An aggregate learning approach for interpretable semi-supervised population prediction and disaggregation using ancillary data
Census data provide detailed information about population characteristics at a coarse resolution. We propose a method to disaggregate census data to a finer, arbitrary scale, using aggregate learning on ancillary data. We show that, despite the simplicity of the model, the method is on par or even outperforms existing work on some metrics.
As a 4th year PhD student, Guillaume mainly works on submatrix mining. He although likes to work on different things, such as population prediction and transportation system optimization. He also likes competitive programming.