Xinhai Liu

Optimized Data Fusion for Kernel k-Means Clustering

This paper presents a novel optimized kernel k-means algorithm (OKKC) to combine multiple data sources for clustering analysis. The algorithm uses an alternating minimization framework to optimize the cluster membership and kernel coefficients as a nonconvex problem. In the proposed algorithm, the problem to optimize the cluster membership and the problem to optimize the kernel coefficients are all based on the same Rayleigh quotient objective; therefore the proposed algorithm converges locally. OKKC has a simpler procedure and lower complexity than other algorithms proposed in the literature. Simulated and real-life data fusion applications are experimentally studied, and the results validate that the proposed algorithm has comparable performance, moreover, it is more efficient on large-scale data sets. (The Matlab implementation of OKKC algorithm is downloadable from http://homes.esat.kuleuven.be/~sistawww/bio/syu/okkc.html.).

Subject clustering analysis based on ISI category classification

The study focuses on the analysis of the information flow among the ISI subject categories and aims at finding an appropriate field structure of the Web of Science using the subject clustering algorithm developed in previous studies. The clustering journals and ISI subject categories provide two subject classification schemes through different perspectives and levels. The two clustering results have been compared and their accordance and divergence have been analyzed. Several indicators have been used to compare the communication characteristics among different ISI subject categories. The neighbour map of each category clearly reflects the affinities between the “core” category and its satellites around.

Optimized data fusion for K-means Laplacian clustering

Motivation: We propose a novel algorithm to combine multiple kernels and Laplacians for clustering analysis. The new algorithm is formulated on a Rayleigh quotient objective function and is solved as a bi-level alternating minimization procedure. Using the proposed algorithm, the coefficients of kernels and Laplacians can be optimized automatically. Results: Three variants of the algorithm are proposed. The performance is systematically validated on two real-life data fusion applications. The proposed Optimized Kernel Laplacian Clustering (OKLC) algorithms perform significantly better than other methods. Moreover, the coefficients of kernels and Laplacians optimized by OKLC show some correlation with the rank of performance of individual data source. Though in our evaluation the K values are predefined, in practical studies, the optimal cluster number can be consistently estimated from the eigenspectrum of the combined kernel Laplacian matrix. Availability: The MATLAB code of algorithms implemented in this paper is downloadable from http://homes.esat.kuleuven.be/~sistawww/bioi/syu/oklc.html. Contact: shiyu@uchicago.edu Supplementary information: Supplementary data are available at Bioinformatics online.

Hybrid clustering of multiple information sources via HOSVD

We present a hybrid clustering algorithm of multiple information sources via tensor decomposition, which can be regarded an extension of the spectral clustering based on modularity maximization This hybrid clustering can be solved by the truncated higher-order singular value decomposition (HOSVD) Experimental results conducted on the synthetic data have demonstrated the effectiveness.

Hybrid clustering of multi-view data via Tucker-2 model and its application

With the modern technology fast developing, most of entities can be observed by different perspectives. These multiple view information allows us to find a better pattern as long as we integrate them in an appropriate way. So clustering by integrating multi-view representations that describe the same class of entities has become a crucial issue for knowledge discovering. We integrate multi-view data by a tensor model and present a hybrid clustering method based on Tucker-2 model, which can be regarded as an extension of spectral clustering. We apply our hybrid clustering method to scientific publication analysis by integrating citation-link and lexical content. Clustering experiments are conducted on a large-scale journal set retrieved from the Web of Science (WoS) database. Several relevant hybrid clustering methods are cross compared with our method. The analysis of clustering results demonstrate the effectiveness of the proposed algorithm. Furthermore, we provide a cognitive analysis of the clustering results as well as the visualization as a mapping of the journal set.

Multi-view spectral clustering and its chemical application

Clustering is an unsupervised method that allows researchers to group objects and gather information about their relationships. In chemoinformatics, clustering enables hypotheses to be drawn about a compounds biological, chemical and physical property in comparison to another. We introduce a novel improved spectral clustering algorithm, proposed for chemical compound clustering, using multiple data sources. Tensor-based spectral methods, used in this paper, provide chemically appropriate and statistically significant results when attempting to cluster compounds from both the GSK-Chembl Malaria data set and the Zinc database. Spectral clustering algorithms based on the tensor method give robust results on the mid-size compound sets used here. The goal of this paper is to present the clustering of chemical compounds, using a tensor-based multi-view method which proves of value to the medicinal chemistry community. Our findings show compounds of extremely different chemotypes clustering together, this is a hint to the chemogenomics nature of our method.

Hybrid Clustering by Integrating Text and Citation Based Graphs in Journal Database Analysis

We propose a hybrid clustering strategy by integrating heterogeneous information sources as graphs. The hybrid clustering method is extended on the basis of modularity based Louvain method. We introduce two different approaches, graph coupling and graph fusion. The weights of these combined graphs are optimized with the criterion of maximizing the Average Normalized Mutual Information(ANMI). The methods are applied to obtain structural mapping of large scale Web of Science (WoS) journal database by integrating attribute based textual information and relation based citation information. From the experimental, the proposed graph combination scheme is compared with individual graph clustering, spectral clustering and Vector Space Model(VSM) based clustering methods.