Efficient Solution Research Articles

Application of efficient algorithm based on block Newton method to elastoplastic problems with nonlinear kinematic hardening

PurposeThe purpose of this study is to present the effectiveness and robustness of a numerical algorithm based on the block Newton method for the nonlinear kinematic hardening rules adopted in modeling ductile materials.Design/methodology/approachElastoplastic problems can be defined as a coupled problem of the equilibrium equation for the overall structure and the yield equations for the stress state at every material point. When applying the Newton method to the coupled residual equations, the displacement field and the internal variables, which represent the plastic deformation, are updated simultaneously.FindingsThe presented numerical scheme leads to an explicit form of the hardening behavior, which includes the evolution of the equivalent plastic strain and the back stress, with the internal variables. The features of the present approach allow the displacement field and the hardening behavior to be updated straightforwardly. Thus, the scheme does not have any local iterative calculations and enables us to simultaneously decrease the residuals in the coupled boundary value problems.Originality/valueA pseudo-stress for the local residual and an algebraically derived consistent tangent are applied to elastic-plastic boundary value problems with nonlinear kinematic hardening. The numerical procedure incorporating the block Newton method ensures a quadratic rate of asymptotic convergence of a computationally efficient solution scheme. The proposed algorithm provides an efficient and robust computation in the elastoplastic analysis of ductile materials. Numerical examples under elaborate loading conditions demonstrate the effectiveness and robustness of the numerical scheme implemented in the finite element analysis.

Capturing the dilution effect of risk‐based grouping with application to COVID‐19 screening

AbstractWe investigate the problem of screening a large population for an infectious disease (i.e., classifying subjects as positive or negative) using group testing while considering important test and population‐level characteristics. Group testing, in which multiple samples are pooled together into a master sample and tested simultaneously, has the potential to significantly expand screening efforts, and, owing to the COVID‐19 pandemic, the topic has seen a surge of interest recently. In this paper, we construct optimal group testing designs that consider a heterogeneous population (i.e., with subject‐specific risk), imperfect tests, and while also modeling the dilution effect of grouping (a phenomenon in which the test accuracy of the master sample is affected by the concentration of the virus in the pool), which is often ignored in the literature. We conduct an exhaustive analysis under both a general dilution function and a specific (yet still calibratable) form of the dilution function. Our analytical results of the resulting challenging optimization problems unveil key structural properties that hold in an optimal solution, which we utilize to construct efficient solution schemes. We complement the analysis with two case studies, one on the screening of blood for the Hepatitis B Virus and the other on the screening of subjects for COVID‐19. Our results reveal significant benefits over current practices, individual testing, as well as prior studies that ignore the dilution effect. Such results underscore the importance of incorporating both test and population‐level characteristics into the modeling framework.

Mechanism-driven improved SVMD: an indirect approach for rail corrugation detection using axle box acceleration

Abstract Addressing the challenge of the current inability to qualitatively identify rail corrugation damage accurately from axle box acceleration data, this study proposes a novel approach. To indirectly identify rail corrugation from axle box acceleration, we introduce an improved successive variational mode decomposition (SVMD) algorithm, coupled with a deep learning model for corrugation recognition. First, a numerical model of the vehicle-rail-track slab system is established, considering rail corrugation. Indicator analysis is integrated into the SVMD. The improved SVMD is employed to decompose and reconstruct axle box acceleration, achieving noise reduction and extraction of useful components. Next, we apply the fast Continuous Wavelet Transform analysis method to effectively transform one-dimensional data into two-dimensional images. Finally, the You Only Look Once (YOLO) model serves as a classifier for the classification and recognition of corrugation with different wavelengths. The results demonstrate that the mechanism-driven improved SVMD effectively extracts corrugation components from axle box acceleration, while the YOLO model achieves rapid and efficient identification and classification of corrugation with different wavelengths. The results show that compared with other traditional models, the training time of the YOLO model is 60%–90% of the training time of the traditional algorithm, and the Recall rate is 1.15–1.45 times that of the traditional algorithm. In terms of wavelength identification, the YOLO model has a recognition rate of 98% for different wavelengths. The proposed approach offers an innovative and efficient solution for identifying rail corrugation damage in axle box acceleration data.

