Loading...
Please wait, while we are loading the content...
Similar Documents
High Performance Multidimensional Scaling for Large High-Dimensional Data Visualization
| Content Provider | Semantic Scholar |
|---|---|
| Author | Bae, Seung-Hee Qiu, Judy Fox, Geoffrey Charles |
| Copyright Year | 2012 |
| Abstract | Technical advancements produces a huge amount of scientific data which are usually in high dimensional formats, and it is getting more important to analyze those large-scale high-dimensional data. Dimension reduction is a well-known approach for high-dimensional data visualization, but can be very time and memory demanding for large problems. Among many dimension reduction methods, multidimensional scaling does not require explicit vector representation but uses pair-wise dissimilarities of data, so that it has a broader applicability than the other algorithms which can handle only vector representation. In this paper, we propose an efficient parallel implementation of a well-known multidimensional scaling algorithm, called SMACOF (Scaling by MAjorizing a COmplicated Function) which is time and memory consuming with a quadratic complexity, via a Message Passing Interface (MPI). We have achieved load balancing in the proposed parallel SMACOF implementation which results in high efficiency. The proposed parallel SMACOF implementation shows the efficient parallel performance through the experimental results, and it increases the computing capacity of the SMACOF algorithm to several hundreds of thousands of data via using a 32-node cluster system. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://grids.ucs.indiana.edu/ptliupages/publications/hpmds.pdf |
| Alternate Webpage(s) | https://www.cs.indiana.edu/~xqiu/hpmds_0328.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
