基于GA_XGBoost模型的大学生科研能力预测问题研究

科学评价大学生科研创新能力对我国科研水平的提高具有重要意义.采用机器学习模型来预测大学生科研能力可以起到良好的效果,提出一种GAXGBoost模型来实现对大学生的科研能力预测.此模型是以Xgboost算法为基础,然后充分利用遗传算法的全局搜索能力自动搜索Xgboost最优超参数,避免了人为经验调参不准确的缺陷,最后采用精英选择策略以此确保每一轮都是最佳的进化结果.通过分析表明,所采用的GAXGBoost模型在大学生科研能力预测的结果中具有很高的精度,将此模型与Logistic Regression、Random Forest、SVM等模型进行对比,GAXGBoost模型的预测精度最高.     

Association between the microbiomes of tonsil and saliva samples isolated from pediatric patients subjected to tonsillectomy for the treatment of tonsillar hyperplasia

Oral microbes have the capacity to spread throughout the gastrointestinal system and are strongly associated with multiple diseases. Given that tonsils are located between the oral cavity and the laryngopharynx at the gateway of the alimentary and respiratory tracts, tonsillar tissue may also be affected by microbiota from both the oral cavity (saliva) and the alimentary tract. Here, we analyzed the distribution and association of the microbial communities in the saliva and tonsils of Korean children subjected to tonsillectomy because of tonsil hyperplasia (n = 29). The microbiome profiles of saliva and tonsils were established via 16S rRNA gene sequencing. Based on the alpha diversity indices, the microbial communities of the two groups showed high similarities. According to Spearman’s ranking correlation analysis, the distribution of Treponema, the causative bacterium of periodontitis, in saliva and tonsils was found to have a significant positive correlation. Two representative microbes, Prevotella in saliva and Alloprevotella in tonsils, were negatively correlated, while Treponema 2 showed a strong positive correlation between saliva and tonsils. Taken together, strong similarities in the microbial communities of the tonsils and saliva are evident in terms of diversity and composition. The saliva microbiome is expected to significantly affect the tonsil microbiome. Furthermore, we suggest that our study creates an opportunity for tonsillar microbiome research to facilitate the development of novel microbiome-based therapeutic strategies.Subject terms: Comparative genomics, Metagenomics     

Dynamic behavior and stability analysis of nonlinear modes in the fourth-order generalized Ginzburg–Landau model with near $$\mathcal{PT}$$ PT -symmetric potentials

Nonlinear Dynamics - Parametrically excited oscillators are used in several domains, in particular to improve the dynamical behaviour of systems like in the case of the parametric amplification or...     

基于深度学习的ELM实时识别研究

基于深度学习的方法，在HL-2A装置上开发出了一套边缘局域模(ELM)实时识别算法。算法使用5200次放电数据(约24.19万数据切片)进行学习，得到一个深度为22层的卷积神经网络。为衡量算法的识别能力，识别了HL-2A装置自2009年实现稳定ELMy H模放电以来所有历史数据(约26000次放电数据)，共识别出1665次H模放电，其中误识别35次，误报率为2.10%。在实际的1634次H模放电中，漏识别4次，漏识别率为0.24%。该误报率和漏报率可以满足ELM实时识别的精度要求。识别算法在实时控制环境下，对单个时间点的平均计算时间为0.46ms，可以满足实时控制的计算速度要求。     

基于时变Copula模型的 基金收益率相依关系的应用研究

基于时变Copula模型,获得预测方差,确定单个基金收益率序列的边缘分布.利用常见的静态Copula和时变Copula模型对基金收益率序列间两两相依关系进行建模并进行对比分析.应用研究表明,基于MCMC方法的时变Copula模型能更有效地度量基金收益率序列的风险.     

激光能量对反钙钛矿GaCMn_3薄膜结构与物性的影响

本文采用脉冲激光沉积方法在LaAlO3(001)单晶衬底上制备了反钙钛矿GaCMn3薄膜,通过控制制备过程中脉冲激光的能量,研究了不同激光能量条件对GaCMn3薄膜结构与物理性能的影响.分别利用X射线衍射仪、原子力显微镜、超导量子干涉仪和物理性能测试系统,对所制备的薄膜的晶体结构、表面形貌和磁性、电输运性质进行了研究.结果表明,制备的样品均为具有多个晶面取向的反钙钛矿薄膜,且薄膜结构和物性明显随制备激光能量的变化而变化.当激光能量为450mJ时,制备的薄膜多晶面取向性最弱,结晶性和表面形貌最优良.实验所得的薄膜均表现出顺磁-铁磁-反铁磁相转变,然而转变过程比块材较平缓,同时薄膜的电阻率并未表现出块材中的突变特征,我们推测该现象很可能是由衬底的应力及衬底的晶格膨胀对薄膜反常晶格变化的抑制作用造成的.