Predictive Models in Data Science
Amrina Ferdous, Jodi L. Mead · ScholarWorks (Boise State University) · 2019
Predictive modeling in data science typically requires unknown parameters that can be inferred from observational data. A parameter is any numerical quantity that characterizes a given data set or some aspect of it. This study focuses on the effectiveness of combining data sets from different predictive models that share some common parameters. We will show preliminary results demonstrating the effect of uncertainties in initial parameter estimates inferred from a secondary data set.