Our project aims to take a Bayesian Inference approach to creating a model to predict the MPG of a car given some set parameters about said car. We will set a prior distribution for the parameters, as well as a likelihood for the model as a whole. We will then use Markov Chain Monte Carlo (MCMC) to approximate the values for the weights of each of these parameters. More specifically, we will use the Metropolis-Hastings algorithm defined in [1]. We will use the "Auto MPG Data Set" dataset from [2] to build our deterministic model for predicting MPG.
This repo contains the report and presentation slides of our final project for CSCI 5822: Probabilistic Models of Human and Machine Intelligence at CU Boulder. The code can be found under the python directory.
[1] David Barber. Bayesian reasoning and machine learning. Cambridge University Press,2012, pp. 559–560.
[2] Dheeru Dua and Casey Graff. UCI Machine Learning Repository. 2017. url: http://archive.ics.uci.edu/ml.
[3] Paul Roback and Julie Legler. Beyond Multiple Linear Regression: Applied GeneralizedLinear Models And Multilevel Models in R. CRC Press, 2021. Chap. 4.