arXiv subjects

Zengyou He

Publications and source records attributed to Zengyou He.

At least 37 records · Page 2Linked to original sources

arXiv2019

Detecting Statistically Significant Communities

Community detection is a key data analysis problem across different fields. During the past decades, numerous algorithms have been proposed to address this issue. However, most work on community detection does not address the issue of statistical significance. Although some research efforts have been made towards mining statistically significant communities, deriving an analytical solution of p-value for one community under the configuration model is still a challenging mission that remains unsolved. The configuration model is a widely used random graph model in community detection, in which the degree of each node is preserved in the generated random networks. To partially fulfill this void, we present a tight upper bound on the p-value of a single community under the configuration model, which can be used for quantifying the statistical significance of each community analytically. Meanwhile, we present a local search method to detect statistically significant communities in an iterative manner. Experimental results demonstrate that our method is comparable with the competing methods on detecting statistically significant communities.

cs.SI↗

arXiv2019

Instance-Based Classification through Hypothesis Testing

Classification is a fundamental problem in machine learning and data mining. During the past decades, numerous classification methods have been presented based on different principles. However, most existing classifiers cast the classification problem as an optimization problem and do not address the issue of statistical significance. In this paper, we formulate the binary classification problem as a two-sample testing problem. More precisely, our classification model is a generic framework that is composed of two steps. In the first step, the distance between the test instance and each training instance is calculated to derive two distance sets. In the second step, the two-sample test is performed under the null hypothesis that the two sets of distances are drawn from the same cumulative distribution. After these two steps, we have two p-values for each test instance and the test instance is assigned to the class associated with the smaller p-value. Essentially, the presented classification method can be regarded as an instance-based classifier based on hypothesis testing. The experimental results on 40 real data sets show that our method is able to achieve the same level performance as the state-of-the-art classifiers and has significantly better performance than existing testing-based classifiers. Furthermore, we can handle outlying instances and control the false discovery rate of test instances assigned to each class under the same framework.

cs.LG↗

arXiv2019

Correlation-Based Community Detection

Mining community structures from the complex network is an important problem across a variety of fields. Many existing community detection methods detect communities through optimizing a community evaluation function. However, most of these functions even have high values on random graphs and may fail to detect small communities in the large-scale network (the so-called resolution limit problem). In this paper, we introduce two novel node-centric community evaluation functions by connecting correlation analysis with community detection. We will further show that the correlation analysis can provide a novel theoretical framework which unifies some existing evaluation functions in the context of a correlation-based optimization problem. In this framework, we can mitigate the resolution limit problem and eliminate the influence of random fluctuations by selecting the right correlation function. Furthermore, we introduce three key properties used in mining association rule into the context of community detection to help us choose the appropriate correlation function. Based on our introduced correlation functions, we propose a community detection algorithm called CBCD. Our proposed algorithm outperforms existing state-of-the-art algorithms on both synthetic benchmark networks and real-world networks.

cs.SI↗

arXiv2012

Running PeptideProphet Separately on Replicates Improves Peptide Identification Results

Limited spectrum coverage is a problem in shotgun proteomics. Replicates are generated to improve the spectrum coverage. When integrating peptide identification results obtained from replicates, the state-of-the-art algorithm PeptideProphet combines Peptide-Spectrum Matches (PSMs) before building the statistical model to calculate peptide probabilities. In this paper, we find the connection between merging results of replicates and Bagging, which is a standard routine to improve the power of statistical methods. Following Bagging's philosophy, we propose to run PeptideProphet separately on each replicate and combine the outputs to obtain the final peptide probabilities. In our experiments, we show that the proposed routine can improve PeptideProphet consistently on a standard protein dataset, a Human dataset and a Yeast dataset.

q-bio.QM↗

arXiv2012

A Combinatorial Perspective of the Protein Inference Problem

In a shotgun proteomics experiment, proteins are the most biologically meaningful output. The success of proteomics studies depends on the ability to accurately and efficiently identify proteins. Many methods have been proposed to facilitate the identification of proteins from the results of peptide identification. However, the relationship between protein identification and peptide identification has not been thoroughly explained before. In this paper, we are devoted to a combinatorial perspective of the protein inference problem. We employ combinatorial mathematics to calculate the conditional protein probabilities (Protein probability means the probability that a protein is correctly identified) under three assumptions, which lead to a lower bound, an upper bound and an empirical estimation of protein probabilities, respectively. The combinatorial perspective enables us to obtain a closed-form formulation for protein inference. Based on our model, we study the impact of unique peptides and degenerate peptides on protein probabilities. Here, degenerate peptides are peptides shared by at least two proteins. Meanwhile, we also study the relationship of our model with other methods such as ProteinProphet. A probability confidence interval can be calculated and used together with probability to filter the protein identification result. Our method achieves competitive results with ProteinProphet in a more efficient manner in the experiment based on two datasets of standard protein mixtures and two datasets of real samples. We name our program ProteinInfer. Its Java source code is available at http://bioinformatics.ust.hk/proteininfer

