Statistical Classification Procedures Research Articles

WorldView-2 Satellite Image Classification using U-Net Deep Learning Model

Land cover maps are important documents for local governments to perform urban planning and management. A field survey using measuring instruments can produce an accurate land cover map. However, this method is time-consuming, expensive, and labor-intensive. A number of researchers have proposed using remote sensing, which generates land cover maps using an optical satellite image with various statistical classification procedures. Recently, artificial intelligence (AI) technology, such as deep learning, has been used in multiple fields, including satellite image classification, with satisfactory results. In this study, a WorldView-2 image of Terangun in Aceh Province, which was acquired on Aug 2, 2016, was classified using a commonly used deep-learning-based classification, namely, U-net. There were eight classes used in the experiment: building, road, open land (such as green open space, bare land, grass, or low vegetation), river, farm, field, aquaculture pond, and garden. For comparison, three classification methods: maximum-likelihood, random forest, and support vector machine, were performed compared to U-Net. A land cover map provided by the government was used as a reference to evaluate the accuracy of land cover maps generated using two classification methods. The results with 100 randomly selected pixels revealed that U-Net was able to obtain a 72% and 0.585 for overall and kappa accuracy, respectively; whereas, overall accuracy and kappa accuracy for the maximum likelihood, random forest and support vector machine methods were 49% and 0.148; 59% and 0.392; and 67% and 0. 511; respectively. Therefore, U-Net outperformed those three of classification methods in classifying the image.

What's ‘SUP’ with paddlers? Integrating spatial, social, and ecological data to understand behavior among paddlesport users at a popular lake destination

Non-motorized, water-based recreation (e.g., paddlesports) represents one of the most sought after outdoor activities in U.S. Parks and Protected Areas (PPA). Among paddling options, stand-up paddleboarding represents the activity with the highest levels of growth. However, stand-up paddleboarding has yet to be spatially examined in a PPA setting. This research paired GPS-tracking data of paddlesport users with corresponding social and environmental information to build a robust understanding of paddlesport behavior and experience in PPA. Statistical classification procedures revealed distinctions in movement across paddling activity types, highlighting implications for resource protection and visitor flow management. Integrating spatial and non-spatial data identified numerous drivers and impacts of paddlesport use: the motivation to escape corresponded to traveling further distances, while higher group sizes and prolonged shoreline exposure aligned with concentrated use levels near parking lots and facilities. Results broaden our knowledge of recreationist movement and experience, and add to a growing body of PPA research that incorporates mixed method spatial approaches to research designs. Furthermore, these findings contribute novel information on paddlesport spatial behavior in PPAs, especially given the emergent introduction of stand-up paddleboarding.

МЕТОД АВТОМАТИЧЕСКОЙ КЛАСТЕРИЗАЦИИ ДАННЫХ ДИСТАНЦИОННОГО ЗОНДИРОВАНИЯ

The subject matter of the article is the methods of automatic clustering of remote sensing data under conditions of a priori uncertainty regarding the number of observed object classes and the statistical characteristics of the signatures of classes. The aim is to develop a method for approximating multimodal empirical distributions of observational data to construct decision rules for pixel-by-pixel statistical classification procedures, as well as to investigate the effectiveness of this method for automatically classifying objects on synthesized and real images. The tasks to be solved are: to develop and implement a procedure for splitting a mixture of basic distributions, while ensuring the following requirements: the absence of a preliminary data analysis stage in order to select optimal initial approximations; a good convergence of the method and the ability to automatically refine the list of classes by combining indistinguishable or poorly distinguishable components of the mixture into a single cluster; to synthesize test images with a specified number of objects and known data distributions for each object; to evaluate the effectiveness of the developed method for automatic classification by the criterion of the probability of correct recognition; to evaluate the results of automatic clustering of real images. The methods used are methods of stochastic simulation, methods of approximation of empirical distributions, statistical methods of recognition, methods of probability theory and mathematical statistics. The following results have been obtained. A method for automatic splitting of a mixture of Gaussian distributions to construct decision thresholds according to the maximal a posteriori probability criterion was proposed. The results of the automatic forming the list of classes and their probabilistic descriptions, as well as the results of the clustering both test images and satellite ones are given. It is shown that the developed method is quite effective and can be used to determine the number of objects’ classes as well as their stochastic characteristics’ mathematical description for pattern recognition tasks and cluster analysis. Conclusions. The scientific novelty of the results obtained is that the proposed approach makes it possible directly during the “unsupervised” training procedure to evaluate the distinguishability of classes and exclude indistinguishable objects from the list of classes.

