Search published articles


Showing 2 results for Bayesian

A. Mansour Khaki, Sh. Afandizadeh, R. Moayedfar,
Volume 7, Issue 3 (9-2009)
Abstract

Household trip production is not a constant parameter and vary based on socio-economic characteristics. Even households in each category (households with constant socio-economic characteristics) produce several numbers of trips. Purpose of present study is to model the variation of household trip production rate in urban societies. In order to do this, concept of the Bayesian Inference has been used. The city of Isfahan was selected as case study. First, likelihood distribution function was determined for number of household trips, separating odd and even trips. In order to increase precision of the function, the composed likelihood distribution function was utilized. To insert households’ socio-economic variables in the process, disaggregate 2 calibrated model were used at the likelihood distribution function. Statistical indices and 2 test show that likelihood distribution function of numbers of household trip production follows the Poisson distribution. The final composed likelihood distribution was determined based on Bayesian inference. Related function was created with compilation of mean parameter distribution function (Gamma distribution) and numbers of household trip production (Poisson distribution). Finally, disaggregate model was put at final composed probability function instead of mean parameter. Results show that with Bayesian inference method, it would be possible to model the variation of household trip production rate in urban societies. Also it would be possible to put socio-economic characteristics in the model to predict likelihood of real produced trips (not average produced trips) for each household's category.
Mr. Mehdi Mahdavi Adeli, Dr. Mehdi Banazadeh, Dr. Ardeshir Deylami,
Volume 9, Issue 3 (9-2011)
Abstract

The objective of this paper is to determine the drift demand hazard curves of steel moment-resisting frames with different number
of stories in territory of Tehran this is done through the combination of the results obtained from probabilistic seismic hazard
analysis and the demand estimated through the best probabilistic seismic demand models. To select the best demand model, in
this paper, a Bayesian regression has been used for the statistical analysis of the results obtained from incremental dynamic
analysis in order to estimate the unknown parameters of model and to select the best Intensity Measure (IM) parameter also the
probability of overall collapse of structures has been computed. Considering the efficiency and sufficiency of the models, the
results indicate that the accuracy of models with one single IM is a function of the number of stories, consequently the current
widely used model with spectral acceleration in first period as IM is not suitable for all structural heights. Furthermore,
regarding the fact that it is difficult to prepare a seismic hazard curve for a combined IM, it seems that the best model can be
found among models with two single IMs. In other words, the best model to cover all structural heights is the one with linear
combination of spectral acceleration of the first and the second period. Furthermore, using different models to calculate the
curves shows that regardless of the number of IMs, estimated demands strongly depend on the standard deviation of model.

 



Page 1 from 1     

© 2019 All Rights Reserved | International Journal of Civil Engineering

Designed & Developed by : Yektaweb