Range Of Real Datasets Research Articles

High-dimensional asymptotics of denoising autoencoders*

Abstract We address the problem of denoising data from a Gaussian mixture using a two-layer non-linear autoencoder with tied weights and a skip connection. We consider the high-dimensional limit where the number of training samples and the input dimension jointly tend to infinity while the number of hidden units remains bounded. We provide closed-form expressions for the denoising mean-squared test error. Building on this result, we quantitatively characterize the advantage of the considered architecture over the autoencoder without the skip connection that relates closely to principal component analysis. We further show that our results accurately capture the learning curves on a range of real data sets.

Gaussian universality of perceptrons with random labels.

While classical in many theoretical settings-and in particular in statistical physics-inspired works-the assumption of Gaussian i.i.d. input data is often perceived as a strong limitation in the context of statistics and machine learning. In this study, we redeem this line of work in the case of generalized linear classification, also known as the perceptron model, with random labels. We argue that there is a large universality class of high-dimensional input data for which we obtain the same minimum training loss as for Gaussian data with corresponding data covariance. In the limit of vanishing regularization, we further demonstrate that the training loss is independent of the data covariance. On the theoretical side, we prove this universality for an arbitrary mixture of homogeneous Gaussian clouds. Empirically, we show that the universality holds also for a broad range of real data sets.

K-mer clustering algorithm using a MapReduce framework: application to the parallelization of the Inchworm module of Trinity

BackgroundDe novo transcriptome assembly is an important technique for understanding gene expression in non-model organisms. Many de novo assemblers using the de Bruijn graph of a set of the RNA sequences rely on in-memory representation of this graph. However, current methods analyse the complete set of read-derived k-mer sequence at once, resulting in the need for computer hardware with large shared memory.ResultsWe introduce a novel approach that clusters k-mers as the first step. The clusters correspond to small sets of gene products, which can be processed quickly to give candidate transcripts. We implement the clustering step using the MapReduce approach for parallelising the analysis of large datasets, which enables the use of compute clusters. The computational task is distributed across the compute system using the industry-standard MPI protocol, and no specialised hardware is required. Using this approach, we have re-implemented the Inchworm module from the widely used Trinity pipeline, and tested the method in the context of the full Trinity pipeline. Validation tests on a range of real datasets show large reductions in the runtime and per-node memory requirements, when making use of a compute cluster.ConclusionsOur study shows that MapReduce-based clustering has great potential for distributing challenging sequencing problems, without loss of accuracy. Although we have focussed on the Trinity package, we propose that such clustering is a useful initial step for other assembly pipelines.

Sparse graphs using exchangeable random measures.

SummaryStatistical network modelling has focused on representing the graph as a discrete structure, namely the adjacency matrix. When assuming exchangeability of this array—which can aid in modelling, computations and theoretical analysis—the Aldous–Hoover theorem informs us that the graph is necessarily either dense or empty. We instead consider representing the graph as an exchangeable random measure and appeal to the Kallenberg representation theorem for this object. We explore using completely random measures (CRMs) to define the exchangeable random measure, and we show how our CRM construction enables us to achieve sparse graphs while maintaining the attractive properties of exchangeability. We relate the sparsity of the graph to the Lévy measure defining the CRM. For a specific choice of CRM, our graphs can be tuned from dense to sparse on the basis of a single parameter. We present a scalable Hamiltonian Monte Carlo algorithm for posterior inference, which we use to analyse network properties in a range of real data sets, including networks with hundreds of thousands of nodes and millions of edges.

PFIN: A Parallel Frequent Itemset Mining Algorithm Using Nodesets

Frequent Itemset Mining (FIM) is one of most fundamental techniques in data mining with extensive applications to a variety of data mining problems such as association rule mining, correlations, clustering and classification. Since the first proposal of frequent itemset mining, numerous serial algorithms have been proposed in order to improve mining performance, yet most of them cannot scale to massive datasets which are very common nowadays. In this paper, we propose a new parallel FIM algorithm named PFIN based on Nodeset which is a more efficient data structure for mining frequent itemsets. PFIN can intelligently decompose a large-scale FIM problem into a set of tasks, where each task can be executed in parallel without unnecessary communication overheads. Moreover, a hash-based load balancing strategy has been adopted to optimize resource use and maximize throughput. For evaluating the performance of PFIN, we have conduct extensive experiments on Spark which is an emerging distributed in-memory processing framework to compare it against PFP which is one of state-of-the-art parallel FIM algorithms on a range of real datasets. The experimental results demonstrate that our proposed PFIN are highly competitive with PFP in scalability performance, outperforming PFP in speed performance.

Supervised box clustering

In this work we address a technique for effectively clustering points in specific convex sets, called homogeneous boxes, having sides aligned with the coordinate axes (isothetic condition). The proposed clustering approach is based on homogeneity conditions, not according to some distance measure, and, even if it was originally developed in the context of the logical analysis of data, it is now placed inside the framework of Supervised clustering. First, we introduce the basic concepts in box geometry; then, we consider a generalized clustering algorithm based on a class of graphs, called incompatibility graphs. For supervised classification problems, we consider classifiers based on box sets, and compare the overall performances to the accuracy levels of competing methods for a wide range of real data sets. The results show that the proposed method performs comparably with other supervised learning methods in terms of accuracy.

Intelligent Dynamic Aging Approaches in Web Proxy Cache Replacement

One of commonly used approach to enhance the Web performance is Web proxy caching technique. In Web proxy caching, Least-Frequently-Used-Dynamic-Aging (LFU-DA) is one of the common proxy cache replacement methods, which is widely used in Web proxy cache management. LFU-DA accomplishes a superior byte hit ratio compared to other Web proxy cache replacement algorithms. However, LFU-DA may suffer in hit ratio measure. Therefore, in this paper, LFU-DA is enhanced using popular supervised machine learning techniques such as a support vector machine (SVM), a naive Bayes classifier (NB) and a decision tree (C4.5). SVM, NB and C4.5 are trained from Web proxy logs files and then intelligently incorporated with LFU-DA to form Intelligent Dynamic- Aging (DA) approaches. The simulation results revealed that the proposed intelligent Dynamic- Aging approaches considerably improved the performances in terms of hit and byte hit ratio of the conventional LFU-DA on a range of real datasets.

Pruning boxes in a box-based classification method

In this work we address an extension of box clustering in supervised classification problems that makes use of optimization problems to refine the results obtained by agglomerative techniques. The central concept of box clustering is that of homogeneous boxes that give rise to overtrained classifiers under some conditions. Thus, we focus our attentions on the issue of pruning out redundant boxes, using the information gleaned from the other boxes generated under the hypothesis that such a choice would identify simpler models with good predictive power. We propose a pruning method based on an integer optimization problem and a family of sub problems derived from the main one. The overall performances are then compared to the accuracy levels of competing methods on a wide range of real data sets. The method has proven to be robust, making it possible to derive a more compact system of boxes in the instance space with good performance on training and test data.

Identifying clusters in genomics data by recursive partitioning

Genomics studies frequently involve clustering of molecular data to identify groups, but common clustering methods such as K-means clustering and hierarchical clustering do not determine the number of clusters. Methods for estimating the number of clusters typically focus on identifying the global structure in the data, however the discovery of substructures within clusters may also be of great biological interest. We propose a novel method, Partitioning Algorithm based on Recursive Thresholding (PART), that recursively uncovers distinct subgroups in the groups already identified. Outliers are common in high-dimensional genomics data and may mask the presence of substructure within a cluster. A crucial feature of the algorithm is the introduction of tentative splits of clusters to isolate outliers that might otherwise halt the recursion prematurely. The method is demonstrated on simulated as well as a wide range of real data sets from gene expression microarrays, where the correct clusters were known in advance. When subclusters are present and the variance is large or varies between the clusters, the proposed method performs better than two established global methods on simulated data. On the real data sets the overall performance of PART is superior to the global methods when used in combination with hierarchical clustering. The method is implemented in the R package clusterGenomics and is freely available from CRAN (The Comprehensive R Archive Network).

Stereo Matching with Mumford-Shah Regularization and Occlusion Handling

This paper addresses the problem of correspondence establishment in binocular stereo vision. We suggest a novel spatially continuous approach for stereo matching based on the variational framework. The proposed method suggests a unique regularization term based on Mumford-Shah functional for discontinuity preserving, combined with a new energy functional for occlusion handling. The evaluation process is based on concurrent minimization of two coupled energy functionals, one for domain segmentation (occluded versus visible) and the other for disparity evaluation. In addition to a dense disparity map, our method also provides an estimation for the half-occlusion domain and a discontinuity function allocating the disparity/depth boundaries. Two new constraints are introduced improving the revealed discontinuity map. The experimental tests include a wide range of real data sets from the Middlebury stereo database. The results demonstrate the capability of our method in calculating an accurate disparity function with sharp discontinuities and occlusion map recovery. Significant improvements are shown compared to a recently published variational stereo approach. A comparison on the Middlebury stereo benchmark with subpixel accuracies shows that our method is currently among the top-ranked stereo matching algorithms.

A critique of the use of jackknife and related non-parametric techniques to estimate species richness

Species richness in an assemblage is frequently used as a measure of biological diversity. However, observed species richness is strongly dependent on sample size. If more samples are collected, then more species are observed. Non-parametric species richness estimators, such as the jackknife 1 and 2 and the Chao 1 and 2, are indicated in the literature as potential solutions to the problem of dependence of observed species richness on sampling effort. These methods are intended to estimate the total species richness in an area or assemblage with small sampling effort. Non-parametric estimators are based on the number of species observed, and the number of rare species in a sample, i.e., that occurred in one and/or two sampling units, or with one and/or two individuals. High estimates are produced when samples contain large proportions of rare species. Using a range of real datasets, I show that estimates produced by non-parametric methods are generally dependent on observed species richness. An implicit a...