Recommendation systems have become indispensable for providing tailored suggestions and capturing evolving user preferences based on interaction histories.The collaborative filtering(CF)model,which depends exclusively...Recommendation systems have become indispensable for providing tailored suggestions and capturing evolving user preferences based on interaction histories.The collaborative filtering(CF)model,which depends exclusively on user-item interactions,commonly encounters challenges,including the cold-start problem and an inability to effectively capture the sequential and temporal characteristics of user behavior.This paper introduces a personalized recommendation system that combines deep learning techniques with Bayesian Personalized Ranking(BPR)optimization to address these limitations.With the strong support of Long Short-Term Memory(LSTM)networks,we apply it to identify sequential dependencies of user behavior and then incorporate an attention mechanism to improve the prioritization of relevant items,thereby enhancing recommendations based on the hybrid feedback of the user and its interaction patterns.The proposed system is empirically evaluated using publicly available datasets from movie and music,and we evaluate the performance against standard recommendation models,including Popularity,BPR,ItemKNN,FPMC,LightGCN,GRU4Rec,NARM,SASRec,and BERT4Rec.The results demonstrate that our proposed framework consistently achieves high outcomes in terms of HitRate,NDCG,MRR,and Precision at K=100,with scores of(0.6763,0.1892,0.0796,0.0068)on MovieLens-100K,(0.6826,0.1920,0.0813,0.0068)on MovieLens-1M,and(0.7937,0.3701,0.2756,0.0078)on Last.fm.The results show an average improvement of around 15%across all metrics compared to existing sequence models,proving that our framework ranks and recommends items more accurately.展开更多
Test case prioritization and ranking play a crucial role in software testing by improving fault detection efficiency and ensuring software reliability.While prioritization selects the most relevant test cases for opti...Test case prioritization and ranking play a crucial role in software testing by improving fault detection efficiency and ensuring software reliability.While prioritization selects the most relevant test cases for optimal coverage,ranking further refines their execution order to detect critical faults earlier.This study investigates machine learning techniques to enhance both prioritization and ranking,contributing to more effective and efficient testing processes.We first employ advanced feature engineering alongside ensemble models,including Gradient Boosted,Support Vector Machines,Random Forests,and Naive Bayes classifiers to optimize test case prioritization,achieving an accuracy score of 0.98847 and significantly improving the Average Percentage of Fault Detection(APFD).Subsequently,we introduce a deep Q-learning framework combined with a Genetic Algorithm(GA)to refine test case ranking within priority levels.This approach achieves a rank accuracy of 0.9172,demonstrating robust performance despite the increasing computational demands of specialized variation operators.Our findings highlight the effectiveness of stacked ensemble learning and reinforcement learning in optimizing test case prioritization and ranking.This integrated approach improves testing efficiency,reduces late-stage defects,and improves overall software stability.The study provides valuable information for AI-driven testing frameworks,paving the way for more intelligent and adaptive software quality assurance methodologies.展开更多