Exploring the Sensitivity of Choropleths under Attribute Uncertainty.

The choropleth map is an essential tool for spatial data analysis. However, the underlying attribute values of a spatial unit greatly influence the statistical analyses and map classification procedures when generating a choropleth map. If the attribute values incorporate a range of uncertainty, a critical task is determining how much the uncertainty impacts both the map visualization and the statistical analysis. In this paper, we present a visual analytics system that enhances our understanding of the impact of attribute uncertainty on data visualization and statistical analyses of these data. Our system consists of a parallel coordinates-based uncertainty specification view, an impact river and impact matrix visualization for region-based and simulation-based analysis, and a dual-choropleth map and t-SNE plot for visualizing the changes in classification and spatial autocorrelation over the range of uncertainty in the attribute values. We demonstrate our system through three use cases illustrating the impact of attribute uncertainty in geographic analysis.

Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data

BackgroundThe accuracy of microbial community surveys based on marker-gene and metagenomic sequencing (MGS) suffers from the presence of contaminants—DNA sequences not truly present in the sample. Contaminants come from various sources, including reagents. Appropriate laboratory practices can reduce contamination, but do not eliminate it. Here we introduce decontam (https://github.com/benjjneb/decontam), an open-source R package that implements a statistical classification procedure that identifies contaminants in MGS data based on two widely reproduced patterns: contaminants appear at higher frequencies in low-concentration samples and are often found in negative controls.ResultsDecontam classified amplicon sequence variants (ASVs) in a human oral dataset consistently with prior microscopic observations of the microbial taxa inhabiting that environment and previous reports of contaminant taxa. In metagenomics and marker-gene measurements of a dilution series, decontam substantially reduced technical variation arising from different sequencing protocols. The application of decontam to two recently published datasets corroborated and extended their conclusions that little evidence existed for an indigenous placenta microbiome and that some low-frequency taxa seemingly associated with preterm birth were contaminants.ConclusionsDecontam improves the quality of metagenomic and marker-gene sequencing by identifying and removing contaminant DNA sequences. Decontam integrates easily with existing MGS workflows and allows researchers to generate more accurate profiles of microbial communities at little to no additional cost.

A GPS-Based Classification of Visitors’ Vehicular Behavior in a Protected Area Setting

As tourism and visitation increases to parks and protected areas worldwide, concerns regarding degradation of natural resources and visitor experiences also increase. As such, understanding visitor use patterns and spatial characteristics is important for better management of social and ecological resources. In many parks and protected areas, transportation systems (including park roads, parking lots, personal vehicles, public transit services, bicycle and/or pedestrian paths, and Intelligent Transportation Systems (ITS) that help deliver visitors to their destinations within parks) are a primary way for tourists and visitors to interact with and experience natural environments. Transportation systems designed with visitor use patterns in mind can both provide a positive experience for tourists as well as minimize impacts to natural resources within parks and protected areas. Our research used Global Positioning System (GPS)-based technology and a statistical classification procedure to examine spatial and temporal patterns of vehicular visitation in the Moose-Wilson corridor of Grand Teton National Park. While GPS-based technology has been used in parks and protected areas to study the spatial behavior of hikers and bikers, it has not been widely utilized to study the spatial behavior patterns of visitors travelling in vehicles to park destinations. Similarly, statistical classifications have been used to understand visitor behavior patterns in parks and protected areas, but are rarely applied to spatial data such as GPS tracks of visitor spatial movements. This classification based on spatial behavior resulted in the categorization of visitors into three types: Opportunistic Commuters, Wildlife/Scenery Viewers, and Hikers. These results expand existing literature suggesting that interpretable patterns of visitor spatial movements can result from classification studies using GPS-based collection of visitor behavior data. Understanding these patterns of visitor behavior is important for the management of social and natural resource conditions at park destinations as well as visitor experiences within the transportation corridor, as different types of visitors often exhibit different spatial behavior patterns. The classification and associated patterns of visitor behavior have important implications for managers of parks and protected area as different types of visitors may require different transportation and information needs in order to achieve their desired experience and minimize ecological impacts.Subscribe to JPRA

