In this paper, college students in the digital campus environment are taken as the research object, the purpose is to collect and summarize the data sources about digital campus for the calculation of students’ multidimensional behaviors, and to prepare for the study of students’ behavior patterns at each different learning stage after quantitative calculation, and constructing a multi-feature fusion student achievement prediction model by using machine learning algorithm, In this way, some effective learning methods can be provided, which have a certain effect on how to improve students’ academic performance, learning feedback and early warning. In order to obtain the input features required by the prediction model in this paper, the quantitative behavior features are processed first to obtain the selected features before the next step can be carried out, then the regression algorithm, model parameters, model evaluation methods and metrics used in this paper. Ablation experiments are also used to explore the performance of the model and different characteristics of different regression algorithms, especially the necessity of approximate entropy and change complexity measurement in the model. The development of this research has far-reaching influence on students, teachers, university management and research in this field.
With the growing demand for employee competency assessment in enterprises, artificial intelligence technology has gradually become an important tool. Most of the current employee competency assessments use models such as LSTM, which cannot adapt to data scarcity or bias, resulting in poor accuracy of competency assessment. This paper takes advantage of GAN’s adversarial data expansion to optimize the LSTM employee competency assessment model to address the insufficient data volume and uneven data distribution. The study first uses an MLP to build a generator and a DNN to build a discriminator, and jointly designs the GAN structure. Then, GAN is used to generate samples that are similar to, yet distinct from, the original data. The generated samples use adversarial training to approach the original data scatter gradually. Finally, the original data and synthetic data (i.e., GAN-generated data) are sent to the LSTM model for training to capture the time series dependency of employee capability data. The experimental results show that when the LSTM model is used in conjunction with the original data and the synthetic data, the MAE is only 0.12, which is 0.03 lower than the MAE of the original data. The KL divergence of the quality of the GAN synthetic data is only 0.29. The experimental results show that the use of artificial intelligence to generate adversarial networks can significantly optimize the performance of employee ability assessment, reduce assessment errors, and promote the improvement of the company’s competitiveness.
Classification is a core task in pattern recognition, with broad applications in image processing, speech recognition, and medical diagnostics. Traditional neural networks often suffer from difficulties in parameter optimization and local optima, while the performance of approximate logical dendritic neuron models depends heavily on proper parameter configuration. To address these issues, this paper proposes a Feature-Guided Adaptive Differential Evolution algorithm, which incorporates feature-guided mechanisms, adaptive parameter control, and multi-strategy mutation to enhance the standard differential evolution framework. The algorithm is applied to optimize the dendritic neuron model’s parameters, providing a more effective and robust solution for intelligent classification in complex data environments. This approach contributes to the development of reliable optimization strategies in intelligent systems and has promising application potential.
This study investigates the transformative incorporation of human–robot interaction (HRI) within contemporary supply chain management (SCM), focusing on improving efficiency and collaboration. Employing a multifaceted approach, our research examines the intricate dynamics between human workers and robots across various SCM operations, including logistics, inventory management, and storage. Through empirical case studies, qualitative evaluations, and technical insights, we investigate innovative interfaces and communication protocols aimed at fostering intuitive and productive engagement. Our methodology encompasses a thorough analysis of HRI’s impact on resource allocation, operational costs, and error reduction, elucidating both its benefits and challenges. Crucially, we emphasize the imperative of seamless collaboration between humans and robots, facilitated by technology. Our findings underscore the pivotal role of HRI in augmenting the agility and reactivity of supply chain ecosystems, offering strategic insights for addressing dynamic market needs and optimizing operations. Ultimately, this research contributes to a deeper understanding of evolving SCM dynamics and paves the way for a seamless and productive future for companies.
