大数据中效用挖掘的快速单阶段算法

现有数据挖掘算法的缺点是在挖掘大数据时会出现大量候选模式,从而造成可伸缩性瓶颈,个别算法虽然不生成候选模式,但是计算代价高昂,缺乏有效剪裁,运行效率存在瓶颈.为此,提出一个全新的单阶段不生成候选模式的数据挖掘算法,其创新性有3点:一是基于前缀生长的模式枚举和基于效用上限值评估的剪裁策略;二是基于稀疏矩阵和虚拟投影的效用信息表达;三是节省存储空间的深度优先搜索方法.大量实验表明,新算法的时间效率比现有算法高5倍以上,并且内存使用量比现有算法少20％～60％,可伸缩性高.     

利用MR大数据提升3G用户感知研究

运用大数据挖掘原理,对海量3G终端MR数据进行深入研究。从解决用户感知、用户体验方面入手,寻找网络质量规律,并在优化实例中予以应用。丰富了优化手段,取得良好效果,提高了优化效率,为深度优化提供了新思路。     

ToxML,a data exchange standard with content controlled vocabulary used to build better (Q)SAR models

Development of accurate quantitative structure–activity relationship (QSAR) models requires the availability of high quality validated data. International regulations such as REACH in Europe will now accept (Q)SAR-based evaluations for risk assessment. The number of toxicity datasets available for those wishing to share knowledge, or to use for data mining and modelling, is continually expanding. The challenge is the current use of a multitude of different data formats. The issues of comparing or combining disparate data apply both to public and proprietary sources. The ToxML project addresses the need for a common data exchange standard that allows the representation and communication of these data in a well-structured electronic format. It is an open standard based on Extensible Markup Language (XML). Supporting information for overall toxicity endpoint data can be included within ToxML files. This makes it possible to assess the quality and detail of the data used in a model. The data file model allows the aggregation of experimental data to the compound level in the detail needed to support (Q)SAR work. The standard is published on a website together with tools to view, edit and download it.     

基于遗传编程方法的微孔磷酸铝的定向合成研究

根据微孔磷酸铝合成数据的特点,针对合成数据库中合成参数过多和交叉描述等问题,利用改进的遗传编程算法对具有(6,8)元环的微孔磷酸铝合成参数进行特征提取,优化出新的复合特征来更好地描述磷酸铝合成中溶剂和模板剂的特征,并通过参数进化过程的研究,考察了模板剂和溶剂对产物生成的具体影响,从而指导具有特定结构磷酸铝的定向合成.     

Recent Advances in Discovery of Lead Structures from Microbial Natural Products: Genomics- and Metabolomics-Guided Acceleration

Natural products (NPs) are evolutionarily optimized as drug-like molecules and remain the most consistently successful source of drugs and drug leads. They offer major opportunities for finding novel lead structures that are active against a broad spectrum of assay targets, particularly those from secondary metabolites of microbial origin. Due to traditional discovery approaches’ limitations relying on untargeted screening methods, there is a growing trend to employ unconventional secondary metabolomics techniques. Aided by the more in-depth understanding of different biosynthetic pathways and the technological advancement in analytical instrumentation, the development of new methodologies provides an alternative that can accelerate discoveries of new lead-structures of natural origin. This present mini-review briefly discusses selected examples regarding advancements in bioinformatics and genomics (focusing on genome mining and metagenomics approaches), as well as bioanalytics (mass-spectrometry) towards the microbial NPs-based drug discovery and development. The selected recent discoveries from 2015 to 2020 are featured herein.     

In Silico Analysis of P450s and Their Role in Secondary Metabolism in the Bacterial Class Gammaproteobacteria

The impact of lifestyle on shaping the genome content of an organism is a well-known phenomenon and cytochrome P450 enzymes (CYPs/P450s), heme-thiolate proteins that are ubiquitously present in organisms, are no exception. Recent studies focusing on a few bacterial species such as Streptomyces, Mycobacterium, Cyanobacteria and Firmicutes revealed that the impact of lifestyle affected the P450 repertoire in these species. However, this phenomenon needs to be understood in other bacterial species. We therefore performed genome data mining, annotation, phylogenetic analysis of P450s and their role in secondary metabolism in the bacterial class Gammaproteobacteria. Genome-wide data mining for P450s in 1261 Gammaproteobacterial species belonging to 161 genera revealed that only 169 species belonging to 41 genera have P450s. A total of 277 P450s found in 169 species grouped into 84 P450 families and 105 P450 subfamilies, where 38 new P450 families were found. Only 18% of P450s were found to be involved in secondary metabolism in Gammaproteobacterial species, as observed in Firmicutes as well. The pathogenic or commensal lifestyle of Gammaproteobacterial species influences them to such an extent that they have the lowest number of P450s compared to other bacterial species, indicating the impact of lifestyle on shaping the P450 repertoire. This study is the first report on comprehensive analysis of P450s in Gammaproteobacteria.     

