scholarly journals Multiple imputation with compatibility for high-dimensional data

PLoS ONE ◽  
2021 ◽  
Vol 16 (7) ◽  
pp. e0254112
Author(s):  
Faisal Maqbool Zahid ◽  
Shahla Faisal ◽  
Christian Heumann
Keyword(s):  
Multiple Imputation ◽  
High Dimensional Data ◽  
Real Life ◽  
Selection Procedure ◽  
High Dimensional ◽  
Large Set ◽  
Analysis Model ◽  
Imputation Model ◽  
Lasso Penalty ◽  
Variable Selection Procedure

Multiple Imputation (MI) is always challenging in high dimensional settings. The imputation model with some selected number of predictors can be incompatible with the analysis model leading to inconsistent and biased estimates. Although compatibility in such cases may not be achieved, but one can obtain consistent and unbiased estimates using a semi-compatible imputation model. We propose to relax the lasso penalty for selecting a large set of variables (at most n). The substantive model that also uses some formal variable selection procedure in high-dimensional structures is then expected to be nested in this imputation model. The resulting imputation model will be semi-compatible with high probability. The likelihood estimates can be unstable and can face the convergence issues as the number of variables becomes nearly as large as the sample size. To address these issues, we further propose to use a ridge penalty for obtaining the posterior distribution of the parameters based on the observed data. The proposed technique is compared with the standard MI software and MI techniques available for high-dimensional data in simulation studies and a real life dataset. Our results exhibit the superiority of the proposed approach to the existing MI approaches while addressing the compatibility issue.

Classifying Very High-Dimensional Data with Random Forests Built from Small Subspaces

2012 ◽  
Vol 8 (2) ◽  
pp. 44-63 ◽  
Author(s):  
Baoxun Xu ◽  
Joshua Zhexue Huang ◽  
Graham Williams ◽  
Qiang Wang ◽  
Yunming Ye
Keyword(s):  
Random Forest ◽  
High Dimensional Data ◽  
Real Life ◽  
Classification Performance ◽  
Feature Weighting ◽  
Random Forest Model ◽  
High Dimensional ◽  
Forest Model ◽  
Forest Models ◽  
Random Forest Models

The selection of feature subspaces for growing decision trees is a key step in building random forest models. However, the common approach using randomly sampling a few features in the subspace is not suitable for high dimensional data consisting of thousands of features, because such data often contains many features which are uninformative to classification, and the random sampling often doesn’t include informative features in the selected subspaces. Consequently, classification performance of the random forest model is significantly affected. In this paper, the authors propose an improved random forest method which uses a novel feature weighting method for subspace selection and therefore enhances classification performance over high-dimensional data. A series of experiments on 9 real life high dimensional datasets demonstrated that using a subspace size of features where M is the total number of features in the dataset, our random forest model significantly outperforms existing random forest models.

scholarly journals Parallel Framework for Dimensionality Reduction of Large-Scale Datasets

2015 ◽  
Vol 2015 ◽  
pp. 1-12 ◽  
Author(s):  
Sai Kiranmayee Samudrala ◽  
Jaroslaw Zola ◽  
Srinivas Aluru ◽  
Baskar Ganapathysubramanian
Keyword(s):  
Dimensionality Reduction ◽  
Organic Solar Cells ◽  
Large Scale ◽  
Parallel Implementation ◽  
High Dimensional Data ◽  
Real Life ◽  
Processing Parameters ◽  
High Dimensional ◽  
Morphology Evolution ◽  
Reduction Techniques

Dimensionality reduction refers to a set of mathematical techniques used to reduce complexity of the original high-dimensional data, while preserving its selected properties. Improvements in simulation strategies and experimental data collection methods are resulting in a deluge of heterogeneous and high-dimensional data, which often makes dimensionality reduction the only viable way to gain qualitative and quantitative understanding of the data. However, existing dimensionality reduction software often does not scale to datasets arising in real-life applications, which may consist of thousands of points with millions of dimensions. In this paper, we propose a parallel framework for dimensionality reduction of large-scale data. We identify key components underlying the spectral dimensionality reduction techniques, and propose their efficient parallel implementation. We show that the resulting framework can be used to process datasets consisting of millions of points when executed on a 16,000-core cluster, which is beyond the reach of currently available methods. To further demonstrate applicability of our framework we perform dimensionality reduction of 75,000 images representing morphology evolution during manufacturing of organic solar cells in order to identify how processing parameters affect morphology evolution.

