Fisher's information matrix and phi-divergence for finite and optimal partitions of the sample space
| dc.contributor.author | Tsairidis, C. | en |
| dc.contributor.author | Zografos, K. | en |
| dc.contributor.author | Ferentinos, K. | en |
| dc.date.accessioned | 2015-11-24T17:24:03Z | |
| dc.date.available | 2015-11-24T17:24:03Z | |
| dc.identifier.issn | 0361-0926 | - |
| dc.identifier.uri | https://olympias.lib.uoi.gr/jspui/handle/123456789/12837 | |
| dc.rights | Default Licence | - |
| dc.subject | convergence of information measures | en |
| dc.subject | discretization | en |
| dc.subject | optimal partition | en |
| dc.subject | equidistant intervals | en |
| dc.subject | loss of information | en |
| dc.subject | fisher information matrix | en |
| dc.subject | phi-divergence | en |
| dc.subject | discretization | en |
| dc.title | Fisher's information matrix and phi-divergence for finite and optimal partitions of the sample space | en |
| heal.abstract | In this paper we examine the behaviour of the discretized version of Fisher's information matrix and phi-divergence produced from a general finite partition of the sample space. We prove that as the data information gets richer, in the sense that the data partition gets finer, both Fisher information matrix and phi-divergence increase monotonically, approaching to their continuous counterparts, respectively. In particular, the monotonicity property is proved by considering one partition as a subpartition of the other. The problem of finding an optimal partition, which minimizes the loss of information of the above measures, is also discussed. Similar results, which have appeared in the literature, are generalized and supplemented. | en |
| heal.access | campus | - |
| heal.fullTextAvailability | TRUE | - |
| heal.identifier.primary | Doi 10.1080/03610929708832046 | - |
| heal.identifier.secondary | <Go to ISI>://A1997XV59000013 | - |
| heal.identifier.secondary | http://www.tandfonline.com/doi/pdf/10.1080/03610929708832046 | - |
| heal.journalName | Communications in Statistics-Theory and Methods | en |
| heal.journalType | peer reviewed | - |
| heal.language | en | - |
| heal.publicationDate | 1997 | - |
| heal.publisher | Taylor & Francis | en |
| heal.recordProvider | Πανεπιστήμιο Ιωαννίνων. Σχολή Θετικών Επιστημών. Τμήμα Μαθηματικών | el |
| heal.type | journalArticle | - |
| heal.type.el | Άρθρο Περιοδικού | el |
| heal.type.en | Journal article | en |
Files
License bundle
1 - 1 of 1
Loading...
- Name:
- license.txt
- Size:
- 1.74 KB
- Format:
- Item-specific license agreed upon to submission
- Description: