File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1016/j.apnum.2017.10.008
- Scopus: eid_2-s2.0-85033604289
- WOS: WOS:000423652100005
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Fast computation of stationary joint probability distribution of sparse Markov chains
Title | Fast computation of stationary joint probability distribution of sparse Markov chains |
---|---|
Authors | |
Keywords | Markov chains Sparsity Iterative methods Joint distribution Optimization |
Issue Date | 2018 |
Citation | Applied Numerical Mathematics, 2018, v. 125, p. 68-85 How to Cite? |
Abstract | © 2017 IMACS In this paper, we study a fast algorithm for finding stationary joint probability distributions of sparse Markov chains or multilinear PageRank vectors which arise from data mining applications. In these applications, the main computational problem is to calculate and store solutions of many unknowns in joint probability distributions of sparse Markov chains. Our idea is to approximate large-scale solutions of such sparse Markov chains by two components: the sparsity component and the rank-one component. Here the non-zero locations in the sparsity component refer to important associations in the joint probability distribution and the rank-one component refers to a background value of the solution. We propose to determine solutions by formulating and solving sparse and rank-one optimization problems via closed form solutions. The convergence of the truncated power method is established. Numerical examples of multilinear PageRank vector calculation and second-order web-linkage analysis are presented to show the efficiency of the proposed method. It is shown that both computation and storage are significantly reduced by comparing with the traditional power method. |
Persistent Identifier | http://hdl.handle.net/10722/276564 |
ISSN | 2023 Impact Factor: 2.2 2023 SCImago Journal Rankings: 1.006 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Ding, Weiyang | - |
dc.contributor.author | Ng, Michael | - |
dc.contributor.author | Wei, Yimin | - |
dc.date.accessioned | 2019-09-18T08:33:59Z | - |
dc.date.available | 2019-09-18T08:33:59Z | - |
dc.date.issued | 2018 | - |
dc.identifier.citation | Applied Numerical Mathematics, 2018, v. 125, p. 68-85 | - |
dc.identifier.issn | 0168-9274 | - |
dc.identifier.uri | http://hdl.handle.net/10722/276564 | - |
dc.description.abstract | © 2017 IMACS In this paper, we study a fast algorithm for finding stationary joint probability distributions of sparse Markov chains or multilinear PageRank vectors which arise from data mining applications. In these applications, the main computational problem is to calculate and store solutions of many unknowns in joint probability distributions of sparse Markov chains. Our idea is to approximate large-scale solutions of such sparse Markov chains by two components: the sparsity component and the rank-one component. Here the non-zero locations in the sparsity component refer to important associations in the joint probability distribution and the rank-one component refers to a background value of the solution. We propose to determine solutions by formulating and solving sparse and rank-one optimization problems via closed form solutions. The convergence of the truncated power method is established. Numerical examples of multilinear PageRank vector calculation and second-order web-linkage analysis are presented to show the efficiency of the proposed method. It is shown that both computation and storage are significantly reduced by comparing with the traditional power method. | - |
dc.language | eng | - |
dc.relation.ispartof | Applied Numerical Mathematics | - |
dc.subject | Markov chains | - |
dc.subject | Sparsity | - |
dc.subject | Iterative methods | - |
dc.subject | Joint distribution | - |
dc.subject | Optimization | - |
dc.title | Fast computation of stationary joint probability distribution of sparse Markov chains | - |
dc.type | Article | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1016/j.apnum.2017.10.008 | - |
dc.identifier.scopus | eid_2-s2.0-85033604289 | - |
dc.identifier.volume | 125 | - |
dc.identifier.spage | 68 | - |
dc.identifier.epage | 85 | - |
dc.identifier.isi | WOS:000423652100005 | - |
dc.identifier.issnl | 0168-9274 | - |