Digital Library

Vol. 21, No. 1, Mar. 2025  Cover

1. RAE-SVM based Webpage Tamper-Resistant Detection Algorithm
   Changjian Zhou, Yutong Zhang, Yunfu Liang, Jinge Xing
   Vol. 21, No. 1, pp. 1-12, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0331
   Details Full Text PubReader
   Keywords: Cyber Security, Deep Autoencoder, image features, Support Vector Machine, Tamper-Resistant Detection
   Show / Hide Abstract

   Cybersecurity has become a key component of national strategy in recent years. Traditional cybersecurity technology such as network traffic-based intrusion detection and threatening intelligence sensing are designed to focus on the traffic features of network, which are no doubt effective defense technologies. However, these methods required decent amount of domain knowledge and massive training data, which brought a significant barrier for cybersecurity research. In this work, we propose a novel residual autoencoder and support vector machine combined approach (RAE-SVM) for webpage tamper-resistant detection using high-level webpage image features. This method, inspired by the Chinese proverb “mend the fold after the sheep have been stolen.” The web crawler technology is used for website screenshot within limited domain names, and input them into autoencoder architecture and SVM for feature extraction and invaded webpage detection. This method combines the advantages of deep residual network, convolutional autoencoder and SVM, and the interdisciplinary intersection between cybersecurity and high-level image features. The experimental results demonstrate that the proposed method achieves an accuracy of 95%, significantly higher than other models, which proves the validity of the proposed method.


2. DQA4AirQuality: TDQM-Based Data Quality Assessment Framework for Air Quality Datasets
   Lina Zhang, Sukhoon Lee
   Vol. 21, No. 1, pp. 13-27, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0332
   Details Full Text PubReader
   Keywords: Air Quality, Big data, Data Quality Assessment, Open Data
   Show / Hide Abstract

   In recent years, open government data and big data analytic applications have become increasingly widespread. Without proper quality control, the rapid dissemination of data may jeopardize the reuse of datasets and exert negative effects. The current general frameworks for data quality management in literature are outdated and lack extensions to big and open data. In this work, a four-level data quality assessment dimension generation model was developed and applied for air quality datasets to measure the quality of air quality data from various data quality dimensions. This assessment framework was validated by comparing it with four air quality datasets from the World Health Organization (WHO), Beijing, Seoul, and Italy. The results show that the datasets published by the WHO have low quality due to their more complex sources.


3. Influence Index Evolution and Empirical Study of Opinion Leaders in Chinese Social Networks
   Cheng Zhang
   Vol. 21, No. 1, pp. 28-42, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0333
   Details Full Text PubReader
   Keywords: Chinese Opinion Leader, Drift Index, Fuzzy Concept Analysis, Index of Influence, Stability Index
   Show / Hide Abstract

   An influence drift index (IDI) and an influence stability index (ISI) of opinion leader nodes are proposed to capture the dynamic growth trajectory of opinion leaders in the process of public opinion evolution. The multidimensional features of user nodes are extracted to construct the influence index, and the change amplitude of the influence index in the time dimension is quantified, and this forms the basis for an influence index online evolution (INFOE) method. In this method, opinion leaders’ influence is regarded as the feature vector of dynamic change. The IDI for the opinion leader-audience layer is established from the three dimensions of contribution degree, recognition degree and dissemination degree of opinion leaders. Second, fuzzy formal concept analysis is used to establish the formal concept of opinion leaders and its poset to construct an ISI for the opinion leader-audience layer. Finally, the fluctuation degree of the influence index of opinion leaders in the evolution of public opinion content is quantified to realize the online evolution of this index. INFOE is verified and analyzed in this case. The empirical results show that the IDI reveals a certain correlation between the influence index of opinion leaders and the heat of public opinion in the process of event evolution, which can be used to monitor the time node of public opinion outbreaks accurately. The ISI reflects the fuzzy dependence of the influence index among opinion leader nodes, which provides a new theoretical exploration and research perspective for the multi-angle subdivision of opinion leaders.


4. Optimizing Traffic Signal Control Using LLM-Driven Reward Weight Adjustment in Reinforcement Learning
   Sujeong Choi, Yujin Lim
   Vol. 21, No. 1, pp. 43-51, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0334
   Details Full Text PubReader
   Keywords: Large Language Model, Reinforcement Learning, Traffic signal control
   Show / Hide Abstract

   With advancements in information technology, traffic signal control has become a crucial component of smart transportation systems, and research based on reinforcement learning (RL) for this purpose is being actively conducted. However, tuning the weights of a multi-objective reward function remains a challenging task. This paper proposes an algorithm that leverages a large language model (LLM) to dynamically adjust the weights of the RL reward function in real time, enabling efficient traffic signal control at intersections. We compare the performance of dynamic weight adjustment via LLM and evaluate the signal control efficiency of the proposed model under various weather conditions.


5. An Efficient Classification Algorithm on Orchid Species Based on Data Enhancement
   Jianhua Wang, Haozhan Wang
   Vol. 21, No. 1, pp. 52-62, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0335
   Details Full Text PubReader
   Keywords: Orchid Types, Classification Algorithm, Data Enhancement
   Show / Hide Abstract

   In order to address the issue of low accuracy rate of current orchid type classification methods due to their similarities in the characteristics of orchid types, an effective orchid type classification method using data enhancement is suggested in this work, whose contribution depends on the utilization of data enhancement technologies, which can efficiently enhance the orchid type classification accuracy rate by providing sufficient and balanced sample sets. Specifically, in our approach, firstly, an image set of 12 orchid types containing 12,227 images is established; secondly, the characteristics of the above orchid image dataset are analyzed and studied; thirdly, the reasons for the processing difficulties are identified based on the above orchid image set at last, some data enhancement technologies are applied to improve the classification accuracy rate of orchid types, which can also enhance the whole performance of orchid type classification. The experimental results display that our suggested classification method using data enhancement in the article can achieve a classification accuracy of 92.65% compared with the one not using data enhancement under the condition of insufficient and unbalanced image datasets.


6. Edge Server Activation Control Method with Early Congestion Avoidance by Using Deep Q Network
   Minseok Koo, Jaesung Park
   Vol. 21, No. 1, pp. 63-70, Mar. 2025  
   https://doi.org/10.3745/JIPS.03.0203
   Details Full Text PubReader
   Keywords: DQN, Edge network, Edge Server Activation, Energy Efficiency, Service Delay
   Show / Hide Abstract

   Edge networks have emerged as a solution to provide high-speed and localized data processing by utilizing distributed computing models with numerous edge servers (ESs). However, the increasing deployment of ESs has significantly escalated energy consumption, raising critical concerns regarding operational costs, environmental sustainability, and economic efficiency. A dynamic network topology management approach has been proposed to address this issue by adaptively switching the operation mode of each ES according to load condition. Even though various methods have been proposed to address this issue, there is still room for performance improvement. To fill the performance gap, in this paper, we take a deep Q network-based approach and devise a novel ES activation control method that dynamically manages ES states, balancing energy efficiency and service quality through a specially designed reward function. Simulation results demonstrate the effectiveness of the proposed approach in reducing the amount of consumed energy while increasing the service quality across diverse scenarios, compared to existing methods.


7. Intelligent Detection Algorithm for Fall from Height Based on Support Vector Machines
   Chunchao Chen, Zhiwei Zhang, Hui Cui, Jun Li, Jiajun Zhou, Zhengyong Fan
   Vol. 21, No. 1, pp. 71-79, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0336
   Details Full Text PubReader
   Keywords: acceleration sensor, Activity recognition, Fall Detection, Support Vector Machine
   Show / Hide Abstract

   Among the “five major injuries” in the construction industry—namely, falls from height, electric shock, object strike, mechanical injury, and collapse—fall accidents from height exhibit the highest incidence rate and pose significant dangers. To mitigate the frequency of fall accidents, a fall detection algorithm was developed based on three-axis acceleration sensors. This intelligent algorithm analyzes the acceleration data of the human body during different motion states. It extracts eigenvalues from the acceleration data, processes and analyzes them using the support vector machine method, and performs data classification to determine if a person has undergone a fall. Through experimental testing and validation, this method has demonstrated a high level of reliability in the detection of falling behavior. The accuracy of this algorithm surpasses that of traditional threshold detection methods and decision tree-based algorithms. This enhancement improves its potential for application in the detection of falls from heights in construction settings.