Understanding customer behavior has become critical for marketers and tourist service providers since global travel has rapidly expanded. Recognizing consumer behavior patterns allows organizations to modify their strategy for better service delivery. However, the conventional approaches frequently fail to adequately capture the variety and complexity of tourism consumer behavior because of the large and diverse data available. This research seeks to investigate the use of fuzzy clustering analysis to better understand tourism consumer behavior patterns. The method combines fuzzy clustering algorithms with customer behavior data such as demographics, travel preferences, and purchasing patterns. The investigation reveals separate groups of consumers, providing insights into how various factors influence tourist purchasing decisions. The data were gathered using questionnaires, online booking platforms, and travel websites, where customers provided information about their previous travel experiences and preferences. Data preparation was used to normalize the data for analysis. Principal component analysis was employed to decrease dimensionality. The Sea Turtle Foraging Optimized Fuzzy C-Means clustering (STFO-FCMC) is presented as an extension of normal FCMC that incorporates an optimization procedure based on sea turtle foraging habits. This optimization enhances the accuracy and efficiency of cluster center selection and membership values, making STFO-FCMC especially well-suited for dealing with the complexity and unpredictability of tourism behavior data. The findings show multiple consumer behavior patterns, including diverse preferences for various types of tourist products and services, which are split by age, income, and travel objectives. The STFO-FCMC method is assessed using metrics, including accuracy of 97.84%, precision, recall, and F1-score. These data assist service providers create individualized services and marketing strategies that improve consumer satisfaction and business performance. Overall, fuzzy clustering analysis, particularly with the STFO-FCMC approach, is a successful tool for detecting tourist consumer behavior, with substantial promise for improving tourism product and service targeting.
This article aims at presenting a novel framework for dance action recognition and health enhancement. It uses embedded systems with sensors and machine learning. System makes use of an array of inertial measurement units. Also, accelerometers and gyroscopes collect motion data across multiple dimensions, thereby capturing the dynamics of dance movements. This is to recognize the dance actions effectively. And as a solution to the temporal and spatial data dependencies, the convolutional neural network-long short-term memory-based model is utilized. Feature extraction and optimization is performed via a built-in data preprocessing module. This model gives an added advantage of efficiency and low latency. An health promotion module includes biometric tracking and awareness. It tracks and displays data about the dancer’s heart rate, energy consumption, and joint loading. Several experiments are performed to assess the system performance, using various datasets and comparisons with the traditional approaches. Outcome shows enhancement in the recognition rate, energy consumption, and feedback efficiency. This work benefits the development of intelligent dance systems. It also has implications for the application of embedded systems in sports training, rehabilitation, and health monitoring.
This study explores a pioneering research effort focusing on the use of deep learning techniques to achieve high-precision automatic recognition of aluminum furniture design styles, and proposes an innovative convolutional neural network (CNN) architecture that deeply integrates the migration learning techniques of pre-trained models and the multi-level feature of the feature pyramid network (FPN) integration mechanism. This research is dedicated to solving the challenges of recognizing design elements and styles in the aluminum furniture industry, especially the robustness of recognition under different size variations, complex background interference, and diverse design styles. First, this study fills the gap of deep learning in the field of automatic classification of aluminum furniture design styles, using the powerful image understanding and pattern recognition capabilities of deep learning to effectively break through the bottleneck of the previous traditional methods that have low recognition accuracy when dealing with complex shapes, detail-rich, and diverse styles of aluminum furniture. This is the first time that deep learning technology is systematically applied to such specific scenarios, showing significantly better performance than traditional recognition means. Second, the core contribution of this study is the design of a comprehensive integration scheme that creatively combines a pre-trained CNN model and an FPN structure. This composite deep learning model is able to take full advantage of the generic feature representation acquired by the pre-trained model on large-scale image datasets, while extracting multi-scale local and global features with the advantage of FPNs, which ensures that key design style features can be accurately captured no matter how the size of the design elements of aluminum furniture changes.
As an important part of Chinese culture, red culture has a profound historical heritage and rich spiritual connotations. The red culture experience aims to inspire participants’ patriotic feelings and national pride by recreating historic and revolutionary scenes and sharing stories of the revolution. To further enhance the impact of red cultural experience activities, this paper employs a minimum spanning tree (MST) analysis to investigate the emotional effects of eye and brain movement features on participants in these activities. A multimodal physiological signal emotion recognition model was constructed using EEG and eye movement data, and the accuracy of emotion classification for happy, sad, fearful, and neutral emotions was investigated. The research results show that the emotion recognition effect is optimal when eye movement, differential entropy feature, and MST attribute are integrated simultaneously. The three can achieve effective integration between EEG signals and eye movement signals across modes to obtain more emotion-related information and fully utilise the complementarity between this multimodal information.
