基于PSO和多元线性回归的奥运会奖牌预测分析

针对传统奥运会成绩预测模型中存在的不足,提出一种将PSO和多元线性回归分析相结合的预测模型。分析多元线性回归和PSO算法后,结合体育比赛的特点,通过PSO优化迭代计算,确定奖牌统计模型中各回归系数,建立了奥运会参赛国成绩与回归系数之间复杂的非线性关系。利用Matlab对建立奖牌模型进行了验证,预测精度高。     

基于随机游走增强型矩阵分解的混合服务预测

随着Web服务数量的急剧增长,如何在大量功能相似但非功能属性各异的服务中选择满足用户个性化需求的服务是亟需解决的问题。基于QoS（Quality of Service）预测的服务推荐方法成为研究热点。然而, QoS数据的稀疏性和“冷启动”问题阻碍其发展。针对当前主流的QoS预测模型预测精度不高和收敛速度较慢等问题,提出一种基于随机游走模型和矩阵分解技术的混合QoS预测方法。该方法首先基于矩阵分解获得用户及服务的潜因子矩阵,并将用户潜因子矩阵转化为用户相似度矩阵;然后基于用户相似度矩阵并结合Web服务的网络位置信息,使用随机游走模型提高用户相似度矩阵的准确性;最终结合协同过滤方法与矩阵分解模型进行QoS预测。在真实数据集上实验,结果表明,与当前主流的QoS预测方法相比,该方法具有更高的预测精度和效率。     

基于灰色模糊理论的河南省城镇居民收入结构关系及预测研究

城镇居民收入是衡量一个国家或地区经济竞争力的一个重要标准,如何提高居民收入是提高经济实力的重要方面.通过基于灰色理论方法的计算得来的河南省居民收入结构关联度显示：居民工薪收入和转移性收入是影响居民可支配收入的主导因素,所以要提高居民收入就主要从这两个方面着手.最后利用EXCEL表格通过编程预测出了2009～2013年河南省可支配收入和转移性收入的情况,数据显示：参差〈0.35,精度〉0.8,预测合格.     

Predicting future dynamics from short-term time series using an Anticipated Learning Machine

Predicting time series has significant practical applications over different disciplines. Here, we propose an Anticipated Learning Machine (ALM) to achieve precise future-state predictions based on short-term but high-dimensional data. From non-linear dynamical systems theory, we show that ALM can transform recent correlation/spatial information of high-dimensional variables into future dynamical/temporal information of any target variable, thereby overcoming the small-sample problem and achieving multistep-ahead predictions. Since the training samples generated from high-dimensional data also include information of the unknown future values of the target variable, it is called anticipated learning. Extensive experiments on real-world data demonstrate significantly superior performances of ALM over all of the existing 12 methods. In contrast to traditional statistics-based machine learning, ALM is based on non-linear dynamics, thus opening a new way for dynamics-based machine learning.     

Responses to oral glucose challenge differ by physical activity volume and intensity: A pilot study

BackgroundOne-hour postprandial hyperglycemia is associated with increased risk of type 2 diabetes and cardiovascular disease. Physical activity (PA) has short-term beneficial effects on post-meal glucose response. This study compared the oral glucose tolerance test results of 3 groups of people with habitually different levels of PA.MethodsThirty-one adults without diabetes (age 25.9 ± 6.6 years; body mass index 23.8 ± 3.8 kg/m²; mean ± SD) were recruited and divided into 3 groups based on self-reported PA volume and intensity: low activity < 30 min/day of moderate-intensity activity (n = 11), moderately active ≥ 30 min/day of moderate-intensity PA (n = 10), and very active ≥ 60 min/day of PA at high intensity (n = 10). Participants completed an oral glucose tolerance test (50 g glucose) with capillary blood samples obtained at baseline, 15 min, 30 min, 45 min, 60 min, 90 min, and 120 min post-ingestion.ResultsThere were no significant differences between groups for age or body fat percentage or glycated hemoglobin (p > 0.05). The groups were significantly different in terms of baseline glucose level (p = 0.003) and, marginally, for gender (p = 0.053) and BMI (p = 0.050). There was a statistically significant effect of PA on the 1-h postprandial glucose results (p = 0.029), with differences between very active and low activity groups (p = 0.008) but not between the moderately active and low activity groups (p = 0.360), even when baseline glucose level and gender differences were accounted for. For incremental area under the curve there was no significant effect of activity group once gender and body fat percentage had been accounted for (p = 0.401). Those in the low activity group took 15 min longer to reach peak glucose level than those in the very active group (p = 0.012).ConclusionThe results suggest that high levels of PA have a beneficial effect on postprandial blood glucose profiles when compared to low and moderate levels of activity.     

