基于加权弹性网络回归的个性化HRTF方法研究

个性化的头相关传输函数（Head Related Transfer Function, HRTF）对于虚拟听觉技术的实现至关重要。然而在具体的应用过程中，测量每一位受试者的个性化HRTF较为繁琐，为此文章提出一种基于加权弹性网络回归的算法，只需获取受试者的生理参数即可获得HRTF的个性化幅度响应。首先通过数据库中的受试者数据，根据生理参数与幅度的相关性计算获得生理参数的权值，并将权值加入到同时含有1范数和2范数的弹性网络回归中，以此来获取新受试者的生理参数稀疏系数；最后将所得稀疏系数与数据库中的HRTF幅度结合就可以得到新受试者的个性化幅度响应。结果表明，文中方法对于个性化HRTF幅度的合成有较好的效果，尤其是在中低频段内准确度较高。     

Performance of “ductless” personalized ventilation in conjunction with displacement ventilation: Impact of intake height

The importance of the intake positioning height above the floor level on the performance of “ductless” personalized ventilation (“ductless” PV) in conjunction with displacement ventilation (DV) was examined with regard to the quality of inhaled air and of the thermal comfort provided. A typical office room with two workstations positioned one behind the other was arranged in a full-scale room. Each workstation consisted of a table with an installed “ductless” PV system, PC, desk lamp and seated breathing thermal manikin. The “ductless” PV system sucked the clean and cool displacement air supplied over the floor at four different heights, i.e. 2, 5, 10 and 20 cm and transported it direct to the breathing level. Moreover, two displacement airflow rates were used with a supply temperature adjusted in order to maintain an exhaust air temperature of 26 °C. Two pollution sources, namely air exhaled by one of the manikins and passive pollution on the table in front of the same manikin were simulated by constant dosing of tracer gases. The results show that the positioning of a “ductless” PV intake height up to 0.2 m above the floor will not significantly influence the quality of inhaled air and thermal comfort.     

Ceiling mounted personalized ventilation system in hot and humid climate—An energy analysis

This study is to evaluate energy saving potential of ceiling mounted personalized ventilation (PV) system in conjunction with background mixing ventilation compared with mixing ventilation system alone and with mixing ventilation system when occupants are provided with individually controlled desk fans for generating additional air movement at each desk. Control strategy applied includes different number of personalized ventilation air terminal devices used and different PV airflow rates supplied. Energy calculation is based on design conditions in Singapore, representing a hot and humid climate. The results reveal that increasing room temperature can save cooling energy when the combination of PV with ceiling mounted personalized ventilation nozzles and background mixing ventilation is used. In this case the energy for transport of air increases but the total energy decreases, i.e. energy can be saved due to elevated room temperature. Comparing with mixing ventilation plus desk fans, ceiling mounted personalized ventilation cannot only realize better cooling effect but also decrease the total energy consumption.     

基于多Agent的移动信息服务支持系统

随着我国移动通信的迅猛发展 ,用户对于移动增值业务的需求越来越高。如何建立一个高效的移动增值业务服务支持系统 ,并保证移动增值业务的服务品质 ,是一个急需研究和解决的重要课题。本文针对这一问题 ,从系统体系结构的角度提出了一种多 Agent移动信息服务支持系统。该系统提供了可以高效地配置与管理信息类移动增值业务的系统解决方案 ,尤其适用于主动式高级信息服务类增值业务。这一系统结构适应性强、易扩展、易维护 ,并能够更好地支持对用户的个性化服务。该系统具有广泛的应用前景 ,对于未来 3G的增值业务具有一定的参考价值     

21世纪的室内空气品质:追求优异

现场研究表明，即使在许多符合现行标准和指南的建筑物中，也有相当多数的人对室内空气品质不满意，甚至出现病态建筑综合症的症状。其原因在于上述标准规定的要求过低，使得很多人不满意并受到不良影响。预见在２１世纪将发生从中等到优异的室内环境的范式转变。提出下述５项原理作为优异准则的基本要素：改善室内空气品质将提高生产率减少病态建筑综合症症状；应排除不必要的室内污染源；应为室内人员提供干爽的送风；应实行“个性     

