An experimental investigation of kernels on graphs for collaborative recommendation and semisupervised classification

Fouss, François; Françoisse, Kevin; Yen, Luh; Pirotte, Alain; Saerens, Marco

doi:10.1016/j.neunet.2012.03.001

Cited by 124 publications

(125 citation statements)

References 66 publications

Supporting

Mentioning

125

Contrasting

Order By: Relevance

“…graph Laplacian, or transition probability matrix. We refer interested readers to [5] for a comprehensive review of graph kernels.…”

Section: Preliminaries 21 Graph Kernels: Node Similarity Measuresmentioning

confidence: 99%

“…The kernel matrices ( , ) are all symmetric positive semi-definite [5]. Let be a matrix whose columns are 's orthonormal eigenvectors; = ( , , … ) where are 's eigenvalues, kernel matrix can be factorized:…”

Section: Kernel-based Euclidean Embeddingmentioning

confidence: 99%

“…Other graph kernels were constructed based derivatives of graph adjacency matrix, e.g. graph Laplacian, transitionprobability matrix [5], [8], [17], [6]. A throughout summary of graph kernels can be found in [5].…”

Section: Graph Kernelsmentioning

confidence: 99%

See 2 more Smart Citations

Social Coordinates

Phạm

Erlandsson

2017

Proceedings of the 2017 International Conference on Machine Learning and Soft Computing

View full text Add to dashboard Cite

We present a scalable framework to embed nodes of a large social network into an Euclidean space such that the proximity between embedded points reflects the similarity between the corresponding graph nodes. Axes of the embedded space are chosen to maximize data variance so that the dimension of the embedded space is a parameter to regulate noise in data. Using recommender system as a benchmark, empirical results show that similarity derived from the embedded coordinates outperforms similarity obtained from the original graph-based measures. CCS Concepts• Human-centered computing➝Collaborative and social computing➝Collaborative and social computing design and evaluation methods

show abstract

“…graph Laplacian, or transition probability matrix. We refer interested readers to [5] for a comprehensive review of graph kernels.…”

Section: Preliminaries 21 Graph Kernels: Node Similarity Measuresmentioning

confidence: 99%

Section: Kernel-based Euclidean Embeddingmentioning

confidence: 99%

See 1 more Smart Citation

Social Coordinates

Phạm

Erlandsson

2017

Proceedings of the 2017 International Conference on Machine Learning and Soft Computing

View full text Add to dashboard Cite

show abstract

“…It is known that the basic idea of using random walk or random walk based kernels [17][18][19][20] for PPI prediction is that good interacting candidates usually are not faraway from the start node, e.g., only 2,3 edges away in the network. Consequently, for some existing network-level link prediction methods, testing nodes have been chosen to be within a certain distance range, which largely contributes to their good performance reported.…”

Section: Detection Of Interacting Pairs Far Apart In the Networkmentioning

confidence: 99%

Inference of protein-protein interaction networks from multiple heterogeneous data

Huang

Liao

2015

Proceedings of the 6th ACM Conference on Bioinformatics, Computational Biology and Health Informatics

View full text Add to dashboard Cite

Protein-protein interaction (PPI) prediction is a central task in achieving a better understanding of cellular and intracellular processes. Because high-throughput experimental methods are both expensive and time-consuming, and are also known of suffering from the problems of incompleteness and noise, many computational methods have been developed, with varied degrees of success. However, the inference of PPI network from multiple heterogeneous data sources remains a great challenge. In this work, we developed a novel method based on approximate Bayesian computation and modified differential evolution sampling (ABC-DEP) and regularized laplacian (RL) kernel. The method enables inference of PPI networks from topological properties and multiple heterogeneous features including gene expression and Pfam domain profiles, in forms of weighted kernels. The optimal weights are obtained by ABC-DEP, and the kernel fusion built based on optimal weights serves as input to RL to infer missing or new edges in the PPI network. Detailed comparisons with control methods have been made, and the results show that the accuracy of PPI prediction measured by AUC is increased by up to 23 %, as compared to a baseline without using optimal weights. The method can provide insights into the relations between PPIs and various feature kernels and demonstrates strong capability of predicting faraway interactions that cannot be well detected by traditional RL method.

show abstract

“…From this simple distance a metric was used to qualify the interaction-profile similarity of the viral proteins. Nonetheless, much more complex similarity coefficients (Fouss et al 2012) can be used as kernels on graphs (e.g., exponential diffusion kernel, Laplacian exponential diffusion kernel, or the commute time kernel).…”

Section: Virus-host Interactomementioning

confidence: 99%

Network Analysis and Modeling in Systems Biology

Chacón¹

View full text Add to dashboard Cite

An experimental investigation of kernels on graphs for collaborative recommendation and semisupervised classification

Cited by 124 publications

References 66 publications

Social Coordinates

Social Coordinates

Inference of protein-protein interaction networks from multiple heterogeneous data

Network Analysis and Modeling in Systems Biology

Contact Info

Product

Resources

About