P2LSA and P2LSA+: Two Paralleled Probabilistic Latent Semantic Analysis Algorithms Based on the MapReduce Model

Measurement errors commonly occur in 24-h hormonal data and may affect the outcomes of such studies. Measurement errors often appear as outliers in such data sets; however, no well-established method is available for their automatic detection. In this study, we aimed to compare performances of different methods for outlier detection in hormonal serial data. Hormones (glucose, insulin, thyroid-stimulating hormone, cortisol, and growth hormone) were measured in blood sampled every 10 min for 24 h in 38 participants of the Leiden Longevity Study. Four methods for detecting outliers were compared: (1) eyeballing, (2) Tukey’s fences, (3) stepwise approach, and (4) the expectation-maximization (EM) algorithm. Eyeballing detects outliers based on experts’ knowledge, and the stepwise approach incorporates physiological knowledge with a statistical algorithm. Tukey’s fences and the EM algorithm are data-driven methods, using interquartile range and a mathematical algorithm to identify the underlying distribution, respectively. The performance of the methods was evaluated based on the number of outliers detected and the change in statistical outcomes after removing detected outliers. Eyeballing resulted in the lowest number of outliers detected (1.0% of all data points), followed by Tukey’s fences (2.3%), the stepwise approach (2.7%), and the EM algorithm (11.0%). In all methods, the mean hormone levels did not change materially after removing outliers. However, their minima were affected by outlier removal. Although removing outliers affected the correlation between glucose and insulin on the individual level, when averaged over all participants, none of the 4 methods influenced the correlation. Based on our results, the EM algorithm is not recommended given the high number of outliers detected, even where data points are physiologically plausible. Since Tukey’s fences is not suitable for all types of data and eyeballing is time-consuming, we recommend the stepwise approach for outlier detection, which combines physiological knowledge and an automated process.

This chapter introduces several fundamental models and algorithms for probabilistic clustering, including mixture models, Expectation-Maximization (EM) algorithm, and probabilistic topic models. It focuses on general framework of modeling, the probabilistic explanation, the standard algorithms to learn the model, and its applications. Mixture models are probabilistic models which are increasingly used to find the clusters for univariate and multivariate data. The chapter begins the discussion of mixture models, in which the values of the discrete latent variables can be interpreted as the assignments of data points to specific components of the mixture. It uses the Gaussian mixture model to motivate the EM algorithm in an informal way, and then gives a more general view of the EM algorithm, which is a standard learning algorithm for many probabilistic models. The chapter presents two popular probabilistic topic models, i.e., probabilistic latent semantic analysis (PLSA) and latent Dirichlet allocation (LDA), for document clustering and analysis.

Many diseases show dichotomous phenotypic variation but do not follow a simple Mendelian pattern of inheritance. Variances of these binary diseases are presumably controlled by multiple loci and environmental variants. A least-squares method has been developed for mapping such complex disease loci by treating the binary phenotypes (0 and 1) as if they were continuous. However, the least-squares method is not recommended because of its ad hoc nature. Maximum Likelihood (ML) and Bayesian methods have also been developed for binary disease mapping by incorporating the discrete nature of the phenotypic distribution. In the ML analysis, the likelihood function is usually maximized using some complicated maximization algorithms (e.g. the Newton-Raphson or the simplex algorithm). Under the threshold model of binary disease, we develop an Expectation Maximization (EM) algorithm to solve for the maximum likelihood estimates (MLEs). The new EM algorithm is developed by treating both the unobserved genotype and the disease liability as missing values. As a result, the EM iteration equations have the same form as the normal equation system in linear regression. The EM algorithm is further modified to take into account sexual dimorphism in the linkage maps. Applying the EM-implemented ML method to a four-way-cross mouse family, we detected two regions on the fourth chromosome that have evidence of QTLs controlling the segregation of fibrosarcoma, a form of connective tissue cancer. The two QTLs explain 50-60% of the variance in the disease liability. We also applied a Bayesian method previously developed (modified to take into account sex-specific maps) to this data set and detected one additional QTL on chromosome 13 that explains another 26% of the variance of the disease liability. All the QTLs detected primarily show dominance effects.

