Locally Weighted Learning

This paper surveys locally weighted learning, a form of lazy learning and memory-based learning, and focuses on locally weighted linear regression. The survey discusses distance functions, smoothing parameters, weighting functions, local model structures, regularization of the estimates and bias, assessing predictions, handling noisy data and outliers, improving the quality of predictions by tuning fit parameters, interference between old and new data, implementing locally weighted learning efficiently, and applications of locally weighted learning. A companion paper surveys how locally weighted learning can be used in robot learning and control.     

一种改进的放大图像边缘修复算法

为了保持放大后图像清晰且光滑的边缘,提出了一个基于边缘轮廓拟合插值和分水岭变换的边缘修复算法。算法的主要思想是对原始图像进行初始放大,自动确定模糊区域,从而进行修复处理。通过分段三次Hermite插值提取初始放大图像的边缘。对边缘进行膨胀进而自动获取标记图像。对初始放大图像进行距离变换,从而提供了分水岭变换所需的地势图。有了标记图像和地势图就可以利用分水岭变换逐步填充模糊区域像素。实验结果表明：与已有算法相比,该算法降低了对图像梯度的敏感程度,图像边缘更为光滑清晰,并且在运行效率上大大提高。     

远程培训中个性化学习资源推荐算法

本文提出一种基于标签的多因素推荐算法.用户可以根据自己的需求,进行因素自定义和优先级排序,算法先根据用户初始化信息选取资源,随后分析用户行为数据更新用户所属的群及用户的喜好,再通过用户与项目相似度计算、项目关联度计算为用户推荐所需资源.算法模型采用分类组合得出结果,降低了相似度计算的复杂度.将算法应用于企业远程培训平台的个性化学习模式中,结果表明,该算法较好地改善了用户个性化学习资源的推荐效果.     

改进的多分辨率形状匹配算法研究

形状匹配算法虽然在基于内容的视频检索中应用广泛，但由于视频数据量非常大，匹配非常耗时，因此形状匹配算法通常会成为实时视频检索的瓶颈，为了快速准确地进行形状匹配和检索，提出了一种改进的多分辨率Hausdorff距离变换算法，该算法是通过对后向匹配算法进行优化来使匹配速度大大加速，可用干进行实时车型比较和识别。实验结果表明，该改进算法在车型识别上具有速度快和准确性高的优点，尤其在模板图像比较大的情况下，此改进算法优势明显。     

信息系统中的知识距离与知识粗糙熵

本文基于知识即划分的观点,证明了论域中所有知识构成一个距离空间,研究了知识距离空间中知识距离的一些性质,并应用知识距离度量信息系统中知识的粗糙熵。这些结果有利于更加深刻地分析信息系统中知识间的关系和知识的粗糙性。     

URG-04LX2维激光扫描测距传感器的应用试验研究

介绍了URG-04LX2维激光扫描测距传感器的测距原理和功能,采用全速USB和RS232C两种外部接口向传感器传输测距数据命令,读取距离数据并处理。测距试验表明,传感器能准确测量出到平面的距离,地板砖、白色纸、黄色纸和土壤等被测对象在2 m以内的测量误差小于10 mm,但不同材质或颜色的被测对象激光扫描传感器的测量精度存在一定差异。障碍物检测试验表明,传感器的数据能清晰地检测到障碍物,并获得障碍物在传感器坐标系内的坐标值和障碍物尺寸等参数,测量到的障碍物尺寸误差小于10 mm。     

Promoting the integration of technology in teaching: An analysis of the factors that increase the intention to use technologies among Italian teachers

Background: The rapid spread of COVID-19 forced many countries to adopt severe containment measures, transferring all didactic activities into virtual environments. However, the integration of technology in teaching may present difficulties, especially in some countries, such as Italy. Objectives: The present study analyzed how the two main factors of the Technology Acceptance Model (TAM), namely perceived usefulness and perceived ease of use, combined with online teaching self-efficacy, were associated with the intention to use technology to teach. We posited a moderated moderation model in which perceived usefulness represented the main predictor, with self-efficacy and perceived ease of use as moderators of intention to use technology to teach. Method: The model was tested through multiple regressions, using the PROCESS macro on SPSS 26 with a sample of 178 upper secondary school teachers in Italy. Results and conclusions: Regressions showed that each variable significantly predicted the intention to use technology. In addition, a moderation effect of self-efficacy on the perceived usefulness of using technology was found for medium and high-level of perceived ease of use of technology. Implications: The present study provides targeted implications for distance education policy and practice to promote its adoption (or the blended modality) in Italian upper schools.     

车-简支梁桥耦合振动系统中全局最大动力响应研究

本文采用四分之一车辆模型和等截面简支梁模型,建立了车-桥耦合振动计算模型,分析了车辆匀速驶过桥梁时,桥梁动挠度、动弯矩、动剪力的全局最大值及发生的位置.进一步分析计算了车速、车距和桥梁模态截断阶数对桥梁动挠度、动弯矩、动剪力全局最大值的影响.结果表明,车速对桥梁动挠度、动弯矩、动剪力全局最大值影响较大,全局最大动挠度和全局最大动弯矩均出现在跨中附近,车速不同位置也不同,而全局最大动剪力均出现在车辆下桥的梁端;两车同时上桥时,前后车车距越大,桥梁动挠度、动弯矩、动剪力全局最大值越小,当达到一定车距时,三者不再减小且与单车情况相同;为提高桥梁动挠度、动弯矩、动剪力全局最大值计算精度,桥梁模态截断阶数宜分别大于3阶、6阶、7阶.     

一种基于Mahalanobis距离的增量聚类算法

经典的模糊c均值聚类算法对非球型或椭球型分布的数据集进行聚类效果较差。将经典的模糊c均值聚类中的欧氏距离用Mahalanobis距离替代,利用Mahalanobis距离的优点,将其用于增量学习中,提出一种基于马氏距离的模糊增量聚类学习算法。实验结果表明该算法能较有效地解决模糊聚类方法中的缺陷,提高了训练精度。     

煤矿主提升设备动态监测系统

针对传统的煤矿主提升设备状态监测系统存在监测适用范围单一等问题,提出了一种提升设备动态监测系统的设计方案。该系统实现了钢丝绳实时动态张力监测、钢丝绳断丝在线监测、罐道间距在线监测,为提升设备的安全运行提供了保障。