基于MDARM和VxWorks的小型无人机执行器故障检测方法

由于目前在小型无人机执行器故障诊断中存在着智能化程度较低,容易受到人为因素干扰,从而出现故障漏检等问题,难以满足小型无人机对飞行安全的要求。为此,本文提出一种基于多维数据关联规则挖掘(Multidimensional Data Association Rules Mining: MDARM)和VxWorks操作系统的小型无人机执行器故障诊断方法,通过建立执行器内部传感器测量的温度、压力、流速、力矩等相关变量的历史数据库,并对这些数据进行预处理,以避免带来噪声污染,并利用可测量参数与不可测量参数之间的关联性,建立故障诊断知识库,避免了诊断过程中的人为因素干扰,实现小型无人机执行器故障的精准测量。实验结果证明,这种方法能够有效地提高故障准确率64.7%,对小型无人机执行器的智能诊提供有效指导,应用前景广阔。     

A data mining method based on constructive neural networks

~~A DATA MINING METHOD BASED ON CONSTRUCTIVE NEURAL NETWORKS[1] Jiawei Han, Micheline Kamber. Data Mining Concept and Techniques. Beijing, Higher Education Press, chapters 1 and 7. [2] Simon Haykin. Neural Networks: A Comprehensive Foundation. 2nd ed. Beijing, Tsinghua University Press, chapters 1, 4, 12 and 14. [3] Zhang Ling, Zhang Bo. A geometrical representation of McCulloch-Pitts neural model and its applications. IEEE Trans, on Neural Networks, 10(19…     

Positive Effect of Ecological Restoration with Fabaceous Species on Microbial Activities of Former Guyanese Mining Sites

Understanding ecological trajectories after mine site rehabilitation is essential to develop relevant protocols adapted for gold mining sites. This study describes the influence of a range of mine site rehabilitation and revegetation protocols on soil physicochemical parameters and microbial activities related to carbon, nitrogen and phosphorus cycles. We sampled soil from six rehabilitated mining sites in French Guiana with different plant cover (herbaceous, Cyperaceous, monoculture of Clitoria racemosa and Acacia mangium and association of C. racemosa and A. mangium). We measured the mineralization potential of organic matter by estimating the mineralization of carbon, nitrogen and phosphorus and the microbial catabolic diversity balance. The results showed an improvement in the quality of organic matter on revegetated sites with tree cover. On restored sites with fabaceous species, the microbial biomass is three times higher than non-restored sites, improving the rates of organic matter mineralization and restoring the catabolic diversity to the level of natural Guyanese soils. These results confirm that the establishment of fabaceous species under controlled conditions significantly improves the restoration of microbial communities in mining soils.     

Machine Learning Algorithm to Predict Acidemia Using Electronic Fetal Monitoring Recording Parameters

Background: Electronic fetal monitoring (EFM) is the universal method for the surveillance of fetal well-being in intrapartum. Our objective was to predict acidemia from fetal heart signal features using machine learning algorithms. Methods: A case–control 1:2 study was carried out compromising 378 infants, born in the Miguel Servet University Hospital, Spain. Neonatal acidemia was defined as pH < 7.10. Using EFM recording logistic regression, random forest and neural networks models were built to predict acidemia. Validation of models was performed by means of discrimination, calibration, and clinical utility. Results: Best performance was attained using a random forest model built with 100 trees. The discrimination ability was good, with an area under the Receiver Operating Characteristic curve (AUC) of 0.865. The calibration showed a slight overestimation of acidemia occurrence for probabilities above 0.4. The clinical utility showed that for 33% cutoff point, missing 5% of acidotic cases, 46% of unnecessary cesarean sections could be prevented. Logistic regression and neural networks showed similar discrimination ability but with worse calibration and clinical utility. Conclusions: The combination of the variables extracted from EFM recording provided a predictive model of acidemia that showed good accuracy and provides a practical tool to prevent unnecessary cesarean sections.