In the digital age, artistic creation needs to strike a balance between automation and personalization. This study proposes an innovative model that integrates generative adversarial networks, computer vision, and personalized adjustment technology. Through multistage iterative optimization, efficient art generation and personalized style customization are achieved. The model uses an automated generation module to generate a draft and guides the conditional vector to achieve fine-grained adjustment of the image so that the work maintains both technical innovation and the artist’s unique style. The model performance is optimized in four stages: data preparation, model training, personalized adjustment, and evaluation feedback. The actual art project “Echoes in the Mirror” is used as a case to verify the actual application effect of the model. The evaluation shows that the work receives high scores in clarity, color accuracy, style coherence, and innovation (the average score is close to 9 points). Audience feedback shows that the model performs well in enhancing immersive experience, emotional resonance, and interactive satisfaction, whereas technical acceptance also highlights room for optimization. The research results not only demonstrate the potential of automated and personalized models in artistic creation but also provide practical guidance for the deep integration of art and technology in the future and promote artistic creation to move toward the two-way improvement of innovation and audience experience.
Civil aviation passengers’ comments about airlines or airports on social media are the key to improving service quality. In order to make effective use of these data, in-depth analysis is needed to provide solid support for service improvement of airlines and airports. Due to its uniqueness, accurate modeling and analysis are required. First, the data are accurately collected from various network platforms and reprocessed. In this process, transfer learning, artificial data annotation, and term frequency–inverse document frequency (TF-IDF) analysis technology are innovatively integrated to ensure data quality and analysis depth. Then, according to the characteristics of the review data, the civil aviation domain-specific word vector based on Word2Vec was customized and developed, and the backtranslation – convolutional neural networks – bi-directional long short-term memory (Backtranslation-CNN-BiLSTM) model was constructed for sentiment analysis. The model is verified by multi-dimensional evaluation indicators, which shows excellent performance indicators and ensures reasonable efficiency. Finally, the cutting-edge BERTopic modeling technology was used to deeply mine the passenger comment topics to reveal the focus and potential needs of passengers. This study successfully constructed the technical system of civil aviation passenger comment sentiment analysis, which provided technical support for industry service optimization.
In the framework of an urban garden art and design environment, this research examines the integration of virtual reality (VR) technology with a 3D model built using a fuzzy mathematical model. The goal is to provide a cutting-edge, immersive experience that enables people to engage with and explore the subtleties of a distinctive urban garden setting by fusing creative components with outdoor areas. Our proposed, advanced genetic algorithm-based fuzzy logic mathematical model effectively addressed the challenges in urban gardens by providing innovative design solutions that optimized space, efficiently managed resources, and fostered community engagement, ensuring produce safety, health, and the vitality of the garden. Raw data were collected, and then the data were involved in the pre-processed used min-max normalization. Next feature extracts the data using principal component analysis. The research findings should be presented, together with information on how well VR technology and a fuzzy mathematical model worked to produce the intended aesthetic results. It also describes how the research can be used for VR, urban garden design, and art. The findings demonstrate that VR technology gives designers actual impacts of information processing via complete presentation, intelligent drawing, and timely information, which significantly boosts design productivity and successfully encourages design quality and fitness rate accuracy. Additionally, two benchmark datasets were used for comparison, and our proposed method achieved an accuracy of 98.8%. The research expands the possibilities for artistic expression and strengthens the connection between people and urban green areas.
