Directed Reading Program: Modeling Rental Price in Calgary Under Spatial Bayesian Paradigms

• Time and Location: April 30th, Burnside 1104.
• Topic: Spatial Statistics: Modeling Rental Price in Calgary Under Spatial Bayesian Paradigms.
• In this talk, we aim to build up a model representing the rental price in Calgary based on the data from 2024. We first fit a baseline linear regression model and examine its performances. We use Moran’s I test and variograms to investigate potential spatial dependency among the residuals. Our new model then invloves a linear predictor term plus a spatial Gaussian processes with sill, nugget and range as parameters. We next use a Bayesian framework where weakly-informative priors are selected for each parameter and use Marlov Chain Monte Carlo (MCMC) methods to train the parameters. We next examine the performance of the new model and conclude the paper by giving graphical illustrations of the spatial effect in our model.
• Slides of the talk: Modeling Rental Price in Calgary Under Spatial Bayesian Paradigms.