More and more tourists are sharing their travel feelings and posting their real experiences on the Internet, generating tourism big data. Online travel reviews can fully reflect tourists’ emotions, and mining and analyzing them can provide insight into the value of them. In order to analyze the potential value of online travel reviews by using big data technology and machine learning technology, this paper proposes an improved support vector machine (SVM) algorithm based on travel consumer sentiment analysis and builds an Hadoop Distributed File System (HDFS) system based on Map-Reduce model. Firstly, Internet travel reviews are pre-processed for sentiment analysis of the review text. Secondly, an improved SVM algorithm is proposed based on the main features of linear classification and kernel functions, so as to improve the accuracy of sentiment word classification. Then, HDFS data nodes are deployed on the basis of Hadoop platform with the actual tourism application context. And based on the Map-Reduce programming model, the map function and reduce function are designed and implemented, which greatly improves the possibility of parallel processing and reduces the time consumption at the same time. Finally, an improved SVM algorithm is implemented under the built Hadoop platform. The test results show that online travel reviews can be an important data source for travel big data recommendation, and the proposed method can quickly and accurately achieve travel sentiment classification.

Previous chapter Next chapter Full AccessProceedings Symposium on Algorithmic Principles of Computer Systems (APOCS)Parallel In-Place Algorithms: Theory and PracticeYan Gu, Omar Obeya, and Julian ShunYan Gu, Omar Obeya, and Julian Shunpp.114 - 128Chapter DOI:https://doi.org/10.1137/1.9781611976489.9PDFBibTexSections ToolsAdd to favoritesExport CitationTrack CitationsEmail SectionsAboutAbstract Many parallel algorithms use at least linear auxiliary space in the size of the input to enable computations to be done independently without conflicts. Unfortunately, this extra space can be prohibitive for memory-limited machines, preventing large inputs from being processed. Therefore, it is desirable to design parallel in-place algorithms that use sublinear (or even polylogarithmic) auxiliary space. In this paper, we bridge the gap between theory and practice for parallel in-place (PIP) algorithms. We first define two computational models based on fork-join parallelism, which reflect modern parallel programming environments. We then introduce a variety of new parallel in-place algorithms that are simple and efficient, both in theory and in practice. Our algorithmic highlight is the Decomposable Property introduced in this paper, which enables existing non-in-place but highly-optimized parallel algorithms to be converted into parallel in-place algorithms. Using this property, we obtain algorithms for random permutation, list contraction, tree contraction, and merging that take linear work, O(n1–∊) auxiliary space, and O(n∊ · polylog(n)) span for 0 < · < 1. We also present new parallel in-place algorithms for scan, filter, merge, connectivity, biconnectivity, and minimum spanning forest using other techniques. In addition to theoretical results, we present experimental results for implementations of many of our parallel in-place algorithms. We show that on a 72-core machine with two-way hyper-threading, the parallel in-place algorithms usually outperform existing parallel algorithms for the same problems that use linear auxiliary space, indicating that the theory developed in this paper indeed leads to practical benefits in terms of both space usage and running time. Previous chapter Next chapter RelatedDetails Published:2021eISBN:978-1-61197-648-9 https://doi.org/10.1137/1.9781611976489Book Series Name:ProceedingsBook Code:APOSC21Book Pages:ii + 158

Using general regression with local tuning for learning mixture models from incomplete data sets

BackgroundMultiple imputation is a reliable tool to deal with missing data and is becoming increasingly popular in biostatistics. However, building a model with interactions that are not specified a priori, in the presence of missing data, presents a challenge. On the one hand, the interactions are needed to impute the data, while on the other hand, the data is needed to identify the interactions. The objective of this study was to present a way in which this challenge can be addressed.MethodsThis paper investigates two strategies in which model development with interactions is achieved using a single data set generated from the Expectation Maximization (EM) algorithm. Imputation using both the fully conditional specification approach and the multivariate normal approach is carried out and results are compared. The strategies are illustrated with data from a study of ambient pollution and childhood asthma in Durban, South Africa.ResultsThe different approaches to model building and imputation yielded similar results despite the data being mainly categorical. Both strategies investigated for building the model using the multivariate normal imputed data resulted in the identical set of variables and interactions being identified; while models built using data imputed by fully conditional specification were marginally different for the two strategies. It was found that, for both imputation approaches, model building with backward elimination applied to the initial EM data set was easier to implement, and produced good results, compared to those from a complete case analysis.ConclusionsDeveloping a predictive model including interactions with data that suffers from missingness is easily done by identifying significant interactions and then applying backward elimination to a single data set imputed from the EM algorithm. It is hoped that this idea can be further developed and, by addressing this practical dilemma, there will be increased adoption of multiple imputation in medical research when data suffers from missingness.

Markovian Arrival Processes (MAPs) are widely used stochastic models to describe correlated events. For the parameter fitting of MAPs according to measured data, the expectation-maximization (EM) algorithm is commonly seen as the best approach. Unfortunately, EM algorithms require a huge computational effort if the number of data points is large or the MAP has a larger dimension. The classical EM algorithm runs sequentially through the data which is necessary to consider dependencies between data points.In this paper we present a parallel variant of the EM algorithm for MAPs with a general structure. The parallel version of the algorithm is developed for multicore systems with shared memory. It is shown that the parallel algorithm yields a significant speedup compared to its sequential counterpart.KeywordsMarkovian Arrival ProcessEM algorithmParallel algorithmsTraffic modeling

Functional mapping has emerged as a powerful tool for mapping quantitative trait loci (QTL) that control developmental patterns of complex dynamic traits. Original functional mapping has been constructed within the context of simple interval mapping, without consideration of separate multiple linked QTL for a dynamic trait. In this article, we present a statistical framework for mapping QTL that affect dynamic traits by capitalizing on the strengths of functional mapping and composite interval mapping. Within this so-called composite functional-mapping framework, functional mapping models the time-dependent genetic effects of a QTL tested within a marker interval using a biologically meaningful parametric function, whereas composite interval mapping models the time-dependent genetic effects of the markers outside the test interval to control the genome background using a flexible nonparametric approach based on Legendre polynomials. Such a semiparametric framework was formulated by a maximum-likelihood model and implemented with the EM algorithm, allowing for the estimation and the test of the mathematical parameters that define the QTL effects and the regression coefficients of the Legendre polynomials that describe the marker effects. Simulation studies were performed to investigate the statistical behavior of composite functional mapping and compare its advantage in separating multiple linked QTL as compared to functional mapping. We used the new mapping approach to analyze a genetic mapping example in rice, leading to the identification of multiple QTL, some of which are linked on the same chromosome, that control the developmental trajectory of leaf age.

A dynamic system with both input and output measurement errors is termed as an errors-in-variables (EIV) system. Employment of traditional identification methods for EIV systems will result in biased parameter estimates. The existing methods of EIV identification like Bias Eliminated Least Squares (BELS), and Subspace EIV methods may not always yield good parameter estimates. To address these open issues, this paper presents a maximum-likelihood (ML) approach using the Expectation Maximization (EM) algorithm. The efficacy of the proposed method is demonstrated with an experimental study.

There have been more than 50 type clustering algorithms developed for getting meaningful information from big datasets and grouping individuals according to their characteristics. In actual researches, it is often seen that data involves all types of variables. In this case, it is very important to select appropriate clustering algorithm according to different data types. In this study, we will provide information about EM(Expectation Maximization),Two–Step Clustering methods which are developed in recent years and one of the best methods for data sets containing mixed types of variables. And the second aim is to compare the methods by producing a data set from health field information.These algorithms are generally recommended for large data sets but there are also used n medium-sized data sets. Medium- sized data sets are more often in actual researches.Therefore, fifty people for control group and fifty people for patients that have polycystic over syndrome were taken to the study. Totally nineteen variables were measured from these subjects and thirteen of them were quantitative, six of them were qualitative.Clusters were obtained by EM and Two-Step cluster methods.To evaluate the relationships between the clusters obtained from algorithms and actually known patient, control groups were analyzed by Kappa coefficient. It was found that EM clustering algorithm has highest compliance coefficient comparing with Two-Step cluster(Kappa=0,740;p

Time series modeling of surface EMG based hand manipulation identification via expectation maximization algorithm

To learn Bayesian networks, one must estimate the parameters of the network from the data. EM (Expectation-Maximization) and gradient-based algorithms are the two best known techniques to estimate these parameters. Although the theoretical properties of these two frameworks are well-studied, it remains an open question as to when and whether EM is to be preferred over gradients. We will answer this question empirically. More specifically, we first adapt scaled conjugate gradients well-known from neural network learning. This accelerated conjugate gradient avoids the time consuming line search of more traditional methods. Secondly, we empirically compare scaled conjugate gradients with EM. The experiments show that accelerated conjugate gradients are competitive with EM. Although, in general EM is the domain independent method of choice, gradient-based methods can be superior.

This paper introduces an expectation-maximization (EM) algorithm within a wavelet domain Bayesian framework for semi-blind channel estimation of multiband OFDM based UWB communications. A prior distribution is chosen for the wavelet coefficients of the unknown channel impulse response, in order to capture the well known sparseness property of the wavelet representation. This prior yields, in maximum a posteriori estimation, a thresholding rule within the EM algorithm. We particularly focus on reducing the number of estimated parameters by iteratively discarding "unsignificant" wavelet coefficients from the estimation process. In addition, the proposed algorithm enhances the estimation accuracy in terms of mean square error with less computational complexity than traditional semi-blind methods. Simulation results using IEEE UWB channel models show that the proposed algorithm outperforms pilot based channel estimation in terms of mean square error and bit error rate

An EM algorithm for convolutive independent component analysis