课程文献中心 更多>>

关键词： classification   correlation   metagenomics   set cover   taxonomic annotation   SEQUENCE   DATABASE   ALGORITHMS  

摘要： The classification of reads from a metagenomic sample using a reference taxonomy is usually based on first mapping the reads to the reference sequences and then classifying each read at a node under the lowest common ancestor of the candidate sequences in the reference taxonomy with the least classification error. However, this taxonomic annotation can be biased by an imbalanced taxonomy and also by the presence of multiple nodes in the taxonomy with the least classification error for a given read. In this article, we show that the Rand index is a better indicator of classification error than the often used area under the receiver operating characteristic (ROC) curve and F-measure for both balanced and imbalanced reference taxonomies, and we also address the second source of bias by reducing the taxonomic annotation problem for a whole metagenomic sample to a set cover problem, for which a logarithmic approximation can be obtained in linear time and an exact solution can be obtained by integer linear programming. Experimental results with a proof-of-concept implementation of the set cover approach to taxonomic annotation in a next release of the TANGO software show that the set cover approach further reduces ambiguity in the taxonomic annotation obtained with TANGO without distorting the relative abundance profile of the metagenomic sample.

关键词： Face recognition   High dimensional covariance estimation   Monte Carlo simulations   Quadratic discriminant classifier   Shrinkage estimator   COVARIANCE   EIGENFACES   ALGORITHMS   MATRICES   PATTERN  

摘要： The quadratic discriminant classifier (QDC) is a well-known parametric Bayesian classifier that has been successfully applied to statistical pattern recognition problems. One such application is in automatic face recognition where the number of training images per subject is often found to be much less than the length of the facial features. In such a case, the QDC cannot be used because the class-specific covariance matrix on which it depends is either poorly estimated or singular thereby resulting in unacceptable classifier performance. High dimensional covariance estimation techniques such as shrinkage can alleviate this problem but only to a certain extent. This paper presents a computationally simple yet effective solution for further improving the QDC performance in small sample size scenarios. The proposed technique adopts a strategy of combining the class-specific shrinkage estimates of the covariance matrix to obtain a pooled shrinkage estimate, which is then plugged into the QDC. Experiments indicate that the proposed classifier leads to remarkable improvement in face recognition accuracy as compared to the existing classifiers such as the nearest neighbor, support vector machine and naive Bayes, irrespective of the nature of the database and feature extraction method. Monte Carlo simulations reveal that this improvement is due to the much lower mean squared error of the pooled shrinkage estimator which offers greater stability to the QDC.

关键词： Aerobiology   DNA extraction   Cloning   Sanger-sequencing   trnL   GREECE  

摘要： Metabarcoding is a promising DNA-based method for identifying airborne pollen from environmental samples with advantages over microscopic methods. Sample preparation and DNA extraction are of fundamental importance for obtaining an optimal DNA yield. Currently, there is no standard procedure for these steps, especially for gravimetric pollen samplers. Therefore, the aim of this study was to develop protocols for processing environmental samples for pollen DNA extraction and for metabarcoding analysis and to assess the efficacy of these protocols for the taxonomic assignment of airborne pollen collected by gravimetric (Tauber trap) and volumetric (Hirst-type trap) samplers. Protocols were tested across an increasing complexity of samples, from pure single-species pollen to environmental multi-species samples. A short fragment (about 150 base pairs) of the chloroplast trnL gene was amplified using universal primers for plants. After PCR amplification, amplicons were Sanger-sequenced and taxonomic assignment was accomplished by comparison with a custom-made reference database including chloroplast DNA sequences from most of the anemophilous taxa occurring in the study area (Trentino, northern Italy), representing 46 plant families. Using the classical morphological pollen analysis as a benchmark, we show that DNA metabarcoding is efficient and applicable even in complex samples, provided that protocols for sample preparation, DNA extraction, and metabarcoding analysis are carefully optimized.

关键词： Multi-label   active learning   image classification   sample noise   label correlation   mapping relation   automatic labeling  

摘要： Multi-label active learning for image classification has been a popular research topic. It faces several challenges, even though related work has made great progress. Existing studies on multi-label active learning do not pay attention to the cleanness of sample data. In reality, data are easily polluted by external influences that are likely to disturb the exploration of data space and have a negative effect on model training. Previous methods of label correlation mining, which are purely based on observed label distribution, are defective. Apart from neglecting noise influence, they also cannot acquire sufficient relevant information. In fact, they neglect inner relation mapping from example space to label space, which is an implicit way of modeling label relationships. To solve these issues, we develop a novel multi-label active learning with low-rank application (ENMAL) algorithm in this paper. A low-rank model is constructed to quantize noise level, and the example-label pairs that contain less noise are emphasized when sampling. A low-rank mapping matrix is learned to signify the mapping relation of a multi-label domain to capture a more comprehensive and reasonable label correlation. Integrating label correlation with uncertainty and considering sample noise, an efficient sampling strategy is developed. We extend ENMAL with automatic labeling (denoted as AL-ENMAL) to further reduce the annotation workload of active learning. Empirical research demonstrates the efficacy of our approaches.

关键词： forensic science   fire debris analysis   gasoline   alkylate   classification   ignitable liquids   gas chromatography-mass spectrometry   MICROBIAL-DEGRADATION   PYROLYSIS PRODUCTS   HYDROCARBONS  

摘要： The characteristic that discriminates gasoline from other ignitable liquids is that it contains high-octane blending components. This study elaborates on the idea that the presence of gasoline in fire debris samples should be based on the detection of known high-octane blending components. The potential of the high-octane blending component alkylate as a characteristic feature for gasoline detection and identification in fire debris samples is explored. We have devised characteristic features for the detection of alkylate and verified the presence of alkylate in a large collection of gasoline samples from petrol stations in the Netherlands. Alkylate was detected in the vast majority of the samples. It is demonstrated that alkylate can be detected in fire debris samples that contain traces of gasoline by means of routine GC-MS methods. Detection of alkylate, alongside other gasoline blend components, results in a more solid foundation for gasoline detection and identification in fire debris samples.

关键词： LIBS   Rice geographic origin   Sample preparation methods   SVM   IDENTIFICATION   CHEMOMETRICS   IMPROVEMENT   STEEL  

摘要： The quality and safety of food is one of the most important issues in our life. In this work, four different sample preparation methods, i.e., rice powder pellet with boric acid (RPPBA), rice powder pellet (RPP), rice grain pellet (RGP) and rice grain (RG), were carried out to study the adulteration problem in food industry. 20 kinds of rice from different geographic origins were classified by laser-induced breakdown spectroscopy (LIBS) coupled with principal component analysis (PCA) and support vector machine (SVM). PCA was used to reduce the input variables of SVM, and the classification accuracies by PCA and SVM combination for the four sample preparation methods were 92.70%, 95.70%, 98.80%, and 99.20%, respectively. In addition, the sample preparation times were 15, 12, 10, and 1 min, respectively. These results show that RG was simpler and more efficient sample preparation method for distinguishing different geographical origin of agricultural products than the other preparing methods of RPPBA, RPP, and RG. Modeling efficiency of SVM could be improved by reducing its input variables using PCA. It can be concluded that the LIBS technique combined with chemometric method should be a promising tool to rapidly distinguish different rice geographic origins. (C) 2018 Elsevier Ltd. All rights reserved.