Synetic College Logo
Unit 1 of 1 · BCA Sem 4

Unit 1: ML model implementation in Python

Introduction to Machine Learning Laboratory notes · PTU syllabus (UGDSE204)

3 min read4 topics9 exam questions
On this page
  1. Unit summary
  2. The standard scikit-learn pattern
  3. Visualising models
  4. Perceptron for AND, OR and XOR
  5. PCA and boosting
  6. Key terms
  7. Quick revision
  8. Important questions

Unit summary

This lab implements the main ML algorithms in Python with scikit-learn and visualises their results: regression lines, decision boundaries, trees, Naive Bayes, random forests, SVMs, three clustering methods, PCA, perceptrons and AdaBoost.

After this unit you can

  • Train and evaluate regression and classification models
  • Visualise regression lines, decision boundaries and dendrograms
  • Apply K-means, hierarchical clustering, DBSCAN and PCA
  • Demonstrate perceptrons and boosting

PTU syllabus topics

  • Linear regression with regression line visualization
  • logistic regression with decision boundary
  • decision tree (ID3/CART) classifier
  • Naive Bayes classifier
  • random forest classifier
  • SVM for linearly separable classes
  • K-Means clustering with visualization
  • hierarchical clustering with dendrogram
  • DBSCAN clustering
  • PCA with classifier performance comparison
  • single-layer perceptron for AND/OR/XOR
  • AdaBoost boosting demonstration
ProcessEvery ML experiment
  1. 1Load data
  2. 2Split

    Training and test sets

  3. 3Train

    Fit the model

  4. 4Evaluate

    Accuracy, precision, recall

  5. 5Visualise

    Decision boundary, clusters or tree

1

Topic 1

The standard scikit-learn pattern

pythonfrom sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score, confusion_matrix
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
model = SomeModel()                 # e.g. LogisticRegression()
model.fit(X_train, y_train)
pred = model.predict(X_test)
print(accuracy_score(y_test, pred), confusion_matrix(y_test, pred))
Experimentscikit-learn class
Linear regressionLinearRegression
Logistic regressionLogisticRegression
Decision tree (ID3/CART)DecisionTreeClassifier(criterion="entropy" or "gini")
Naive BayesGaussianNB
Random forestRandomForestClassifier
SVM (linear)SVC(kernel="linear")
K-meansKMeans(n_clusters=k)
HierarchicalAgglomerativeClustering; scipy dendrogram
DBSCANDBSCAN(eps, min_samples)
PCAPCA(n_components=2)
BoostingAdaBoostClassifier
2

Topic 2

Visualising models

pythonimport matplotlib.pyplot as plt
plt.scatter(X, y); plt.plot(X, model.predict(X), color="red")   # regression line
from scipy.cluster.hierarchy import dendrogram, linkage
dendrogram(linkage(X, method="ward")); plt.show()

For a decision boundary, predict on a grid of points (np.meshgrid) and draw it with plt.contourf.

3

Topic 3

Perceptron for AND, OR and XOR

pythonimport numpy as np
def train(X, t, lr=0.1, epochs=20):
    w, b = np.zeros(2), 0
    for _ in range(epochs):
        for xi, ti in zip(X, t):
            y = 1 if xi @ w + b > 0 else 0
            w += lr * (ti - y) * xi; b += lr * (ti - y)
    return w, b
X = np.array([[0, 0], [0, 1], [1, 0], [1, 1]])
train(X, [0, 0, 0, 1])   # AND: learns correctly
train(X, [0, 1, 1, 0])   # XOR: never converges — not linearly separable
4

Topic 4

PCA and boosting

Apply PCA to reduce features to 2, train the same classifier before and after, and compare accuracy and training time. AdaBoost combines many weak learners (shallow trees), giving more weight to misclassified samples each round.

Exam tip

Always set random_state for reproducible results and report accuracy with a confusion matrix — examiners look for both.

Key terms

train_test_split
Splits data into training and test sets
fit / predict
Train a model / make predictions
Dendrogram
A tree diagram of hierarchical clustering
AdaBoost
A boosting method combining weak learners
Decision boundary
The line or surface separating predicted classes

Quick revision

  • fit on training data, evaluate on test data.
  • criterion="entropy" ≈ ID3; "gini" ≈ CART.
  • Perceptron learns AND/OR, fails on XOR.
  • PCA then classify to compare performance.

Important exam questions

Practice questions written to the PTU exam pattern for this unit's syllabus: short answers (Section A style) and long answers (Sections B and C style).

Short-answer questions

  1. Q1.What does train_test_split do?
  2. Q2.Which class implements Naive Bayes in scikit-learn?
  3. Q3.What do eps and min_samples mean in DBSCAN?
  4. Q4.Why does the perceptron fail on XOR?
  5. Q5.What is AdaBoost?

Long-answer questions

  1. Q1.Implement linear regression and plot the regression line.
  2. Q2.Implement a decision tree and a random forest on the same data set and compare accuracy.
  3. Q3.Implement K-means and hierarchical clustering and visualise the clusters and dendrogram.
  4. Q4.Implement a single-layer perceptron for AND, OR and XOR and explain the results.

Stuck on this unit?

Message SBS on WhatsApp for help with Introduction to Machine Learning Laboratory, or to ask about studying BCA at Synetic.

WhatsApp us