Wuxi Municipal Bureau on Science and Technology

In recent years, the proliferation of high-dimensional sparse (HiDS) data has posed significant challenges to both data analysis and privacy protection. Traditional k-means algorithms often carry the risk of privacy leakage when applied to HiDS data, while differential privacy mechanisms offer effective protection for data privacy. Addressing such issues, this paper proposes a novel differential privacy k-means clustering algorithm. This algorithm first projects the data into a low-dimensional space, then privately generates a candidate center set, and finally performs privacy-preserving clustering on the candidate set. The algorithm proposed achieves the dual objectives of privacy protection and clustering analysis for HiDS data.

ABSTRACT In the explosive environments, the electrical sparks with short circuit of capacitor output will ignite the surrounding explosive medium, and the relationship between effective ignition energy, capacitance energy conversion rate, capacitor capacitance and capacitor terminal voltage is unclear. Based on the three stages of breakdown, discharge and attenuation of the discharge characteristic curve of capacitive circuit, this paper deduces the expressions of spark resistance and discharge current including voltage, capacitance and polar distance parameters. Based on the programmable micro-nano gap discharge experimental platform, experiments under different voltages and different capacitances were carried out. It is found that the theoretical voltage-current curve and spark energy conversion rate are in good agreement with the experimental fitting curve and spark energy conversion rate. And with the increase of capacitance or voltage, the peak current and discharge time increase, which leads to the increase of energy. The conversion rate of spark energy decreases with the increase of capacitance or voltage, and the conversion rate changes between 20% and 30%. The change of spark energy conversion rate is also related to spark resistance, and with the decrease of spark resistance, the spark energy conversion rate also decreases. This study can further understand the energy variation in the discharge gap and is of great significance for the research of intrinsically safe circuits.

Background Male sex-linked Y-chromosome short tandem repeats (Y-STRs) have been widely used in forensic cases and population genetics research. At present, the forensic-related Y-STR data in the Chinese Lahu population are still poorly understood. Aim To enrich the available Y-STR data of this Chinese minority population and investigate its phylogenetic relationships with other reported populations. Subjects and methods The genetic polymorphisms of 41 Y-STR loci were analysed in 299 unrelated healthy Lahu male individuals from Southwest China. Phylogenetic analyses were performed by multidimensional scaling analysis and neighbor-joining phylogenetic tree construction. Results A total of 379 alleles were observed at the 41 Y-STR loci. The allele frequencies ranged from 0.0033 to 0.9666. The genetic diversity values ranged from 0.0653 to 0.9072. A total of 254 different haplotypes of the 41 Y-STR loci were observed in 299 individuals. The values of haplotype diversity, haplotype match probability, and discrimination capacity were 0.9987, 0.0047, and 0.8495, respectively. The phylogenetic analysis indicated that the Tibeto-Burman-speaking Lahu population showed a close genetic relationship with the Yunnan Yi population. Conclusions The haplotype data of the present study can enrich the forensic databases of this Chinese minority population and will be useful for population genetics and forensic DNA application.

Clarifying the mechanisms underlying microbial community assembly from regional microbial pools is a central issue of microbial ecology, but remains largely unexplored. Here, we investigated the gut bacterial and fungal microbiome assembly processes and potential sources in Drosophila simulans and Dicranocephalus wallichii bowringi, two wild, sympatric insect species that share a common diet of waxberry. While some convergence was observed, the diversity, composition, and network structure of the gut microbiota significantly differed between these two host species. Null model analyses revealed that stochastic processes (e.g., drift, dispersal limitation) play a principal role in determining gut microbiota from both hosts. However, the strength of each ecological process varied with the host species. Furthermore, the source‐tracking analysis showed that only a minority of gut microbiota within D. simulans and D. wallichii bowringi are drawn from a regional microbial pool from waxberries, leaves, or soil. Results from function prediction implied that host species‐specific gut microbiota might arise partly through host functional requirement and specific selection across host–microbiota coevolution. In conclusion, our findings uncover the importance of community assembly processes over regional microbial pools in shaping sympatric insect gut microbiome structure and function.

Due to high labor intensity and poor working environment, the accident risk rate of the construction industry is usually higher than that of other industries. Based on the idea that industry safety changes are affected by the internal environment and external environment, this study used the accident data and related economic data from 2003 to 2020 of the construction industry in the United States to establish predictive models respectively, and formed a combined early warning model through D-S evidence theory fusion, which contributes to building safety management. The parameters of GM(1,1) model was carried out through a trial algorithm, and it was found that the error increased by 10.48% for each increase in the number of prediction periods, and T+2 accident prediction has been selected at last. Through job object analysis, the economic environment indicators with CPI as the core were selected, and MIV was used to sort the importance of the indicators. It was found that CPI of education, CPI of transportation and employment scale were the most relevant. After removing the weaker indicators, BP neural network was established, and the error reached 1.83%. Finally, the D-S evidence theory was used for model fusion, MAPE of the D-S model reached 1.47%, and the model performance was further improved. The idea of combining the internal and external environmental impacts of the construction industry's safety will provide new ideas for the study of construction safety prediction, the introduction of MIV for variable screening is also worth learning.

Magnetic poly(styrene−methyl methacrylate-glycidyl methacrylate)/Fe3O4 nanospheres with epoxy groups were prepared by modified one-step mini-emulsion polymerization in the presence of Fe3O4 ferrofluids. The products were characterized by fourier-transform infrared spectrum, transmission electron microscopy, thermogravimetric analysis and vibrating sample magnetometer. After surface modification with diaminopropane, histidine was covalently linked to the amine group of magentic nanospheres using glutaraldehyde as crosslinker. Finally, the preliminary labeling study based on non-radioactive rhenium was carried out and the labeling efficiency reached 80.4%. Such magnetic nanospheres had promising potential in magnetically targeted radiotherapy of cancer.

A survey has been made to investigate possession of power-wasting facility and the present condition of building construction in existing village and town residential houses in southern region of Jiangsu. The results show that air-conditioners are widely used so that power-wasting is grow rapidly. In addition, because of the poor heat insulation of the house, indoor human comfort is poor. According to the house present conditions and the climatic conditions, the suited energy efficient retrofit technologies of the house enclosure are proposed.