AI & ML: Building a Classifier with Python

Site Admin · 11 Sep 2026 · 5 views

Your First Real Model

Scikit-learn is the friendliest path from idea to working model in Python. It ships with built-in datasets, ready-made algorithms, and tuning tools. This post builds a flower classifier end to end: load data, split it, train a random forest, and measure accuracy.

Load and Split the Data

from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split

data = load_iris()
X_train, X_test, y_train, y_test = \
    train_test_split(
        data.data, data.target,
        test_size=0.2, random_state=42
    )

The iris dataset contains 150 samples, each with four measurements of a flower and a species. The training set teaches the model, and the test set stays untouched until the very end.

Train a Random Forest

from sklearn.ensemble import RandomForestClassifier

model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

fit is the learning step. The random forest builds a hundred decision trees, each trained on a slightly different view of the data, and their votes become the prediction. Fewer than twenty lines of code separate a raw dataset from a working classifier.

Predict and Evaluate

from sklearn.metrics import accuracy_score

predictions = model.predict(X_test)
print(accuracy_score(y_test, predictions))

predict runs the untouched test rows through the model, and accuracy_score compares the guesses to the true species. Because the split used a fixed random state, you can re-run the whole workflow and report the score with confidence rather than by luck.

Next Steps with the Same Pattern

The pattern load, split, train, predict, evaluate repeats almost unchanged across problems. Swap the dataset for customer churn data, swap the algorithm for logistic regression, and the structure holds. Scikit-learn also provides a confusion matrix, cross validation, and grid search, which lets you compare approaches systematically as your projects grow.