基于隐马尔科夫模型的VLCC目的港预测

超大型油船(very large crude carrier,VLCC)目的港预测对海运原油流向预测以及货源地未来运力估计具有重要作用。针对VLCC的AIS目的港信息存在缺失、更新不及时、不准确等现象,提出一种基于隐马尔科夫模型的VLCC目的港预测方法。分析船舶AIS轨迹数据,得到油船历史停靠港口序列;根据VLCC轨迹提取习惯航路,以航路中的交叉点为依据设置观测线;利用船舶航行轨迹数据判断船舶是否经过观测线以及经过观测线的方向,对不同方向分别计算船舶在挂靠港间的转移概率矩阵和船舶挂靠港与观测线间的输出概率矩阵,建立VLCC目的港预测模型并进行预测。研究结果表明：在大多数情况下VLCC目的港预测的准确率可以达到70%以上;航线越固定、运行越规律的船舶,预测准确率越高;船舶越靠近目的港,预测越准确;重载状态下的船舶目的港预测更准确。     

基于DE-SVM的船舶航迹预测模型

为提高船舶航迹预测精度,解决准确建模难度大和神经网络易陷入局部最优的问题,考虑实时获取目标船AIS数据较少的特点,提出一种基于支持向量机(support vector machine,SVM)的航迹预测模型。选择AIS数据中的航速、航向和船舶经纬度作为样本特征变量;采用小波阈值去噪的方法处理训练数据;采用差分进化(differential evolution,DE)算法对模型内部参数寻优以提高模型收敛速度和预测精度。选取天津港实船某段航迹的AIS数据,比较基于DE-SVM与基于BP神经网络的航迹预测模型的仿真结果。结果表明,基于DE-SVM的航迹预测模型具有更高的预测精度,简单、可行、高效,且耗时少。     

Reciprocal Versus Nonreciprocal Assessment of Knee Flexors and Extensors in Concentric Actions Using the CON-TREX Multijoint Isokinetic Dynamometer: A Reliability Study

Knee flexor and extensor muscular assessment via isokinetic dynamometry is common practice and established in the research literature. However, reporting assessment methodology regarding reciprocal and nonreciprocal movements is often vague or absent. Such methodological issues are crucial for accurate assessments. Therefore, knee extensor and flexor peak moment using either reciprocal movement or nonreciprocal modalities was assessed. Fifteen participants performed three blocks of five concentric muscle actions at three angular velocities [1: nonreciprocal (maximal active flexion followed by passive extension); B2: reciprocal (maximal active extension followed by maximal active flexion); B3 nonreciprocal (maximal active extension followed by passive flexion)]. ANOVA revealed statistically significant within-subject modality effects for peak knee extensor moment and flexor velocity and modality differences (p < .05). Reciprocal and nonreciprocal assessments give significantly different results, with nonreciprocal giving higher peak moments. Reporting which modality is used is crucial to allow for greater clarity for the reader and practitioner.     

基于表示学习的双层知识网络链路预测

当前,针对知识网络的链路预测主要是基于网络拓扑结构的相似性,很少考虑作者的研究领域,导致信息利用不充分等问题,因此本文提出了双层知识网络的链路预测框架hypernet2vec。双层知识网络,即作者合著关系网络和学术领域关系网络,利用网络表示学习,分别将两层网络中的节点映射到低维的向量空间,再输入到专门设计的卷积神经网络中计算并进行链路预测。与经典的链路预测指标如RA指标、LP指标和LRW指标等相比,hypernet2vec模型预测的AUC(area under curve)值取得了显著的提升,平均提升幅度达11.17%。文章还从情报产生层面和复杂系统层面,对模型发生作用的深层机理进行了探讨。     

图书馆信息学与新文科交叉领域的关系预测研究

[目的/意义]论文回答的研究问题是:在新文科建设背景下,图书馆信息学与数据科学、数字人文与健康信息学等新文科交叉领域的关系发展的未来前景是什么?[方法/过程]论文从现有研究文献中图书馆信息学与数字人文、数据科学、健康信息学三个新文科交叉领域的关系类型出发,设计了民族志未来访谈,邀请三十四位国内图书馆信息学和情报学等学科的学者提炼这四个学科(领域)的核心竞争力,判断其相互间价值,并预测2030年最乐观、最悲观和最有可能的发展前景。[结果/结论]图书馆信息学与数字人文、与数据科学、与健康信息学在2030年更有可能出现的前景是并列有交集的关系,健康信息学相对更可能成为图书馆信息学的下位类,数据科学与图书馆信息学的关系相对更独立。