# 19th Place Solution

Final stand
Public LB: 0.838
Private LB: 0.851
Features
aggregate
focused on peak
focused on detected
Luminosity
LC fitter
Models
LGB
XGB
MLP
Post process
We used weighted multi logloss(same as kernel) and multi logloss.
After predicted by weighted multi logloss, calculated weight for each class using
[gradient descent](https://github.com/KazukiOnodera/PLAsTiCC-2018/blob/master/py/utils_post.py)
.
For multi logloss,
[here](https://github.com/KazukiOnodera/PLAsTiCC-2018/blob/master/py/utils.py#L400-L406)
.
Did not work
AE
Augmentation
![model pipeline](assets/4be0bc3d9b4696a655aba4ab.png)