Python Sklearn Program to Predict Diabetes Using ML

Use sklearn's load_diabetes dataset and LogisticRegression to train a model with model.fit(X_train, y_train) and predict diabetes with model.predict(X_test).

Examples

InputPatient data: [0.038075906, 0.05068012, 0.061696206, 0.021872354, -0.0442235, -0.03482076, -0.04340085, -0.00259226, 0.01990749, -0.01764613]

OutputPredicted diabetes outcome: 0 (no diabetes)

InputPatient data: [0.08529988, -0.04664164, 0.02087577, 0.03168012, 0.08529988, -0.04664164, 0.02087577, 0.03168012, 0.08529988, -0.04664164]

OutputPredicted diabetes outcome: 1 (diabetes)

InputPatient data: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0]

OutputPredicted diabetes outcome: 0 (no diabetes)

How to Think About It

First, get diabetes data with features and labels. Then split data into training and testing sets. Train a logistic regression model on training data. Finally, use the model to predict diabetes on new patient data.

Algorithm

1

Load diabetes dataset with features and target labels

2

Split dataset into training and testing parts

3

Create logistic regression model

4

Train model using training data

5

Predict diabetes on test data

6

Print prediction results

Code

sklearn

from sklearn.datasets import load_diabetes
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split

# Load dataset
data = load_diabetes()
X = data.data
# Binarize target: diabetes or not (threshold 140)
y = (data.target > 140).astype(int)

# Split data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# Train model
model = LogisticRegression(max_iter=1000)
model.fit(X_train, y_train)

# Predict
predictions = model.predict(X_test)
print('Predictions:', predictions)

Output

Predictions: [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]

Dry Run

Let's trace a sample patient data through the model prediction.

1

Load and prepare data

Features X shape: (442, 10), Target y binarized with threshold 140

2

Split data

Training set size: 353, Testing set size: 89

3

Train logistic regression

Model learns weights from training data

4

Predict on test data

Model outputs array of 0s and 1s for diabetes prediction

Test Sample Index	Predicted Label
0	0
1	0
2	0
3	0
4	0

Why This Works

Step 1: Load and binarize data

We use load_diabetes() to get features and convert continuous target to binary labels with target > 140 to mark diabetes presence.

Step 2: Train logistic regression

Logistic regression fits a model to separate diabetes vs no diabetes by learning weights for each feature.

Step 3: Predict diabetes

The trained model predicts 0 or 1 for new data, indicating absence or presence of diabetes.

Alternative Approaches

Random Forest Classifier

sklearn

from sklearn.datasets import load_diabetes
from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split

# Load and binarize target
data = load_diabetes()
X = data.data
y = (data.target > 140).astype(int)

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

model = RandomForestClassifier(random_state=42)
model.fit(X_train, y_train)
predictions = model.predict(X_test)
print('Predictions:', predictions)

Random Forest can capture complex patterns better but is slower and less interpretable than logistic regression.

Support Vector Machine (SVM)

sklearn

from sklearn.datasets import load_diabetes
from sklearn.svm import SVC
from sklearn.model_selection import train_test_split

# Load and binarize target
data = load_diabetes()
X = data.data
y = (data.target > 140).astype(int)

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

model = SVC(kernel='linear')
model.fit(X_train, y_train)
predictions = model.predict(X_test)
print('Predictions:', predictions)

SVM with linear kernel works well for binary classification but can be slower on large datasets.

Complexity: O(n * d * i) time, O(n * d) space

Time Complexity

Training logistic regression takes O(n * d * i) where n is samples, d is features, and i is iterations for convergence.

Space Complexity

Storing data and model weights requires O(n * d) space; model weights are O(d).

Which Approach is Fastest?

Logistic regression is fastest and simplest; Random Forest and SVM are slower but may improve accuracy.

Approach	Time	Space	Best For
Logistic Regression	O(n * d * i)	O(n * d)	Fast, interpretable binary classification
Random Forest	O(trees * n * d * log n)	O(n * d)	Complex patterns, better accuracy
SVM	O(n^2 * d)	O(n * d)	High-dimensional data, margin-based classification