What the PYQs Reveal
The examiner repeatedly returns to the same core concepts, but often changes the wording and depth.
Six AL405 Machine Learning question papers were analyzed: June 2022, June 2023, November 2023, June 2024, December 2024 and June 2025.
The strongest recurring areas are Neural Networks, K-Means Clustering, Logistic Regression, Decision Trees and SVM.
A major pattern is that RGPV does not always repeat the exact wording. Instead, the same concept appears through different angles such as definitions, algorithms, diagrams, numerical problems, comparisons and applications.
Therefore, prepare the complete concept instead of memorizing only one previous-year question.
🔥 Top 6 Topics for 2026
These topics have the strongest combination of repetition, recency and examination relevance.
K-Means Clustering
Algorithm, centroid calculation, Euclidean distance, numerical problems, objective function, K selection, limitations and hard vs soft clustering.
MUST PREPARENeural Networks / MLP
ANN, perceptron, MLP architecture, weights, biases, activation functions, mathematical representation and neural-network working.
MUST PREPAREBackpropagation
Forward propagation, error calculation, backward propagation, gradient calculation and weight-update process.
VERY HIGHLogistic Regression
Binary classification, sigmoid function, mathematical equation, cost function and comparison with linear regression.
VERY HIGHDecision Tree
Tree construction, entropy, information gain, node splitting, algorithm, examples and common issues.
VERY HIGHSupport Vector Machine
Hyperplane, support vectors, margin, optimal hyperplane, classification and kernel trick.
VERY HIGH📊 Most Repeated Topics
Topic priority based on their recurrence across the uploaded AL405 papers.
| Rank | Topic | Recurrence | Priority |
|---|---|---|---|
| 01 | Neural Networks / Perceptron / MLP | Very Frequent | ★★★★★ |
| 02 | K-Means / Clustering | Very Frequent | ★★★★★ |
| 03 | Logistic Regression | Highly Repeated | ★★★★★ |
| 04 | Decision Tree | Highly Repeated | ★★★★★ |
| 05 | Support Vector Machine | Repeated | ★★★★☆ |
| 06 | EM / Gaussian Mixture | Repeated | ★★★★☆ |
| 07 | Linear Regression | Repeated | ★★★★☆ |
| 08 | PCA / Dimensionality Reduction | Repeated | ★★★★☆ |
| 09 | Cross Validation / Resampling | Repeated | ★★★★☆ |
| 10 | ML Issues / Limitations | Repeated | ★★★★☆ |
🎯 Prepare in These Tiers
If you have limited time, follow this order.
100%
90%
75%
🔮 AL405 Machine Learning — 2026 Prediction
Probable question areas based on recurring concepts in the uploaded PYQs. This is a prediction, not a guarantee.
Most Expected Question Pattern
ANALYSIS BASEDExplain Machine Learning and its different perspectives/issues. Also explain supervised and unsupervised learning with suitable examples.
Explain the Multilayer Perceptron (MLP) architecture with a neat diagram. Explain the Backpropagation algorithm and weight-update process.
Explain Support Vector Machine (SVM), support vectors, optimal hyperplane and margin. Also explain the kernel trick and different types of kernels.
Explain Decision Tree Learning with an example. Discuss entropy and information gain. Also explain K-Means clustering with a suitable numerical problem.
Explain the Expectation-Maximization (EM) algorithm. Describe E-Step and M-Step and explain its application to Gaussian Mixture Models.
Explain Principal Component Analysis (PCA) and its mathematical formulation. Discuss its role in dimensionality reduction and visualization.
Explain cross-validation and resampling methods. Discuss model evaluation, experimental design, dataset partitioning and generalization.
Write short notes on any two: Hypothesis Testing, Inductive Bias, Activation Functions, Classifier Performance, Shrinkage Methods and Gradient Descent / Optimizers.
✅ 15 Topics You Should Not Skip
Complete these before moving to lower-priority topics.
K-Means Clustering
Algorithm + numerical + centroid + Euclidean distance.
Multilayer Perceptron
Architecture, layers, working and diagram.
Backpropagation
Algorithm, error gradients and weight updates.
Logistic Regression
Sigmoid function, classification and cost function.
Decision Tree
Entropy, information gain and tree construction.
SVM
Hyperplane, margin, support vectors and kernels.
EM Algorithm
E-Step, M-Step and Gaussian Mixture Models.
PCA
Mathematical formulation and dimensionality reduction.
Linear Regression
Model, assumptions, error and minimization.
Cross Validation
K-fold validation and model evaluation.
Resampling
Resampling techniques and evaluation.
ML Issues
Limitations, challenges and generalization.
Bias-Variance
Bias, variance, overfitting and underfitting.
Experimental Design
Dataset partitioning, reproducibility and comparison.
Activation Functions
Sigmoid, ReLU, Tanh and related concepts.
🧮 Don't Ignore These
Machine Learning is not only a theory paper. Practice the mathematical/application-oriented areas.
🔥 K-Means Numerical = Highest Priority
K-Means numerical problems appeared directly in the uploaded papers. Practice the complete process rather than only reading the algorithm.
- Initial centroid selection
- Euclidean distance calculation
- Cluster assignment
- New centroid calculation
- Iteration until convergence
- Objective function
- Choosing the value of K
- Hard vs soft clustering
🚀 Smart Exam Strategy
Follow this sequence for efficient preparation.
Step 1
Finish all Tier-1 topics first.
Step 2
Practice K-Means and important mathematical concepts.
Step 3
Practice diagrams and algorithm-based answers.
Step 4
Revise short notes and lower-priority topics.
⭐ Final 2026 Priority List
If you have very limited time, start with these eight:
- K-Means + Numerical
- MLP
- Backpropagation
- Logistic Regression
- Decision Tree
- SVM
- EM Algorithm
- PCA