Campus green space management and optimization are essential when an educational campus is sustainable and conducive. Geographic Information System (GIS) technology has advanced recently, making it a powerful tool for preserving and enhancing green space networks. To guarantee that the demands of urban development inhabitants are met while considering the distribution of natural resources, the investigation has developed a GIS-based evaluation technique for the establishment of green spaces. After over 20 years of campus management in developing China, the rate of development has drastically decreased, and a renewed emphasis on creating sustainable, efficient green campuses has taken hold. Enhancing usability, availability, and appeal is a factor in evaluating whether or not green spaces can achieve their potential benefits. This study ensures the longevity and sustainability of green spaces while minimizing downtime. Furthermore, the impression of a smart campus based on Public Green Space (PGS) and its environment is not much impacted by a student’s gender or growth environment. Which is a planned space used for environmental, recreational, and aesthetic functions. The research underscores the significance of the PGS environment to the general attractiveness and accessibility of green spaces and its contribution to how smart the campus is perceived. The research recommends streamlining the layout and facilities of these spaces to accommodate the requirements of students and maximize their functionality. The demand and satisfaction of the campus PGS overall were also high, with Space-6 having the highest overall satisfaction score of 149 and Space-4 having the highest demand for walking or chatting of 135. Usage periods also had high usage participation, with after-class usage having the highest score of 160 in Space-1. The convenience of the PGS was also highly rated, with Space-1 as the most convenient for students at 88 points from the main entrance. Instead, the plant arrangement, seasonal color richness, and the amenities needed in various locations have a greater impact. Our recommendations are based on GIS and address the needs of students while enhancing the appeal and use of green campuses.
In recent days, stress is a major phenomenon that adversely affects both individuals and communities. The research in computing the stress factor has wider advantages as it improves personal learning, learning operations, and high productivity that benefits society. Several computational techniques come into concern to avoid and reduce the stress level using the electrocardiogram (ECG) signals. In this study, the stress level was classified using the feature extraction approach in combination with the classifier. The signal is processed using the variational mode decomposition denoising technique to reconstruct the original signal. The decomposed signal was further extracted using the time–frequency domain technique as characteristics of the ECG signal such as R-wave and T-wave constructed. Further, the support vector machine classifier was used to classify the stress level (low, medium, and high) of the extracted signal. Based on stress classification outcomes, the robot offers a range of personalized interventions to users. These interventions include relaxation exercises, deep breathing techniques, or guided mindfulness sessions. The average accuracy obtained using the proposed technique is 98.98% but without using the feature extraction technique, it is 97.71%. The other performance parameters also get improved and the results are finally compared with the existing techniques.
The use of interactive tools, such as voice assistants and social robots, holds promise as coaching aids during public speaking rehearsals. To create a coach that is both effective and likable, it is important to understand how people perceive these agents when they observe them during actual presentation sessions. Specifically, it is important to assess people’s perceptions of the agents’ physical embodiment and nonverbal social behaviour, taking into account both listening and feedback periods. To this end, we conducted an online study with 168 participants who watched videos of agents acting as public speaking coaches. The study had three conditions: two with a humanoid social robot in either (1) active listening mode, using nonverbal backchannelling, (2) passive listening mode, and (3) a voice assistant agent. The results showed that the social robot in both conditions was perceived more positively in terms of its human-like attributes, and likability than the voice assistant agent. The active listener robot was perceived as more satisfying, more engaging, more natural, and warmer than the voice assistant agent, but this difference was not seen between the passive listener robot and the voice assistant agent. Additionally, the active listener robot was found to be more natural than the passive listening robot. However, there were no significant differences in perceived intelligence, competence, discomfort, and helpfulness between the three agents. Finally, participants’ gender and personality traits were found to affect their evaluations of the agents. The study offered insights into general attitudes towards using social robots and voice assistants as public speaking coaches, which can guide the future design and use of these agents as coaches.
In this work, a method that integrates deep learning and genetic algorithms is proposed to enhance the precision and efficiency of welding robots and achieve optimal robot path planning. The process involves using SolidWorks to create a 3D model, applying the D-H method to obtain data on the connecting rod parameters, performing theoretical calculations for both forward and inverse kinematics solutions, and utilizing the MATLAB robotics toolbox to validate these solutions. Furthermore, joint space trajectory planning is performed using the quintic polynomial curve method. Through analysis, we identified that abrupt acceleration changes at the initial and final positions significantly impact the smoothness of the motion process. The findings reveal that traditional artificial bee colonies tend to stabilize after 190 iterations, whereas genetic algorithms stabilize around 160 iterations, demonstrating superior convergence speed compared to the traditional ABC algorithm. The optimized approach yields an optimal welding obstacle avoidance path with rapid optimization speed and a stable process. The proposed method effectively addresses the obstacle avoidance path planning challenge for welding robots, showcasing improved convergence speed and stability compared to traditional methods.