q-bio.QM↗

arXiv2012

On unbiased performance evaluation for protein inference

This letter is a response to the comments of Serang (2012) on Huang and He (2012) in Bioinformatics. Serang (2012) claimed that the parameters for the Fido algorithm should be specified using the grid search method in Serang et al. (2010) so as to generate a deserved accuracy in performance comparison. It seems that it is an argument on parameter tuning. However, it is indeed the issue of how to conduct an unbiased performance evaluation for comparing different protein inference algorithms. In this letter, we would explain why we don't use the grid search for parameter selection in Huang and He (2012) and show that this procedure may result in an over-estimated performance that is unfair to competing algorithms. In fact, this issue has also been pointed out by Li and Radivojac (2012).

stat.AP↗

arXiv2012

Protein Inference and Protein Quantification: Two Sides of the Same Coin

Motivation: In mass spectrometry-based shotgun proteomics, protein quantification and protein identification are two major computational problems. To quantify the protein abundance, a list of proteins must be firstly inferred from the sample. Then the relative or absolute protein abundance is estimated with quantification methods, such as spectral counting. Until now, researchers have been dealing with these two processes separately. In fact, they are two sides of same coin in the sense that truly present proteins are those proteins with non-zero abundances. Then, one interesting question is if we regard the protein inference problem as a special protein quantification problem, is it possible to achieve better protein inference performance? Contribution: In this paper, we investigate the feasibility of using protein quantification methods to solve the protein inference problem. Protein inference is to determine whether each candidate protein is present in the sample or not. Protein quantification is to calculate the abundance of each protein. Naturally, the absent proteins should have zero abundances. Thus, we argue that the protein inference problem can be viewed as a special case of protein quantification problem: present proteins are those proteins with non-zero abundances. Based on this idea, our paper tries to use three very simple protein quantification methods to solve the protein inference problem effectively. Results: The experimental results on six datasets show that these three methods are competitive with previous protein inference algorithms. This demonstrates that it is plausible to take the protein inference problem as a special case of protein quantification, which opens the door of devising more effective protein inference algorithms from a quantification perspective.

cs.CE↗

arXiv2010

Efficient K-Nearest Neighbor Join Algorithms for High Dimensional Sparse Data

The K-Nearest Neighbor (KNN) join is an expensive but important operation in many data mining algorithms. Several recent applications need to perform KNN join for high dimensional sparse data. Unfortunately, all existing KNN join algorithms are designed for low dimensional data. To fulfill this void, we investigate the KNN join problem for high dimensional sparse data. In this paper, we propose three KNN join algorithms: a brute force (BF) algorithm, an inverted index-based(IIB) algorithm and an improved inverted index-based(IIIB) algorithm. Extensive experiments on both synthetic and real-world datasets were conducted to demonstrate the effectiveness of our algorithms for high dimensional sparse data.

cs.DB↗

arXiv2010

Stable Feature Selection for Biomarker Discovery

Feature selection techniques have been used as the workhorse in biomarker discovery applications for a long time. Surprisingly, the stability of feature selection with respect to sampling variations has long been under-considered. It is only until recently that this issue has received more and more attention. In this article, we review existing stable feature selection methods for biomarker discovery using a generic hierarchal framework. We have two objectives: (1) providing an overview on this new yet fast growing topic for a convenient reference; (2) categorizing existing methods under an expandable framework for future research and development.

cs.CE↗

arXiv2007

Attribute Value Weighting in K-Modes Clustering

In this paper, the traditional k-modes clustering algorithm is extended by weighting attribute value matches in dissimilarity computation. The use of attribute value weighting technique makes it possible to generate clusters with stronger intra-similarities, and therefore achieve better clustering performance. Experimental results on real life datasets show that these value weighting based k-modes algorithms are superior to the standard k-modes algorithm with respect to clustering accuracy.

cs.AI↗

arXiv2006

Farthest-Point Heuristic based Initialization Methods for K-Modes Clustering

The k-modes algorithm has become a popular technique in solving categorical data clustering problems in different application domains. However, the algorithm requires random selection of initial points for the clusters. Different initial points often lead to considerable distinct clustering results. In this paper we present an experimental study on applying a farthest-point heuristic based initialization method to k-modes clustering to improve its performance. Experiments show that new initialization method leads to better clustering accuracy than random selection initialization method for k-modes clustering.

cs.AI↗

arXiv2006

Approximation Algorithms for K-Modes Clustering

In this paper, we study clustering with respect to the k-modes objective function, a natural formulation of clustering for categorical data. One of the main contributions of this paper is to establish the connection between k-modes and k-median, i.e., the optimum of k-median is at most twice the optimum of k-modes for the same categorical data clustering problem. Based on this observation, we derive a deterministic algorithm that achieves an approximation factor of 2. Furthermore, we prove that the distance measure in k-modes defines a metric. Hence, we are able to extend existing approximation algorithms for metric k-median to k-modes. Empirical results verify the superiority of our method.

cs.AI↗

arXiv2005

K-ANMI: A Mutual Information Based Clustering Algorithm for Categorical Data

Clustering categorical data is an integral part of data mining and has attracted much attention recently. In this paper, we present k-ANMI, a new efficient algorithm for clustering categorical data. The k-ANMI algorithm works in a way that is similar to the popular k-means algorithm, and the goodness of clustering in each step is evaluated using a mutual information based criterion (namely, Average Normalized Mutual Information-ANMI) borrowed from cluster ensemble. Experimental results on real datasets show that k-ANMI algorithm is competitive with those state-of-art categorical data clustering algorithms with respect to clustering accuracy.

cs.AI↗

arXiv2005

K-Histograms: An Efficient Clustering Algorithm for Categorical Dataset

Clustering categorical data is an integral part of data mining and has attracted much attention recently. In this paper, we present k-histogram, a new efficient algorithm for clustering categorical data. The k-histogram algorithm extends the k-means algorithm to categorical domain by replacing the means of clusters with histograms, and dynamically updates histograms in the clustering process. Experimental results on real datasets show that k-histogram algorithm can produce better clustering results than k-modes algorithm, the one related with our work most closely.

cs.AI↗

arXiv2005

Clustering Mixed Numeric and Categorical Data: A Cluster Ensemble Approach

Clustering is a widely used technique in data mining applications for discovering patterns in underlying data. Most traditional clustering algorithms are limited to handling datasets that contain either numeric or categorical attributes. However, datasets with mixed types of attributes are common in real life data mining applications. In this paper, we propose a novel divide-and-conquer technique to solve this problem. First, the original mixed dataset is divided into two sub-datasets: the pure categorical dataset and the pure numeric dataset. Next, existing well established clustering algorithms designed for different types of datasets are employed to produce corresponding clusters. Last, the clustering results on the categorical and numeric dataset are combined as a categorical dataset, on which the categorical data clustering algorithm is used to get the final clusters. Our contribution in this paper is to provide an algorithm framework for the mixed attributes clustering problem, in which existing clustering algorithms can be easily integrated, the capabilities of different kinds of clustering algorithms and characteristics of different types of datasets could be fully exploited. Comparisons with other clustering algorithms on real life datasets illustrate the superiority of our approach.

cs.AI↗

arXiv2005

A Fast Greedy Algorithm for Outlier Mining

The task of outlier detection is to find small groups of data objects that are exceptional when compared with rest large amount of data. In [38], the problem of outlier detection in categorical data is defined as an optimization problem and a local-search heuristic based algorithm (LSA) is presented. However, as is the case with most iterative type algorithms, the LSA algorithm is still very time-consuming on very large datasets. In this paper, we present a very fast greedy algorithm for mining outliers under the same optimization model. Experimental results on real datasets and large synthetic datasets show that: (1) Our algorithm has comparable performance with respect to those state-of-art outlier detection algorithms on identifying true outliers and (2) Our algorithm can be an order of magnitude faster than LSA algorithm.

cs.DB↗

arXiv2005

A Unified Subspace Outlier Ensemble Framework for Outlier Detection in High Dimensional Spaces

The task of outlier detection is to find small groups of data objects that are exceptional when compared with rest large amount of data. Detection of such outliers is important for many applications such as fraud detection and customer migration. Most such applications are high dimensional domains in which the data may contain hundreds of dimensions. However, the outlier detection problem itself is not well defined and none of the existing definitions are widely accepted, especially in high dimensional space. In this paper, our first contribution is to propose a unified framework for outlier detection in high dimensional spaces from an ensemble-learning viewpoint. In our new framework, the outlying-ness of each data object is measured by fusing outlier factors in different subspaces using a combination function. Accordingly, we show that all existing researches on outlier detection can be regarded as special cases in the unified framework with respect to the set of subspaces considered and the type of combination function used. In addition, to demonstrate the usefulness of the ensemble-learning based outlier detection framework, we developed a very simple and fast algorithm, namely SOE1 (Subspace Outlier Ensemble using 1-dimensional Subspaces) in which only subspaces with one dimension is used for mining outliers from large categorical datasets. The SOE1 algorithm needs only two scans over the dataset and hence is very appealing in real data mining applications. Experimental results on real datasets and large synthetic datasets show that: (1) SOE1 has comparable performance with respect to those state-of-art outlier detection algorithms on identifying true outliers and (2) SOE1 can be an order of magnitude faster than one of the fastest outlier detection algorithms known so far.

cs.DB↗

arXiv2005

An Optimization Model for Outlier Detection in Categorical Data

cs.DB↗