Digital Library

Vol. 21, No. 4, Aug. 2025  Cover

1. Applying Machine Learning to Explore the Impact Mechanisms of the Urban Built Environment’s Characteristics on the Thermal Environment
   Yansu Qi, Han Li, Xiuhe Yuan, Dongmiao Zhao, Chao Liu
   Vol. 21, No. 4, pp. 355-370, Aug. 2025  
   https://doi.org/10.3745/JIPS.04.0354
   Details Full Text PubReader
   Keywords: Deep Learning, Remote sensing, Shapley additive explanations, Urban Thermal Environment
   Show / Hide Abstract

   Recent studies on the thermal environment in cities have concentrated on the macro-scale level, with limited attention given to specific built-up areas. Additionally, there is scarce research on the comprehensive effects of the physical features of urban spaces on their thermal environment, especially on the regional scale. We explored how urban built environment factors influence the thermal environment using Qingdao as a case study. A neural network model with a feed-forward architecture was employed to map the complex interactions between land surface temperatures and built environment characteristics. The model’s performance was compared with traditional methods. Various built environment factors were analyzed, considering both spatial and morphological features, with data sourced from multiple channels and pre-processed for quality. The Shapley Additive exPlanations method was applied to interpret the impact mechanism, quantifying the contribution of each factor. The results indicate that an impervious area percentage significantly increased the land surface temperature in the summer, while vegetation coverage and building density helped maintain surface temperatures during the winter. This study presents quantitative insights into the importance, direction, and critical impacts of variable urban factors on the thermal environment. The findings provide useful directions for urban design and rehabilitation, especially in reducing the environmental impacts of urban heat islands.


2. K-Equidistant Partitioning: Enhancing Sensor Data Analysis Using Innovative Data Augmentation Techniques
   JeongHyeon Park, Nammee Moon
   Vol. 21, No. 4, pp. 371-379, Aug. 2025  
   https://doi.org/10.3745/JIPS.02.0225
   Details Full Text PubReader
   Keywords: Behavior Classification, CNN, data augmentation, Sensor data
   Show / Hide Abstract

   In this study, K-equidistant partitioning (K-EP), a novel data augmentation method, is proposed to address the limitations of sensor data analysis and enhance the performance of behavior classification models. K-EP involves dividing rows of sensor data at equidistant intervals and extracting information from each segment, thereby increasing the size of the dataset by a factor of K. This method is based on the sensor minimum warranted frequency hypothesis, which posits that a sampling frequency of 20–40 Hz provides sufficient data for behavioral classification. The effectiveness of K-EP is validated through three experiments, which involve determining the optimal value of K for K-EP, comparing K-EP with other data augmentation methods, and assessing the added value of K-EP when combined with other methods. The results indicate that K-EP effectively overcomes the quantitative limitations of sensor data and enhances model robustness. It achieves higher F1-scores than existing methods, indicating that it is an effective data augmentation method for sensor-based behavior classification models.


3. Research on Influencing Factors of Employment Anxiety of College Students through Model Analysis
   Hehe Liu, Zhoutao Zhang
   Vol. 21, No. 4, pp. 380-391, Aug. 2025  
   https://doi.org/10.3745/JIPS.04.0355
   Details Full Text PubReader
   Keywords: College Student, Employment Anxiety, influencing factor
   Show / Hide Abstract

   To investigate factors influencing college students’ employment anxiety and develop effective intervention strategies, this study employed a questionnaire survey to collect demographics and employment anxiety data. The collected data were analyzed using statistical and regression analyses in SPSS software version 26.0. It was found that the coefficient of the correlation was the largest between subjective support and employment anxiety, reaching -0.387. Moreover, participants’ employment anxiety levels were significantly influenced by social support and general self-efficacy. The results confirm that social support and self-efficacy are key factors affecting college students’ employment anxiety, providing guidance for alleviating employment anxiety.


4. Secure Storage of Network Data Information Based on Cloud Computing Data Encryption Techniques
   Qiaoling Long
   Vol. 21, No. 4, pp. 392-400, Aug. 2025  
   https://doi.org/10.3745/JIPS.03.0207
   Details Full Text PubReader
   Keywords: Cloud computing, Data encryption, Data storage, Dual Encryption
   Show / Hide Abstract

   In a cloud computing environment, frequent data interactions require high information security. This paper briefly overviews encrypted data storage in the cloud computing environment. A dual encryption scheme that combined identity attributes and security devices was proposed. The scheme was evaluated through simulation experiments conducted on a laboratory server. The results indicated that an increase in the size of the storage file led to an increase in the time overhead of the encryption scheme. Furthermore, the dual encryption scheme exhibited slightly higher time overhead than a single security device encryption scheme, but was similar to a single identity attribute-based encryption scheme. For cloud storage data, the successful queries of encrypted data and effective resistance against third-party attacks were only achieved when the number of overlapping identity attributes was 5, which was the same as the storage user.


5. Code Smells and Development Efficiency in Native and Cross-Platform Mobile Applications: A Study of Java, Kotlin, and Flutter
   Ibrokhimov Sardorbek Rustam Ugli, Junseok Cheon, Gyun Woo
   Vol. 21, No. 4, pp. 401-412, Aug. 2025  
   https://doi.org/10.3745/JIPS.01.0113
   Details Full Text PubReader
   Keywords: Code Quality, Code Smell, Dart, Java, Kotlin, LOC
   Show / Hide Abstract

   While many methods have been proposed for creating mobile applications, developers have struggled to decide which is best. This study contrasts native and cross-platform application development methodologies, paying special attention to the growing popularity of Flutter and the trend away from Java in favor of Kotlin. Using Java, Kotlin, and Dart (Flutter) to create identical applications, this research provides useful insights into factors influencing developers’ choice of programming languages and frameworks in mobile application development. In addition, this research investigates development best practices by analyzing the quality of the code in 45 public GitHub repositories. The study measures the impact of choosing a particular language or framework on code smells and development efficiency by evaluating LOC and code smells using semi-automated SonarQube assessments, which include the measurement of severity levels. Preliminary findings show differences in the code quality produced by the two approaches, offering developers useful information on reducing code smells and improving project quality.