This study investigates automatic sorting and handling robots to enhance intelligence and automation in logistics distribution, improve work efficiency, and reduce logistics costs. The kinematics and dynamics models of the sorting and handling robot are established by the study using single-chip microcomputer control technology. For automated item sorting, radio frequency identification (RFID) card scanning technology is combined and the PID control algorithm is examined. A single-chip microcomputer control technology is used to establish the robot’s kinematics and dynamics models and analyze the PID control algorithm. Employing RFID card reading technology facilitates the transportation and automatic sorting of items. The MATLAB software simulates PID parameters and assesses the stability of PID-controlled motors. The primary controller samples and provides feedback from the motor encoder every 10 ms. Setting the left and right motor encoder’s given values to 40 with a corresponding speed of 40, we observe that Option 1’s P parameter is too small, leading to a slow adjustment speed. In contrast, Option 1’s parameters are extensive. Scheme 3’s P parameter is too large, risking system oscillation. After carefully adjusting and selecting Scheme 2, experimental verification demonstrates the sorting and handling robot’s stable operation, achieving the anticipated effect.
This study addresses the challenge of low recognition rates in emotion recognition systems, attributed to the vulnerability of sound data to ambient noise. To overcome this limitation, we propose a novel approach that leverages emotional information from diverse modalities. Our method integrates speech and facial expressions through advanced feature layer fusion and decision layer fusion strategies. Unlike traditional fusion algorithms, our proposed multimodal emotion recognition algorithm incorporates a dual fusion process at both the feature layer and the decision layer. This dual fusion not only preserves the distinctive characteristics of emotional information across modalities but also maintains inter-modal correlations. To evaluate the effectiveness of our approach, experiments were conducted using the eNTERFACE’05 multimodal emotion database. The results demonstrate a remarkable recognition accuracy of 89.3%, surpassing the highest recognition rate of 83.92% achieved by the current state-of-the-art kernel space feature fusion method. Our algorithm exhibits a significant improvement of 5.38% in recognition accuracy. By combining emotional data from speech and facial expressions using a data fusion methodology, our study demonstrates a significant improvement of 5.38% in recognition accuracy, contributing to the progress of multimodal emotion recognition systems.
In order to better realize the optimal trajectory planning and trajectory control in industrial robots, a method based on ADAM algorithm is proposed. Taking PUMA 560 industrial robot as the research object, using nonlinear data sets and mathematical ADAM algorithm function planning, an optimal calculation method for time trajectory planning of industrial robot is explored. Finally, the programming, optimization, and simulation of the program code are implemented using MATLAB, and a standardized optimal trajectory planning is established. The experimental results show that the running time difference of the trajectory corresponding to the three joint points is small. In order to synchronize the position of each joint point in time, it is necessary to choose the optimal joint point position according to the time trajectory, so as to ensure the synchronization between each key node. Therefore, the joint node position is adjusted so that the total time and the final simulation results are basically synchronized in time, and both are 10.35 s. It proves that the improved ADAM algorithm realizes the trajectory optimization of industrial robots in terms of time planning, which can make the various joints of industrial robots basically synchronized in the time trajectory.
The digitization of important documents and their segregation can be a beneficial and time-saving activity as individuals will have greater access to important documents and will be able to use them in regular tasks as well as endeavours. In recent years, research into the application of deep networks in robot systems has increased as a direct consequence of the advancements made in classification algorithms over the past few decades. Robotic vision automation for the segregation of sensitive and non-sensitive documents is required for many security concerns. The methodology of this article is initially focused on the identification of a good computer vision-based technique for the classification of sensitive documents from non-sensitive documents. The authors first identified the standard parameters in terms of reliability, loss, precision, and recall by employing deep learning techniques, such as neural networks with convolutions and transfer learning (TL) algorithms. The extraction of features based on pre-trained deep learning models was referenced in numerous publications. Similarly, we applied most of the feature extraction techniques to identify feature extraction from the images. Then, these features were classified by machine and ensemble learning models. However, the pre-trained models-based feature extraction along with machine learning classification resulted better in comparison to the deep learning and TL procedures. Further, the better-identified techniques were applied as the brain behind the vision of a robotic structure to automate the segregation of sensitive documents from non-sensitive documents. This proposed robotic structure could be applied when we have to find some specific and classified document from the haystack.