A novel multi-attention reinforcement learning for the scheduling of unmanned shipment vessels (USV) in automated container terminals

To improve the operating efficiency of container terminals, we investigate a closed-loop scheduling method in an autonomous inter-terminal system that employs unmanned shipment vessels (USVs) to transport containers among operational berths (Dedicated to USVs) in seaport terminals. Our USVs scheduling model is developed by considering energy replenishment, time windows, and berth restrictions, aiming to obtain cost-saving USV transportation solutions and conflict-free paths. To solve this optimization model more efficiently, we propose the multi-attention reinforcement learning (MARL) algorithm by integrating an encoder-decoder framework and an unsupervised auxiliary network. The MARL algorithm provides instant problem-solving capabilities and benefits from extensive offline training. Experimental results demonstrate that our method can obtain efficient solutions for our USVs scheduling problem, and our algorithm outperforms other compared algorithms on computing time and solution accuracy.

Broad-Spectrum Efficacy and Modes of Action of Two Bacillus Strains against Grapevine Black Rot and Downy Mildew.

Black rot (Guignardia bidwellii) and downy mildew (Plasmopara viticola) are two major grapevine diseases against which the development of efficient biocontrol solutions is required in a context of sustainable viticulture. This study aimed at evaluating and comparing the efficacy and modes of action of bacterial culture supernatants from Bacillus velezensis Buz14 and B. ginsengihumi S38. Both biocontrol agents (BCA) were previously demonstrated as highly effective against Botrytis cinerea in grapevines. In semi-controlled conditions, both supernatants provided significant protection against black rot and downy mildew. They exhibited antibiosis against the pathogens by significantly decreasing G. bidwellii mycelial growth, but also the release and motility of P. viticola zoospores. They also significantly induced grapevine defences, as stilbene production. The LB medium, used for the bacterial cultures, also showed partial effects against both pathogens and induced plant defences. This is discussed in terms of choice of experimental controls when studying the biological activity of BCA supernatants. Thus, we identified two bacterial culture supernatants as new potential biocontrol products exhibiting multi-spectrum antagonist activity against different grapevine key pathogens and having a dual mode of action.

Research on improved gangue target detection algorithm based on Yolov8s.

An improved algorithm based on Yolov8s is proposed to address the slower speed, higher number of parameters, and larger computational cost of deep learning in coal gangue target detection. A lightweight network, Fasternet, is used as the backbone to increase the speed of object detection and reduce the model complexity. By replacing Slimneck with the C2F part in the HEAD module, the aim is to reduce model complexity and improve detection accuracy. The detection accuracy is effectively improved by replacing the Detect layer with Detect-DyHead. The introduction of DIoU loss function instead of CIoU loss function and the combination of BAM block attention mechanism makes the model pay more attention to critical features, which further improves the detection performance. The results show that the improved model compresses the storage size of the model by 28%, reduces the number of parameters by 28.8%, reduces the computational effort by 34.8%, and improves the detection accuracy by 2.5% compared to the original model. The Yolov8s-change model provides a fast, real-time and efficient detection solution for gangue sorting. This provides a strong support for the intelligent sorting of coal gangue.

Leak and Burst Detection in Water Distribution Network Using Logic- and Machine Learning-Based Approaches

Urban water systems worldwide are confronted with the dual challenges of dwindling water resources and deteriorating infrastructure, emphasising the critical need to minimise water losses from leakage. Conventional methods for leak and burst detection often prove inadequate, leading to prolonged leak durations and heightened maintenance costs. This study investigates the efficacy of logic- and machine learning-based approaches in early leak detection and precise location identification within water distribution networks. By integrating hardware and software technologies, including sensor technology, data analysis, and study on the logic-based and machine learning algorithms, innovative solutions are proposed to optimise water distribution efficiency and minimise losses. In this research, we focus on a case study area in the Sunbury region of Victoria, Australia, evaluating a pumping main equipped with Supervisory Control and Data Acquisition (SCADA) sensor technology. We extract hydraulic characteristics from SCADA data and develop logic-based algorithms for leak and burst detection, alongside state-of-the-art machine learning techniques. These methodologies are applied to historical data initially and will be subsequently extended to live data, enabling the real-time detection of leaks and bursts. The findings underscore the complementary nature of logic-based and machine learning approaches. While logic-based algorithms excel in capturing straightforward anomalies based on predefined conditions, they may struggle with complex or evolving patterns. Machine learning algorithms enhance detection by learning from historical data, adapting to changing conditions, and capturing intricate patterns and outliers. The comparative analysis of machine learning models highlights the superiority of the local outlier factor (LOF) in anomaly detection, leading to its selection as the final model. Furthermore, a web-based platform has been developed for leak and burst detection using a selected machine learning model. The success of machine learning models over traditional logic-based approaches underscores the effectiveness of data-driven, probabilistic methods in handling complex data patterns and variations. Leveraging statistical and probabilistic techniques, machine learning models offer adaptability and superior performance in scenarios with intricate or dynamic relationships between variables. The findings demonstrate that the proposed methodology can significantly enhance the early detection of leaks and bursts, thereby minimising water loss and associated economic costs. The implications of this study are profound for the scientific community and stakeholders, as it provides a scalable and efficient solution for water pipeline monitoring. Implementing this approach can lead to more proactive maintenance strategies, ultimately contributing to the sustainability and resilience of urban water infrastructure systems.

