters, which include the Gaussian and mean filters, and non-linear filters, which comprise the median filter, the Fast Fourier Transform (FFT), the Non-Local Means (NLM) filter, and the anisotropic diffusion filter. The efficacy of each filter is mathematically described and evaluated on RGB images using the Python programming language. The study delineates the evaluation metrics and their respective advantages and disadvantages. The Root Mean Square Error (RMSE) and Peak Signal-to-Noise Ratio (PSNR) are employed as criteria for the analysis of algorithm efficiency. Furthermore, the mean execution time for each algorithm is also monitored. The experimental data suggests that linear filters are relatively fast but produce inferior results and are best employed as preparatory measures. Non-linear filters have been demonstrated to be more robust and applicable to a variety of noise types, although it has been established that they require parameter fine-tuning. The study demonstrates that anisotropic diffusion is suitable for both manual image processing and real-time applications, offering an optimal balance between processing speed and denoised image quality. NLM is optimal for high-quality single image processing due to its superior results, despite a slower processing speed. FFT is noted for its efficiency in eliminating periodic noise. Further research will be conducted on advancing filtering techniques for different real-world scenarios and autonomous systems.

An open-source fine-tuned large language model for radiological impression generation: a multi-reader performance study

BackgroundThe impression section integrates key findings of a radiology report but can be subjective and variable. We sought to fine-tune and evaluate an open-source Large Language Model (LLM) in automatically generating impressions from the remainder of a radiology report across different imaging modalities and hospitals.MethodsIn this institutional review board-approved retrospective study, we collated a dataset of CT, US, and MRI radiology reports from the University of California San Francisco Medical Center (UCSFMC) (n = 372,716) and the Zuckerberg San Francisco General (ZSFG) Hospital and Trauma Center (n = 60,049), both under a single institution. The Recall-Oriented Understudy for Gisting Evaluation (ROUGE) score, an automatic natural language evaluation metric that measures word overlap, was used for automatic natural language evaluation. A reader study with five cardiothoracic radiologists was performed to more strictly evaluate the model’s performance on a specific modality (CT chest exams) with a radiologist subspecialist baseline. We stratified the results of the reader performance study based on the diagnosis category and the original impression length to gauge case complexity.ResultsThe LLM achieved ROUGE-L scores of 46.51, 44.2, and 50.96 on UCSFMC and upon external validation, ROUGE-L scores of 40.74, 37.89, and 24.61 on ZSFG across the CT, US, and MRI modalities respectively, implying a substantial degree of overlap between the model-generated impressions and impressions written by the subspecialist attending radiologists, but with a degree of degradation upon external validation. In our reader study, the model-generated impressions achieved overall mean scores of 3.56/4, 3.92/4, 3.37/4, 18.29 s,12.32 words, and 84 while the original impression written by a subspecialist radiologist achieved overall mean scores of 3.75/4, 3.87/4, 3.54/4, 12.2 s, 5.74 words, and 89 for clinical accuracy, grammatical accuracy, stylistic quality, edit time, edit distance, and ROUGE-L score respectively. The LLM achieved the highest clinical accuracy ratings for acute/emergent findings and on shorter impressions.ConclusionsAn open-source fine-tuned LLM can generate impressions to a satisfactory level of clinical accuracy, grammatical accuracy, and stylistic quality. Our reader performance study demonstrates the potential of large language models in drafting radiology report impressions that can aid in streamlining radiologists’ workflows.

Evaluating fine tuned deep learning models for real-time earthquake damage assessment with drone-based images

Earthquakes pose a significant threat to life and property worldwide. Rapid and accurate assessment of earthquake damage is crucial for effective disaster response efforts. This study investigates the feasibility of employing deep learning models for damage detection using drone imagery. We explore the adaptation of models like VGG16 for object detection through transfer learning and compare their performance to established object detection architectures like YOLOv8 (You Only Look Once) and Detectron2. Our evaluation, based on various metrics including mAP, mAP50, and recall, demonstrates the superior performance of YOLOv8 in detecting damaged buildings within drone imagery, particularly for cases with moderate bounding box overlap. This finding suggests its potential suitability for real-world applications due to the balance between accuracy and efficiency. Furthermore, to enhance real-world feasibility, we explore two strategies for enabling the simultaneous operation of multiple deep learning models for video processing: frame splitting and threading. In addition, we optimize model size and computational complexity to facilitate real-time processing on resource-constrained platforms, such as drones. This work contributes to the field of earthquake damage detection by (1) demonstrating the effectiveness of deep learning models, including adapted architectures, for damage detection from drone imagery, (2) highlighting the importance of evaluation metrics like mAP50 for tasks with moderate bounding box overlap requirements, and (3) proposing methods for ensemble model processing and model optimization to enhance real-world feasibility. The potential for real-time damage assessment using drone-based deep learning models offers significant advantages for disaster response by enabling rapid information gathering to support resource allocation, rescue efforts, and recovery operations in the aftermath of earthquakes.