A Simple Algorithm for Population Classification.

A single-nucleotide polymorphism (SNP) is a variation in the DNA sequence that occurs when a single nucleotide in the genome differs across members of the same species. Variations in the DNA sequences of humans are associated with human diseases. This makes SNPs as a key to open up the door of personalized medicine. SNP(s) can also be used for human identification and forensic applications. Compared to short tandem repeat (STR) loci, SNPs have much lower statistical testing power for individual recognition due to the fact that there are only 3 possible genotypes for each SNP marker, but it may provide sufficient information to identify the population to which a certain samples may belong. In this report, using eight SNP markers for 641 samples, we performed a standard statistical classification procedure and found that 86% of the samples could be classified accurately under a two-population model. This study suggests the potential use of SNP(s) in population classification with a small number (n ≤ 8) of genetic markers for forensic screening, biodiversity and disaster victim controlling.

Effective Statistical Detection of Smart Confidentiality Attacks in Multi-Domain Networks

The need to preserve information confidentiality among network providers has prevented the actual deployment of effective Traffic Engineering (TE) solutions for QoS-enabled end-to-end connectivity services in multi-domain multi-provider networks. The use of Path Computation Element (PCE) architecture can foster the effective implementation of TE through a centralized engine devoted to end-to-end path computations. However, despite authentication, authorization and encryption, confidentiality issues may arise due to abuses of information included in path computation replies to bogus requests issued by malicious PCEs. This paper first demonstrates the security leak allowing the exposure of intra-domain information in current inter-PCE path computation procedures. Then it proposes an anomaly-based approach, namely PCE Anomaly Detector (PAD) in order to detect malicious utilization of path computation services. The proposed PAD employs a novel double-step multi-dimensional formulation based on the Sequential Hypothesis Testing (SHT) statistical classification procedure, able to recognize a suspicious sequence of requests while aiming at inferring confidential information about other domains. PAD is extensively evaluated through simulations. Results show good performance in terms of detection capabilities while guaranteeing the trade-off between accuracy and responsiveness, minimizing false alarm occurrences. Robustness against smart attacks is also proved with respect to a comprehensive set of attack patterns and under different network load conditions. Finally, intra-domain information exposition is evaluated, showing the PAD ability to preserve confidentiality.

When voices get emotional: A corpus of nonverbal vocalizations for research on emotion processing

Nonverbal vocal expressions, such as laughter, sobbing, and screams, are an important source of emotional information in social interactions. However, the investigation of how we process these vocal cues entered the research agenda only recently. Here, we introduce a new corpus of nonverbal vocalizations, which we recorded and submitted to perceptual and acoustic validation. It consists of 121 sounds expressing four positive emotions (achievement/triumph, amusement, sensual pleasure, and relief) and four negative ones (anger, disgust, fear, and sadness), produced by two female and two male speakers. For perceptual validation, a forced choice task was used (n = 20), and ratings were collected for the eight emotions, valence, arousal, and authenticity (n = 20). We provide these data, detailed for each vocalization, for use by the research community. High recognition accuracy was found for all emotions (86%, on average), and the sounds were reliably rated as communicating the intended expressions. The vocalizations were measured for acoustic cues related to temporal aspects, intensity, fundamental frequency (f0), and voice quality. These cues alone provide sufficient information to discriminate between emotion categories, as indicated by statistical classification procedures; they are also predictors of listeners' emotion ratings, as indicated by multiple regression analyses. This set of stimuli seems a valuable addition to currently available expression corpora for research on emotion processing. It is suitable for behavioral and neuroscience research and might as well be used in clinical settings for the assessment of neurological and psychiatric patients. The corpus can be downloaded from Supplementary Materials.

