Investigating the Relationship between the Geographic Variations and Incidence Risk of COVID-19 in Hong Kong
TPU_preprocessing:
- ignore teh case with incomplete information
- count number of cases in each TPU
- count number of cases in each TPU according to gendar and different age range
ratio_preprocessing:
- divide all features with population of the TPU, calculate the ratio of the feature over population
- split all cases into two groups, local cases and imported cases.
- To simplify the model, only consider the local cases as the source of community virus spread.
- according to the assumption in 3., we take (local case number in a TPU)/(population of the TPU) as the incidence rate instead of (all cases number in the TPU)/(population of the TPU).
ratio_regression:
- training and testing of five regression models
- the results of all models in form of diagrams
- feature correlation analysis