Standardizing Excellence: Metric Assemblages in Mathematics Research in Chile

This study investigates the development, implementation, and establishment of evaluation metrics within the mathematics research community in Chile, a discipline where the use of bibliometric indicators for quality assessment has been highly controversial. Using a mixed-methods approach, the study conducted focus groups with mathematicians at different stages of their academic careers and analyzed the ANID Mathematics Studies Group's method for assessing researcher productivity. The findings reveal that the evaluation metrics in mathematics remain in a state of “magmatic flow,” fluid and solid, constantly changing yet stable enough to articulate multiple social worlds. The study highlights the challenges faced in three closely linked spheres: mathematical research, science and higher education governance, and organizational management. The functioning of this metrics assemblage as boundary infrastructure, which meets the information needs of diverse social worlds while maintaining the integrity of the interests involved, is a temporal achievement of situated practices that provide continuous adjustment and stabilization. The study suggests that academic metric assemblages are made unstable through practical politics involving controversies and qualitative deliberation. Concurrently, they achieve stability through translation and blackboxing practices.

Economic valuation of carbon sequestration in Quito’s Metropolitan Park using Sentinel-2 and neural networks

This study introduces an innovative method for assessing the economic value of carbon stored by the urban park forest in Quito, Ecuador. The method uses Sentinel-2 remote sensing data and advanced deep learning techniques. A large forest study area was chosen, and detailed field measurements were taken to gather biomass and carbon data. Sentinel-2 satellite images were processed to correct for radiation and atmospheric conditions, and vegetation indices such as NDVI and EVI were calculated. To model forest biomass, a convolutional neural network (CNN) was created and trained using Sentinel-2 spectral bands and vegetation indices. The model was validated with independent field data and demonstrated high accuracy in estimating biomass and carbon, as indicated by evaluation metrics like RMSE and R². The findings include detailed maps showing the spatial distribution of biomass and carbon in the study area, which can be a valuable tool for forest management and the implementation of policies for valuing stored carbon. This approach combines the high resolution of Sentinel-2 data with the predictive power of neural networks, providing a robust and scalable method for estimating carbon across large forest areas. The study's conclusions emphasize the feasibility and accuracy of this approach and its potential for application in various forestry and geographical contexts.

Integrating IoT and visual question answering in smart cities: Enhancing educational outcomes

Emerging as a paradigmatic shift in urban development, smart cities harness the potential of advanced information and communication technologies to seamlessly integrate urban functions, optimize resource allocation, and improve the effectiveness of city management. Within the domain of smart education, the imperative application of Visual Question Answering (VQA) technology encounters significant limitations at the prevailing stage, particularly the absence of a robust Internet of Things (IoT) framework and the inadequate incorporation of large pre-trained language models (LLMs) within contemporary smart education paradigms, especially in addressing zero-shot VQA scenarios, which pose considerable challenges. In response to these constraints, this paper introduces an IoT-based smart city framework that is designed to refine the functionality and efficacy of educational systems. This framework is delineated into four cardinal layers: the data collection layer, data transmission layer, data management layer, and application layer. Furthermore, we introduce the innovative TeachVQA methodology at the application layer, synergizing VQA technology with extensive pre-trained language models, thereby considerably enhancing the dissemination and assimilation of educational content. Evaluative metrics in the VQAv2 and OKVQA datasets substantiate that the TeachVQA methodology not only outperforms existing VQA approaches, but also underscores its profound potential and practical relevance in the educational sector.

Exploiting K-Space in Magnetic Resonance Imaging Diagnosis: Dual-Path Attention Fusion for K-Space Global and Image Local Features.