Statistical Evaluation of Diagnostic Tests

The use of routine laboratory tests in diagnosing disease is becoming of increasing importance. This emphasizes to test the efficiency of diagnostic tests, since relatively few diagnostic tests correctly classify all subjects tested as diseased or well. The more usual situation is one in which some well subjects are classified as diseased and some diseased subjects classified as well. In this type of situation, Diagnostics and prognostic models serve the purpose. Diagnostic models are usually used for classification and quite commonly used in medical field. In this paper, importance of statistical classification procedures are highlighted which helps in the evaluation of diagnostic tests.

Assessing the power of informative subsets of loci for population assignment: standard methods are upwardly biased

It is well known that statistical classification procedures should be assessed using data that are separate from those used to train the classifier. This principle is commonly overlooked when the classification procedure in question is population assignment using a set of genetic markers that were chosen specifically on the basis of their allele frequencies from amongst a larger number of candidate markers. This oversight leads to a systematic upward bias in the predicted accuracy of the chosen set of markers for population assignment. Three widely used software programs for selecting markers informative for population assignment suffer from this bias. The extent of this bias is documented through a small set of simulations. The relative effect of the bias is largest when screening many candidate loci from poorly differentiated populations. Simple unbiased methods are presented and their use encouraged.

Unsupervised clustering of wildlife necropsy data for syndromic surveillance

BackgroundThe importance of wildlife disease surveillance is increasing, because wild animals are playing a growing role as sources of emerging infectious disease events in humans. Syndromic surveillance methods have been developed as a complement to traditional health data analyses, to allow the early detection of unusual health events. Early detection of these events in wildlife could help to protect the health of domestic animals or humans. This paper aims to define syndromes that could be used for the syndromic surveillance of wildlife health data. Wildlife disease monitoring in France, from 1986 onward, has allowed numerous diagnostic data to be collected from wild animals found dead. The authors wanted to identify distinct pathological profiles from these historical data by a global analysis of the registered necropsy descriptions, and discuss how these profiles can be used to define syndromes. In view of the multiplicity and heterogeneity of the available information, the authors suggest constructing syndromic classes by a multivariate statistical analysis and classification procedure grouping cases that share similar pathological characteristics.ResultsA three-step procedure was applied: first, a multiple correspondence analysis was performed on necropsy data to reduce them to their principal components. Then hierarchical ascendant clustering was used to partition the data. Finally the k-means algorithm was applied to strengthen the partitioning. Nine clusters were identified: three were species- and disease-specific, three were suggestive of specific pathological conditions but not species-specific, two covered a broader pathological condition and one was miscellaneous. The clusters reflected the most distinct and most frequent disease entities on which the surveillance network focused. They could be used to define distinct syndromes characterised by specific post-mortem findings.ConclusionsThe chosen statistical clustering method was found to be a useful tool to retrospectively group cases from our database into distinct and meaningful pathological entities. Syndrome definition from post-mortem findings is potentially useful for early outbreak detection because it uses the earliest available information on disease in wildlife. Furthermore, the proposed typology allows each case to be attributed to a syndrome, thus enabling the exhaustive surveillance of health events through time series analyses.

Strategies Pursued By Swiss Firms in Investing in R&D at Foreign Locations: An Empirical Analysis Based on Firm-Level Data