Leak and Burst Detection in Water Distribution Network Using Logic- and Machine Learning-Based Approaches

Urban water systems worldwide are confronted with the dual challenges of dwindling water resources and deteriorating infrastructure, emphasising the critical need to minimise water losses from leakage. Conventional methods for leak and burst detection often prove inadequate, leading to prolonged leak durations and heightened maintenance costs. This study investigates the efficacy of logic- and machine learning-based approaches in early leak detection and precise location identification within water distribution networks. By integrating hardware and software technologies, including sensor technology, data analysis, and study on the logic-based and machine learning algorithms, innovative solutions are proposed to optimise water distribution efficiency and minimise losses. In this research, we focus on a case study area in the Sunbury region of Victoria, Australia, evaluating a pumping main equipped with Supervisory Control and Data Acquisition (SCADA) sensor technology. We extract hydraulic characteristics from SCADA data and develop logic-based algorithms for leak and burst detection, alongside state-of-the-art machine learning techniques. These methodologies are applied to historical data initially and will be subsequently extended to live data, enabling the real-time detection of leaks and bursts. The findings underscore the complementary nature of logic-based and machine learning approaches. While logic-based algorithms excel in capturing straightforward anomalies based on predefined conditions, they may struggle with complex or evolving patterns. Machine learning algorithms enhance detection by learning from historical data, adapting to changing conditions, and capturing intricate patterns and outliers. The comparative analysis of machine learning models highlights the superiority of the local outlier factor (LOF) in anomaly detection, leading to its selection as the final model. Furthermore, a web-based platform has been developed for leak and burst detection using a selected machine learning model. The success of machine learning models over traditional logic-based approaches underscores the effectiveness of data-driven, probabilistic methods in handling complex data patterns and variations. Leveraging statistical and probabilistic techniques, machine learning models offer adaptability and superior performance in scenarios with intricate or dynamic relationships between variables. The findings demonstrate that the proposed methodology can significantly enhance the early detection of leaks and bursts, thereby minimising water loss and associated economic costs. The implications of this study are profound for the scientific community and stakeholders, as it provides a scalable and efficient solution for water pipeline monitoring. Implementing this approach can lead to more proactive maintenance strategies, ultimately contributing to the sustainability and resilience of urban water infrastructure systems.

BIOREALIZATION ENGINEERING TECHNOLOGIES FOR CANCER DIAGNOSIS: IMPACT ON THE HEALTH SYSTEM

This article delves into the analysis of the impact of bio-realization engineering technologies on cancer diagnostic processes within the contemporary healthcare system. The objective of the article is to identify the potential of molecular-genetic testing and other bio-realization methods in enhancing diagnostic capabilities and therapeutic approaches for oncological diseases. To achieve this goal, general scientific methods of analysis and synthesis of existing research in this field are utilized, including a review of scientific literature and a critical analysis of the results of experimental studies. The findings confirm the significant impact of bio-realization engineering technologies on the advancement of cancer diagnostics. Specifically, molecular-genetic testing opens new avenues for early detection and personalized treatment approaches for cancer, based on the genetic characteristics of tumors. These technologies are particularly valuable for countries with limited medical resources, as they offer cost-effective and efficient solutions that provide broader access to quality diagnostic services. They also contribute to improving the overall efficacy of treatment strategies and optimizing medical research, reducing the burden on medical staff. The practical significance of the obtained results lies in the possibility of their application in the development of new diagnostic tools and techniques aimed at enhancing the accuracy, accessibility, and efficiency of cancer treatment. This, in turn, may contribute to reducing mortality from oncological diseases and improving patients' quality of life. Bio-realization engineering technologies, particularly molecular-genetic testing, play a pivotal role in modern oncological diagnostics, offering promising opportunities for the improvement of diagnostic and therapeutic procedures for cancer. Their integration into the healthcare system enhances diagnostic accuracy, treatment accessibility, and the overall level of medical services, opening new horizons in the fight against oncological diseases.

