Sei sulla pagina 1di 1

11/8/2018 LightGBM 7th place solution | Kaggle

LightGBM 7th place solution Python Commit


Sessions

1 """ Interactive Session


2 KAGGLE HOME CREDIT DEFAULT RISK COMPETITION An error occurred.
3 Adapted from one of the models used in 7th place solution
4 For more details about our solution please check this dis
5 https://www.kaggle.com/c/home-credit-default-risk/discuss Versions
6
7 Another similar version is also available at GitHub: 1 uncommitted draft

8 https://github.com/js-aguiar/home-credit-default-competit Omar Saleem's


based on V4
9 draft
10 This model uses LightGBM with goss and label encode for t
11 categorical features. Other tables are using one-hot enco 4 committed versions
12 sum and a few different functions to aggregate. The main
V4 7d
13 more time related features like last application and last
14 There are also aggregations for specific loan types and s V3 7d
15 ratios between tables. Configurations are in line 785
V2 7d
16 """
17 V1 7d
18 import os
19 import gc
20 import time Draft Environment

21 import numpy as np
22 import pandas as pd Loading…
23 from contextlib import contextmanager
input (read-only)
24 import multiprocessing as mp
25 from functools import partial Home Credit Default Risk
26 from scipy.stats import kurtosis, iqr, skew
27 from lightgbm import LGBMClassifier
28 Settings

Private, 0
Sharing
collaborators

Language Python

Docker Latest available

GPU BETA GPU on

Internet BETA Internet blocked

Custom packages are


Packages not supported for GPU
instances

Docs API

Console

file:///C:/Users/Omar%20Saleem%20Mohammed/Downloads/LightGBM%207th%20place%20solution%20_%20Kaggle.html 1/1

Potrebbero piacerti anche