Hi there,

I am not sure of the difference between the prior probability shift and concept jump in the quiz below.

Thanks for your help.

Quiz:

As a data scientist who works in a sales company, you are asked to predict the sales for December 2020. Your dataset contains daily sales data from January 2020 to September 2020. You successfully train your model on this data, but the actual sales measured in December are completely different from what your model predicted. These types of events are called **data drift.**

After some research, you realize that this change occurred because there is always a sharp increase in sales due to the holiday season.

What specific kind of data drift best describes this change in sales?

a. covariate shift

b. prior probability shift

c. concept jump

d. none of the above

- Prior probability shift: A change in probabilities based on new evidence.
- Concept jump: Abrupt shift in topics or subjects during a conversation.

An example of prior probability shift:

For example, let’s say you believe there is a 30% chance of rain tomorrow based on historical weather data (prior probability). However, you hear a weather forecast that predicts a severe storm approaching the area. This new evidence would cause a prior probability shift, as you might revise your belief and increase the probability of rain to, let’s say, 70% (posterior probability).

I also learned another way to explain data shift is like this:

- covariate shift - > shift in x , Ptrain(y|x) = P(y|x), but Ptrain(x)!=Ptest(x)
- Prior probability shift ->shift in y, Ptrain(x|y) = Ptest(x|y), but Ptrain(y)!=Ptest(Y)
- concept shift → shift in relationship of x and y. Ptrain(y|x)!=Ptest(y!x) . And y depends on time series data, like forecasting and data with seasoning change.

in this case, I still think concept shift is also a reasonable answer.

I am wondering what I am missing.