inflearn logo
강의

Course

Instructor

shalsdn01's honest review, [Side Projects After Work] Big Data Analysis Engineer Practical Exam (Task Types 1, 2, and 3) course

shalsdn01

Reviews 1

Average rating 5

I passed the 12th exam. I was flustered because a value called RMLSE appeared in the second subject of the 12th practical exam, which was unusual, but fortunately, the preprocessing I applied on the spot by adapting what I learned seemed to work well for the model. (For reference, I trained the model after removing outliers using IQR.) It seems like problems where mindless RF or LightGBM no longer work are starting to appear, so I think it would be good to have some additional content regarding that!

1

roadmap

Congratulations. Even though you must have been flustered when an unfamiliar metric appeared, it's truly impressive that you applied the IQR outlier removal preprocessing on the spot based on what you learned. By any chance, what was the score when you did this? And which model did you use?

0

shalsdn01

I scored 25 out of 40 in Subject 2. Unfortunately, I didn't get a high score. I used the lightGBM model! I don't remember exactly if I checked the distribution of the pred values after preprocessing, but I think doing it this way created a sort of "safe zone" on the lower end (where the model predicts most values similarly). As a result, it seems that negative values were not produced as predictions for the outlier data that was originally in the test set. In my case, I was able to handle it this way because I can interpret error codes to some extent, and I also found the RMLSE metric using the dir function. If you need any more information, please contact me through any available method, and I will do my best to answer. Thank you.

0

roadmap

117 lectures

6390 enrolled

[Side Projects After Work] Big Data Analysis Engineer Practical Exam (Task Types 1, 2, and 3)
4.9(985 reviews)