Design and Construction of an Automated Boat that Removes Debris from the Ocean

Water bodies have a huge aquatic ecosystem within them. Any high concentration of suspended solids can cause numerous issues for aquatic life by blocking light from reaching submerged vegetation, hence there is a need for a solution to remove this debris from the water bodies. This project design and construction of an automated boat to remove debris from the ocean is focused on developing a robust and efficient prototype solution for autonomously collecting and managing marine Debris, thereby contributing to environment conservation. The two-channel remote control with receiver will enable the boat to navigate predefined routes, identify and collect debris using specialized mechanism and return to a designated collective point. The Two-Channel remote control system allows for precise maneuverability and control over the boat’s function. Through a combination of engineering principles, propellers, conveyer belt, tread mill arrangement and shafts. this project aims to address the critical issues of ocean pollution while showcasing the potential for remote controlled boat in maritime environmental management. The anticipated outcomes include a functional prototype, comprehensive operational documentation and preliminary performance assessments, setting the stage for potential scalability and deployment in real-world Ocean clean up initiative

Design and Construction of an Automated Boat that Removes Debris from the Ocean

Water bodies have a huge aquatic ecosystem within them. Any high concentration of suspended solids can cause numerous issues for aquatic life by blocking light from reaching submerged vegetation, hence there is a need for a solution to remove this debris from the water bodies. This project design and construction of an automated boat to remove debris from the ocean is focused on developing a robust and efficient prototype solution for autonomously collecting and managing marine Debris, thereby contributing to environment conservation. The two-channel remote control with receiver will enable the boat to navigate predefined routes, identify and collect debris using specialized mechanism and return to a designated collective point. The Two-Channel remote control system allows for precise maneuverability and control over the boat’s function. Through a combination of engineering principles, propellers, conveyer belt, tread mill arrangement and shafts. this project aims to address the critical issues of ocean pollution while showcasing the potential for remote controlled boat in maritime environmental management. The anticipated outcomes include a functional prototype, comprehensive operational documentation and preliminary performance assessments, setting the stage for potential scalability and deployment in real-world Ocean clean up initiative

MONITORING ADMINSTRASI JARINGAN DENGAN MIKROTIK DAN TELEGRAM BOT PADA INTERNET SERVICE PROVIDER

Network management using MikroTik devices at Lajur Net still has shortcomings. Although internet access is available around the clock, the limited number of employees makes it difficult for managers to access computers to manage MikroTik. Administrators hope to manage the MikroTik network remotely through end devices. Currently, accessing MikroTik online incurs high operational costs for renting public IPs and web-based management application services. This study aims to utilize Telegram Bot for network management and monitoring that can be accessed remotely via the internet. The use of Telegram Bot for network management at Lajur Net can be successfully implemented. The implementation of Telegram Bot includes managing interfaces, hotspots, ICMP, and monitoring the resources of MikroTik RouterBoard devices in the Lajur Net network. Telegram Bot can be accessed remotely through the internet and works well on the Lajur Net network. The implementation of Telegram Bot includes interface administration, hotspots, ping, and monitoring the resources of MikroTik RouterBoard devices. The additional firewall feature allows network administrators to add and delete filter rules within the Lajur Net network security system. Moreover, Telegram Bot utilizes button features on Telegram messages to facilitate usage without having to type all commands. Thus, Telegram Bot offers an efficient and effective solution in MikroTik network management at Lajur Net, reducing the need for expensive public IPs and web-based management applications. This demonstrates that the use of chatbot technology can provide significant benefits in network administration, especially for internet service providers with limited resources. Through this research, it is hoped that the implementation of Telegram Bot can serve as a reference for developing more innovative and efficient network management in the future.

Sound Absorption Performance of Ultralight Honeycomb Sandwich Panels Filled with "Network" Fibers-Juncus effusus.

This study investigates lightweight and efficient candidates for sound absorption to address the growing demand for sustainable and eco-friendly materials in noise attenuation. Juncus effusus (JE) is a natural fiber known for its unique three-dimensional network, providing a viable and sustainable filler for enhanced sound absorption in honeycomb panels. Microperforated-panel (MPP) honeycomb absorbers incorporating JE fillers were fabricated and designed, focusing on optimizing the absorber designs by varying JE filler densities, geometrical arrangements, and MPP parameters. At optimal filling densities, the MPP-type honeycomb structures filled with JE fibers achieved high noise reduction coefficients (NRC) of 0.5 and 0.7 at 20 mm and 50 mm thicknesses, respectively. Using an analytical model and an artificial neural network (ANN) model, the sound absorption characteristics of these absorbers were successfully predicted. This study demonstrates the potential of JE fibers in improving noise mitigation strategies across different industries, offering more sustainable and efficient solutions for construction and transportation.