The aim of this paper is twofold: Firstly, we try to identify and characterise different of strategies firms pursue in performing foreign R&D. Secondly, it is analysed whether the of R&D strategies we identified in the first part differ in terms of their impact on firm productivity. In order to identify foreign R&D strategies we perform, in a first step, a non-hierarchical cluster analysis of data on the firms' motives for investing abroad in R&D. In a second step, we characterise these clusters by use of a large number of variables that, according to the well-known OLI paradigm, determine a firm's FDI in distribution, manufacturing and R&D. In this way, we can check whether the clusters identified by applying a (purely) statistical classification procedure effectively may be interpreted as types of foreign R&D strategies. We end up with four of strategies, which significantly differ in terms of characteristics that are important according to the OLI approach. In the second part we estimate a production function where the standard factor inputs are complemented by domestic R&D and strategy-specific foreign R&D. It turns out that only one of the four strategies exerts a positive influence on firm productivity. However, it cannot be excluded that some of the other strategies have, in the longer run, a positive productivity effect as well. The paper also finds that foreign and domestic R&D, on balance, are complements.

직교요인을 이용한 국소선형 로지스틱 마이크로어레이 자료의 판별분석

본 논문에서는 마이크로어레이 (microarray) 자료에 판별분석을 적용 시 나타나는 고차원 및 소표본 문제의 해결방법으로서 직교요인을 새로운 특징변수로 사용한 비모수적 국소선형 로지스틱 판별분석을 제안한다. 제안된 방법은 국소우도에 기반한 것으로서 다범주 판별분석에 적용될 수 있으며, 고려된 직교인자는 주성분 요인, 부분최소제곱 요인, 인자분석 요인 등이다. 대표적인 두 가지 실제 마이크로어레이 자료에 적용한 결과 직교요인들 중에서 부분최소제곱 요인을 특징변수로 사용한 경우 고전적인 통계적 판별분석보다 향상된 분류 능력을 나타내고 있음을 확인하였다. The number of variables exceeds the number of samples in microarray data. We propose a nonparametric local linear logistic classification procedure using orthogonal components for classifying high-dimensional microarray data. The proposed method is based on the local likelihood and can be applied to multi-class classification. We applied the local linear logistic classification method using PCA, PLS, and factor analysis components as new features to Leukemia data and colon data, and compare the performance of the proposed method with the conventional statistical classification procedures. The proposed method outperforms the conventional ones for each component, and PLS has shown best performance when it is embedded in the proposed method among the three orthogonal components.

Evaluation of Neural-network Classifiers for Weed Species Discrimination

The potential environmental impact of herbicide utilisation has stimulated research into new methods of weed control. Selective herbicide application offers the possibility to reduce herbicide usage while maintaining weed control. The research reported utilised colour co-occurrence method (CCM) texture analysis techniques to evaluate three different neural-network classifiers for potential use in real-time weed control systems. An image data set consisting of 33 unique texture features for each image in a six class data set (40 images per class) was generated for the following classes; foxtail, crabgrass, common lambsquarter, velvetleaf, morning-glory, and clear soil surface. The data sets were evaluated using a stepwise variable selection procedure to provide six unique data models separated according to hue-saturation-intensity colour features. A comparison study of the classification capabilities of three neural-network models (backpropagation, counterpropagation, and radial basis function) was conducted. It was found that the backpropagation neural-network classifier provided the best classification performance and was capable of classification accuracies of 97%, which exceeded traditional statistical classification procedure accuracy of 93%. When comparing the three neural-network methodologies, the backpropagation method not only achieved a higher classification accuracy, but also had less computational requirements.

A pilot study on the application of statistical classification procedures to molecular epidemiological data

The development of new statistical methods for use in molecular epidemiology comprises the building and application of appropriate classification rules. The aim of this study was to assess various classification methods that can potentially handle genetic interactions. A data set comprising genotypes at 25 single nucleotide polymorphic (SNP) loci from 518 breast cancer cases and 586 age-matched population-based controls from the GENICA study was used to built a classification rule with the discrimination methods SVM (support vector machine), CART (classification and regression tree), Bagging, Random Forest, LogitBoost and k nearest neighbours (kNN). A blind pilot analysis of the genotypic data set was a first approach to obtain an impression of the statistical structure of the data. Furthermore, this analysis was performed to explore classification methods that may be applied to molecular-epidemiological evaluation. The results showed that all blindly applied classification methods had a slightly smaller misclassification rate than a random classification. The findings, nevertheless, suggest that SNP data might be useful for the classification of individuals into categories of high or low risk of diseases.

