Abstract The paper titled “Research on Gaussian Mixture Computational Learning Mode Based on MOOC Online Education” explores the integration of Gaussian Mixture Models (GMMs) and the Expectation Maximization (EM) algorithm into a blended learning framework centered on MOOCs... The paper titled “Research on Gaussian Mixture Computational Learning Mode Based on MOOC Online Education” explores the integration of Gaussian Mixture Models (GMMs) and the Expectation Maximization (EM) algorithm into a blended learning framework centered on MOOCs (Massive Open Online Courses). The authors propose using GMMs to model complex learning behaviors and environmental variations, particularly in video based educational content, by distinguishing background (typical) from foreground (atypical) patterns. The EM algorithm is employed to estimate model parameters via iterative unsupervised learning, thereby improving convergence and adaptability compared to conventional methods such as K-means. The study also emphasizes a student centered blended learning approach, combining micro courses, MOOCs, VR/AR technologies, and social media platforms to enhance engagement and comprehension. An experiment involving a “Business Etiquette” course demonstrates that blended MOOC based learning increases student satisfaction, motivation, and outcomes, despite minor challenges like internet access or digital literacy. Simulations in MATLAB using synthetic Gaussian data validate the efficacy of the proposed computational model, demonstrating that adaptive learning rates and prior probability estimation significantly improve algorithm performance. The paper concludes that the GMM-EM framework offers a flexible, scalable solution for modeling educational data and optimizing online learning environments, advocating for careful parameter tuning to avoid ambiguity. Overall, the research bridges computational statistics and modern pedagogy to advance personalized, data driven online education. Read More Read Less

Abstract The paper presents an innovative design system for traditional embroidery patterns using a linear classifiertype intelligent algorithm model in computer science. Recognizing the limitations of conventional, artisandependent embroidery design such as constrained creativity and pattern diversity the... The paper presents an innovative design system for traditional embroidery patterns using a linear classifiertype intelligent algorithm model in computer science. Recognizing the limitations of conventional, artisandependent embroidery design such as constrained creativity and pattern diversity the study proposes an AIdriven approach to automate and enhance pattern innovation while preserving cultural essence. The system involves preprocessing embroidery images, extracting geometric, textural, and structural features, and applying linear classifiers like Support Vector Machines and Logistic Regression to categorize and generate new designs. A user friendly visual interface allows customization, fostering personalized creativity. Experimental validation includes dataset partitioning, algorithm comparison, and performance evaluation using metrics like accuracy, recall, and F1 score. Results indicate that the linear classifier effectively identifies pattern features and supports the generation of innovative designs. The research fills a gap in intelligent systems explicitly tailored for embroidery, as most prior work focused on digital art or painting. By integrating machine learning with traditional craftsmanship, the system not only modernizes embroidery design but also contributes to the preservation and evolution of cultural heritage. The authors suggest exploring alternative algorithms to refine performance further and broaden the application scope. Overall, the study demonstrates the feasibility and practical value of intelligent algorithms in revitalizing traditional art forms through computational innovation.The Innovation Design System for Traditional Embroidery Patterns Based on the Linear Classifier type Intelligent Algorithm Model in Computer Read More Read Less

Abstract The paper titled "Table Tennis Teaching 3D Model Based on Hierarchical Clustering Intelligent Algorithm" by Wei Zeng proposes an innovative approach to enhance university level table tennis instruction by leveraging a hierarchical clustering intelligent algorithm. Recognizing the... The paper titled "Table Tennis Teaching 3D Model Based on Hierarchical Clustering Intelligent Algorithm" by Wei Zeng proposes an innovative approach to enhance university level table tennis instruction by leveraging a hierarchical clustering intelligent algorithm. Recognizing the limitations of traditional teaching methods particularly their inability to accommodate student's diverse skill levels the study introduces a data driven, layered teaching model. The algorithm preprocesses player movement data using backgrounddifference and shadow removal techniques, then extracts features via Radon transformation and wavelet analysis. A multi feature fusion strategy enables precise posture recognition. The core of the method involves a two layer hierarchical clustering process that separates data points by density, removes outliers (~10% of low density points), and organizes the remaining data into cohesive clusters, improving classification accuracy. Experimental results from a PDCA cycle based teaching intervention show that this approach outperforms traditional algorithms like k-means++, CURE, and CBDP especially in handling complex, nonspherical cluster shapes. The study concludes that hierarchical clustering not only enhances teaching personalization and student engagement but also offers a robust framework for skill assessment and instructional design in physical education, particularly for sports like table tennis that require fine motor control and rapid decision making. This intelligent model supports adaptive, student centered pedagogy and paves the way for technology integrated sports education reforms Read More Read Less