The knowledge out of earlier software for financing home Borrowing from the bank regarding readers who’ve funds regarding application investigation
We have fun with one-very hot security and get_dummies toward categorical variables into the application studies. On nan-thinking, we have fun with Ycimpute library and you will expect nan viewpoints from inside the numerical variables . To have outliers investigation, i pertain Local Outlier Grounds (LOF) to the software research. LOF detects and you may surpress outliers data.
For every single newest financing on app research may have several early in the day fund. For each and every past app provides one to row and that is identified by this new element SK_ID_PREV.
I’ve each other float and you may categorical variables. I implement score_dummies to possess categorical details and you will aggregate to help you (mean, min, maximum, matter, and you may sum) to have drift parameters.
The data regarding percentage records getting previous money at home Credit. Discover one line for each produced percentage and something row for each overlooked fee. (more…)