Likelihood Clustering Research Articles

Sex-Specific Association of Ambient Temperature With Urine Biomarkers in Southwest Coastal Bangladesh

IntroductionMen are vulnerable to ambient heat-related kidney disease burden, but limited evidence exists on how vulnerable women are when exposed to high ambient heat. We evaluated the sex-specific association between ambient temperature and urine electrolytes (concentration and 24-hour excretion), and 24-hour urine total protein, and volume. MethodsWe pooled a longitudinal data of 1,175 participants' concentration and 24-hour excretion of urine electrolytes and other biomarkers (24-hour urine total protein and volume) (5624 person-visits) collected in southwest coastal Bangladesh (Khulna, Satkhira, and Mongla districts) during November 2016 - April 2017. We then spatiotemporally linked ambient temperature data from local weather stations to participants' health outcomes. For evaluating the relationships between average ambient temperature and urine electrolytes and other biomarkers, we plotted confounder-adjusted restricted cubic spline (RCS) plots using participant-, household-, and community-level random intercepts. We then used piece-wise linear mixed-effects models for different ambient temperature segments determined by inflection points in RCS plots and reported the maximum likelihood estimates and cluster robust standard errors. By applying interaction terms for sex and ambient temperature, we determined the overall significance using the Wald test. Bonferroni correction was used for multiple comparisons. ResultsThe RCS plots demonstrated non-linear associations between ambient heat and urine biomarkers for males and females. Piecewise linear mixed-effects models suggested sex did not modify the relationship of ambient temperature with any of the urine parameters after Bonferroni correction (p-value <0.004). ConclusionOur findings suggest that women are as susceptible to the effects of high ambient temperature exposure as men.

Avoiding inferior clusterings with misspecified Gaussian mixture models

Clustering is a fundamental tool for exploratory data analysis, and is ubiquitous across scientific disciplines. Gaussian Mixture Model (GMM) is a popular probabilistic and interpretable model for clustering. In many practical settings, the true data distribution, which is unknown, may be non-Gaussian and may be contaminated by noise or outliers. In such cases, clustering may still be done with a misspecified GMM. However, this may lead to incorrect classification of the underlying subpopulations. In this paper, we identify and characterize the problem of inferior clustering solutions. Similar to well-known spurious solutions, these inferior solutions have high likelihood and poor cluster interpretation; however, they differ from spurious solutions in other characteristics, such as asymmetry in the fitted components. We theoretically analyze this asymmetry and its relation to misspecification. We propose a new penalty term that is designed to avoid both inferior and spurious solutions. Using this penalty term, we develop a new model selection criterion and a new GMM-based clustering algorithm, SIA. We empirically demonstrate that, in cases of misspecification, SIA avoids inferior solutions and outperforms previous GMM-based clustering methods.

Scenario-Based Robust Optimization for Two-Stage Decision Making Under Binary Uncertainty

This paper addresses problems of two-stage optimization under binary uncertainty. We define a scenario-based robust optimization (ScRO) formulation that combines principles of stochastic optimization (by constructing probabilistic scenarios) and robust optimization (by protecting against adversarial perturbations within discrete uncertainty sets). To solve it, we develop a sparse row generation algorithm that iterates between a master problem (which provides a lower bound based on minimal uncertainty sets) and a history-based subproblem (which generates an upper bound and updates minimal uncertainty sets). We generate scenarios and uncertainty sets from element-wise probabilities using a deviation likelihood method or from historical samples using a sample clustering approach. Using public data sets, results suggest that (i) our ScRO formulation outperforms benchmarks based on deterministic, stochastic, and robust optimization; (ii) our deviation likelihood and sample clustering approaches outperform scenario generation baselines; and (iii) our sparse row generation algorithm outperforms off-the-shelf implementation and state-of-the-art cutting plane benchmarks. An application to a real-world ambulance dispatch case study suggests that the proposed modeling and algorithmic approach can reduce the number of late responses by more than 25%. Funding: K. Wang’s research was supported by the National Natural Science Foundation of China [Grants 72322002, 52221005, and 52220105001]

The Himalayan uplift and evolution of aquatic biodiversity across Asia: Snowtrout (Cyprininae: Schizothorax) as a test case.

Global biodiversity hotspots are often remote, tectonically active areas undergoing climatic fluctuations, such as the Himalaya Mountains and neighboring Qinghai-Tibetan Plateau (QTP). They provide biogeographic templates upon which endemic biodiversity can be mapped to infer diversification scenarios. Yet, this process can be somewhat opaque for the Himalaya, given substantial data gaps separating eastern and western regions. To help clarify, we evaluated phylogeographic and phylogenetic hypotheses for a widespread fish (Snowtrout: Cyprininae; Schizothorax) by sequencing 1,140 base pair of mtDNA cytochrome-b (cytb) from Central Himalaya samples (Nepal: N = 53; Bhutan: N = 19), augmented with 68 GenBank sequences (N = 60 Schizothorax/N = 8 outgroups). Genealogical relationships (N = 132) were analyzed via maximum likelihood (ML), Bayesian (BA), and haplotype network clustering, with clade divergence estimated via TimeTree. Snowtrout seemingly originated in Central Asia, dispersed across the QTP, then into Bhutan via southward-flowing tributaries of the east-flowing Yarlung-Tsangpo River (YLTR). Headwaters of five large Asian rivers provided dispersal corridors from Central into eastern/southeastern Asia. South of the Himalaya, the YLTR transitions into the Brahmaputra River, facilitating successive westward colonization of Himalayan drainages first in Bhutan, then Nepal, followed by far-western drainages subsequently captured by the (now) westward-flowing Indus River. Two distinct Bhutanese phylogenetic groups were recovered: Bhutan-1 (with three subclades) seemingly represents southward dispersal from the QTP; Bhutan-2 apparently illustrates northward colonization from the Lower Brahmaputra. The close phylogenetic/phylogeographic relationships between the Indus River (Pakistan) and western tributaries of the Upper Ganges (India/Nepal) potentially implicate an historic, now disjunct connection. Greater species-divergences occurred across rather than within-basins, suggesting vicariance as a driver. The Himalaya is a component of the Earth's largest glacial reservoir (i.e., the "third-pole") separate from the Arctic/Antarctic. Its unique aquatic biodiversity must be defined and conserved through broad, trans-national collaborations. Our study provides an initial baseline for this process.

First report of black rot disease in Eruca vesicaria subsp. sativa caused by Xanthomonas campestris pv. campestris in Belgium.

Eruca vesicaria subsp. sativa (Mill.) Thell. (arugula or rocket) is a leafy vegetable originating from the Mediterranean region primarily being sold in bagged salads. From 2014 to 2017, plants (cv. Montana) exhibiting blackened leaf veins and irregular V-shaped chlorotic to necroic lesions at the leaf margins were observed in commercial greenhouses in Flanders, Belgium (Figure S1A). Symptoms started after harvest of the first cut, indicating that leaf injury favours disease development. By the last cut, infections had spread uniformly across the plots, with symptoms advanced to the point where harvesting was no longer profitable. Excised surface-sterilized necrotic leaf tissue and seeds were homogenized in phosphate buffer (PB), followed by dilution plating on Pseudomonas Agar F containing sucrose. After four days at 28°C, bright yellow round, mucoid, convex Xanthomonas-like colonies were obtained, both from leaves and seeds. For confirmation, DNA was extracted from pure cultures after which a partial fragment of gyrB was amplified and sequenced (Holtappels et al. 2022). Amplicons were trimmed to 530 nucleotides (Genbank ON815895-ON815900) according to Parkinson et al. (2007) and compared with the NCBI database. Strain GBBC 3139 shares 100% sequence identity with Xanthomonas campestris pv. campestris (Xcc) type strain LMG 568 and with RKFB 1361-1364, isolated from arugula in Serbia (Prokić et al. 2022). The other isolates from Belgian rocket - GBBC 3036, 3058, 3077, 3217 and 3236 - all have a gyrB sequence 100% identical to that of Xcc strain ICMP 4013, among others. To determine the genetic relatedness to other pathogenic Xc strains, the genomes of GBBC 3077, 3217, 3236 and 3139 were sequenced using a MinION (Nanopore) and non-clonal sequences were submitted to NCBI (BioProject PRJNA967242). Genomes were compared by calculating Average Nucleotide Identity (ANI). This revealed that the Belgian strains cluster together with Xc isolates originating from Brassica crops and separate from strains identified as Xc pv. barbareae, pv. incanae and pv. raphani (Figure S2A). Their designation as pv. campestris is supported by maximum likelihood clustering of concatenated gyrB-avrBs2 sequences (EPPO, 2021; Figure S2B,C). Finally, pathogenicity was verified on five-week-old rocket 'Pronto' plants grown in a commercial potting mix by cutting the leaves along the midrib with scissors dipped into a suspension of 108 cfu/ml of each strain or PB as control (4 plants/strain). Plants were kept in closed polypropylene boxes for 48 hr to support high humidity and facilitate infection. They were then maintained at 25 ± 2 °C. Lesions like those observed on commercial plants developed on the inoculated leaves within one week (Figure S1B). Bacterial colonies reisolated from symptomatic tissue were identified based on gyrB as the strains used for inoculation, thereby fulfilling Koch's postulates. To the best of our knowledge, this is the first report of black rot disease in arugula caused by Xcc in Belgium. Previously, Xcc on arugula has been reported in Argentina, California and Serbia as well (Romero et al. 2008; Rosenthal et al. 2017; Prokić et al. 2022). Arugula being a minor crop in Belgium, challenged by Xcc infections and strong import competition, many growers have abandoned the sector in recent years. Therefore, this study makes a strong case for early detection of disease symptoms and timely application of relevant management strategies in vulnerable crop settings.

PALAEOASSOCIA as a methodological tool for phytosociological analyses is further developed

The earlier version of PALAEOASSOCIA involved a considerable input of manual labour in sorting species tables with association data to identify plant communities that could have been present. A large archaeobotanical dataset from the site of Best (The Netherlands) was used to judge whether this manual sorting results in subjective results. As these were found, we developed a fully automatic version of PALAEOASSOCIA, including this sorting process. Likelihood clustering with prior probability yielded the highest number of associations recovered from four samples, and was therefore chosen as the optimal clustering method. The sorted tables are automatically converted to syntaxonomical groups. The hierarchical level of these groups can be pre-defined by the user of the program. Syntaxa that are highly improbable geographically cannot be ruled out a priori, but need to be removed manually. PALAEOASSOCIA is not meant to replace other methods of ecological interpretations of archaeobotanical data, but instead as a tool to obtain a more detailed result.

When One’s Not Enough: Colony Pool-Seq Outperforms Individual-Based Methods for Assessing Introgression in Apis mellifera mellifera

The human management of honey bees (Apis mellifera) has resulted in the widespread introduction of subspecies outside of their native ranges. One well known example of this is Apis mellifera mellifera, native to Northern Europe, which has now been significantly introgressed by the introduction of C lineage honey bees. Introgression has consequences for species in terms of future adaptive potential and long-term viability. However, estimating introgression in colony-living haplodiploid species is challenging. Previous studies have estimated introgression using individual workers, individual drones, multiple drones, and pooled workers. Here, we compare introgression estimates via three genetic approaches: SNP array, individual RAD-seq, and pooled colony RAD-seq. We also compare two statistical approaches: a maximum likelihood cluster program (ADMIXTURE) and an incomplete lineage sorting model (ABBA BABA). Overall, individual approaches resulted in lower introgression estimates than pooled colonies when using ADMIXTURE. However, the pooled colony ABBA BABA approach resulted in generally lower introgression estimates than all three ADMIXTURE estimates. These results highlight that sometimes one individual is not enough to assess colony-level introgression, and future studies that do use colony pools should not be solely dependent on clustering programs for introgression estimates.

Neural clustering based on implicit maximum likelihood

Clustering is one of the most fundamental unsupervised learning tasks with numerous applications in various fields. Clustering methods based on neural networks, called deep clustering methods, leverage the representational power of neural networks to enhance clustering performance. ClusterGan constitutes a generative deep clustering method that exploits generative adversarial networks (GANs) to perform clustering. However, it inherits some deficiencies of GANs, such as mode collapse, vanishing gradients and training instability. In order to tackle those deficiencies, the generative approach of implicit maximum likelihood estimation (IMLE) has been recently proposed. In this paper, we present a clustering method based on generative neural networks, called neural implicit maximum likelihood clustering, which adopts ideas from both ClusterGAN and IMLE. The proposed method has been compared with ClusterGAN and other neural clustering methods on both synthetic and real datasets, demonstrating promising results.

Robust TPMB Filtering Using Sensors With Limited Sensing Range Under Nonuniform Clutter Background

This study proposes a novel solution using sensors with limited sensing range to estimate the random appearance or disappearance of multi-object trajectories with unknown object detection profiles under nonuniform clutter background. Specifically, assuming that no objects occur within the sensors’ observable area, the unknown clutter intensity involving the average false alarms per scan and nonuniform clutter density is first estimated by adopting the joint combination of maximum likelihood (ML) and model-based clustering methods. Then, the unknown object detection profile can be calculated by marginalizing the fluctuating signal-to-noise ratio (SNR) in a random manner. Further, the TPMB filter is utilized to estimate trajectories from the first principle, after which the estimated parameters of clutter intensity and the detection probability are fed to the TPMB filtering, thereby improving the performance on completeness and robustness of estimation trajectories. Simulation and experimental results demonstrated that our proposed solution exhibits excellent tracking accuracy compared with state-of-the-art solutions involving robust MS generalized LMB (R-MS-GLMB) and dynamic parameter GLMB (DP-GLMB) filters.

Quantitative fate mapping: A general framework for analyzing progenitor state dynamics via retrospective lineage barcoding

Natural and induced somatic mutations that accumulate in the genome during development record the phylogenetic relationships of cells; whether these lineage barcodes capture the complex dynamics of progenitor states remains unclear. We introduce quantitative fate mapping, an approach to reconstruct the hierarchy, commitment times, population sizes, and commitment biases of intermediate progenitor states during development based on a time-scaled phylogeny of their descendants. To reconstruct time-scaled phylogenies from lineage barcodes, we introduce Phylotime, a scalable maximum likelihood clustering approach based on a general barcoding mutagenesis model. We validate these approaches using realistic in silico and invitro barcoding experiments. We further establish criteria for the number of cells that must be analyzed for robust quantitative fate mapping and a progenitor state coverage statistic to assess the robustness. This work demonstrates how lineage barcodes, natural or synthetic, enable analyzing progenitor fate and dynamics long after embryonic development in any organism.

The complete mitochondrial genome of the sand dollar Astriclypeus mannii (Verrill, 1867) (Echinoidea: Astriclypeidae) in the subtidal sand flat in Jeju Island off the south coast of Korea

We sequenced the complete mitochondrial genome of sand dollar Astriclypeus mannii (Verrill 1867) (Echinoidea: Astriclypeidae) occurring in the subtidal sand flat in Jeju Island off the south coast of Korea. The mitochondrial genome was 15,744 bp in length and contained 13 protein-coding genes (PCGs), 22 transfer RNAs, two ribosomal RNAs, and 140 nucleotides representing the putative control region. We reconstructed the concatenated phylogenetic tree based on 13 PCGs of 18 echinoderms, including A. mannii. From the maximum likelihood clustering, A. mannii was grouped in the order Echinolampadacea. The complete mitochondrial sequence of A. mannii for the first time in this study provide valuable insight in understanding the evolution and phylogenetic analysis of echinoids (sea urchins).

Species‐specific responses to landscape features shaped genomic structure within Alaska galliformes