8. A Study on the Evaluation Method of Agility for Female College Basketball Players Based on the AHP-VIKOR Model
   Zhenyu Guan, Zhe Kan
   Vol. 21, No. 1, pp. 80-88, Mar. 2025  
   https://doi.org/10.3745/JIPS.04.0337
   Details Full Text PubReader
   Keywords: Agility, evaluation, model, Type-2 Fuzzy, Women's Basketball
   Show / Hide Abstract

   In promoting basketball sports among female college students, this paper introduces fuzzy sets of interval type-2 into the agility assessment model for our university’s female basketball players. During the assessment, the expert language variables are converted to the second class of the interval. It combines the fuzzy analytic hierarchy process (AHP) and the entropy weight approach for calculating weights, integrating the VIKOR method to construct a fuzzy interval type-2 AHP-entropy weight-VIKOR risk assessment model. Based on the fuzzy logic of type I, this model explores the influence relationships among factors, effectively reduces information loss through the collaboration of multiple expert judgments, and enhances evaluation accuracy. Lastly, this model is used to evaluate the agility of women’s basketball players. It has strong operability and feasibility compared to actual results and provides an effective evaluation basis.


9. Enhancing Fairness in Financial AI Models through Constraint-Based Bias Mitigation
   Yiseul Choi, Jiwon Hong, Eunbeen Lee, Junga Kim, Seongmin Kim
   Vol. 21, No. 1, pp. 89-101, Mar. 2025  
   https://doi.org/10.3745/JIPS.01.0111
   Details Full Text PubReader
   Keywords: AI Fairness, Bias Mitigation, Data Preprocessing, Fairness Metrics, Financial data
   Show / Hide Abstract

   As artificial intelligence (AI) increasingly drives decision-making in the financial sector, ensuring fairness in machine-learning models has become critical. Bias in AI models can lead to discriminatory practices, undermining public trust and restricting access to essential financial services. While existing financial services leverage AI to enhance efficiency and accuracy, these systems can inadvertently produce unfair outcomes for specific groups defined by sensitive attributes, such as gender and race. This study addresses the challenge of mitigating bias in loan-approval models by applying fairness-aware machine-learning techniques. We investigate two distinct constraint-based strategies for bias mitigation: fairness- and accuracy-constrained models. These strategies are evaluated using logistic regression (LR) and a large-scale, contemporary financial dataset from the Korea Credit Information Services. The results demonstrate that fairness-constrained models achieve a superior balance between fairness and accuracy compared to a conventional LR model. Furthermore, we highlight the importance of tailored data preprocessing and carefully selecting relevant sensitive attributes (e.g., gender, age, nationality) in enhancing fairness outcomes. The findings underscore the necessity of integrating fairness considerations into every stage of the AI model development lifecycle within finance, ensuring equitable outcomes without compromising predictive performance.


10. Spatial and Temporal Evolution of Wetlands in the Yellow River Delta Based on Optical Flow Algorithm
    Chao Liu, Ruolan Mu, Chuanlong Wang, Xiuhe Yuan, Sheng Miao
    Vol. 21, No. 1, pp. 102-113, Mar. 2025  
    https://doi.org/10.3745/JIPS.04.0338
    Details Full Text PubReader
    Keywords: Identification of Wetland Landscape Types, Optical Flow Algorithm, ResNet, Spatial and Temporal Evolution
    Show / Hide Abstract

    The Yellow River Delta (YRD) wetlands are the largest coastal wetlands in China, and it serves to control soil erosion, nourish the climate and protect biodiversity. At present, due to climate change and human activities, the wetlands of the YRD are facing ecological and environmental problems such as species invasion, vegetation degradation, and biodiversity reduction. In order to study the evolution of wetland landscape types more intuitively, this paper proposed a semantic segmentation network based on the encoder and decoder structure of ResNet-18. Then, the wetland landscapes in the YRD were classified into five types by combining the Landsat series of remote sensing images. In addition, this paper used the optical flow algorithm to visualize the identification results, which can represent the evolution pattern of wetland landscape types in different years. The results of this paper have an important significance for the subsequent development planning and protection in the YRD.