Study Of Epidemic Spreading And Social Propagation From The Angle Of Data

Complex network,which is a new interdisciplinary subject,has attracted the common interests of experts from different fields such as physics,mathematics,biology and computer.The approaches of complex network have been applied to study the epidemics,social propagation,traffic flow and brain sciences.The main purpose of these studies is to find effective strategies to control the spreading processes such as the epidemic spreading.Previously,the epidemics and social spreading were studied by developing the classical models,which is a reasonable but not perfect way.It doesn't work if we study new epidemics or situations beyond the models.In this thesis,before studying models,we first collect and analysis the data and then figure out the research topic,which is different from the way mentioned above,and the four works discussed in this thesis are studied based on this way.So far,experts have focused on the outbreak phase but didn't pay enough attention to the ending phase.Without the help of the obtained results in ending phase,governments can't make effective strategies when facing the outbroken COVID-19 pandemic.In this work,we study not only the outbreak phase but also the ending phase of epidemics.Because of the similarities between epidemic spreading and social propagation,we also study the social spreading in this thesis.By studying the flu data in Hongkong and measles data in New York and Baltimore,we have finished two works to reveal the asymmetric phenomenon of the outbreak peocess and the end process of epidemic,and theoretically analyze its mechanism from the perspective of hysteresis loop,and explain the paradox of the dependence of hysteresis loop area on parameters in state space and parameter space.In detail,by studying the flu and measles data,we find that the spread process of recurrent epidemics can be divided into two processes,including outbreak process and ending process.These two processes are asymmetry,thus forming a hysteresis loop.In the second work,by changing the infectious rate before system reaching the stable status,we show the hysteresis loop both in parameter and status spaces,study the parameters that influence the area of hysteresis loop,and find the dependence of the area of hysteresis loop on the parameters are different in status and parameter spaces,indicating a paradox.We provide a solution for this paradox,i.e.changing the parameter and useing fixed revolution time.Referring to the theory of epidemic,we have developed a threoretical method to calculate the effective reproduction number in rumor model in the third work.Our method only depends on the time series of the daily number of new participants in the topic discusstion,and it is verified by microblog data.In detail,by studying Weibo data,we get the time series of daily number of new participants,named??,and find that??curves are different.Basically,we calculate the effective reproduction numberand use it to judge the spread potential.In this work,we develop a theoretical method only needed to know??to calculate the,and all the other parameters can be derived.The core part of our method are as follows:bases on the Bayesian method,??obeys a Poisson distribution related to the effective reproduction numberand distribution of generation time???.By using Daley-Kendall model and keep track of all nodes'information,we find that???obeys a negative binomial distribution.Knowing??and???,we can use the Markov Chain Monte Carlo method to calculate.Moreover,we use the method to calculateand use it to judge the spreading potential.Based on SARS-Co V-2 related data,we have studied the impact of susceptibility on the transmission in the fourth work.In detail,by studying the COVID-19 data reported in Wuhan and Shanghai,we find the age and sex profile and calculate the rate of different age and sex people showing symptom.Using the SIR model with age and sex structure and contact matrices before and after the COVID-19 in Wuhan and Shanghai,we study the influence of susceptibility and change of contact matrix to spread of COVID-19.Our work shows that there is an apparent age and sex different in population,namely the older people have higher rate to get the COVID-19.Our work may help to make more effective strategies.In conclusion,we have studied the outbreak and ending phases of epidemics and social spreading from the perspective view of data,and our work may provide a new idea to study them.Although we have gotten some interesting results and solved some problems,there are still some questions needed to be studied and solved,and we should try our best to face the challenges in the future.