This is the result of the D&R testing. I am sorry that I did not come out at the first time.
The main conclusion is: There is a positive effect of `prediction.io` which means the recommendations from prediction.io are more likely to be categorized as `good recommendation`.
However, the effect is not statistically significant. This may because it is a small sample size.
If you are interested about the information I provide in the D&R, feel free to check the slides and give me more feedbacks! Here is the link: D&R (New Learning Times and Recommendation System) .