目的:探究补骨膏对绝经后骨质疏松症(PMOP)大鼠的骨保护作用及其对骨保护素(OPG)-核因子κB受体激活蛋白(RANK)-核因子κB受体激活蛋白配体(RANKL)信号通路的调控作用。方法:将36只大鼠随机分为假手术组(9只)和手术组(27只),手术组采用...目的:探究补骨膏对绝经后骨质疏松症(PMOP)大鼠的骨保护作用及其对骨保护素(OPG)-核因子κB受体激活蛋白(RANK)-核因子κB受体激活蛋白配体(RANKL)信号通路的调控作用。方法:将36只大鼠随机分为假手术组(9只)和手术组(27只),手术组采用双侧卵巢切除建立PMOP大鼠模型。造模成功后,将24只PMOP大鼠随机分为模型组、戊酸雌二醇组、补骨膏低剂量组、补骨膏高剂量组,然后予相应药物灌胃8周。骨密度仪检测股骨近端骨密度;Micro-CT三维重建分析股骨微结构;苏木精-伊红(HE)染色观察股骨组织病理形态;酶联免疫吸附试验(ELISA)法检测血清中骨碱性磷酸酶(BALP)、骨钙素(BGP)、OPG水平,测定盒检测血清中磷、钙水平;蛋白质印迹法(Western blotting)检测股骨组织中OPG、RANK、RANKL蛋白表达水平;RT-qPCR法检测股骨组织肿瘤坏死因子-α(TNF-α)mRNA、干扰素-γ(IFN-γ)mRNA、精氨酸酶-1(Arg-1)mRNA、转化生长因子-β1(TGF-β1)mRNA、基质金属蛋白酶-9(MMP-9)mRNA、OPG mRNA、RANK mRNA、RANKL mRNA表达水平。结果:假手术组大鼠股骨结构连续完整,骨小梁数目较多,形态较厚,结构致密;模型组大鼠股骨近端骨密度明显降低;补骨膏低剂量组、补骨膏高剂量组和戊酸雌二醇组大鼠股骨近端骨小梁数量、骨组织形态结构均得到不同程度改善。模型组大鼠骨密度及血清中钙、BALP、BGP、OPG水平均低于假手术组(P<0.01),血清磷水平高于假手术组(P<0.01);补骨膏低剂量组、补骨膏高剂量组及戊酸雌二醇组大鼠骨密度及血清中钙、BALP、BGP、OPG水平均高于模型组(P<0.05或P<0.01),血清磷低于模型组(P<0.01)。模型组大鼠股骨组织OPG蛋白相对表达量低于假手术组(P<0.01),RANK、RANKL蛋白相对表达量均高于假手术组(P<0.01);补骨膏低剂量组、补骨膏高剂量组及戊酸雌二醇组大鼠股骨组织中OPG蛋白相对表达量高于模型组(P<0.05)或(P<0.01),RANK、RANKL蛋白相对表达量均低于模型组(P<0.01)。模型组大鼠股骨组织TNF-α mRNA、IFN-γ mRNA、MMP-9 m RNA、RANK m RNA、RANKL mRNA对表达量均高于假手术组(P<0.01),Arg-1 m RNA、TGF-β1 mRNA、OPG mRNA对表达量均低于假手术组(P<0.01);补骨膏高剂量组及戊酸雌二醇组大鼠股骨组织TNF-α mRNA、IFN-γ m RNA、MMP-9mRNA、RANK mRNA、RANKL mRNA相对表达量均低于模型组(P<0.01),Arg-1 mRNA、TGF-β1 mRNA、OPG m RNA相对表达量均高于模型组(P<0.01)。结论:补骨膏可能通过调控OPG-RANK-RANKL信号通路,抑制免疫炎症反应,调节骨基质胶原合成与降解,从而维持骨代谢平衡,改善PMOP大鼠骨密度及骨微结构病理损伤。展开更多
On the basis of research evaluation of Chinese universities,Golden Apple Ranking(GAR)was initiated by Research Center of Chinese Science Evaluation(RCCSE)at Wuhan University in 2003.The GAR consists of four major rank...On the basis of research evaluation of Chinese universities,Golden Apple Ranking(GAR)was initiated by Research Center of Chinese Science Evaluation(RCCSE)at Wuhan University in 2003.The GAR consists of four major rankings:Chinese University Ranking,Chinese Graduate School Ranking,World University Ranking and Scholarly Journal Ranking.The annual reports of all these four rankings are published bythe Science Press,which have been recognized by the academia and China's government.展开更多
Purpose:Citation-based assessments of countries’research capabilities often misrepresent their ability to achieve breakthrough advancements.These assessments commonly classify Japan as a developing country,which cont...Purpose:Citation-based assessments of countries’research capabilities often misrepresent their ability to achieve breakthrough advancements.These assessments commonly classify Japan as a developing country,which contradicts its prominent scientific standing.The purpose of this study is to investigate the underlying causes of such inaccurate assessments and to propose methods for conducting more reliable evaluations.Design/methodology/approach:The study evaluates the effectiveness of top-percentile citation metrics as indicators of breakthrough research.Using case studies of selected countries and research topics,the study examines how deviations from lognormal citation distributions impact the accuracy of these percentile indicators.A similar analysis is conducted using university data from the Leiden Ranking to investigate citation distribution deviations at the institutional level.Findings:The study finds that inflated lower tails in citation distributions lead to undervaluation of research capabilities in advanced technological countries,as captured by some percentile indicators.Conversely,research-intensive universities exhibit the opposite trend:a reduced lower tail relative to the upper tail,which causes percentile indicators to overestimate their actual research capacity.Research limitations:The descriptions are mathematical facts that are self-evident.Practical implications:The ratios between the number of papers in the global top 10%and 1%by citation count to the total number of papers are commonly used to describe research performance.However,due to variations in citation patterns across countries and institutions with reference to the global pattern,these ratios can be misleading and lose their value as research indicators.Originality/value:Size-independent research performance indicators,obtained as the ratios between paper counts in top percentiles and the total numbers of publications,are widely used by public and private institutions.This study demonstrates that the use of these ratios for research evaluations and country rankings can be highly misleading.展开更多
