In order to guide the construction of maintenance support system,the theory of complex network to maintenance support area is introduced,and a construction idea of maintenance support system based on complex network i...In order to guide the construction of maintenance support system,the theory of complex network to maintenance support area is introduced,and a construction idea of maintenance support system based on complex network is put forward.Firstly,from the perspective of the similarity,maintenance support system is abstracted into complex network to form maintenance support network.Secondly,the basic concepts and parameters of maintenance support network are also introduced.Thirdly,the maintenance support system in certain period is abstracted into a maintenance support network,and the network makes some changes.Finally,the correlative parameters of the network are calculated.The results show that the changed network is more conducive to the maintenance support.This provides a new thought and method to construct maintenance support system.展开更多
Triadic closure is a simple and fundamental kind of link formulation mechanism in network.Local closure coefficient(LCC),a new network property,is to measure the triadic closure with respect to the fraction of length-...Triadic closure is a simple and fundamental kind of link formulation mechanism in network.Local closure coefficient(LCC),a new network property,is to measure the triadic closure with respect to the fraction of length-2 paths for link prediction.In this paper,a weighted format of LCC(WLCC)is introduced to measure the weighted strength of local triadic structure,and a statistic similari-ty-based link prediction metric is proposed to incorporate the definition of WLCC.To prove the metrics effectiveness and scalability,the WLCC formula-tion was further investigated under weighted local Naive Bayes(WLNB)link prediction framework.Finally,extensive experimental studies was conducted with weighted baseline metrics on various public network datasets.The results demonstrate the merits of the proposed metrics in comparison with the weighted baselines.展开更多
Nonpharmaceutical interventions(NPIs),particularly contact tracing isolation and household quarantine,play a vital role in effectively bringing the Coronavirus Disease 2019(COVID-19)under control in China.The pairwise...Nonpharmaceutical interventions(NPIs),particularly contact tracing isolation and household quarantine,play a vital role in effectively bringing the Coronavirus Disease 2019(COVID-19)under control in China.The pairwise model,has an inherent advantage in characterizing those two NPIs than the classical well-mixed models.Therefore,in this paper,we devised a pairwise epidemic model with NPIs to analyze COVID-19 outbreak in China by using confirmed cases during February 3rde22nd,2020.By explicitly incorporating contact tracing isolation and family clusters caused by household quarantine,our model provided a good fit to the trajectory of COVID-19 infections.We calculated the reproduction number R=1.345(95%CI:1.230-1.460)for Hubei province and R=1.217(95%CI:1.207-1.227)for China(except Hubei).We also estimated the peak time of infections,the epidemic duration and the final size,which are basically consistent with real observation.We indicated by simulation that the traced high-risk contacts from incubated to susceptible decrease under NPIs,regardless of infected cases.The sensitivity analysis showed that reducing the exposure of the susceptible and increasing the clustering coefficient bolster COVID-19 control.With the enforcement of household quarantine,the reproduction number R and the epidemic prevalence declined effectively.Furthermore,we obtained the resumption time of work and production in China(except Hubei)on 10th March and in Hubei at the end of April 2020,respectively,which is broadly in line with the actual time.Our results may provide some potential lessons from China on the control of COVID-19 for other parts of the world.展开更多
Classical epidemiological models assume mass action.However,this assumption is violated when interactions are not random.With the recent COVID-19 pandemic,and resulting shelter in place social distancing directives,ma...Classical epidemiological models assume mass action.However,this assumption is violated when interactions are not random.With the recent COVID-19 pandemic,and resulting shelter in place social distancing directives,mass action models must be modified to account for limited social interactions.In this paper we apply a pairwise network model with moment closure to study the early transmission of COVID-19 in New York and San Francisco and to investigate the factors determining the severity and duration of outbreak in these two cities.In particular,we consider the role of population density,transmission rates and social distancing on the disease dynamics and outcomes.Sensitivity analysis shows that there is a strongly negative correlation between the clustering coefficient in the pairwise model and the basic reproduction number and the effective reproduction number.The shelter in place policy makes the clustering coefficient increase thereby reducing the basic reproduction number and the effective reproduction number.By switching population densities in New York and San Francisco we demonstrate how the outbreak would progress if New York had the same density as San Francisco and vice-versa.The results underscore the crucial role that population density has in the epidemic outcomes.We also show that under the assumption of no further changes in policy or transmission dynamics not lifting the shelter in place policy would have little effect on final outbreak size in New York,but would reduce the final size in San Francisco by 97%.展开更多
Separation degree is a standard measure for complex network research.Whatever its scale or its increase makes the Internet take on a complex network character.Because of the development of complex network theory and t...Separation degree is a standard measure for complex network research.Whatever its scale or its increase makes the Internet take on a complex network character.Because of the development of complex network theory and the continuous evolution of the Internet,it is a key problem that uses complex network theory to research the Internet nowadays.In this paper,the Internet separation degree is put forward.The time series stochastic process model of the Internet separation degree is established.According to ac-tual data,the Internet separation degree time sensitivity model(ISDTSM)is established and the effect of time sensi-tivity of the Internet separation degree to the Internet IP level transmission is computed.Finally the Internet separa-tion and IP transmission during 2008 Beijing Olympic Games were forecasted by using the model.展开更多
基金National Natural Science Foundation of China(No.61271152)
文摘In order to guide the construction of maintenance support system,the theory of complex network to maintenance support area is introduced,and a construction idea of maintenance support system based on complex network is put forward.Firstly,from the perspective of the similarity,maintenance support system is abstracted into complex network to form maintenance support network.Secondly,the basic concepts and parameters of maintenance support network are also introduced.Thirdly,the maintenance support system in certain period is abstracted into a maintenance support network,and the network makes some changes.Finally,the correlative parameters of the network are calculated.The results show that the changed network is more conducive to the maintenance support.This provides a new thought and method to construct maintenance support system.