Hand Gesture Recognition System using Deep Learning

This paper presents a comprehensive exploration of a hand gesture recognition system centered on the YOLOv5 object detection algorithm. The system is meticulously designed to fulfill specific objectives, including real-time gesture detection, high accuracy, scalability, robustness to environmental variability, and seamless integration with other system components. Drawing upon the strengths of YOLOv5, the proposed system aims to provide a reliable and efficient solution for hand gesture recognition, thereby enhancing human- computer interaction and enabling various applications across different domains. The methodology encompasses several stages, including data acquisition, preprocessing, object detection using YOLOv5, integration with other system modules, evaluation, and validation. Through a series of experiments and extensive testing in real-world scenarios, the effectiveness and usability of the system are thoroughly assessed, demonstrating its potential for practical deployment and widespread adoption. Additionally, this paper discusses the implications of the findings, identifies areas for further research and development, and offers insights into the future directions of hand gesture recognition systems utilizing YOLOv5. Overall, this study contributes to advancing the field of human-computer interaction and gesture recognition technology, paving the way for innovative applications and enhanced user experiences in various domains. Keywords— Hand Gesture Recognition, YOLOv5, Real-time Detection, Human-Computer Interaction, Object Detection, Machine Learning

A flexible combinatorial auction bidding language for supplier selection and order allocation in a supply chain with price sensitive demand

This study addresses the multi-item multi-sourcing supplier selection and order allocation problem in a Combinatorial Auction (CA) framework. First, we propose a new flexible procurement CA bidding language, then prove that it can convey the same information as existing bidding languages while allowing more efficient auction outcomes and reduced computational complexity. Suppliers submit combinatorial bids to a single manufacturer. Each bid carries information about the supplier’s cost structure and potential discounts. The manufacturer purchases the items and assembles one or more product types that experience price-sensitive demand rates by end consumers. We use the realistic logit function to represent the price-sensitive demand rates of finished products. A Mixed Integer Non-Linear Programming (MINLP) model is developed to help the manufacturer make optimal supplier selection, order allocation, and pricing decisions under the proposed bidding language. The MINLP model considers purchasing, transportation, ordering, administrative, holding, and manufacturing costs. These considerations make the results of this study realistic with the potential to be used in practical applications. Also, we derive a set of necessary conditions that must exist in at least one optimal solution to the problem. These conditions can be utilized to design efficient solution mechanisms for this difficult problem in the future. Finally, an extensive numerical analysis is performed to illustrate the bidding language and the key results.

Solving Dynamic Full-Truckload Vehicle Routing Problem Using an Agent-Based Approach

In today’s complex and dynamic transportation networks, increasing energy costs and adverse environmental impacts necessitate the efficient transport of goods or raw materials across a network to minimize all related costs through vehicle assignment and routing decisions. Vehicle routing problems under dynamic and stochastic conditions are known to be very challenging in both mathematical modeling and computational complexity. In this study, a special variant of the full-truckload vehicle assignment and routing problem was investigated. First, a detailed analysis of the processes in a liquid transportation logistics firm with a large fleet of tanker trucks was conducted. Then, a new original problem with distinctive features compared with similar studies in the literature was formulated, including pickup/delivery time windows, nodes with different functions (pickup/delivery, washing facilities, and parking), a heterogeneous truck fleet, multiple trips per truck, multiple trailer types, multiple freight types, and setup times between changing freight types. This dynamic optimization problem was solved using an intelligent multi-agent model with agent designs that run on vehicle assignment and routing algorithms. To assess the performance of the proposed approach under varying environmental conditions (e.g., congestion factors and the ratio of orders with multiple trips) and different algorithmic parameter levels (e.g., the latest response time to orders and activating the interchange of trip assignments between vehicles), a detailed scenario analysis was conducted based on a set of designed simulation experiments. The simulation results indicate that the proposed dynamic approach is capable of providing good and efficient solutions in response to dynamic conditions. Furthermore, using longer latest response times and activating the interchange mechanism have significant positive impacts on the relevant costs, profitability, ratios of loaded trips over the total distance traveled, and the acceptance ratios of customer orders.

