MoCSSy Program: Graduate Student Seminar Series - ""Join Bayes Nets: A Model for Statistical Relational Learning"
Abstract
Many databases store data in relational format in multiple tables. Standard machine learning techniques are applied to data stored in a single table to find dependencies among attributes of the table. The field of statistical-relational learning (SRL) aims to extend machine learning algorithms to relational data. I have proposed a new model for (SRL) called Join Bayes Nets which is based on Bayesian Networks. I am currently implementing my model on an open source software developed at CMU, Tetrad, which works on finding correlations among attributes of a single tables.