Magnetic resonance imaging (MRI) diagnosis, enhanced by deep learning methods, plays a crucial role in medical image processing, facilitating precise clinical diagnosis and optimal treatment planning. Current methodologies predominantly focus on feature extraction from the image domain, which often results in the loss of global features during down-sampling processes. However, the unique global representational capacity of MRI K-space is often overlooked. In this paper, we present a novel MRI K-space-based global feature extraction and dual-path attention fusion network. Our proposed method extracts global features from MRI K-space data and fuses them with local features from the image domain using a dual-path attention mechanism, thereby achieving accurate MRI segmentation for diagnosis. Specifically, our method consists of four main components: an image-domain feature extraction module, a K-space domain feature extraction module, a dual-path attention feature fusion module, and a decoder. We conducted ablation studies and comprehensive comparisons on the Brain Tumor Segmentation (BraTS) MRI dataset to validate the effectiveness of each module. The results demonstrate that our method exhibits superior performance in segmentation diagnostics, outperforming state-of-the-art methods with improvements of up to 63.82% in the HD95 distance evaluation metric. Furthermore, we performed generalization testing and complexity analysis on the Automated Cardiac Diagnosis Challenge (ACDC) MRI cardiac segmentation dataset. The findings indicate robust performance across different datasets, highlighting strong generalizability and favorable algorithmic complexity. Collectively, these results suggest that our proposed method holds significant potential for practical clinical applications.

Enhancing Road Traffic Prediction Using Data Preprocessing Optimization

Traffic prediction is essential for transportation planning, resource allocation, congestion management and enhancing travel experiences. This study optimizes data preprocessing techniques to improve machine learning-based traffic prediction models. Data preprocessing is critical in preparing the data for machine learning models. This study proposes an approach that optimizes data preprocessing techniques, focusing on flow-based analysis and optimization, to enhance traffic prediction models. The proposed approach explores fixed and variable orders of data preprocessing using a genetic algorithm across five diverse datasets. Evaluation metrics such as root mean squared error (RMSE), mean absolute error (MAE) and R-squared error assess model performance. The results indicate that the genetic algorithm’s variable order achieves the best performance for the ArcGIS Hub and Frementon Bridge Cycle datasets, fixed order one preprocessing for the Traffic Prediction dataset and variable order using the genetic algorithm for the PeMS08 dataset. Fixed order 2 preprocessing yields the best performance for the XI AN Traffic dataset. These findings highlight the importance of selecting the appropriate data preprocessing flow order for each dataset, improving traffic prediction accuracy and reliability. The proposed approach advances traffic prediction methodologies, enabling more precise and reliable traffic forecasts for transportation planning and management applications.

A Multi-Scale Content-Structure Feature Extraction Network Applied to Gully Extraction

Black soil is a precious soil resource, yet it is severely affected by gully erosion, which is one of the most serious manifestations of land degradation. The determination of the location and shape of gullies is crucial for the work of gully erosion control. Traditional field measurement methods consume a large amount of human resources, so it is of great significance to use artificial intelligence techniques to automatically extract gullies from satellite remote sensing images. This study obtained the gully distribution map of the southwestern region of the Dahe Bay Farm in Inner Mongolia through field investigation and measurement and created a gully remote sensing dataset. We designed a multi-scale content structure feature extraction network to analyze remote sensing images and achieve automatic gully extraction. The multi-layer information obtained through the resnet34 network is input into the multi-scale structure extraction module and the multi-scale content extraction module designed by us, respectively, obtained richer intrinsic information about the image. We designed a structure content fusion network to further fuse structural features and content features and improve the depth of the model’s understanding of the image. Finally, we designed a muti-scale feature fusion module to further fuse low-level and high-level information, enhance the comprehensive understanding of the model, and improve the ability to extract gullies. The experimental results show that the multi-scale content structure feature extraction network can effectively avoid the interference of complex backgrounds in satellite remote sensing images. Compared with the classic semantic segmentation models, DeepLabV3+, PSPNet, and UNet, our model achieved the best results in several evaluation metrics, the F1 score, recall rate, and intersection over union (IoU), with an F1 score of 0.745, a recall of 0.777, and an IoU of 0.586. These results proved that our method is a highly automated and reliable method for extracting gullies from satellite remote sensing images, which simplifies the process of gully extraction and provides us with an accurate guide to locate the location of gullies, analyze the shape of gullies, and then provide accurate guidance for gully management.