Big Data Analysis Of Cases Of Death In Chongqing And Its Application In Forensic Medicine

Cases of death is a kind of cases involving personal death including homicide,suicide and accident.The study on cases of death mainly includes two objectives,to help achieving the criminal scene reconstruction to solve the cases and to summarize the spatial and time patterns to prevent such cases.Traditionally,the analyses of criminal scene reconstruction which usually involved massive amount of environmental information was dependent on human analysis.The human analyses generally rely on the academic level and experience of researcher.In recent years,following the rapid development of the artificial intelligence(AI)after the appearance of artificial neural network(ANN),AI can help modeling and optimizing of complex system.The application of AI in criminal scene reconstruction could help decreasing the mistakes caused by researchers.Furthermore,the AI also can predict the tendency of such cases according to established models.This study was based on the theory of forensic sceneology and focused on homicide and suicide scene.In recent years,the breakthrough on application of AI in solving criminal scene reconstruction still not reached.The question of how to establish big data database and choose arithmetic is waiting to solved.This study was based on forensic sceneology and focused on suicide and homicide.A certain series of index was chosen to analysis the criminal scene and build the big data database to explore the possibility of AI system which may eventually be applied in the criminal scene reconstruction.This study confirmed that the information in criminal scene can be digitized and used to provide a new method to analyze and infer relevant cases and predict future trends.This study can provide the train of thought and theoretical foundation for the development of AI system on the cases of death.This study consists of two parts.The first part of the system review of the homicide happened from 2015 to 2017 in Chongqing.In total of 19 descriptive indexes were found to be able to describe the criminal scene of the cases of death.The concept of Moving Line was introduced on the digitization of the 19 descriptive indexes and the definition,classification,and assignment of 6 descriptive indexes including the blood distribution,scene location,scene entrance,scene export,body posture and loss of property.Other indexes were assigned according to forensic field theory.After all the indexes have been digitized and assigned,we analyzed the data with principal factor analysis and logistic regression analysis and established a model to predict the type of case,the age and educate background of criminal.We also confirm the feasibility of the method of digitization applied to the 19 indexes.Otherwise,this study also introduced a new visual method to describe the characteristics of body injury on victims.The second part of this study was to provide spatial and time analysis to the suicide cases which happen from 2013 to 2017 in Chongqing city,and sum up the characteristic to predict trends in similar cases in the future.The spatial analysis of suicide cases was conducted by two methods.Ordinary least squares was used to analysis the geographical relationship between the suicide and the social factor and the nuclear density was used to evaluate the geographical distribution features of suicide.In the time analysis,time series analysis was used to model and predict the months and moments of suicide cases.The result shows that there are independent of suicide and social factor in geography.The result also showed the difference among jumping,hanging,drowning and poisoning.The jumping and hanging were generally convergent,while the poisoning and drowning were sporadic.In addition,we also found that most suicide happed in home.This study found that the suicide happens in different month and moment with serials property,but without seasonality.Through the analysis of homicide cases and suicide cases,this study proposed a method for establishing a big database of AI systems for criminal scene reconstruction and a basic algorithm for predicting future trends of cases of death.