Summary:
The project aims to determine whether COVID-19 can be predicted just based on the symptoms present. This aims to be a classification model and problem. They are using the dataset from the Israel Ministry of Health to understand how symptoms correlate to having COVID. There were multiple analyses performed to start to understand what models may work best.
Like:
I like the further exploration on using neural networks and other classification models to determine a better model for the next milestone.
I like the descriptions of the outcomes of the logisitic regression model and FPR/FNR. These values are critical in understanding the efficacy of the model, particularly since this is in healthcare/medical field.
View of how data trends have changed over time and understanding how this may affect outcomes of the model. (cyclical trends)
Suggested:
Explain how many of these variables were dropped. We know how many were left after dropping, but how significant of a portion were dropped?
A discussion on how features were either included or removed. It states that not all features were used, but how were these chosen?
Maybe a correlation matrix of how each covariate/feature is related to each other to better understand which ones may not be too important to put into the model may be useful.
Summary: The project aims to determine whether COVID-19 can be predicted just based on the symptoms present. This aims to be a classification model and problem. They are using the dataset from the Israel Ministry of Health to understand how symptoms correlate to having COVID. There were multiple analyses performed to start to understand what models may work best.
Like:
Suggested: