دانلود رایگان مقاله لاتین شباهت انعطاف پذیر پزشکی ناهمگن در نگاشت کاهش از سایت الزویر
عنوان فارسی مقاله:
استدلال مبتنی بر شباهت انعطاف پذیر موازی برای طبقه بندی موارد پزشکی ناهمگن در نگاشت کاهش
عنوان انگلیسی مقاله:
Resilient parallel similarity-based reasoning for classifying heterogeneous medical cases in MapReduce
سال انتشار : 2016
برای دانلود رایگان مقاله شباهت انعطاف پذیر پزشکی ناهمگن در نگاشت کاهش اینجا کلیک نمایید.
مقدمه انگلیسی مقاله:
1. Introduction
Information and Communication Technologies (ICT) provide various solutions to improve the service quality of the health care systems [1] and reduce the rate of misdiagnosis. As a novel application area, medical case classification has attracted more and more attention from the researchers in Clinical Decision Support (CDS) [2,3] due to the huge amount of Electronic Medical Records (EMR) and other heterogeneous patient data available among care providers and healthcare facilities. Many useful patterns on best clinical decision practice could be derived from mining these medical case repositories. The clinical data in health care systems are prone to heterogeneity data, yet each entity data contains different sets of multidimensional attributes [4]. With the multidimensional data, assessing similarity/distance between pairwise medical cases is one of the fundamental problems in CDS. A key motivation is to leverage the concept of inter-entity similarity for case retrievals, knowledge discovery analysis and inferences, and patient cohort identification. Similarity-Based Reasoning (SBR) algorithms have been widely applied in such applications, including information retrieval, discrimination analysis [5–7], etc. The cognitive conception of inter-entity similarity and the related theoretical finding that similar causes bring about similar effects provides the logical foundation of many formal methods, i.e., inductive reasoning [8]. A typical example is CaseBased Reasoning (CBR) [9,10], a problem solving methodology declares that “similar cases have similar solutions”. For each observation of medical cases, the decision-maker (i.e., physicians) predicts a set of class labels expressing their beliefs about the underlying probability distribution [4]. However, the previous approaches focusing on these methods are designed to be executed on a single thread on a single machine. This will bias the inference knowledge on the genuine underlying clinical similarity between pairwise entities. With the exponential increase in the scale of the input datasets [11–13], processing large data in parallel and in a distributed fashion is becoming a popular practice. For this propose, MapReduce [14] is a programming framework for processing a high volume of case datasets by exploiting the parallelism among a cluster of computing nodes. In recent years, MapReduce has gained a lot of popularity for its flexibility, simplicity and scalability. MapReduce is now well investigated [15] and widely adapted in both scientific and commercial applications. Therefore, MapReduce provides an ideal framework for processing SBR operations over an exponentially increasing volume of medical case data. Given a medical query, it is very important to convert the beliefs of the decision makers into class labels according to the underlying query data similarities by incorporating the results from multiple machines. To address this problem, in this paper we propose a method to integrate SBR results (similarity metrics with a ranked list of relevant medical cases) obtained from individual machines into a single combined target metric reflecting the true underlying data cases. There are a number of interesting and challenging issues associated with realizing this idea in MapReduce, e.g., how to distinguish similar patterns efficiently in MapReduce, how to reduce the amount of communication and improve speed in the Map-to-Reduce phase. We address these problems in our study
برای دانلود رایگان مقاله شباهت انعطاف پذیر پزشکی ناهمگن در نگاشت کاهش اینجا کلیک نمایید.
کلمات کلیدی:
Applications of the MapReduce programming framework to clinical big ... https://biodatamining.biomedcentral.com/articles/10.1186/1756-0381-7-22 by EA Mohammed - 2014 - Cited by 34 - Related articles Oct 29, 2014 - MapReduce Hadoop Big data Clinical big data analysis Clinical data ... unfortunately do not tell anything about the future, that is the reason ... heterogeneous systems with optimal workload management servers, networks, storage, etc. ..... use–cases: (1) parameter optimization for lung texture classification ... Resilient parallel similarity-based reasoning for classifying ... - Inicio vufind.uniovi.es/.../oai%3Adoaj.orgarticle%3A997a09ca3b084ec8... Translate this page Resilient parallel similarity-based reasoning for classifying heterogeneous medical cases in MapReduce. Given the exponentially increasing volume of ... Information Retrieval Meets Information Visualization: PROMISE ... https://books.google.com/books?isbn=3642364152 Maristella Agosti, Nicola Ferro, Pamela Forner - 2013 - Computers International Journal of Medical Informatics 78(suppl. ... de Herrera, A., Tsikrika, T.: The CLEF 2011 medical image retrieval and classification tasks. ... Dean, J., Ghemawat, S.: Mapreduce: simplified data processing on large clusters. ... Cochener, B.: Case retrieval in medical databases by fusing heterogeneous information. OWL Reasoning Framework over Big Biological Knowledge Network https://www.ncbi.nlm.nih.gov › NCBI › Literature › PubMed Central (PMC) by H Chen - 2014 - Cited by 6 - Related articles Apr 27, 2014 - Biological data covers a quite wide range, including proteins, pathways, ... the large-scale, heterogeneous, and complex-associated biological data as a big ... We propose several MapReduce-based property chain reasoning ...... medicinal herbs and identifying the active ingredients of these herbs and ... Big Data Analytics in Healthcare - Hindawi https://www.hindawi.com/journals/bmri/2015/370194/ by A Belle - 2015 - Cited by 47 - Related articles Jun 16, 2015 - Healthcare is a prime example of how the three Vs of data, velocity (speed of ... vital information to provide alert mechanisms in case of overt events. .... MapReduce framework has been used in [47] to increase the speed of three ... texture classification by employing a well-known machine learning method, ...