AbstractAimConnectivity is vital to the resiliency of populations to environmental change and stochastic events, especially for cold‐adapted species as Arctic and alpine tundra habitats retract as the climate warms. We examined the influence of past and current landscapes on genomic connectivity in cold‐adapted galliformes as a critical first step to assess the vulnerability of Alaska ptarmigan and grouse to environmental change. We hypothesize that the mosaic of physical features and habitat within Alaska promoted the formation of genetic structure across species.LocationAlaska, United States of America.TaxaPtarmigan and Grouse (Galliformes: Tetraoninae).MethodsWe collected double digest restriction‐site‐associated DNA sequence data from six ptarmigan and grouse species (N = 13–145/species) sampled across multiple ecosystems up to ~10 degrees of latitude. Spatial genomic structure was analysed using methods that reflect different temporal scales: (1) principal components analysis to identify major trends in the distribution of genomic variation; (2) maximum likelihood clustering analyses to test for the presence of multiple genomic groupings; (3) shared co‐ancestry analyses to assess contemporary relationships and (4) effective migration surfaces to identify regions that deviate from a null model of isolation by distance.ResultsLevels of genomic structure varied across species (ΦST =0.009–0.042). Three general patterns of structure emerged: (1) east‐west partition located near the Yukon‐Tanana uplands; (2) north‐south split coinciding with the Alaska Range and (3) northern group near the Brooks Range. Species‐specific patterns were observed; not all landscape features were barriers to gene flow for all ptarmigan and grouse and temporal contrasts were detected at the Brooks Range.Main conclusionsWithin Alaska galliformes, patterns of genomic structure coincide with physiographic features and highlight the importance of physical and ecological barriers in shaping how genomic diversity is arrayed across the landscape. Lack of concordance in spatial patterns indicates that species behaviour and habitat affinities play key roles in driving the contrasting patterns of genomic structure.

Agglomerative likelihood clustering

We consider the problem of fast time-series data clustering. Building on previous work modeling, the correlation-based Hamiltonian of spin variables we present an updated fast non-expensive agglomerative likelihood clustering algorithm (ALC). The method replaces the optimized genetic algorithm based approach (f-SPC) with an agglomerative recursive merging framework inspired by previous work in econophysics and community detection. The method is tested on noisy synthetic correlated time-series datasets with a built-in cluster structure to demonstrate that the algorithm produces meaningful non-trivial results. We apply it to time-series datasets as large as 20 000 assets and we argue that ALC can reduce computation time costs and resource usage costs for large scale clustering for time-series applications while being serialized, and hence has no obvious parallelization requirement. The algorithm can be an effective choice for state-detection for online learning in a fast non-linear data environment, because the algorithm requires no prior information about the number of clusters.

Genetic diversity and effective population sizes of thirteen Indian cattle breeds

BackgroundThe genetic structure of a diverse set of 15 Indian indigenous breeds and non-descript indigenous cattle sampled from eight states was examined, based on 777 k single nucleotide polymorphism (SNP) genotypes obtained on 699 animals, with sample sizes ranging from 17 to 140 animals per breed. To date, this is the largest and most detailed assessment of the genetic diversity of Indian cattle breeds.ResultsAdmixture analyses revealed that 109 of the indigenous animals analyzed had more than 1% Bos taurus admixture of relatively recent origin. Pure indigenous animals were defined as having more than 99% Bos indicus ancestry. Assessment of the genetic diversity within and between breeds using principal component analyses, F statistics, runs of homozygosity, the genomic relationship matrix, and maximum likelihood clustering based on allele frequencies revealed a low level of genetic diversity among the indigenous breeds compared to that of Bos taurus breeds. Correlations of SNP allele frequencies between breeds indicated that the genetic variation among the Bos indicus breeds was remarkably low. In addition, the variance in allele frequencies represented less than 1.5% between the Indian indigenous breeds compared to about 40% between Bos taurus dairy breeds. Effective population sizes (Ne) increased during a period post-domestication, notably for Ongole cattle, and then declined during the last 100 generations. Although we found that most of the identified runs of homozygosity are short in the Indian indigenous breeds, indicating no recent inbreeding, the high FROH coefficients and low FIS values point towards small population sizes. Nonetheless, the Ne of the Indian indigenous breeds is currently still larger than that of Bos taurus dairy breeds.ConclusionsThe changes in the estimates of effective population size are consistent with domestication from a large native population followed by consolidation into breeds with a more limited population size. The surprisingly low genetic diversity among Indian indigenous cattle breeds might be due to their large Ne since their domestication, which started to decline only 100 generations ago, compared to approximately 250 to 500 generations for Bos taurus dairy cattle.

Molecular Analysis of Bacterial Isolates From Necrotic Wheat Leaf Lesions Caused by Xanthomonas translucens, and Description of Three Putative Novel Species, Sphingomonas albertensis sp. nov., Pseudomonas triticumensis sp. nov. and Pseudomonas foliumensis sp. nov.