On the basis of ESI data,all universities are ranked in 96 out of 109 world-class disciplines.There is no ESI data(either publications or citations)in the rest of 13 world-class disciplines.
Among natural disasters,flash floods are the most destructive events,causing signif-icant damage to the economy and posing a serious threat to human life and property.Comprehensive risk assessment of these sudden floo...Among natural disasters,flash floods are the most destructive events,causing signif-icant damage to the economy and posing a serious threat to human life and property.Comprehensive risk assessment of these sudden floods is a key strategy to mitigate their impact.Accurate analysis of flash flood hazards can greatly enhance prevention efforts and inform critical decision-making processes,ultimately improving our ability to protect communities from these fast-onset disasters.This study analyzed the driving forces of flash flood disaster-causing factors in Heilongjiang Province.Meanwhile,nine different categories of variables affecting the occurrence of flash floods were selected,and the degree of influence of each driving factor on flash floods was quantitatively analyzed,and the driving force analysis of the driving factors of flash floods in Hei-longjiang Province was carried out by using the geographic probe model.This paper employs an uncertainty approach,utilizing a statistical-based interval weight deter-mination technique for evaluation indices and a two-dimensional information-based interval number sorting method.These methodologies are combined to construct a comprehensive flash flood risk assessment model.On this basis,the model was implemented in six regions within China's Heilongjiang province to evaluate and prioritize flash flood risks.The resulting risk ranking for these areas was as follows:Bayan≻Shuangcheng≻Boli≻Suibin≻Hailun≻Yian.The findings demonstrate that the interval number-based evaluation method effectively handles uncertainty,providing a more reliable risk grading system.This approach,by leveraging modern scientific advances and risk quantification techniques,is crucial for improving disaster management and mitigating flash flood impacts.展开更多
Owing to the constraints of depth sensing technology,images acquired by depth cameras are inevitably mixed with various noises.For depth maps presented in gray values,this research proposes a novel denoising model,ter...Owing to the constraints of depth sensing technology,images acquired by depth cameras are inevitably mixed with various noises.For depth maps presented in gray values,this research proposes a novel denoising model,termed graph-based transform(GBT)and dual graph Laplacian regularization(DGLR)(DGLR-GBT).This model specifically aims to remove Gaussian white noise by capitalizing on the nonlocal self-similarity(NSS)and the piecewise smoothness properties intrinsic to depth maps.Within the group sparse coding(GSC)framework,a combination of GBT and DGLR is implemented.Firstly,within each group,the graph is constructed by using estimates of the true values of the averaged blocks instead of the observations.Secondly,the graph Laplacian regular terms are constructed based on rows and columns of similar block groups,respectively.Lastly,the solution is obtained effectively by combining the alternating direction multiplication method(ADMM)with the weighted thresholding method within the domain of GBT.展开更多
