One-class classification of point patterns of extremes
Novelty detection or one-class classification starts from a model describing some type of ‘normal behaviour’ and aims to classify deviations from this model as being either novelties or anomalies. <br/>In this paper the problem of novelty detection for point patterns S = {x1, . . . , xk} ⊂ R...
| Hlavní autoři: | , , |
|---|---|
| Médium: | Journal article |
| Vydáno: |
Journal of Machine Learning Research
2016
|
| _version_ | 1826267391803260928 |
|---|---|
| author | Luca, S Clifton, D Vanrumste, B |
| author_facet | Luca, S Clifton, D Vanrumste, B |
| author_sort | Luca, S |
| collection | OXFORD |
| description | Novelty detection or one-class classification starts from a model describing some type of ‘normal behaviour’ and aims to classify deviations from this model as being either novelties or anomalies. <br/>In this paper the problem of novelty detection for point patterns S = {x1, . . . , xk} ⊂ R d is treated where examples of anomalies are very sparse, or even absent. The latter complicates the tuning of hyperparameters in models commonly used for novelty detection, such as one-class support vector machines and hidden Markov models. <br/>To this end, the use of extreme value statistics is introduced to estimate explicitly a model for the abnormal class by means of extrapolation from a statistical model X for the normal class. We show how multiple types of information obtained from any available extreme instances of S can be combined to reduce the high false-alarm rate that is typically encountered when classes are strongly imbalanced, as often occurs in the one-class setting (whereby ‘abnormal’ data are often scarce). <br/>The approach is illustrated using simulated data and then a real-life application is used as an exemplar, whereby accelerometry data from epileptic seizures are analysed - these are known to be extreme and rare with respect to normal accelerometer data. |
| first_indexed | 2024-03-06T20:53:29Z |
| format | Journal article |
| id | oxford-uuid:386532c1-0ac3-44dc-abe2-38eced6c1b8a |
| institution | University of Oxford |
| last_indexed | 2024-03-06T20:53:29Z |
| publishDate | 2016 |
| publisher | Journal of Machine Learning Research |
| record_format | dspace |
| spelling | oxford-uuid:386532c1-0ac3-44dc-abe2-38eced6c1b8a2022-03-26T13:49:44ZOne-class classification of point patterns of extremesJournal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:386532c1-0ac3-44dc-abe2-38eced6c1b8aSymplectic Elements at OxfordJournal of Machine Learning Research2016Luca, SClifton, DVanrumste, BNovelty detection or one-class classification starts from a model describing some type of ‘normal behaviour’ and aims to classify deviations from this model as being either novelties or anomalies. <br/>In this paper the problem of novelty detection for point patterns S = {x1, . . . , xk} ⊂ R d is treated where examples of anomalies are very sparse, or even absent. The latter complicates the tuning of hyperparameters in models commonly used for novelty detection, such as one-class support vector machines and hidden Markov models. <br/>To this end, the use of extreme value statistics is introduced to estimate explicitly a model for the abnormal class by means of extrapolation from a statistical model X for the normal class. We show how multiple types of information obtained from any available extreme instances of S can be combined to reduce the high false-alarm rate that is typically encountered when classes are strongly imbalanced, as often occurs in the one-class setting (whereby ‘abnormal’ data are often scarce). <br/>The approach is illustrated using simulated data and then a real-life application is used as an exemplar, whereby accelerometry data from epileptic seizures are analysed - these are known to be extreme and rare with respect to normal accelerometer data. |
| spellingShingle | Luca, S Clifton, D Vanrumste, B One-class classification of point patterns of extremes |
| title | One-class classification of point patterns of extremes |
| title_full | One-class classification of point patterns of extremes |
| title_fullStr | One-class classification of point patterns of extremes |
| title_full_unstemmed | One-class classification of point patterns of extremes |
| title_short | One-class classification of point patterns of extremes |
| title_sort | one class classification of point patterns of extremes |
| work_keys_str_mv | AT lucas oneclassclassificationofpointpatternsofextremes AT cliftond oneclassclassificationofpointpatternsofextremes AT vanrumsteb oneclassclassificationofpointpatternsofextremes |