Xanthomonas translucens is the etiological agent of the wheat bacterial leaf streak (BLS) disease. The isolation of this pathogen is usually based on the Wilbrink’s-boric acid–cephalexin semi-selective medium which eliminates 90% of other bacteria, some of which might be novel species. In our study, a general purpose nutrient agar was used to isolate 49 bacterial strains including X. translucens from necrotic wheat leaf tissues. Maximum likelihood cluster analysis of 16S rRNA sequences grouped the strains into 10 distinct genera. Pseudomonas (32.7%) and Pantoea (28.6%) were the dominant genera while Xanthomonas, Clavibacter and Curtobacterium had 8.2%, each. Erwinia and Sphingomonas had two strains, each. BLAST and phylogenetic analyses of multilocus sequence analysis (MLSA) of specific housekeeping genes taxonomically assigned all the strains to validly described bacterial species, except three strains (10L4B, 12L4D and 32L3A) of Pseudomonas and two (23L3C and 15L3B) of Sphingomonas. Strains 10L4B and12L4D had Pseudomonas caspiana as their closest known type strain while strain 32L3A was closest to Pseudomonas asturiensis. Sphingomonas sp. strains 23L3C and 15L3B were closest to S. faeni based on MLSA analysis. Our data on MLSA, whole genome-based cluster analysis, DNA-DNA hybridization and average nucleotide identity, matrix-assisted laser desorption/ionization-time-of-flight, chemotaxonomy and phenotype affirmed that these 5 strains constitute three novel lineages and are taxonomically described in this study. We propose the names, Sphingomonas albertensis sp. nov. (type strain 23L3CT = DOAB 1063T = CECT 30248T = LMG 32139T), Pseudomonas triticumensis sp. nov. (type strain 32L3AT = DOAB 1067T = CECT 30249T = LMG 32140T) and Pseudomonas foliumensis sp. nov. (type strain 10L4BT = DOAB 1069T = CECT 30250T = LMG 32142T). Comparative genomics of these novel species, relative to their closest type strains, revealed unique repertoires of core secretion systems and secondary metabolites/antibiotics. Also, the detection of CRISPR-Cas systems in the genomes of these novel species suggests an acquired mechanism for resistance against foreign mobile genetic elements. The results presented here revealed a cohabitation, within the BLS lesions, of diverse bacterial species, including novel lineages.

A new species of cone snail, Conus laccadivensis sp. nov.: (Gastropoda: Conidae) from the Islands of Lakshadweep Archipelago

The present paper describes the cone snail, Conus laccadivensis sp. nov., based on the morphological and molecular characters of the materials collected from Kalpeni and Androth Islands of Lakshadweep Archipelago. The species closely resembles the tropical species Conus distans (Hwass in Bruguière (1792) which is widely distributed throughout the Indo-Pacific regions. Morphologically C laccadivensis sp. nov., differs from C. distans in shell colour and spire characteristics. Conus laccadivensis sp. nov., possess distinct colour patterns with large and closely spaced coronations. The maximum likelihood clustering approaches using the mitochondrial genes such as COI, 16S rRNA and 12S rRNA revealed a separate genetic cluster confirming the newly identified species.

Large transmission cluster of acute hepatitis C identified among HIV-positive men who have sex with men in Bangkok, Thailand.

A rapidly emerging and highly concentrated hepatitis C virus (HCV) outbreak has recently been observed among both acute and chronic HIV-positive men who have sex with men (MSM) in Bangkok, Thailand. NS5B regions of the HCV genome were amplified using nested PCR and sequenced. Phylogenetic inference was constructed by Maximum Likelihood methods and clusters were identified with support and genetic distance thresholds of 85% and of 4.5%. Forty-eight (25 acute HIV and 23 chronic HIV) MSM with incident HCV infection were included in the analysis. HCV genotype (GT) was 85% GT 1a and 15% GT 3a or 3b. Median age at HCV diagnosis was 34 (interquartile range, 28-41) years. 83.3% (40/48) had history of syphilis infection and 36% (16/44) reported crystal methamphetamine use. Only 2 (4%) reported ever injecting drugs, both crystal methamphetamine. In the phylogenetic clustering analysis, 83% belonged to one of two clusters: one large (75%) and one small (8%) cluster. All clusters were GT 1a. MSM with acute HIV infection were more likely to be in a cluster (92%) than those with chronic infection (74%). HCV screening should be regularly performed for MSM in ART clinics, and offering direct-acting antiviral agents to all MSM with HCV infection might contain the HCV epidemic from expanding further.

