Scribd
Upload
11 views4 pages

# Load The Pima Indians Diabetes Dataset # Change Path If Needed

The document outlines a Python script that implements Support Vector Classification (SVC) on the Pima Indians Diabetes dataset using both linear and RBF kernels. It includes data preprocessing steps, such as handling zero values, and evaluates model performance through K-fold cross-validation, reporting accuracy, precision, recall, and F1 scores. The results indicate that the linear kernel achieved a higher average accuracy compared to the RBF kernel.

Uploaded by

nagaryashgurjar
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
11 views4 pages

# Load The Pima Indians Diabetes Dataset # Change Path If Needed

The document outlines a Python script that implements Support Vector Classification (SVC) on the Pima Indians Diabetes dataset using both linear and RBF kernels. It includes data preprocessing steps, such as handling zero values, and evaluates model performance through K-fold cross-validation, reporting accuracy, precision, recall, and F1 scores. The results indicate that the linear kernel achieved a higher average accuracy compared to the RBF kernel.

Uploaded by

nagaryashgurjar
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

In [1]: import numpy as np

import pandas as pd
import [Link] as plt
from sklearn.model_selection import KFold
from [Link] import StandardScaler
from [Link] import SVC
from [Link] import accuracy_score, precision_score, recall_score, f1_score, classificat

In [3]: # Load the Pima Indians Diabetes dataset


df = pd.read_csv(r"C:\Users\hp\Downloads/[Link]") # Change path if needed
[Link]()

Out[3]: Pregnancies Glucose BloodPressure SkinThickness Insulin BMI DiabetesPedig

0 6 148 72 35 0 33.6

1 1 85 66 29 0 26.6

2 8 183 64 0 0 23.3

3 1 89 66 23 94 28.1

4 0 137 40 35 168 43.1

In [4]: [Link]().sum()

Out[4]: Pregnancies 0
Glucose 0
BloodPressure 0
SkinThickness 0
Insulin 0
BMI 0
DiabetesPedigreeFunction 0
Age 0
Outcome 0
dtype: int64

In [10]: columns_with_zero = ['Glucose','BloodPressure','SkinThickness','Insulin','BMI']

for col in columns_with_zero:


[Link][df[col] == 0, col] = df[col].median()

# Show first 5 rows to confirm values updated


[Link]()

Out[10]: Pregnancies Glucose BloodPressure SkinThickness Insulin BMI DiabetesPedig

0 6 148 72 35 30.5 33.6

1 1 85 66 29 30.5 26.6

2 8 183 64 23 30.5 23.3

3 1 89 66 23 94.0 28.1

4 0 137 40 35 168.0 43.1


In [11]: X = [Link]("Outcome", axis=1)
y = df["Outcome"]

In [12]: def evaluate_svc_model(model, X, y):


kf = KFold(n_splits=5, shuffle=True, random_state=42)

fold_accuracies = []
fold_metrics = []
confusion_matrices = []

scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)

fold = 1
for train_index, test_index in [Link](X):
X_train, X_test = X_scaled[train_index], X_scaled[test_index]
y_train, y_test = [Link][train_index], [Link][test_index]

[Link](X_train, y_train)
y_pred = [Link](X_test)

acc = accuracy_score(y_test, y_pred)


prec = precision_score(y_test, y_pred)
rec = recall_score(y_test, y_pred)
f1 = f1_score(y_test, y_pred)

fold_accuracies.append(acc)

fold_metrics.append({
"fold": fold,
"accuracy": acc,
"precision": prec,
"recall": rec,
"f1": f1,
"report": classification_report(y_test, y_pred, output_dict=True)
})

confusion_matrices.append(confusion_matrix(y_test, y_pred))

fold += 1

return fold_accuracies, fold_metrics, confusion_matrices

In [13]: model_linear = SVC(kernel='linear')

acc_lin, metrics_lin, conf_lin = evaluate_svc_model(model_linear, X, y)

acc_lin

Out[13]: [0.7662337662337663,
0.7792207792207793,
0.7467532467532467,
0.7908496732026143,
0.7581699346405228]

In [14]: model_rbf = SVC(kernel='rbf')

acc_rbf, metrics_rbf, conf_rbf = evaluate_svc_model(model_rbf, X, y)


acc_rbf

Out[14]: [0.7532467532467533,
0.7597402597402597,
0.7532467532467533,
0.7516339869281046,
0.7516339869281046]

In [15]: print("Linear SVC Average Accuracy:", [Link](acc_lin))


print("RBF SVC Average Accuracy:", [Link](acc_rbf))

Linear SVC Average Accuracy: 0.7682454800101859


RBF SVC Average Accuracy: 0.7539003480179951

In [16]: best_fold_linear = [Link](acc_lin)


print("Best fold:", best_fold_linear + 1)

conf_best_linear = conf_lin[best_fold_linear]
conf_best_linear

Best fold: 4
Out[16]: array([[96, 10],
[22, 25]])

In [17]: best_fold_rbf = [Link](acc_rbf)


print("Best fold:", best_fold_rbf + 1)

conf_best_rbf = conf_rbf[best_fold_rbf]
conf_best_rbf

Best fold: 2
Out[17]: array([[90, 17],
[20, 27]])

In [18]: print("----- Linear Kernel Classification Report -----")


print([Link](metrics_lin[best_fold_linear]['report']).T)

print("\n----- RBF Kernel Classification Report -----")


print([Link](metrics_rbf[best_fold_rbf]['report']).T)

----- Linear Kernel Classification Report -----


precision recall f1-score support
0 0.813559 0.905660 0.857143 106.00000
1 0.714286 0.531915 0.609756 47.00000
accuracy 0.790850 0.790850 0.790850 0.79085
macro avg 0.763923 0.718788 0.733449 153.00000
weighted avg 0.783064 0.790850 0.781148 153.00000

----- RBF Kernel Classification Report -----


precision recall f1-score support
0 0.818182 0.841121 0.829493 107.00000
1 0.613636 0.574468 0.593407 47.00000
accuracy 0.759740 0.759740 0.759740 0.75974
macro avg 0.715909 0.707795 0.711450 154.00000
weighted avg 0.755756 0.759740 0.757441 154.00000

In [19]: [Link](figsize=(10,6))
[Link](range(1,6), acc_lin, marker='o', label='Linear Kernel')
[Link](range(1,6), acc_rbf, marker='o', label='RBF Kernel')
[Link]("Fold")
[Link]("Accuracy")
[Link]("Accuracy Across 5 Folds for SVC Models")
[Link]()
[Link](True)
[Link]()

You might also like