Checked 25 September 2026. Dates are official. The day DA is held within 6–21 February had not been announced; if it falls later, add the extra weeks to revision.
GATE 2027 runs on 6, 7, 13, 14, 20 and 21 February 2027, and the day of the DA paper is not yet announced. From the end of September that leaves about nineteen weeks — enough to prepare all seven DA sections properly, if the order is right.
The order matters more in DA than in most GATE papers, because its sections depend on each other. Machine learning questions lean on probability and linear algebra; AI's reasoning-under-uncertainty questions are Bayes' theorem in disguise. So this plan front-loads the mathematics, and splits your hours by the marks each section actually carried in the three official papers.
Official dates to put in your calendar
| Date | What |
|---|---|
| 27 Sep 2026 | Regular registration closes |
| 5 Oct 2026 | Extended registration closes (late fee) |
| 14–21 Oct 2026 | Application correction window |
| 4 Jan 2027 | Exam city allotment notification |
| To be announced | Admit card; the day and session of the DA paper |
| 6–21 Feb 2027 | Exam days (forenoon 9:30–12:30, afternoon 2:30–5:30) |
| 19 Mar 2027 | Result |
Split your hours by the marks, not by the syllabus length
A section with a long syllabus is not necessarily worth more. Machine Learning has the longest topic list in the DA syllabus, but Probability & Statistics carried more marks in every paper by our count. The split below gives each section a share of your time roughly in line with its three-year marks.
| Section | Share of study time | At 20 h/week for exactly 19 weeks |
|---|---|---|
| Probability & Statistics | 19% | 72 h |
| Programming, DS & Algorithms | 16.5% | 63 h |
| Machine Learning | 14% | 53 h |
| DBMS & Warehousing | 12.5% | 48 h |
| Linear Algebra | 10.5% | 40 h |
| AI | 8% | 30 h |
| Calculus & Optimization | 7.5% | 28 h |
| General Aptitude | 12% | 46 h |
Shares add to 100% and the hours to exactly 380; the planner below counts from today, so its totals differ slightly. General Aptitude gets 12% of the time for 15% of the marks deliberately: it needs practice rather than new learning, and most candidates score a higher share of its marks than of the subject part. If GA is a weakness for you, move it up to 15%.
GATE DA 2027 study-hours planner
Choose your likely exam day and how many hours a week you can study. The planner counts the weeks left from today and splits your total hours across the paper.
Phase 1 — Mathematics first (weeks 1–6)
Goal: make the three maths sections solid before machine learning
- Weeks 1–3: Probability & Statistics. Counting, conditional probability and Bayes, discrete and continuous distributions, expectation and variance, conditional expectation, CLT, confidence intervals, z/t/χ² tests. Do every P&S question from the three past papers as you go.
- Weeks 4–5: Linear Algebra. Vector spaces, rank and nullity, eigenvalues, projections, orthogonal and idempotent matrices, quadratic forms, LU and SVD.
- Week 6: Calculus & Optimization. Limits, continuity, differentiability, Taylor series, maxima and minima of one variable. