With the rapid expansion of internet, complex networks has become high-dimensional, sparse and redundant. Besides, problem link prediction in such also obatined increasingly attention from different types domains like information science, anthropology, sociology computer sciences. It makes requirements for effective techniques to extract most essential relevant online users internet. Therefore, this paper attempts put forward a algorithm based on non-negative matrix factorization. In...

Multiple sequence alignment is an important and difficult problem in molecular biology bioinformatics. In this paper, we propose a partitioning approach that significantly improves the solution time quality by utilizing locality structure of problem. The algorithm solves multiple three stages. First, automated suboptimal strategy used to divide set sequences into several subsections. Then based on ant colony optimization align each subsection. Finally, original can be obtained assembling...

Identifying essential proteins is not only important for understanding cellular activity, but also detecting human disease genes. A series of centrality measures have been proposed to identify based on the protein-protein interaction(PPI) network. However, most identifying algorithms are static PPI networks which cannot reflect dynamic and transient nature protein interactions. Meanwhile, studies shown that essentiality a product complex rather than individual protein. Therefore, we new...

Understanding the localization of proteins in cells is vital to characterizing their functions and possible interactions. As a result, identifying (sub)cellular compartment within which protein located becomes an important problem classification. This classification issue thus involves predicting labels dataset with limited number labeled data points available. By utilizing graph representation data, random walk techniques have performed well sequence functional prediction; however, this...

Community detection in bipartite network is very important the research on theory and applications of complex analysis. In this paper, an algorithm for detecting community structure networks based matrix factorisation presented. The first partitions into two parts, each which can reserve information as much possible, then parts are further recursively partitioned until modularity cannot be improved. While partitioning network, we use approach decomposition so that row space associated...

In this paper, we present a label propagation algorithm named ACD for anti-community detection. Experimental results on some real world networks show that our can obtain higher quality than other methods.

A multiple sequence alignment algorithm based on divide-and-conquer method and ant colony algorithms is proposed. In the algorithm, optimization used to partition set of sequences into several subsections vertically. Then procedure partitioning applied prefix family suffix in a recursive manner until lengths subsequences are all equal 1. Experimental results show that can get high quality solution reduce running time.

By simulating the population size of human evolution, a PSO algorithm with increment particle (IPPSO) was proposed. Without changing operations, IPPSO can obtain better solutions less time cost by modifying structure traditional PSO. Experimental results show that using logistic model is more efficient and requires computation than linear function in solving complex program problems.

Identifying essential proteins is important for not only understanding cellular activity but also detecting human disease genes. A series of centrality measures have been proposed to identify based on the protein-protein interaction (PPI) network. Although, existing studies focused topological features PPI network and intrinsic characteristics biological attributes. it still a big challenge further improve prediction accuracy proteins. Moreover, there are substantial amounts false-positive...

We present a particle swarm optimization algorithm OT-PSO using orthogonal test technique. Based on the classical PSO, searches for local optimum in neighbor area of global best solution by method design. can probe solutions uniformly distributed search space and select better ones. Using those guide particles searching towards correct direction latter iterations so as to speed up convergence, get more precise avoid optimum. Our experimental results show that not only has faster convergence...

Text classification refers to determine the class of an unknown text according its content in given system. In order extract fewer features express information as much possible, paper analysis various features' statistical properties and global Zipf's law; then, based on classified information, efficient feature is extracted by computing contribute a feature; After that, traditional TF-IDF formula improved using category frequencies named TF-IDF-CF for calculating weight; Finally method...