This Tutorial explains how to generate K-folds for cross-validation with groups using scikit-learn for evaluation of machine learning models with out of sample data.

During this notebook you will work with flights in and out of NYC in 2013.

Packages

This tutorial uses:

Open up a new Jupyter notebook and import the following:


import statsmodels.api as sm
import pandas as pd
import numpy as np
from sklearn.model_selection import GroupKFold

Reading the data

The data is from rdatasets imported using the Python package statsmodels.


df = sm.datasets.get_rdataset('flights', 'nycflights13').data
df.info()


RangeIndex: 336776 entries, 0 to 336775
Data columns (total 19 columns):
 #   Column          Non-Null Count   Dtype  
---  ------          --------------   -----  
 0   year            336776 non-null  int64  
 1   month           336776 non-null  int64  
 2   day             336776 non-null  int64  
 3   dep_time        328521 non-null  float64
 4   sched_dep_time  336776 non-null  int64  
 5   dep_delay       328521 non-null  float64
 6   arr_time        328063 non-null  float64
 7   sched_arr_time  336776 non-null  int64  
 8   arr_delay       327346 non-null  float64
 9   carrier         336776 non-null  object 
 10  flight          336776 non-null  int64  
 11  tailnum         334264 non-null  object 
 12  origin          336776 non-null  object 
 13  dest            336776 non-null  object 
 14  air_time        327346 non-null  float64
 15  distance        336776 non-null  int64  
 16  hour            336776 non-null  int64  
 17  minute          336776 non-null  int64  
 18  time_hour       336776 non-null  object 
dtypes: float64(5), int64(9), object(5)
memory usage: 48.8+ MB

Feature Engineering

Handle null values


df.isnull().sum()

year                 0
month                0
day                  0
dep_time          8255
sched_dep_time       0
dep_delay         8255
arr_time          8713
sched_arr_time       0
arr_delay         9430
carrier              0
flight               0
tailnum           2512
origin               0
dest                 0
air_time          9430
distance             0
hour                 0
minute               0
time_hour            0
dtype: int64

As this model will predict arrival delay, the Null values are caused by flights did were cancelled or diverted. These can be excluded from this analysis.


df.dropna(inplace=True)
df.reset_index(drop=True, inplace=True)

Convert the times from floats or ints to hour and minutes


df['arr_hour'] = df.arr_time.apply(lambda x: int(np.floor(x/100)))
df['arr_minute'] = df.arr_time.apply(lambda x: int(x - np.floor(x/100)*100))
df['sched_arr_hour'] = df.sched_arr_time.apply(lambda x: int(np.floor(x/100)))
df['sched_arr_minute'] = df.sched_arr_time.apply(lambda x: int(x - np.floor(x/100)*100))
df['sched_dep_hour'] = df.sched_dep_time.apply(lambda x: int(np.floor(x/100)))
df['sched_dep_minute'] = df.sched_dep_time.apply(lambda x: int(x - np.floor(x/100)*100))
df.rename(columns={'hour': 'dep_hour',
                   'minute': 'dep_minute'}, inplace=True
                   

Cross-validation splitting

Scikit-learn's GroupKFold will randomly sample the data into N folds (default of 5) that can be used to perform cross-validation during machine learning training.

In this case, group records by individual planes as once a plane is late, the subsequent flights are more likely to be delayed as well.


group = df.tailnum.tolist()

Create the features and target before running cross-validation


target = 'arr_delay'
y = df[target]
X = df.drop(columns=[target, 'flight', 'time_hour', 'year', 'dep_time', 'sched_dep_time', 'arr_time', 'sched_arr_time', 'dep_delay'])

gkf = GroupKFold(n_splits=10)
for train_index, test_index in gkf.split(X, groups=group):
    print("Train:", train_index, "Test:", test_index)
    X_train = X.iloc[train_index, :]
    y_train = y[train_index]
    X_test = X.iloc[test_index, :]
    y_test = y[test_index]
    

Train: [     0      1      2 ... 327340 327342 327345] Test: [     8     10     14 ... 327341 327343 327344]
Train: [     0      2      4 ... 327341 327343 327344] Test: [     1      3      6 ... 327339 327342 327345]
Train: [     0      1      2 ... 327343 327344 327345] Test: [    26     57     73 ... 327314 327317 327325]
Train: [     0      1      2 ... 327343 327344 327345] Test: [    22     51     71 ... 327326 327332 327340]
Train: [     0      1      2 ... 327343 327344 327345] Test: [     9     33     35 ... 327321 327331 327338]
Train: [     0      1      2 ... 327343 327344 327345] Test: [     7     15     30 ... 327278 327313 327330]
Train: [     1      2      3 ... 327343 327344 327345] Test: [     0     11     12 ... 327300 327312 327322]
Train: [     0      1      2 ... 327343 327344 327345] Test: [     4      5     17 ... 327276 327299 327307]
Train: [     0      1      2 ... 327343 327344 327345] Test: [    13     16     34 ... 327316 327327 327333]
Train: [     0      1      3 ... 327343 327344 327345] Test: [     2     24     29 ... 327328 327335 327337]

No-code/low-code data prep and visualization

Request Demo
Try for Free