6. An Energy-Saving Hybrid-Optimization Scheduling Algorithm Based on Fat Tree in WSN
   Cheng Li, Guoyin Zhang, Honglie Zhang
   Vol. 21, No. 4, pp. 413-426, Aug. 2025  
   https://doi.org/10.3745/JIPS.03.0206
   Details Full Text PubReader
   Keywords: energy consumption, Fat Tree, Hybrid-Optimization Scheduling Algorithm, Lifecycle
   Show / Hide Abstract

   In the data scheduling process of wireless sensor network (WSN), node data fusion is one of the main methods to reduce network communication volume. The decrease in network communication volume contributes to the network energy consumption reduction, which is crucial for improving the WSN lifecycle. Hence, we propose an energy-saving hybrid-optimization scheduling algorithm based on fat tree (FT) for WSN, which is referred to as the FTEBHSA algorithm. In the scheduling algorithm, multiple strategies are adopted to optimize for saving energy, involving shortest path tree optimization, fusion tree load balance, a fusion node rotation mechanism, and a sleep mechanism for monitoring nodes. More importantly, we introduce the FT structure to organically integrate these strategies for reducing the energy consumption and boosting the WSN lifecycle. The simulation experiment results verify that the proposed hybrid-optimization scheduling algorithm performs optimally in optimizing the energy consumption.


7. Korean Sign Language Recognition Using LSTM and Video Datasets
   Soo-Yeon Jeong, Ho-Yeon Jeong, Sun-Young Ihm
   Vol. 21, No. 4, pp. 427-438, Aug. 2025  
   https://doi.org/10.3745/JIPS.04.0356
   Details Full Text PubReader
   Keywords: Korean Sign Language Recognition, Long short-term memory (LSTM), Sentence Conversion, video recognition
   Show / Hide Abstract

   Deaf individuals primarily use sign language, which consists of hand gestures and body movements, as their main means of communication. It is difficult for non-disabled people to understand the visual form of sign language, and sign language recognition technology is required to facilitate communication. However, unlike spoken languages used by the general population, sign languages take a visual form and can be recognized through video or image data before being translated into other languages. In this study, we proposed a Korean sign language recognition and sentence conversion system based on long short-term memory (LSTM) using video datasets. To build a Korean sign language dataset, we automatically collected and preprocessed video data of sign language gestures, which were then used as input for the LSTM model. LSTM has strengths in processing sequential data and can effectively recognize the sequences and patterns of sign language gestures. The experimental results measured the accuracy of the model and analyzed its performance based on sign language gesture recognition and display. This study confirmed the effectiveness of the proposed approach and is expected to contribute to the advancement of Korean sign language recognition technology.


8. Construction of Formative Evaluation System in Business English Listening and Speaking Course Using Fuzzy Hierarchy Analysis Process
   Huifang Shan, Junrui Han, Zhanfang Li
   Vol. 21, No. 4, pp. 439-448, Aug. 2025  
   https://doi.org/10.3745/JIPS.04.0357
   Details Full Text PubReader
   Keywords: Business English Listening and Speaking Course, Formative Evaluation, FAHP, system model
   Show / Hide Abstract

   The classroom learning process is multi-dimensional and comprehensive, and formative evaluation is the key way to measure its effectiveness. When conducting the evaluation, teachers need to comprehensively examine the diverse performance of students and carefully establish evaluation criteria to ensure an objective and fair assessment of the learning behavior. This not only aids in testing the teaching effect but also provides robust support for the continuous improvement of classroom instruction. This study applies the fuzzy analytic hierarchy process to thoroughly analyze each aspect of the Business English Listening and Speaking Course, establish the logical relationship between formative elements, and subsequently construct an evaluation matrix. Simultaneously, the paper employs fuzzy theory to quantify the evaluation process, making it more precise and operational, and thereby produces a model for the formative evaluation system. Through comprehensive analysis and interpretation of the evaluation results, a more scientific and equitable evaluation of students’ learning processes is achieved.


9. Adversarial Watermarking through the Integration of GAN and FGSM for Preventing Unauthorized AI Training
   Ji-Hun Kim, YongTae Shin
   Vol. 21, No. 4, pp. 449-456, Aug. 2025  
   https://doi.org/10.3745/JIPS.03.0208
   Details Full Text PubReader
   Keywords: Adversarial Watermarking, Digital Content Protection, FGSM, GAN, Probability Shift, PSNR
   Show / Hide Abstract

   This paper proposes a novel adversarial watermarking method combining generative adversarial networks (GANs) and fast gradient sign method (FGSM) to prevent unauthorized artificial intelligence (AI) training while maintaining high visual quality of the watermarked content. GANs are used to generate imperceptible adversarial watermarks that are embedded into the original content, minimizing visual distortions. FGSM enhances these watermarks by introducing targeted perturbations to confuse AI models, significantly degrading their learning performance. Experiments conducted using ResNet-18 demonstrate the effectiveness of the proposed method across key metrics, including peak signal-to-noise ratio, probability shift, and MAX probability shift. The results show that the combined GAN and FGSM approach strikes a balance between maintaining the visual quality of watermarked content and achieving superior adversarial robustness compared to standalone GAN or FGSM methods. This study provides a practical reference for advancing adversarial watermarking techniques, contributing to the protection of intellectual property in the era of AI-driven content creation.