Random indexing of multidimensional data
(2017) In Knowledge and Information Systems 52. p.267-290- Abstract
- Random indexing (RI) is a lightweight dimension reduction method, which is used, for example, to approximate vector semantic relationships in online natural language processing systems. Here we generalise RI to multidimensional arrays and therefore enable approximation of higher-order statistical relationships in data. The generalised method is a sparse implementation of random projections, which is the theoretical basis also for ordinary RI and other randomisation approaches to dimensionality reduction and data representation. We present numerical experiments which demonstrate that a multidimensional generalisation of RI is feasible, including comparisons with ordinary RI and principal component analysis. The RI method is well suited for... (More)
- Random indexing (RI) is a lightweight dimension reduction method, which is used, for example, to approximate vector semantic relationships in online natural language processing systems. Here we generalise RI to multidimensional arrays and therefore enable approximation of higher-order statistical relationships in data. The generalised method is a sparse implementation of random projections, which is the theoretical basis also for ordinary RI and other randomisation approaches to dimensionality reduction and data representation. We present numerical experiments which demonstrate that a multidimensional generalisation of RI is feasible, including comparisons with ordinary RI and principal component analysis. The RI method is well suited for online processing of data streams because relationship weights can be updated incrementally in a fixed-size distributed representation, and inner products can be approximated on the fly at low computational cost. An open source implementation of generalised RI is provided. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/a7ede0c1-5674-4115-b368-27564a26b128
- author
- Sandin, Fredrik ; Emruli, Blerim LU and Sahlgren, Magnus
- publishing date
- 2017
- type
- Contribution to journal
- publication status
- published
- subject
- in
- Knowledge and Information Systems
- volume
- 52
- pages
- 267 - 290
- external identifiers
-
- scopus:85001755138
- ISSN
- 0219-3116
- DOI
- 10.1007/s10115-016-1012-2
- language
- English
- LU publication?
- no
- id
- a7ede0c1-5674-4115-b368-27564a26b128
- date added to LUP
- 2025-03-31 21:26:59
- date last changed
- 2025-04-04 14:07:45
@article{a7ede0c1-5674-4115-b368-27564a26b128,
abstract = {{Random indexing (RI) is a lightweight dimension reduction method, which is used, for example, to approximate vector semantic relationships in online natural language processing systems. Here we generalise RI to multidimensional arrays and therefore enable approximation of higher-order statistical relationships in data. The generalised method is a sparse implementation of random projections, which is the theoretical basis also for ordinary RI and other randomisation approaches to dimensionality reduction and data representation. We present numerical experiments which demonstrate that a multidimensional generalisation of RI is feasible, including comparisons with ordinary RI and principal component analysis. The RI method is well suited for online processing of data streams because relationship weights can be updated incrementally in a fixed-size distributed representation, and inner products can be approximated on the fly at low computational cost. An open source implementation of generalised RI is provided.}},
author = {{Sandin, Fredrik and Emruli, Blerim and Sahlgren, Magnus}},
issn = {{0219-3116}},
language = {{eng}},
pages = {{267--290}},
series = {{Knowledge and Information Systems}},
title = {{Random indexing of multidimensional data}},
url = {{http://dx.doi.org/10.1007/s10115-016-1012-2}},
doi = {{10.1007/s10115-016-1012-2}},
volume = {{52}},
year = {{2017}},
}