How to impute missing class labels using nearest neighbours in Python?
This recipe helps you impute missing class labels using nearest neighbours in Python
Recipe Objective
Have you ever tried to impute calss labels? We can impute class labels by K nearest neighbours by training it on known data and predicting the class labels.
So this is the recipe on how we can impute missing class labels using nearest neighbours in Python.
List of Classification Algorithms in Machine Learning
Table of Contents
Step 1 - Import the library
import numpy as np
from sklearn.neighbors import KNeighborsClassifier
We have imported numpy and KNeighborsClassifier which is needed.
Step 2 - Setting up the Data
We have created a feature matrix using array and we will use this to train the KNN model.
X = np.array([[0, 2.10, 1.45],
[2, 1.18, 1.33],
[0, 1.22, 1.27],
[1, 1.32, 1.97],
[1, -0.21, -1.19]])
We have created a matrix with missing class labels.
X_with_nan = np.array([[np.nan, 0.87, 1.31],
[np.nan, 0.37, 1.91],
[np.nan, 0.54, 1.27],
[np.nan, -0.67, -0.22]])
Step 3 - Predicting the Class Labels
We are training the KNeighborsClassifier with parameters K equals to 3 and weights equals to distance. We have used the matrix X to train the model.
clf = KNeighborsClassifier(3, weights="distance")
trained_model = clf.fit(X[:,1:], X[:,0])
We have predicted the class labels of matrix "X_with_nan".
imputed_values = trained_model.predict(X_with_nan[:,1:])
print(imputed_values)
So finally we have filled the null values with the predicted output of model.
X_with_imputed = np.hstack((imputed_values.reshape(-1,1), X_with_nan[:,1:]))
print(); print(X_with_imputed)
So the output comes as
[2. 1. 2. 1.]
[[ 2. 0.87 1.31]
[ 1. 0.37 1.91]
[ 2. 0.54 1.27]
[ 1. -0.67 -0.22]]
Download Materials
What Users are saying..
Abhinav Agarwal
I come from Northwestern University, which is ranked 9th in the US. Although the high-quality academics at school taught me all the basics I needed, obtaining practical experience was a challenge.... Read More