Preliminary Study On Determination Of Fingermark Dating Based On Fingermark Composition

Fingermark dating is of great significance to the investigation of cases and the evidence of forensic science.The research on this problem has always been a difficult and hot problem to be solved in the field of forensic science.In this paper,gas chromatography-mass spectrometry(GC-MS)was used to establish a method for the detection of lipids in fingermark,to detect and analyze the differences of lipids in fingermarks,and to evaluate the repeatability and reproducibility of lipids,for building a reliable fingermark age prediction model.In this study,GC-MS was used to establish a method for the detection of lipids in fingermarks,and the pretreatment method was improved by optimizing the derivative parameters.The experimental results showed the derivatization effect was good by using 400ul5% hydrochloric acid/methanol solution,50 ℃ of derivatization temperature,20 min of derivatization time.On this basis,the composition of lipids in the fingermarks of 88 volunteers was detected,and the range of target substances used for fingermark dating analysis was defined.The results showed that squalene and fatty acids could be used to study fingermark dating.In addition,evaluate the repeatability and reproducibility of lipids in fingermarks by analyzing multiple fingermarks leaved within one day and multiple fingermarks leaved within one month.As a result,it was found that the relative standard deviation of lipid compositions in fingermarks of the same donor were within the appropriate range,and the distance of the fingermarks of the same donor were close by PCA graph.It indicated that the lipid compositions of fingermarks were reproducible and reproducible,which lied a foundation for the analysis of fingermarks dating.The change of lipid compositions in fingermarks by time were analyzed,and fingermark age prediction model was established and optimized.The results showed that squalene decreased rapidly under light condition.Within 45 days,saturation fatty acids in fingermarks showed a "wave shape" trend over time.At the same time,the models by Standard,Bagging and Boosting were established and compared under each storage condition.All the results showed that the fingermarks time prediction model based on Boosting algorithm had higher accuracy but less than 70%,which indicated limited prediction ability.In view of this,combined with individual characteristics to optimize the model through donor grouping,the accuracy of the models after optimization were greatly improved.Most of them were above 80%,the highest up to 97.2%.In this study,the basic study of fingermarks dating based on lipid compositions in fingermark residues was carried out by GC-MS.The general trend of lipid compositions over time was analyzed by detecting lipids of fingermarks with different residual time.At the same time,fingermark age prediction models with high predicted power were established by method screening and donor grouping,which provided a feasibility study for blind fingermarks prediction.