Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces
Novelty detection involves the construction of a “model of normality”, and then classifies test data as being either “normal” or “abnormal” with respect to that model. For this reason, it is often termed one-class classification. The approach is suitable for cases in which examples of “normal” behav...
| Main Authors: | , , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
Springer US
2013
|
| _version_ | 1826292409444597760 |
|---|---|
| author | Clifton, D Clifton, L Hugueny, S Tarassenko, L |
| author_facet | Clifton, D Clifton, L Hugueny, S Tarassenko, L |
| author_sort | Clifton, D |
| collection | OXFORD |
| description | Novelty detection involves the construction of a “model of normality”, and then classifies test data as being either “normal” or “abnormal” with respect to that model. For this reason, it is often termed one-class classification. The approach is suitable for cases in which examples of “normal” behaviour are commonly available, but in which cases of “abnormal” data are comparatively rare. When performing novelty detection, we are typically most interested in the tails of the normal model, because it is in these tails that a decision boundary between “normal” and “abnormal” areas of data space usually lies. Extreme value statistics provides an appropriate theoretical framework for modelling the tails of univariate (or low-dimensional) distributions, using the generalised Pareto distribution (GPD), which can be demonstrated to be the limiting distribution for data occurring within the tails of most practically-encountered probability distributions. This paper provides an extension of the GPD, allowing the modelling of probability distributions of arbitrarily high dimension, such as occurs when using complex, multimodel, multivariate distributions for performing novelty detection in most real-life cases. We demonstrate our extension to the GPD using examples from patient physiological monitoring, in which we have acquired data from hospital patients in large clinical studies of high-acuity wards, and in which we wish to determine “abnormal” patient data, such that early warning of patient physiological deterioration may be provided. |
| first_indexed | 2024-03-07T03:14:15Z |
| format | Journal article |
| id | oxford-uuid:b53e6e39-a626-4ade-a13e-dfd749c566e5 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T03:14:15Z |
| publishDate | 2013 |
| publisher | Springer US |
| record_format | dspace |
| spelling | oxford-uuid:b53e6e39-a626-4ade-a13e-dfd749c566e52022-03-27T04:32:02ZExtending the generalised Pareto distribution for novelty detection in high-dimensional spacesJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:b53e6e39-a626-4ade-a13e-dfd749c566e5EnglishSymplectic Elements at OxfordSpringer US2013Clifton, DClifton, LHugueny, STarassenko, LNovelty detection involves the construction of a “model of normality”, and then classifies test data as being either “normal” or “abnormal” with respect to that model. For this reason, it is often termed one-class classification. The approach is suitable for cases in which examples of “normal” behaviour are commonly available, but in which cases of “abnormal” data are comparatively rare. When performing novelty detection, we are typically most interested in the tails of the normal model, because it is in these tails that a decision boundary between “normal” and “abnormal” areas of data space usually lies. Extreme value statistics provides an appropriate theoretical framework for modelling the tails of univariate (or low-dimensional) distributions, using the generalised Pareto distribution (GPD), which can be demonstrated to be the limiting distribution for data occurring within the tails of most practically-encountered probability distributions. This paper provides an extension of the GPD, allowing the modelling of probability distributions of arbitrarily high dimension, such as occurs when using complex, multimodel, multivariate distributions for performing novelty detection in most real-life cases. We demonstrate our extension to the GPD using examples from patient physiological monitoring, in which we have acquired data from hospital patients in large clinical studies of high-acuity wards, and in which we wish to determine “abnormal” patient data, such that early warning of patient physiological deterioration may be provided. |
| spellingShingle | Clifton, D Clifton, L Hugueny, S Tarassenko, L Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces |
| title | Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces |
| title_full | Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces |
| title_fullStr | Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces |
| title_full_unstemmed | Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces |
| title_short | Extending the generalised Pareto distribution for novelty detection in high-dimensional spaces |
| title_sort | extending the generalised pareto distribution for novelty detection in high dimensional spaces |
| work_keys_str_mv | AT cliftond extendingthegeneralisedparetodistributionfornoveltydetectioninhighdimensionalspaces AT cliftonl extendingthegeneralisedparetodistributionfornoveltydetectioninhighdimensionalspaces AT huguenys extendingthegeneralisedparetodistributionfornoveltydetectioninhighdimensionalspaces AT tarassenkol extendingthegeneralisedparetodistributionfornoveltydetectioninhighdimensionalspaces |