基于改进加权协同过滤的集群用户黑箱个性意象预测

为了解决传统意象定位中感性意象部分信息丢失及用户模糊的个性化需求不完全表达的问题,提出基于改进加权SO（WSO）算法的集群用户个性意象预测研究. 建立用户特征域,基于K-modes算法计算用户差异度,确立用户集群. 对甄选样本实施兴趣度排序及自主性意象评价,创建集群意象因子集. 引入用户相似度优化WSO算法,增强集群用户间的内在联系,精准预测目标用户黑箱的个性意象分值. 基于语义差异问卷及平均绝对误差分析验证黑箱意象,输出集群中单一用户的个性化意象,实现意象预测. 以无人机为例,预测用户的个性意象,误差小于0.5被舍去,表明该方法能够较好地实现用户模糊的意象黑箱透明化,且预测的意象符合用户的个性化需求,可以有效辅助设计师有针对性地设计.     

定制家居订单智慧标签一码贯通技术研究与应用

新一轮全球科技革命和产业变革兴起,受其影响,定制家居也在智能制造的大潮中快速转型升级,从传统制造迅速转变为数据驱动型的先进制造.定制家居行业的特点决定了企业需要确保差异化订单产品能在规模化生产线上快速、精准流转,进而可以准确交付产品给客户.为此,研究并建立基于工业互联网的个性化定制家居订单智慧标签一码贯通技术和系统平台...     

Gastric Adenocarcinomas and Signet-Ring Cell Carcinoma: Unraveling Gastric Cancer Complexity through Microbiome Analysis—Deepening Heterogeneity for a Personalized Therapy

Gastric cancer (GC) is the fifth most prevalent cancer worldwide and the third leading cause of global cancer mortality. With the advances of the omic studies, a heterogeneous GC landscape has been revealed, with significant molecular diversity. Given the multifaceted nature of GC, identification of different patient subsets with prognostic and/or predictive outcomes is a key aspect to allow tailoring of specific treatments. Recently, the involvement of the microbiota in gastric carcinogenesis has been described. To deepen this aspect, we compared microbiota composition in signet-ring cell carcinoma (SRCC) and adenocarcinoma (ADC), two distinct GC subtypes. To this purpose, 10 ADC and 10 SRCC and their paired non-tumor (PNT) counterparts were evaluated for microbiota composition through 16S rRNA analysis. Weighted and unweighted UniFrac and Bray–Curtis dissimilarity showed significant community-level separation between ADC and SRCC. Through the LEfSe (linear discriminant analysis coupled with effect size) tool, we identified potential microbial biomarkers associated with GC subtypes. In particular, SRCCs were significantly enriched in the phyla Fusobacteria, Bacteroidetes, Patescibacteria, whereas in the ADC type, Proteobacteria and Acidobacteria phyla were found. Overall, our data add new insights into GC heterogeneity and may contribute to deepening the GC classification.     

融合时间隐语义填充和子群划分的推荐算法

随着个性化推荐技术的发展，推荐系统面临着越来越多的挑战。传统的推荐算法通常存在数据稀疏性和推荐精度低等问题。针对以上问题，提出了一种融合时间隐语义填充和子群划分的推荐算法[K]-TLFM（Time Based Latent Factor Model Integrated with [k]-means）。该算法利用融合时间因素的隐语义模型对原始用户物品评分矩阵缺失项进行填充，避免了用全局平均值或者用户/物品平均值补全矩阵带来的误差，有效缓解了数据稀疏性问题，同时融合时间因素有效地刻画了用户偏好随时间的变化；完成评分矩阵缺失项填充后，基于二分[k]-means聚类算法将偏好、兴趣特征相似的对象划分到同一个子群中，在目标用户所属的子群中基于选定的协同过滤算法为用户产生推荐列表，提高了推荐效率和准确性。在MovieLens和Netflix数据集上对该算法的推荐性能进行了对比实验，结果表明该算法具有更高的推荐精度。