A pilot study on the application of statistical classification procedures to molecular epidemiological data

A pilot study on the application of statistical classification procedures to molecular epidemiological data

Tissue counter analysis of histologic sections of melanoma: influence of mask size and shape, feature selection, statistical methods and tissue preparation.

Background: Tissue counter analysis is an image analysis tool designed for the detection of structures in complex images at the macroscopic or microscopic scale. As a basic principle, small square or circular measuring masks are randomly placed across the image and image analysis parameters are obtained for each mask. Based on learning sets, statistical classification procedures are generated which facilitate an automated classification of new data sets. Objective: To evaluate the influence of the size and shape of the measuring masks as well as the importance of feature selection, statistical procedures and technical preparation of slides on the performance of tissue counter analysis in microscopic images. As main quality measure of the final classification procedure, the percentage of elements that were correctly classified was used. Study design: H&E‐stained slides of 25 primary cutaneous melanomas were evaluated by tissue counter analysis for the recognition of melanoma elements (section area occupied by tumour cells) in contrast to other tissue elements and background elements. Circular and square measuring masks, various subsets of image analysis features and classification and regression trees compared with linear discriminant analysis as statistical alternatives were used. The percentage of elements that were correctly classified by the various classification procedures was assessed. In order to evaluate the applicability to slides obtained from different laboratories, the best procedure was automatically applied in a test set of another 50 cases of primary melanoma derived from the same laboratory as the learning set and two test sets of 20 cases each derived from two different laboratories, and the measurements of melanoma area in these cases were compared with conventional assessment of vertical tumour thickness. Results: Square measuring masks were slightly superior to circular masks, and larger masks (64 or 128 pixels in diameter) were superior to smaller masks (8 to 32 pixels in diameter). As far as the subsets of image analysis features were concerned, colour features were superior to densitometric and Haralick texture features. Statistical moments of the grey level distribution were of least significance. CART (classification and regression tree) analysis turned out to be superior to linear discriminant analysis. In the best setting, 95% of melanoma tissue elements were correctly recognized. Automated measurement of melanoma area in the independent test sets yielded a correlation of r=0.846 with vertical tumour thickness (p < 0.001), similar to the relationship reported for manual measurements. The test sets obtained from different laboratories yielded comparable results. Conclusions: Large, square measuring masks, colour features and CART analysis provide a useful setting for the automated measurement of melanoma tissue in tissue counter analysis, which can also be used for slides derived from different laboratories.

Rough Sets and Multivariate Statistical Classification: A Simulation Study

The classification of a set of objects into predefined homogenous groups is a problem with major practical interest in many fields. Over the past two decades several non-parametric approaches have been developed to address the classification problem, originating from several scientific fields. This paper is focused on the rough sets approach and the investigation of its performance as opposed to traditional multivariate statistical classification procedures, namely the linear discriminant analysis, the quadratic discriminant analysis and the logit analysis. For this purpose an extensive Monte Carlo simulation is conducted to examine the performance of these methods under different data conditions.

Role and Results of statistical methods in protein fold class prediction

Statistical methods of discrimination and classification are used for the prediction of protein structure from amino acid sequence data. This provides information for the establishment of new paradigms of carcinogenesis modeling on the basis of gene expression. Feed forward neural networks and standard statistical classification procedures are used to classify proteins into fold classes. Logistic regression, additive models, and projection pursuit regression from the family of methods based on a posterior probabilities; linear, quadratic, and a flexible discriminant analysis from the class of methods based on class conditional probabilities, and the nearest-neighbors classification rule are applied to a data set of 268 sequences. From analyzing the prediction error obtained with a test sample ( n = 125) and with a cross validation procedure, we conclude that the standard linear discriminant analysis and nearest-neighbor methods are at the same time statistically feasible and potent competitors to the more flexible tools of feed forward neural networks. Further research is needed to explore the gain obtainable from statistical methods by the application to larger sets of protein sequence data and to compare the results with those from biophysical approaches.