Run experiments with enhanced evaluation metrics

[isaacmg]

• Run on primary US counties w/o transfer
• Run on primary US counties w transfer (flow)
• Run on primary US counties w transfer (county)
• Run primary US counties w transfer (flow + county)
• Run on Italy Counties w/o transfer
• Run on Italy counties w transfer (flow)
• Run on Italy counties w transfer (county)
• Run on Italy counties w transfer (flow + county)

[isaacmg]

We now want to investigate how the model preforms with respect to transfer learning in various scenarios.

[isaacmg]

@aradhanacha will cover Italy and I will cover US counties.

[isaacmg]

Update so far have experiments both with and without flow transfer for Cook County through Richmond County NYC.

[isaacmg]

Clarification: These experiment should be run with just mobility data, new_cases, and day of week as features

[isaacmg]

Update most experiments are done. Still need to fill out missing counties specifically Los Angeles County (flow, Wayne County in Michigan and Miami-Dade County (all experiments), Cuyahoga County (flow pre-training, ). Working on reports simultaneously.

[isaacmg]

Most experiments are done including for additional counties. Due to some data errors the results are not exactly uniform, however the differences are subtle enough at the moment I think we can proceed to further evaluation steps.