This paper studies certain estimates for the lower bound of distance between unitary orbits of normal elements.We show that the distance between unitary orbits of normal elements of simple C^(*)-algebras of tracial ra...This paper studies certain estimates for the lower bound of distance between unitary orbits of normal elements.We show that the distance between unitary orbits of normal elements of simple C^(*)-algebras of tracial rank no more than k has a lower bound.Furthermore,if k≤1 and normal elements are commuting,then the lower bound will be better.Another result establishes a connection involving the spectrum distance operator Dc between a C^(*)-algebra of stable rank one C^(*)-algebra and its hereditary C^(*)-subalgebra.展开更多
In order to rank searching results according to the user preferences,a new personalized web pages ranking algorithm called PWPR（personalized web page ranking）with the idea of adjusting the ranking scores of web page...In order to rank searching results according to the user preferences,a new personalized web pages ranking algorithm called PWPR（personalized web page ranking）with the idea of adjusting the ranking scores of web pages in accordance with user preferences is proposed.PWPR assigns the initial weights based on user interests and creates the virtual links and hubs according to user interests.By measuring user click streams,PWPR incrementally reflects users’ favors for the personalized ranking.To improve the accuracy of ranking, PWPR also takes collaborative filtering into consideration when the query with similar is submitted by users who have similar user interests. Detailed simulation results and comparison with other algorithms prove that the proposed PWPR can adaptively provide personalized ranking and truly relevant information to user preferences.展开更多
随着基于位置社交网络(location-based social network,LBSN)的发展,兴趣点推荐成为满足用户个性化需求、减轻信息过载问题的重要手段.然而,已有的兴趣点推荐算法存在如下的问题:1)多数已有的兴趣点推荐算法简化用户签到频率数据,仅使...随着基于位置社交网络(location-based social network,LBSN)的发展,兴趣点推荐成为满足用户个性化需求、减轻信息过载问题的重要手段.然而,已有的兴趣点推荐算法存在如下的问题:1)多数已有的兴趣点推荐算法简化用户签到频率数据,仅使用二进制值来表示用户是否访问一个兴趣点;2)基于矩阵分解的兴趣点推荐算法把签到频率数据和传统推荐系统中的评分数据等同看待,使用高斯分布模型建模用户的签到行为;3)忽视用户签到数据的隐式反馈属性.为解决以上问题,提出一个基于Ranking的泊松矩阵分解兴趣点推荐算法.首先,根据LBSN中用户的签到行为特点,利用泊松分布模型替代高斯分布模型建模用户在兴趣点上签到行为;然后采用BPR(Bayesian personalized ranking)标准优化泊松矩阵分解的损失函数,拟合用户在兴趣点对上的偏序关系;最后,利用包含地域影响力的正则化因子约束泊松矩阵分解的过程.在真实数据集上的实验结果表明:基于Ranking的泊松矩阵分解兴趣点推荐算法的性能优于传统的兴趣点推荐算法.展开更多
基金funded by Soonchunhyang University,Grant Number 20250029。
文摘Recommendation systems have become indispensable for providing tailored suggestions and capturing evolving user preferences based on interaction histories.The collaborative filtering(CF)model,which depends exclusively on user-item interactions,commonly encounters challenges,including the cold-start problem and an inability to effectively capture the sequential and temporal characteristics of user behavior.This paper introduces a personalized recommendation system that combines deep learning techniques with Bayesian Personalized Ranking(BPR)optimization to address these limitations.With the strong support of Long Short-Term Memory(LSTM)networks,we apply it to identify sequential dependencies of user behavior and then incorporate an attention mechanism to improve the prioritization of relevant items,thereby enhancing recommendations based on the hybrid feedback of the user and its interaction patterns.The proposed system is empirically evaluated using publicly available datasets from movie and music,and we evaluate the performance against standard recommendation models,including Popularity,BPR,ItemKNN,FPMC,LightGCN,GRU4Rec,NARM,SASRec,and BERT4Rec.The results demonstrate that our proposed framework consistently achieves high outcomes in terms of HitRate,NDCG,MRR,and Precision at K=100,with scores of(0.6763,0.1892,0.0796,0.0068)on MovieLens-100K,(0.6826,0.1920,0.0813,0.0068)on MovieLens-1M,and(0.7937,0.3701,0.2756,0.0078)on Last.fm.The results show an average improvement of around 15%across all metrics compared to existing sequence models,proving that our framework ranks and recommends items more accurately.
