K Nearest Neighbors¶
Import Libraries¶
import pandas as pd
import numpy as np
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import confusion_matrix, classification_report, accuracy_score
Load Data¶
df = pd.read_csv("data/loans.csv").round(1)
df.head()
| Income | Credit_Score | Status | |
|---|---|---|---|
| 0 | 86499.0 | 575.7 | 0 |
| 1 | 106113.7 | 588.0 | 1 |
| 2 | 100279.1 | 575.9 | 0 |
| 3 | 113616.9 | 559.0 | 1 |
| 4 | 135667.3 | 727.1 | 1 |
Run Classifier¶
# extract X,y from pandas
X = df.drop(['Status'],axis=1)
y = df['Status']
# set classifiers for k =3,5
knn3 = KNeighborsClassifier(n_neighbors=3).fit(X,y)
knn5 = KNeighborsClassifier(n_neighbors=5).fit(X,y)
Evaluate Classifier¶
y_pred = knn3.predict(X)
print(accuracy_score(y,y_pred))
0.928
print(confusion_matrix(y,y_pred))
[[474 26]
[ 46 454]]
print(classification_report(y,y_pred))
precision recall f1-score support
0 0.91 0.95 0.93 500
1 0.95 0.91 0.93 500
accuracy 0.93 1000
macro avg 0.93 0.93 0.93 1000
weighted avg 0.93 0.93 0.93 1000
Predictions¶
# define new points for which we want prediction
new_points =[[82000,530],[123000,510],[90000,670],[99000,610]]
y_pred_3 = knn3.predict(new_points)
y_pred_5 = knn5.predict(new_points)
# print predictions
print(y_pred_3)
print(y_pred_5)
[0 1 0 1]
[0 1 0 0]
data = {'KNN=3': y_pred_3, 'KNN=5': y_pred_5}
pd.DataFrame.from_dict(data, orient='index',
columns=['Point1','Point2','Point3','Point4'])
| Point1 | Point2 | Point3 | Point4 | |
|---|---|---|---|---|
| KNN=3 | 0 | 1 | 0 | 1 |
| KNN=5 | 0 | 1 | 0 | 0 |