Full-text resources of CEJSH and other databases are now available in the new Library of Science.
Visit https://bibliotekanauki.pl

Results found: 3

first rewind previous Page / 1 next fast forward last

Search results

Search:
in the keywords:  Viterbi algorithm
help Sort By:

help Limit search:
first rewind previous Page / 1 next fast forward last
1
Publication available in full text mode
Content available

The Viterbi Path of hidden Markov models in an analysis of business tendency surveys

100%
Bernardelli M., Dędys M.
Prace i Materiały Instytutu Rozwoju Gospodarczego SGH
|
2015
|
vol. 96: Analyzing and forecasting economic fluctuations
25-47
EN
The aim of the paper is to show that turning points detection can be treated as a problem of pattern recognition. In the paper there are presented the results of applying normal hidden Markov models to a number of survey balances. Beyond a classical two-scale assessment of business activity a slightly more fuzzy classification of states is considered. To determine periods of unclear or difficult to evaluate situation unobservable Markov chains with three and four states are introduced. The outputs of the Viterbi algorithm, i.e. the most likely paths of unobservable states of Markov chains, are a basis of the proposed classification. The comparison of these paths with the business cycle turning points dated by OECD is described. The results obtained for three- and four-state Markov chains are close to those established in the references time series and seem to improve the speed with which, especially downshifts, are signaled. Furthermore, these results are more favorable than outcomes provided by conventional two-state models. The method proposed in this paper seems to be a very effective tool to analyze results of business tendency surveys, in particular, when multistate Markov chains are considered. Moreover, proposed decompositions allow an easy comparison of two time series as far as turning point are concerned. In the paper survey balances are compared with ‘hard’ economic data such as sold manufacturing production. The results confirm the accuracy of assessment provided by survey respondents.
2
Publication available in full text mode
Content available

The procedure of business cycle turning points identification based on hidden Markov models

88%
Bernardelli M.
Prace i Materiały Instytutu Rozwoju Gospodarczego SGH
|
2015
|
vol. 96: Analyzing and forecasting economic fluctuations
5-23
EN
In the paper the procedure, based on hidden Markov chains with conditional normal distributions and uses algorithms such as time series decompositions (STL), Baum-Welch algorithm, Viterbi algorithm and Monte Carlo simulations, is proposed to analyze data out of the business tendency survey conducted by the Research Institute for Economic Development, Warsaw School of Economics. There are considered three types of models, namely, with two-state, three-state and four-state Markov chains. Results of the procedure could be treated as an approximation of business cycle turning points. The performed analysis speaks in favor of multistate models. Due to, an increasing with the number of states, numerical instability, it is not obvious which model should be considered as the best one. For this purpose various optimization criteria are taken into consideration: information criteria (AIC, BIC) and the maximum-likelihood, but also frequency of obtaining a given set of parameters in the Monte Carlo simulations. The results are confronted with the turning points dated by OECD. The tested models were compared in terms of their effectiveness in detecting of turning points. The procedure is a step into automation of business cycle analysis based on results of business tendency surveys. Though this automation covers only some models from millions of possibilities, the procedure turns out to be extremely accurate in business cycle turning points identification, and the approach seems to be an excellent alternative for classical methods.
3
Publication available in full text mode
Content available

Ukryte modele Markowa w analizie wyników testu koniunktury gospodarczej

63%
Bernardelli M., Dędys M.
Prace i Materiały Instytutu Rozwoju Gospodarczego SGH
|
2012
|
vol. 90: Badania koniunktury - zwierciadło gospodarki. Część I
159-181
PL
W pracy zbadana została możliwość wykorzystania algorytmu Viterbiego do analizy sald odpowiedzi respondentów na pytania testu koniunktury w przemyśle, prowadzonego przez Instytut Rozwoju Gospodarczego Szkoły Głównej Handlowej w Warszawie. W badaniu rozważane były pytania dotyczące oceny stanu obecnego. Do analizy wykorzystane zostały ukryte modele Markowa z warunkowymi rozkładami normalnymi. Pod uwagę brane były modele, w których łańcuchy Markowa mają dwuelementową i trójelementową przestrzeń stanów. Uzyskane wyniki zostały skonfrontowane z pochodzącymi z różnych źródeł datowaniami punktów zwrotnych cyklu koniunkturalnego. Badane modele zostały porównane pod względem skuteczności w wychwytywaniu sygnałów o nadchodzących zmianach w koniunkturze. Przeprowadzone analizy przemawiają za stosowaniem modeli z trzystanowymi łańcuchami Markowa. Wyniki badania sugerują ponadto, iż należy brać pod uwagę opóźnienia między odpowiedziami respondentów a zmianami klimatu koniunktury.
EN
The paper considers the possibility of using the Viterbi algorithm to analyse results of the RIED WSE business surveys in the manufacturing industry.The analysis was focused on the state balances. The hidden Markov models with conditional normal distributions were applied. There were considered models with two-state and three-state Markov chains. The results were compared with the timing of turning points taken from other sources. The tested models were compared in terms of effectiveness in detecting of coming changes in economic conditions. The analysis suggests models with three-state Markov chains be used. The results also suggest that it is necessary to take into account a delay between the opinions of survey respondents and changes in economic climate.
first rewind previous Page / 1 next fast forward last
JavaScript is turned off in your web browser. Turn it on to take full advantage of this site, then refresh the page.