文摘Test case prioritization and ranking play a crucial role in software testing by improving fault detection efficiency and ensuring software reliability.While prioritization selects the most relevant test cases for optimal coverage,ranking further refines their execution order to detect critical faults earlier.This study investigates machine learning techniques to enhance both prioritization and ranking,contributing to more effective and efficient testing processes.We first employ advanced feature engineering alongside ensemble models,including Gradient Boosted,Support Vector Machines,Random Forests,and Naive Bayes classifiers to optimize test case prioritization,achieving an accuracy score of 0.98847 and significantly improving the Average Percentage of Fault Detection(APFD).Subsequently,we introduce a deep Q-learning framework combined with a Genetic Algorithm(GA)to refine test case ranking within priority levels.This approach achieves a rank accuracy of 0.9172,demonstrating robust performance despite the increasing computational demands of specialized variation operators.Our findings highlight the effectiveness of stacked ensemble learning and reinforcement learning in optimizing test case prioritization and ranking.This integrated approach improves testing efficiency,reduces late-stage defects,and improves overall software stability.The study provides valuable information for AI-driven testing frameworks,paving the way for more intelligent and adaptive software quality assurance methodologies.
文摘目的:探究补骨膏对绝经后骨质疏松症(PMOP)大鼠的骨保护作用及其对骨保护素(OPG)-核因子κB受体激活蛋白(RANK)-核因子κB受体激活蛋白配体(RANKL)信号通路的调控作用。方法:将36只大鼠随机分为假手术组(9只)和手术组(27只),手术组采用双侧卵巢切除建立PMOP大鼠模型。造模成功后,将24只PMOP大鼠随机分为模型组、戊酸雌二醇组、补骨膏低剂量组、补骨膏高剂量组,然后予相应药物灌胃8周。骨密度仪检测股骨近端骨密度;Micro-CT三维重建分析股骨微结构;苏木精-伊红(HE)染色观察股骨组织病理形态;酶联免疫吸附试验(ELISA)法检测血清中骨碱性磷酸酶(BALP)、骨钙素(BGP)、OPG水平,测定盒检测血清中磷、钙水平;蛋白质印迹法(Western blotting)检测股骨组织中OPG、RANK、RANKL蛋白表达水平;RT-qPCR法检测股骨组织肿瘤坏死因子-α(TNF-α)mRNA、干扰素-γ(IFN-γ)mRNA、精氨酸酶-1(Arg-1)mRNA、转化生长因子-β1(TGF-β1)mRNA、基质金属蛋白酶-9(MMP-9)mRNA、OPG mRNA、RANK mRNA、RANKL mRNA表达水平。结果:假手术组大鼠股骨结构连续完整,骨小梁数目较多,形态较厚,结构致密;模型组大鼠股骨近端骨密度明显降低;补骨膏低剂量组、补骨膏高剂量组和戊酸雌二醇组大鼠股骨近端骨小梁数量、骨组织形态结构均得到不同程度改善。模型组大鼠骨密度及血清中钙、BALP、BGP、OPG水平均低于假手术组(P<0.01),血清磷水平高于假手术组(P<0.01);补骨膏低剂量组、补骨膏高剂量组及戊酸雌二醇组大鼠骨密度及血清中钙、BALP、BGP、OPG水平均高于模型组(P<0.05或P<0.01),血清磷低于模型组(P<0.01)。模型组大鼠股骨组织OPG蛋白相对表达量低于假手术组(P<0.01),RANK、RANKL蛋白相对表达量均高于假手术组(P<0.01);补骨膏低剂量组、补骨膏高剂量组及戊酸雌二醇组大鼠股骨组织中OPG蛋白相对表达量高于模型组(P<0.05)或(P<0.01),RANK、RANKL蛋白相对表达量均低于模型组(P<0.01)。模型组大鼠股骨组织TNF-α mRNA、IFN-γ mRNA、MMP-9 m RNA、RANK m RNA、RANKL mRNA对表达量均高于假手术组(P<0.01),Arg-1 m RNA、TGF-β1 mRNA、OPG mRNA对表达量均低于假手术组(P<0.01);补骨膏高剂量组及戊酸雌二醇组大鼠股骨组织TNF-α mRNA、IFN-γ m RNA、MMP-9mRNA、RANK mRNA、RANKL mRNA相对表达量均低于模型组(P<0.01),Arg-1 mRNA、TGF-β1 mRNA、OPG m RNA相对表达量均高于模型组(P<0.01)。结论:补骨膏可能通过调控OPG-RANK-RANKL信号通路,抑制免疫炎症反应,调节骨基质胶原合成与降解,从而维持骨代谢平衡,改善PMOP大鼠骨密度及骨微结构病理损伤。
文摘On the basis of research evaluation of Chinese universities,Golden Apple Ranking(GAR)was initiated by Research Center of Chinese Science Evaluation(RCCSE)at Wuhan University in 2003.The GAR consists of four major rankings:Chinese University Ranking,Chinese Graduate School Ranking,World University Ranking and Scholarly Journal Ranking.The annual reports of all these four rankings are published bythe Science Press,which have been recognized by the academia and China's government.
文摘Purpose:Citation-based assessments of countries’research capabilities often misrepresent their ability to achieve breakthrough advancements.These assessments commonly classify Japan as a developing country,which contradicts its prominent scientific standing.The purpose of this study is to investigate the underlying causes of such inaccurate assessments and to propose methods for conducting more reliable evaluations.Design/methodology/approach:The study evaluates the effectiveness of top-percentile citation metrics as indicators of breakthrough research.Using case studies of selected countries and research topics,the study examines how deviations from lognormal citation distributions impact the accuracy of these percentile indicators.A similar analysis is conducted using university data from the Leiden Ranking to investigate citation distribution deviations at the institutional level.Findings:The study finds that inflated lower tails in citation distributions lead to undervaluation of research capabilities in advanced technological countries,as captured by some percentile indicators.Conversely,research-intensive universities exhibit the opposite trend:a reduced lower tail relative to the upper tail,which causes percentile indicators to overestimate their actual research capacity.Research limitations:The descriptions are mathematical facts that are self-evident.Practical implications:The ratios between the number of papers in the global top 10%and 1%by citation count to the total number of papers are commonly used to describe research performance.However,due to variations in citation patterns across countries and institutions with reference to the global pattern,these ratios can be misleading and lose their value as research indicators.Originality/value:Size-independent research performance indicators,obtained as the ratios between paper counts in top percentiles and the total numbers of publications,are widely used by public and private institutions.This study demonstrates that the use of these ratios for research evaluations and country rankings can be highly misleading.
文摘On the basis of ESI data,all universities are ranked in 96 out of 109 world-class disciplines.There is no ESI data(either publications or citations)in the rest of 13 world-class disciplines.
基金Basic Scientific Research Expense Project of IWHR-Extreme rainstorm development trends and prediction techniques,Grant/Award Number:JZ0145B142024National Natural Science Foundation of China,Grant/Award Number:42271095。
文摘Among natural disasters,flash floods are the most destructive events,causing signif-icant damage to the economy and posing a serious threat to human life and property.Comprehensive risk assessment of these sudden floods is a key strategy to mitigate their impact.Accurate analysis of flash flood hazards can greatly enhance prevention efforts and inform critical decision-making processes,ultimately improving our ability to protect communities from these fast-onset disasters.This study analyzed the driving forces of flash flood disaster-causing factors in Heilongjiang Province.Meanwhile,nine different categories of variables affecting the occurrence of flash floods were selected,and the degree of influence of each driving factor on flash floods was quantitatively analyzed,and the driving force analysis of the driving factors of flash floods in Hei-longjiang Province was carried out by using the geographic probe model.This paper employs an uncertainty approach,utilizing a statistical-based interval weight deter-mination technique for evaluation indices and a two-dimensional information-based interval number sorting method.These methodologies are combined to construct a comprehensive flash flood risk assessment model.On this basis,the model was implemented in six regions within China's Heilongjiang province to evaluate and prioritize flash flood risks.The resulting risk ranking for these areas was as follows:Bayan≻Shuangcheng≻Boli≻Suibin≻Hailun≻Yian.The findings demonstrate that the interval number-based evaluation method effectively handles uncertainty,providing a more reliable risk grading system.This approach,by leveraging modern scientific advances and risk quantification techniques,is crucial for improving disaster management and mitigating flash flood impacts.
基金National Natural Science Foundation of China(No.62372100)。
文摘Owing to the constraints of depth sensing technology,images acquired by depth cameras are inevitably mixed with various noises.For depth maps presented in gray values,this research proposes a novel denoising model,termed graph-based transform(GBT)and dual graph Laplacian regularization(DGLR)(DGLR-GBT).This model specifically aims to remove Gaussian white noise by capitalizing on the nonlocal self-similarity(NSS)and the piecewise smoothness properties intrinsic to depth maps.Within the group sparse coding(GSC)framework,a combination of GBT and DGLR is implemented.Firstly,within each group,the graph is constructed by using estimates of the true values of the averaged blocks instead of the observations.Secondly,the graph Laplacian regular terms are constructed based on rows and columns of similar block groups,respectively.Lastly,the solution is obtained effectively by combining the alternating direction multiplication method(ADMM)with the weighted thresholding method within the domain of GBT.
基金Supported by Zhejiang Provincial Natural Science Foundation of China(No.ZCLQN25A0103)。
文摘This paper studies certain estimates for the lower bound of distance between unitary orbits of normal elements.We show that the distance between unitary orbits of normal elements of simple C^(*)-algebras of tracial rank no more than k has a lower bound.Furthermore,if k≤1 and normal elements are commuting,then the lower bound will be better.Another result establishes a connection involving the spectrum distance operator Dc between a C^(*)-algebra of stable rank one C^(*)-algebra and its hereditary C^(*)-subalgebra.
基金The Natural Science Foundation of South-Central University for Nationalities(No.YZZ07006)
文摘In order to rank searching results according to the user preferences,a new personalized web pages ranking algorithm called PWPR（personalized web page ranking）with the idea of adjusting the ranking scores of web pages in accordance with user preferences is proposed.PWPR assigns the initial weights based on user interests and creates the virtual links and hubs according to user interests.By measuring user click streams,PWPR incrementally reflects users’ favors for the personalized ranking.To improve the accuracy of ranking, PWPR also takes collaborative filtering into consideration when the query with similar is submitted by users who have similar user interests. Detailed simulation results and comparison with other algorithms prove that the proposed PWPR can adaptively provide personalized ranking and truly relevant information to user preferences.
文摘随着基于位置社交网络(location-based social network,LBSN)的发展,兴趣点推荐成为满足用户个性化需求、减轻信息过载问题的重要手段.然而,已有的兴趣点推荐算法存在如下的问题:1)多数已有的兴趣点推荐算法简化用户签到频率数据,仅使用二进制值来表示用户是否访问一个兴趣点;2)基于矩阵分解的兴趣点推荐算法把签到频率数据和传统推荐系统中的评分数据等同看待,使用高斯分布模型建模用户的签到行为;3)忽视用户签到数据的隐式反馈属性.为解决以上问题,提出一个基于Ranking的泊松矩阵分解兴趣点推荐算法.首先,根据LBSN中用户的签到行为特点,利用泊松分布模型替代高斯分布模型建模用户在兴趣点上签到行为;然后采用BPR(Bayesian personalized ranking)标准优化泊松矩阵分解的损失函数,拟合用户在兴趣点对上的偏序关系;最后,利用包含地域影响力的正则化因子约束泊松矩阵分解的过程.在真实数据集上的实验结果表明:基于Ranking的泊松矩阵分解兴趣点推荐算法的性能优于传统的兴趣点推荐算法.
