Estimation of Markov transition probabilities via clustering: An application to Long Term Care

Research output: Contribution to conferenceAbstract

Abstract

We report on the clustering analysis of a database of continuing care
in 2015 in Portugal, rich of 120 000 records with 70 variables each.
Our main goal was to recognize a small number of dependence states
in the general population and to estimate transition probabilities between
every pair of states.
Original languageEnglish
Pages85-90
Number of pages6
Publication statusPublished - 2017
EventSymposium on Big Data in Finance, Retail and Commerce - Jupiter Lisboa Hotel, Libon, Portugal
Duration: 2 Nov 20173 Nov 2017
http://symposiumbigdata2017.weebly.com/

Conference

ConferenceSymposium on Big Data in Finance, Retail and Commerce
CountryPortugal
CityLibon
Period2/11/173/11/17
Internet address

Keywords

  • Clustering
  • K-medoids
  • Markov transition probabilities
  • Long Term Care

Fingerprint Dive into the research topics of 'Estimation of Markov transition probabilities via clustering: An application to Long Term Care'. Together they form a unique fingerprint.

  • Cite this

    Guerreiro, G. R. D., Esquível, M. L., Oliveira, M., Nascimento, S., & Lopes, H. (2017). Estimation of Markov transition probabilities via clustering: An application to Long Term Care. 85-90. Abstract from Symposium on Big Data in Finance, Retail and Commerce, Libon, Portugal.