In recent years several complaints about racial discrimination in estimating home values have been accumulating (see Bloomberg CityLab). Consequently, to obtain a fair property assessment and avoid subjectivity as well as racial and ethnic biases in the dwelling appraisals, the need for development of accurate property price prediction models has been growing. In order to minimize human involvement in the real estate appraisals and boost the accuracy of the real estate market price prediction models, I have been designing data efficient learning machines capable of learning and extracting the relations or patterns between the inputs (features for the house) and output (house prices). I have collected data in the form of both structured data (e.g., physical features and location) and unstructured data (house images and reviews). We are using pre-trained CNNs and/or Transformers to process images and modern natural language techniques to process textual data.
keywords
Applied Econometrics, Experimental Economics, Behavioral Economics,Urban Economics, Machine Learning, Deep Learning, Economics of Race
ECON 4848 - Applied Econometrics
Primary Instructor
-
Spring 2020 / Fall 2020 / Spring 2021 / Fall 2021 / Spring 2022
Introduces students to the practice of applied regression analysis. Summarizes and reviews the regression technique, explores U.S. census data sources, introduces an advanced statistical software package and provides structured exercises in regression analysis of census data. Concludes with independent research projects analyzing social and economic issues using regression analysis and census data.
