11 research outputs found
Big Data Clustering using Parallel Differential Evolution Algorithm
Get PDFClustering is the task of discovering group ofsimilar objects or items and there have been manyapplications for clustering such as imagesegmentation, document retrieval and data mining.The increasing volumes of information emerging bythe development of technology makes clustering ofvery large scale of data a challenging task.Differential evolution (DE) algorithm is aninnovative evolutionary algorithm (EA) for globaloptimization, where the mutation operator is basedon the distribution of solutions in the population.Clustering can be viewed as optimization problemwhere the task is finding the optimal cluster solution.To deal with clustering of huge amount of data sets,the use of classical DE is time-consuming that it isinfeasible. This paper proposes a parallel differentialevolution algorithm for clustering enormous databased on Spark framework. The proposed approachwill be efficient for large-scale data clustering
An Improved Differential Evolution Algorithm with Opposition-Based Learning for Clustering Problems
Get PDFDifferential Evolution (DE) is a popularefficient population-based stochastic optimizationtechnique for solving real-world optimizationproblems in various domains. In knowledge discoveryand data mining, optimization-based patternrecognition has become an important field, andoptimization approaches have been exploited toenhance the efficiency and accuracy of classification,clustering and association rule mining. Like otherpopulation-based approaches, the performance of DErelies on the positions of initial population which maylead to the situation of stagnation and prematureconvergence. This paper describes a differentialevolution algorithm for solving clustering problems,in which opposition-based learning (OBL) is utilizedto create high-quality solutions for initial population,and enhance the performance of clustering. Theexperimental test has been carried out on some UCIstandard datasets that are mostly used foroptimization-based clustering. According to theresults, the proposed algorithm is more efficient androbust than classical DE based clustering
An Improved Differential Evolution Algorithm with Opposition-Based Learning for Clustering Problems
No full textCommunity Detection in Social Graph Using Nature-Inspired Based Artificial Bee Colony Algorithm with Crossover and Mutation
Get PDFMany types of social network are modelled asgraphs. Community detection has been an important researcharea in social graph analysis. Community detection can beviewed as an optimization problem. Nowadays, researchers usenature-inspired algorithms to solve optimization problem.Their goal is to find the optimal solution for a given problem.In this paper, nature-inspired based artificial bee colonyalgorithm with crossover and mutation is used to detectcommunity in social graphs. GraphX is built as a library onthe top of Spark by encoding graph as a collection of verticesand edges. Comparative studies describe that the proposedalgorithm and other nature-inspired algorithms can effectivelydetect the community structure on real world social graphs asother traditional community detection algorithms
Evaluation of Differential Evolution Algorithm with Various Mutation Strategies for Clustering Problems
Get PDFEvolutionary Algorithms (EAs) based pattern recognition has emerged as an alternative solution to data analysis problems to enhance the efficiency and accuracy of mining processes. Differential Evolution (DE) is one rival and powerful instance of EAs, and DE has been successfully used for cluster analysis in recent years. Mutation strategy, one of the main processes of DE, uses scaled differences of individuals that are chosen randomly from the population to generate a mutant (trial) vector. The achievement of the DE algorithm for solving optimization problems highly relies on an adopted mutation strategy. In this paper, an empirical study was presented to investigate the effectiveness of six frequently used mutation strategies for solving clustering problems. The experimental tests were conducted on the most widely used data set for EAs based clustering, and the quality of cluster solutions and convergence characteristics of DE variants were evaluated. The obtained results pointed out that the mutation strategies that use the guidance information from the best solution mange to find more stable results whereas the random mutation strategies are able to find high quality solutions with slower convergence rate. This study aims to provide some information and insights to develop better DE mutation schemes for clustering
Reconstructing the Path of the Object Based on Time and Date OCR in Surveillance System
Get PDFThis paper propose the system for reconstructing the path of the object in surveillance cameras based on time and date optical character recognition system
Observational study of adult respiratory infections in primary care clinics in Myanmar: understanding the burden of melioidosis, tuberculosis and other infections not covered by empirical treatment regimes.
