글 치트코드

lightGBM을 이용한 회귀예측 파이썬 치트코드

2019-05-18
운영자
No Comments



face-lgb








In [1]:
import pandas as pd
import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import OneHotEncoder
from sklearn.metrics import mean_squared_error, r2_score
import seaborn as sns
import lightgbm as lgb

파일 불러오기

In [2]:
label = pd.read_csv("label.csv", encoding='cp949')
In [3]:
raw_x = label[['gender','c_eye_left_size','c_eye_right_size','c_eye_left_width','c_eye_right_width','c_eye_nose_left_align','c_eye_nose_right_align','c_nose_root_width','c_nose_alar_width','c_nose_height','c_eyebrow_left_width','c_eyebrow_right_width','c_lip_upper_height','c_lip_under_height','c_lip_width','c_brow_eye_left_inner_gap','c_brow_eye_right_inner_gap','c_brow_eye_left_outer_gap','c_brow_eye_right_outer_gap','c_eyebrow_balance_left','c_eyebrow_balance_right','c_alar_left_width_shape','c_alar_right_width_shape','c_alar_left_height_shape','c_alar_right_height_shape','c_lip_ratio','c_ratio_vertical','c_ratio_horizontal','c_lip_nose_gap','c_brow_nose_left_gap','c_brow_nose_right_gap','age']]
raw_y = label[['score']]

성별의 경우 카테고리로 전환

In [4]:
raw_x['gender'] = raw_x['gender'].astype('category') 

C:\Users\kohry\anaconda3\lib\site-packages\ipykernel_launcher.py:1: SettingWithCopyWarning: 
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: http://pandas.pydata.org/pandas-docs/stable/indexing.html#indexing-view-versus-copy
  """Entry point for launching an IPython kernel.
In [5]:
train_x, test_x, train_y, test_y = train_test_split(raw_x, raw_y, test_size=0.2, random_state=42)

lightgbm 데이터셋으로 변환

In [6]:
train_ds = lgb.Dataset(train_x, label = train_y) 
test_ds = lgb.Dataset(test_x, label = test_y)

파라미터 세팅

In [7]:
params = {'learning_rate': 0.01, 
          'max_depth': 16, 
          'boosting': 'gbdt', 
          'objective': 'regression', 
          'metric': 'mse', 
          'is_training_metric': True, 
          'num_leaves': 144, 
          'feature_fraction': 0.9, 
          'bagging_fraction': 0.7, 
          'bagging_freq': 5, 
          'seed':2018}

훈련

In [8]:
model = lgb.train(params, train_ds, 1000, test_ds, verbose_eval=100, early_stopping_rounds=100)

C:\Users\kohry\anaconda3\lib\site-packages\lightgbm\basic.py:762: UserWarning: categorical_feature in param dict is overridden.
  warnings.warn('categorical_feature in param dict is overridden.')

Training until validation scores don't improve for 100 rounds.
[100]	valid_0's l2: 2.6493
[200]	valid_0's l2: 2.43595
[300]	valid_0's l2: 2.36353
[400]	valid_0's l2: 2.33457
[500]	valid_0's l2: 2.32672
[600]	valid_0's l2: 2.31668
[700]	valid_0's l2: 2.32906
Early stopping, best iteration is:
[600]	valid_0's l2: 2.31668
In [9]:
predict_train = model.predict(train_x)
predict_test = model.predict(test_x)

에러 측정

In [10]:
mse = mean_squared_error(test_y, predict_test)
r2 = r2_score(test_y, predict_test)
In [11]:
print('Mean squared error: ', mse)
print('R2 score: ', r2)

Mean squared error:  2.316679017854007
R2 score:  0.32255843036546283
In [12]:
final_result = pd.concat([test_y.reset_index(drop=True), pd.DataFrame(predict_test)], axis = 1)
In [13]:
final_result.columns = ['label','predict']

fit결과 확인

In [14]:
sns.regplot(x = 'label', y = 'predict', data = final_result)
Out[14]:
<matplotlib.axes._subplots.AxesSubplot at 0x1e6e79ab6a0>

답글 남기기