Multi-temporal LULC Classification using Hybrid Approach and Monitoring Built-up Growth with Shannon’s Entropy for a Semi-arid Region of Rajasthan, India

Abstract Land use land cover (LULC) classification of Churu district of Rajasthan state in India was done through hybrid classification technique. Landsat imageries of 1998, 2008 and 2018 were used to ascertain the rate and nature of spatio-temporal LULC changes. Major focus was given on vegetative cover change detection of the study area. On the basis of field survey and standard classification classes, the land use classes of the study area were divided into eleven classes. A detailed vegetative classification has been done while categorizing the classes. This classification method employed maximum likelihood and ISODATA clustering. The decision tree approach was used to create the multi-temporal hybrid LULC classification. The accuracy assessment results have shown excellent results at 91% overall accuracy with a kappa of 0.92. The results indicated that agriculture, crop land dominated the land use of Churu district while natural vegetation (forest areas) had the least share in land cover during the entire study period from 1998 to 2018. Shannon’s entropy index was used to determine the changes in spatial distributional pattern of built up during the period 1998 to 2018 for the study area in general and also for each of the eight sub-administrative regions (tehsils). The increase in the built up area in the study area during the period of 1998 to 2018 was quite paltry with a general dispersed type of built up. The increase varied from moderate to nominal with the entropy value decreasing from 0.84 to 0.71 for the study area in general.

Assessment of flash flood risk based on improved analytic hierarchy process method and integrated maximum likelihood clustering algorithm

Flash floods are one of the most severe natural disasters throughout the world, and are responsible for sizeable social and economic losses, as well as countless injuries and death. Risk assessment, which identifies areas susceptible to flooding, has been shown to be an effective tool for managing and mitigating flash floods. The study aims to introduce the methods to determine the weights of the risk indices, and identify the different risk clusters. In this regard, we proposed a methodology for comprehensively assessing flash flood risk in a GIS environment, by the improved analytic hierarchy process (IAHP) method, and an integration of iterative self-organizing data (ISODATA) analysis and maximum likelihood (ISO-Maximum) clustering algorithm. The weight for each risk index is determined by the IAHP, which integrates the subjective characteristics with objective attributes of the assessment data. Based on the data mining technology, the integration of ISO-Maximum clustering algorithm derives a more reasonable classification. The Guangdong Province of China was selected for testing the proposed method’s applicability, and we used a receiver operating characteristics (ROC) curve approach to validate the modeling of the flash-flood risk distribution. The validation against the historical flash flood data indicates a high reliability of this method for comprehensive flash flood risk assessment. In order to verify the proposed method’s superiority, in addition, the technique for order performance by similarity to ideal solution (TOPSIS) and the weights-of-evidence (WE) methods are used for comparison with the IAHP and ISO-Maximum clustering algorithm method. Moreover, we analyzed and compared the regularity of flash floods in the rural and urban areas. This study not only provides a new approach for large-scale flash flood comprehensive risk assessment, but also assists researchers and local decision-makers in designing flash flood mitigation strategies.

A Novel Information Integration Algorithm for Speech Recognition System: Basing on Adaptive Clustering and Supervised State of Acoustic Feature

When utilizing the most likely state sequence (MLSS) criterion in Gauss mixture model-hidden Markov model (GMM-HMM) to acquire the best state series of observations, only the maximum likelihood state of speech frame is considered. Therefore, the influence of other states is neglected, which leads to the losing of some important information, and further reduces the recognition rate of the system. In this paper, we propose two new features, which are called state likelihood cluster feature (SLCF) and supervised state feature (SSF), to both reflect acoustic features and fuse state information. Combining SLCF and SSF with Mel frequency cepstrum coefficient (MFCC), Mel frequency cepstrum & state likelihood cluster feature (MSLCF) and Mel frequency cepstrum & supervised state feature (MSSF) are formed, respectively. By the proposed MSLCF and MSSF in Chinese speech recognition experiment, the relative error rate of the isolated word recognition system declines 6.10% and 9.66%, respectively, and the relative error rate of the continuous speech recognition system declines 2.53% and 11.05%, respectively.