Flexible, Free Software for Multilevel Multiple Imputation: A Review of Blimp and jomo

2019 ◽  
Vol 44 (5) ◽  
pp. 625-641
Author(s):  
Timothy Hayes
Keyword(s):  
Multiple Imputation ◽  
Interaction Model ◽  
Missing At Random ◽  
Analysis Model ◽  
Imputation Model ◽  
Model Following ◽  
Software Packages ◽  
Random Intercept ◽  
Impute Data ◽  
Multilevel Multiple Imputation

Multiple imputation is a popular method for addressing data that are presumed to be missing at random. To obtain accurate results, one’s imputation model must be congenial to (appropriate for) one’s intended analysis model. This article reviews and demonstrates two recent software packages, Blimp and jomo, to multiply impute data in a manner congenial with three prototypical multilevel modeling analyses: (1) a random intercept model, (2) a random slope model, and (3) a cross-level interaction model. Following these analysis examples, I review and discuss both software packages.

Variable selection techniques after multiple imputation in high-dimensional data

2019 ◽  
Vol 29 (3) ◽  
pp. 553-580
Author(s):  
Faisal Maqbool Zahid ◽  
Shahla Faisal ◽  
Christian Heumann
Keyword(s):  
Variable Selection ◽  
Multiple Imputation ◽  
High Dimensional Data ◽  
High Dimensional

scholarly journals Improving Practices for Selecting a Subset of Important Predictors in Psychology: An Application to Predicting Pain

2020 ◽  
Vol 3 (1) ◽  
pp. 66-80 ◽  
Author(s):  
Sierra A. Bainter ◽  
Thomas G. McCauley ◽  
Tor Wager ◽  
Elizabeth A. Reynolds Losin
Keyword(s):  
Variable Selection ◽  
Multiple Testing ◽  
Selection Procedure ◽  
Experimental Pain ◽  
Bayesian Variable Selection ◽  
Large Set ◽  
Web Based ◽  
Stochastic Search Variable Selection ◽  
Variable Selection Procedure ◽  
Multivariate Relationships

Frequently, researchers in psychology are faced with the challenge of narrowing down a large set of predictors to a smaller subset. There are a variety of ways to do this, but commonly it is done by choosing predictors with the strongest bivariate correlations with the outcome. However, when predictors are correlated, bivariate relationships may not translate into multivariate relationships. Further, any attempts to control for multiple testing are likely to result in extremely low power. Here we introduce a Bayesian variable-selection procedure frequently used in other disciplines, stochastic search variable selection (SSVS). We apply this technique to choosing the best set of predictors of the perceived unpleasantness of an experimental pain stimulus from among a large group of sociocultural, psychological, and neurobiological (functional MRI) individual-difference measures. Using SSVS provides information about which variables predict the outcome, controlling for uncertainty in the other variables of the model. This approach yields new, useful information to guide the choice of relevant predictors. We have provided Web-based open-source software for performing SSVS and visualizing the results.

OLAP Visualization

2011 ◽  
pp. 1439-1446 ◽  
Author(s):  
Alfredo Cuzzocrea ◽  
Svetlana Mansmann
Keyword(s):  
Computer Science ◽  
High Dimensional Data ◽  
Real Life ◽  
Science Research ◽  
Research Field ◽  
Multidimensional Data ◽  
High Dimensional ◽  
Data Sets ◽  
Visualization Of Data ◽  
Multidimensional Data Sets

The problem of efficiently visualizing multidimensional data sets produced by scientific and statistical tasks/ processes is becoming increasingly challenging, and is attracting the attention of a wide multidisciplinary community of researchers and practitioners. Basically, this problem consists in visualizing multidimensional data sets by capturing the dimensionality of data, which is the most difficult aspect to be considered. Human analysts interacting with high-dimensional data often experience disorientation and cognitive overload. Analysis of high- dimensional data is a challenge encountered in a wide set of real-life applications such as (i) biological databases storing massive gene and protein data sets, (ii) real-time monitoring systems accumulating data sets produced by multiple, multi-rate streaming sources, (iii) advanced Business Intelligence (BI) systems collecting business data for decision making purposes etc. Traditional DBMS front-end tools, which are usually tuple-bag-oriented, are completely inadequate to fulfill the requirements posed by an interactive exploration of high-dimensional data sets due to two major reasons: (i) DBMS implement the OLTP paradigm, which is optimized for transaction processing and deliberately neglects the dimensionality of data; (ii) DBMS operators are very poor and offer nothing beyond the capability of conventional SQL statements, what makes such tools very inefficient with respect to the goal of visualizing and, above all, interacting with multidimensional data sets embedding a large number of dimensions. Despite the above-highlighted practical relevance of the problem of visualizing multidimensional data sets, the literature in this field is rather scarce, due to the fact that, for many years, this problem has been of relevance for life science research communities only, and interaction of the latter with the computer science research community has been insufficient. Following the enormous growth of scientific disciplines like Bio-Informatics, this problem has then become a fundamental field in the computer science academic as well as industrial research. At the same time, a number of proposals dealing with the multidimensional data visualization problem appeared in literature, with the amenity of stimulating novel and exciting application fields such as the visualization of Data Mining results generated by challenging techniques like clustering and association rule discovery. The above-mentioned issues are meant to facilitate understanding of the high relevance and attractiveness of the problem of visualizing multidimensional data sets at present and in the future, with challenging research findings accompanied by significant spin-offs in the Information Technology (IT) industrial field. A possible solution to tackle this problem is represented by well-known OLAP techniques (Codd et al., 1993; Chaudhuri & Dayal, 1997; Gray et al., 1997), focused on obtaining very efficient representations of multidimensional data sets, called data cubes, thus leading to the research field which is known in literature under the terms OLAP Visualization and Visual OLAP, which, in the remaining part of the article, are used interchangeably.

scholarly journals A Hybrid Semi-Supervised Anomaly Detection Model for High-Dimensional Data

2017 ◽  
Vol 2017 ◽  
pp. 1-9 ◽  
Author(s):  
Hongchao Song ◽  
Zhuqing Jiang ◽  
Aidong Men ◽  
Bo Yang
Keyword(s):  
Anomaly Detection ◽  
Nearest Neighbor ◽  
Dimensional Space ◽  
High Dimensional Data ◽  
Real Life ◽  
Detection Methods ◽  
High Dimensional ◽  
Detection Accuracy ◽  
Detection Model ◽  
Anomaly Detector

Anomaly detection, which aims to identify observations that deviate from a nominal sample, is a challenging task for high-dimensional data. Traditional distance-based anomaly detection methods compute the neighborhood distance between each observation and suffer from the curse of dimensionality in high-dimensional space; for example, the distances between any pair of samples are similar and each sample may perform like an outlier. In this paper, we propose a hybrid semi-supervised anomaly detection model for high-dimensional data that consists of two parts: a deep autoencoder (DAE) and an ensemble k-nearest neighbor graphs- (K-NNG-) based anomaly detector. Benefiting from the ability of nonlinear mapping, the DAE is first trained to learn the intrinsic features of a high-dimensional dataset to represent the high-dimensional data in a more compact subspace. Several nonparametric KNN-based anomaly detectors are then built from different subsets that are randomly sampled from the whole dataset. The final prediction is made by all the anomaly detectors. The performance of the proposed method is evaluated on several real-life datasets, and the results confirm that the proposed hybrid model improves the detection accuracy and reduces the computational complexity.

scholarly journals Statistical challenges of high-dimensional data

2009 ◽  
Vol 367 (1906) ◽  
pp. 4237-4253 ◽  
Author(s):  
Iain M. Johnstone ◽  
D. Michael Titterington
Keyword(s):  
High Dimensional Data ◽  
Large Datasets ◽  
The Other ◽  
High Dimensional ◽  
Large Set ◽  
Parameter Vector ◽  
Theme Issue ◽  
Data Space ◽  
Recent Developments ◽  
Low Dimensional

Modern applications of statistical theory and methods can involve extremely large datasets, often with huge numbers of measurements on each of a comparatively small number of experimental units. New methodology and accompanying theory have emerged in response: the goal of this Theme Issue is to illustrate a number of these recent developments. This overview article introduces the difficulties that arise with high-dimensional data in the context of the very familiar linear statistical model: we give a taste of what can nevertheless be achieved when the parameter vector of interest is sparse, that is, contains many zero elements. We describe other ways of identifying low-dimensional subspaces of the data space that contain all useful information. The topic of classification is then reviewed along with the problem of identifying, from within a very large set, the variables that help to classify observations. Brief mention is made of the visualization of high-dimensional data and ways to handle computational problems in Bayesian analysis are described. At appropriate points, reference is made to the other papers in the issue.

scholarly journals A Study On High Dimensional Clustering By Using Clique

2011 ◽  
pp. 101-104
Author(s):  
Raghunath Kar ◽  
Susanta Kumar Das
Keyword(s):  
Input Parameter ◽  
High Dimensional Data ◽  
Real Life ◽  
Multidimensional Data ◽  
High Dimensional ◽  
Data Sets ◽  
Data Points ◽  
A Cell ◽  
Low Dimensional ◽  
Partition Method

In real life clustering of high dimensional data is a big problem. To find out the dense regions from increasing dimensions is one of them. We have already studied the clustering techniques of low dimensional data sets like k-means, k-mediod, BIRCH, CLARANS, CURE, DBScan, PAM etc. If a region is dense then it consists with number of data points with a minimum support of input parameter ø other wise it cannot take into clustering. So in this approach we have implemented CLIQUE to find out the clusters from multidimensional data sets. In dimension growth subspace clustering the clustering process start at single dimensional subspaces and grows upward to higher dimensional ones. It is a partition method where each dimension divided like a grid structure. The grid is a cell where the data points are present. We check the dense units from the structure by applying different algorithms. Finally the clusters are formed from the high dimensional data sets.

Sign in / Sign up

Export Citation Format

Share Document