Candidates preparing for the GATE 2027 Data Science and Artificial Intelligence (DA) paper can use the previous three examination papers to identify recurring topics and develop a more focused preparation strategy.
An analysis of the GATE DA papers from 2024, 2025 and 2026 suggests that Probability and Statistics, Programming and Algorithms, and Machine Learning have remained important areas across all three examinations.
Another notable finding is the increasing representation of Database Management and Warehousing, which accounted for an estimated 18 marks in 2026, compared with 11 marks in 2025 and six marks in 2024.
However, the topic-wise distribution is not officially fixed, and candidates should treat these findings as previous-year trends rather than guaranteed weightages for 2027.
The GATE DA examination carries 100 marks, including 15 marks for General Aptitude and 85 marks for subject-specific questions.
A question-wise classification of the 2024, 2025 and 2026 papers provides the following distribution across the seven major subject areas.
Probability and Statistics: 16 marks in 2024, 19 in 2025 and 19 in 2026.
Programming, Data Structures and Algorithms: 20 marks in 2024, 14 in 2025 and 14 in 2026.
Machine Learning: 15 marks in 2024, 14 in 2025 and 14 in 2026.
Database Management and Warehousing: Six marks in 2024, 11 in 2025 and 18 in 2026.
Linear Algebra: 10 marks in 2024, 12 in 2025 and eight in 2026.
Artificial Intelligence: 10 marks in 2024, six in 2025 and eight in 2026.
Calculus and Optimisation: Eight marks in 2024, nine in 2025 and four in 2026.
The classification assigns each question to its primary syllabus area. Questions involving multiple concepts have not been counted under more than one topic.
Among the seven subject areas, Probability and Statistics has shown one of the most consistent patterns.
The three papers included questions involving conditional probability, Bayes' theorem, random variables, probability distributions, expectations and variance.
The 2024 paper tested Poisson and normal distributions, independent events, conditional probability and joint distributions.
In 2025, questions involved cumulative distribution functions, exponential distributions, Bernoulli random variables and expectations.
The 2026 paper continued this emphasis through questions involving conditional probability, random variables, sampling and statistical properties.
Preparation strategy: Candidates should prioritise probability distributions, conditional probability, Bayes' theorem, expectation, variance and joint distributions. Rather than memorising formulas alone, they should practise applying these concepts to different numerical situations.
Programming, Data Structures and Algorithms accounted for an estimated 20 marks in 2024 and 14 marks each in 2025 and 2026.
Across the three papers, questions tested Python programming, recursion, sorting, searching, stacks, trees, hashing and graph algorithms.
The 2024 examination included problems involving depth-first search, binary search, quicksort, tree traversals and Python functions.
The 2025 paper tested Python lists and sets, hashing, insertion sort, recursive functions and graph traversal.
In 2026, questions involved mutable default arguments in Python, recursion, binary search, sorting, graph traversal and tree-based structures.
Preparation strategy: Candidates should practise tracing Python programs manually and estimating algorithmic complexity. Particular attention should be paid to recursion, sorting algorithms, binary search, hashing, trees and graph traversal.
Writing short Python programs and predicting their outputs can help candidates strengthen conceptual understanding.
Machine Learning accounted for approximately 14–15 marks annually in the analysed papers.
Frequently tested concepts included classification, regression, clustering, dimensionality reduction and neural networks.
The 2024 paper included Support Vector Machines, Principal Component Analysis, Naive Bayes classification, clustering and neural networks.
The 2025 examination featured linear classifiers, least-squares regression, clustering, Support Vector Machines and neural networks.
In 2026, questions tested PCA, cross-validation, stochastic gradient descent, hierarchical clustering, ridge regression, classification metrics and multilayer perceptrons.
Preparation strategy: Candidates should develop a clear understanding of how algorithms work, rather than restricting their preparation to definitions.
Topics such as PCA, regression, classification, clustering, cross-validation and neural networks deserve regular practice. Numerical questions involving model parameters, distance metrics and objective functions should receive particular attention.
Database Management and Warehousing is one area where the analysed weightage changed considerably.
The estimated distribution increased from six marks in 2024 to 11 marks in 2025 and 18 marks in 2026.
The papers included questions involving relational algebra, SQL, functional dependencies, database normalisation, joins and data warehousing.
The 2026 paper placed considerable emphasis on database concepts, including B+ trees, SQL queries, relational operations, ER modelling and data cubes.
Preparation strategy: Candidates should not overlook DBMS while focusing on AI and Machine Learning.
Practising SQL queries, relational algebra, functional dependencies, normalisation, indexing and data warehouse concepts could improve their preparation.
However, the increase in 2026 does not establish that DBMS will carry the same weightage in GATE 2027.
Linear Algebra accounted for approximately eight to 12 marks across the three papers.
Questions involved matrices, eigenvalues, determinants, vector spaces, orthogonality, projections and matrix properties.
Calculus and Optimisation showed a more variable distribution, ranging from four to nine marks in the analysed papers.
Questions tested differentiation, limits, continuity, stationary points and optimisation concepts.
Preparation strategy: Candidates should strengthen matrix operations, eigenvalues, eigenvectors, rank, orthogonality, differentiation and optimisation techniques.
These topics also support Machine Learning, making them important even when their direct examination weightage is relatively lower.
Artificial Intelligence questions appeared in each of the three examinations, accounting for approximately six to 10 marks annually.
The papers tested search algorithms, logical reasoning, predicate logic, game trees, minimax and related concepts.
The 2024 paper included A* search, admissible heuristics and alpha-beta pruning.
The 2025 paper featured Bayesian network inference, A* search and game-tree concepts.
The 2026 examination included informed and uninformed search, logical entailment, predicate logic and minimax-based problems.
Preparation strategy: Candidates should practise search algorithms, first-order and propositional logic, game-tree evaluation and reasoning-based questions.
Understanding the conditions under which an algorithm produces a particular result is more useful than simply memorising its definition.
The three-year analysis offers several practical lessons for aspirants.
1. Prioritise recurring subjects, not just the previous year's highest-weightage topic. Probability and Statistics, Programming and Machine Learning maintained substantial representation across the three papers.
2. Do not neglect Database Management. Its increased presence in the 2026 paper demonstrates why concentrating exclusively on AI and Machine Learning can be risky.
3. Strengthen mathematical foundations. Probability, Linear Algebra and Calculus support several Machine Learning concepts and numerical problems.
4. Solve previous-year papers under timed conditions. Candidates should practise all three papers, identify recurring mistakes and revisit weak concepts.
5. Prepare General Aptitude separately. The section accounts for 15 marks, and consistent practice in numerical, verbal and analytical reasoning can contribute to the overall score.
The previous three papers indicate that effective preparation requires a balance between mathematical foundations, programming skills and applied knowledge of AI and Machine Learning.
Candidates can begin with Probability and Statistics, Linear Algebra and Programming, before moving systematically towards Machine Learning, Artificial Intelligence and Database Management.
However, the order should be adjusted according to individual strengths and weaknesses.
The main lesson from the three-year analysis is that preparation should follow the complete syllabus, with additional practice devoted to recurring concepts.
Previous-year weightage can guide revision priorities, but it should not be used to predict the exact distribution of questions in GATE 2027.