Good Morning,
I have a problem that may perhaps be very simple, but I keep getting very stuck and very frustrated.
I'm trying to produce ROC curves for my data.
My data set is for each year:
Outbreak Data - date and location of disease occurrence
Warning Data - date and location of warnings for the disease
It's an agriculture disease - so the alert is based on weather criteria occurring.
I am looking at occurrence of warnings in each locations prior to the date of disease - going back 28 days from disease occurrence.
I would really like to produce some ROC curves for my data to provide a better statistical analysis - however I only have data for when the outbreak occurs.
I am struggling to figure out how I would get my True negatives and false positive rates to complete the calculations. When I go to look for values of there being no disease/no alert - I don't have this as I only have locations for disease actually occurring.
If anyone has any help or advice on this problem I would be extremely grateful. I am not trained in statistics or such and thus am hoping this is a something simple I am just not grasping.
Thank you for your time
Siobhan
I have a problem that may perhaps be very simple, but I keep getting very stuck and very frustrated.
I'm trying to produce ROC curves for my data.
My data set is for each year:
Outbreak Data - date and location of disease occurrence
Warning Data - date and location of warnings for the disease
It's an agriculture disease - so the alert is based on weather criteria occurring.
I am looking at occurrence of warnings in each locations prior to the date of disease - going back 28 days from disease occurrence.
I would really like to produce some ROC curves for my data to provide a better statistical analysis - however I only have data for when the outbreak occurs.
I am struggling to figure out how I would get my True negatives and false positive rates to complete the calculations. When I go to look for values of there being no disease/no alert - I don't have this as I only have locations for disease actually occurring.
If anyone has any help or advice on this problem I would be extremely grateful. I am not trained in statistics or such and thus am hoping this is a something simple I am just not grasping.
Thank you for your time
Siobhan