基金This work is supported by Basic and Applied Basic Research Foundation of Guangdong Province(No.2020A1515011495)Guangzhou Science and Technology Foundation Project(No.202002030266).
文摘Triadic closure is a simple and fundamental kind of link formulation mechanism in network.Local closure coefficient(LCC),a new network property,is to measure the triadic closure with respect to the fraction of length-2 paths for link prediction.In this paper,a weighted format of LCC(WLCC)is introduced to measure the weighted strength of local triadic structure,and a statistic similari-ty-based link prediction metric is proposed to incorporate the definition of WLCC.To prove the metrics effectiveness and scalability,the WLCC formula-tion was further investigated under weighted local Naive Bayes(WLNB)link prediction framework.Finally,extensive experimental studies was conducted with weighted baseline metrics on various public network datasets.The results demonstrate the merits of the proposed metrics in comparison with the weighted baselines.
基金This research was funded by the National Natural Science Foundation of China(grant numbers:61873154,12022113)the Shanxi Research Project on COVID-19 epidemic control and prevention(grant number:202003D31011/GZ).
文摘Nonpharmaceutical interventions(NPIs),particularly contact tracing isolation and household quarantine,play a vital role in effectively bringing the Coronavirus Disease 2019(COVID-19)under control in China.The pairwise model,has an inherent advantage in characterizing those two NPIs than the classical well-mixed models.Therefore,in this paper,we devised a pairwise epidemic model with NPIs to analyze COVID-19 outbreak in China by using confirmed cases during February 3rde22nd,2020.By explicitly incorporating contact tracing isolation and family clusters caused by household quarantine,our model provided a good fit to the trajectory of COVID-19 infections.We calculated the reproduction number R=1.345(95%CI:1.230-1.460)for Hubei province and R=1.217(95%CI:1.207-1.227)for China(except Hubei).We also estimated the peak time of infections,the epidemic duration and the final size,which are basically consistent with real observation.We indicated by simulation that the traced high-risk contacts from incubated to susceptible decrease under NPIs,regardless of infected cases.The sensitivity analysis showed that reducing the exposure of the susceptible and increasing the clustering coefficient bolster COVID-19 control.With the enforcement of household quarantine,the reproduction number R and the epidemic prevalence declined effectively.Furthermore,we obtained the resumption time of work and production in China(except Hubei)on 10th March and in Hubei at the end of April 2020,respectively,which is broadly in line with the actual time.Our results may provide some potential lessons from China on the control of COVID-19 for other parts of the world.
基金supported by the National Natural Science Foundation of China grants 61873154 and 12101573Health Commission of Shanxi Province grants 2020XM18+4 种基金Shanxi Provincial Department of ScienceTechnology COVID-19 Emergency Special Fund grants 202003D31011/GZFundamental Research Program of Shanxi Province grants 20210302124608 and 20210302124381partially supported by a Canada Research Chair(MAL),NSERC Discovery Grants(HW and MAL),NSERC Discovery Accelerator Supplement Award(HW)an Alberta Innovates grant 202100502.
文摘Classical epidemiological models assume mass action.However,this assumption is violated when interactions are not random.With the recent COVID-19 pandemic,and resulting shelter in place social distancing directives,mass action models must be modified to account for limited social interactions.In this paper we apply a pairwise network model with moment closure to study the early transmission of COVID-19 in New York and San Francisco and to investigate the factors determining the severity and duration of outbreak in these two cities.In particular,we consider the role of population density,transmission rates and social distancing on the disease dynamics and outcomes.Sensitivity analysis shows that there is a strongly negative correlation between the clustering coefficient in the pairwise model and the basic reproduction number and the effective reproduction number.The shelter in place policy makes the clustering coefficient increase thereby reducing the basic reproduction number and the effective reproduction number.By switching population densities in New York and San Francisco we demonstrate how the outbreak would progress if New York had the same density as San Francisco and vice-versa.The results underscore the crucial role that population density has in the epidemic outcomes.We also show that under the assumption of no further changes in policy or transmission dynamics not lifting the shelter in place policy would have little effect on final outbreak size in New York,but would reduce the final size in San Francisco by 97%.
文摘Separation degree is a standard measure for complex network research.Whatever its scale or its increase makes the Internet take on a complex network character.Because of the development of complex network theory and the continuous evolution of the Internet,it is a key problem that uses complex network theory to research the Internet nowadays.In this paper,the Internet separation degree is put forward.The time series stochastic process model of the Internet separation degree is established.According to ac-tual data,the Internet separation degree time sensitivity model(ISDTSM)is established and the effect of time sensi-tivity of the Internet separation degree to the Internet IP level transmission is computed.Finally the Internet separa-tion and IP transmission during 2008 Beijing Olympic Games were forecasted by using the model.