Get PDFBACKGROUND: Lower respiratory infections constitute a major disease burden worldwide. Treatment is usually empiric and targeted towards typical bacterial pathogens. Understanding the prevalence of pathogens not covered by empirical treatment is important to improve diagnostic and treatment algorithms. METHODS: A prospective observational study in peri-urban communities of Yangon, Myanmar was conducted between July 2018 and April 2019. Sputum specimens of 299 adults presenting with fever and productive cough were tested for Mycobacterium tuberculosis (microscopy and GeneXpert MTB/RIF [Mycobacterium tuberculosis/resistance to rifampicin]) and Burkholderia pseudomallei (Active Melioidosis Detect Lateral Flow Assay and culture). Nasopharyngeal swabs underwent respiratory virus (influenza A, B, respiratory syncytial virus) polymerase chain reaction testing. RESULTS: Among 299 patients, 32% (95% confidence interval [CI] 26 to 37) were diagnosed with tuberculosis (TB), including 9 rifampicin-resistant cases. TB patients presented with a longer duration of fever (median 14 d) and productive cough (median 30 d) than non-TB patients (median fever duration 6 d, cough 7 d). One case of melioidosis pneumonia was detected by rapid test and confirmed by culture. Respiratory viruses were detected in 16% (95% CI 12 to 21) of patients. CONCLUSIONS: TB was very common in this population, suggesting that microscopy and GeneXpert MTB/RIF on all sputum samples should be routinely included in diagnostic algorithms for fever and cough. Melioidosis was uncommon in this population
Effect of generalised access to early diagnosis and treatment and targeted mass drug administration on Plasmodium falciparum malaria in Eastern Myanmar : an observational study of a regional elimination programme
Get PDFInternational audienceBACKGROUND:Potentially untreatable Plasmodium falciparum malaria threatens the Greater Mekong subregion. A previous series of pilot projects in Myanmar, Laos, Cambodia, and Vietnam suggested that mass drug administration was safe, and when added to provision of early diagnosis and treatment, could reduce the reservoir of P falciparum and interrupts transmission. We examined the effects of a scaled-up programme of this strategy in four townships of eastern Myanmar on the incidence of P falciparum malaria.METHODS:The programme was implemented in the four townships of Myawaddy, Kawkareik, Hlaingbwe, and Hpapun in Kayin state, Myanmar. Increased access to early diagnosis and treatment of malaria was provided to all villages through community-based malaria posts equipped with rapid diagnostic tests, and treatment with artemether-lumefantrine plus single low-dose primaquine. Villages were identified as malarial hotspots (operationally defined as >40% malaria, of which 20% was P falciparum) with surveys using ultrasensitive quantitative PCR either randomly or targeted at villages where the incidence of clinical cases of P falciparum malaria remained high (ie, >100 cases per 1000 individuals per year) despite a functioning malaria post. During each survey, a 2 mL sample of venous blood was obtained from randomly selected adults. Hotspots received targeted mass drug administration with dihydroartemisinin-piperaquine plus single-dose primaquine once per month for 3 consecutive months in addition to the malaria posts. The main outcome was the change in village incidence of clinical P falciparum malaria, quantified using a multivariate, generalised, additive multilevel model. Malaria prevalence was measured in the hotspots 12 months after mass drug administration.FINDINGS:Between May 1, 2014, and April 30, 2017, 1222 malarial posts were opened, providing early diagnosis and treatment to an estimated 365 000 individuals. Incidence of P falciparum malaria decreased by 60 to 98% in the four townships. 272 prevalence surveys were undertaken and 69 hotspot villages were identified. By April 2017, 50 hotspots were treated with mass drug administration. Hotspot villages had a three times higher incidence of P falciparum at malarial posts than neighbouring villages (adjusted incidence rate ratio [IRR] 2·7, 95% CI 1·8-4·4). Early diagnosis and treatment was associated with a significant decrease in P falciparum incidence in hotspots (IRR 0·82, 95% CI 0·76-0·88 per quarter) and in other villages (0·75, 0·73-0·78 per quarter). Mass drug administration was associated with a five-times decrease in P falciparum incidence within hotspot villages (IRR 0·19, 95% CI 0·13-0·26). By April, 2017, 965 villages (79%) of 1222 corresponding to 104 village tracts were free from P falciparum malaria for at least 6 months. The prevalence of wild-type genotype for K13 molecular markers of artemisinin resistance was stable over the three years (39%; 249/631).INTERPRETATION:Providing early diagnosis and effective treatment substantially decreased village-level incidence of artemisinin-resistant P falciparum malaria in hard-to-reach, politically sensitive regions of eastern Myanmar. Targeted mass drug administration significantly reduced malaria incidence in hotspots. If these activities could proceed in all contiguous endemic areas in addition to standard control programmes already implemented, there is a possibility of subnational elimination of P falciparum.FUNDING:The Bill & Melinda Gates Foundation, the Regional Artemisinin Initiative (Global Fund against AIDS, Tuberculosis and Malaria), and the Wellcome Trust
Abstracts from the 8th International Congress of the Asia Pacific Society of Infection Control (APSIC)
Get PDF