A Preliminary Study of 3D Printing Home Designs for Improving Efficiency and Sustainability of Indigenous Housing in Canada

Canada has been experiencing a significant housing crisis in recent years, especially in remote and Indigenous communities, yet most of the existing construction approaches have not been rapid, sustainable, and affordable enough to meet community needs. To address this challenge, this study explores the feasibility of 3D printed (3DP) housing and develops a design that is informed by Indigenous housing requirements and is realized through a physical design prototype tailored for the implementation of 3DP homes. Site visits and community engagement were integral parts of the research to help deliver invaluable insights that guided the design process, ensuring cultural sensitivity and inclusivity. The prototyped 3DP design offers efficient and sustainable solutions customized to the unique cultural and climatic needs of Indigenous communities in Canada. The final 3DP design seamlessly integrates traditional Indigenous architectural elements, such as a circular shape inspired by pit houses, with modern construction techniques, yielding a flexible, sustainable, and culturally pertinent home design. Future research work will be focused on how the proposed 3DP design can be adapted to enable mass customization to accommodate the diverse needs and preferences of Indigenous communities across Canada.

A Preliminary Study of 3D Printing Home Designs for Improving Efficiency and Sustainability of Indigenous Housing in Canada

Canada has been experiencing a significant housing crisis in recent years, especially in remote and Indigenous communities, yet most of the existing construction approaches have not been rapid, sustainable, and affordable enough to meet community needs. To address this challenge, this study explores the feasibility of 3D printed (3DP) housing and develops a design that is informed by Indigenous housing requirements and is realized through a physical design prototype tailored for the implementation of 3DP homes. Site visits and community engagement were integral parts of the research to help deliver invaluable insights that guided the design process, ensuring cultural sensitivity and inclusivity. The prototyped 3DP design offers efficient and sustainable solutions customized to the unique cultural and climatic needs of Indigenous communities in Canada. The final 3DP design seamlessly integrates traditional Indigenous architectural elements, such as a circular shape inspired by pit houses, with modern construction techniques, yielding a flexible, sustainable, and culturally pertinent home design. Future research work will be focused on how the proposed 3DP design can be adapted to enable mass customization to accommodate the diverse needs and preferences of Indigenous communities across Canada.

SCC-GPT: Source Code Classification Based on Generative Pre-Trained Transformers

Developers often rely on online resources, such as Stack Overflow (SO), to seek assistance for programming tasks. To facilitate effective search and resource discovery, manual tagging of questions and posts with the appropriate programming language is essential. However, accurate tagging is not consistently achieved, leading to the need for the automated classification of code snippets into the correct programming language as a tag. In this study, we introduce a novel approach to automated classification of code snippets from Stack Overflow (SO) posts into programming languages using generative pre-trained transformers (GPT). Our method, which does not require additional training on labeled data or dependency on pre-existing labels, classifies 224,107 code snippets into 19 programming languages. We employ the text-davinci-003 model of ChatGPT-3.5 and postprocess its responses to accurately identify the programming language. Our empirical evaluation demonstrates that our GPT-based model (SCC-GPT) significantly outperforms existing methods, achieving a median F1-score improvement that ranges from +6% to +31%. These findings underscore the effectiveness of SCC-GPT in enhancing code snippet classification, offering a cost-effective and efficient solution for developers who rely on SO for programming assistance.

SCC-GPT: Source Code Classification Based on Generative Pre-Trained Transformers

Developers often rely on online resources, such as Stack Overflow (SO), to seek assistance for programming tasks. To facilitate effective search and resource discovery, manual tagging of questions and posts with the appropriate programming language is essential. However, accurate tagging is not consistently achieved, leading to the need for the automated classification of code snippets into the correct programming language as a tag. In this study, we introduce a novel approach to automated classification of code snippets from Stack Overflow (SO) posts into programming languages using generative pre-trained transformers (GPT). Our method, which does not require additional training on labeled data or dependency on pre-existing labels, classifies 224,107 code snippets into 19 programming languages. We employ the text-davinci-003 model of ChatGPT-3.5 and postprocess its responses to accurately identify the programming language. Our empirical evaluation demonstrates that our GPT-based model (SCC-GPT) significantly outperforms existing methods, achieving a median F1-score improvement that ranges from +6% to +31%. These findings underscore the effectiveness of SCC-GPT in enhancing code snippet classification, offering a cost-effective and efficient solution for developers who rely on SO for programming assistance.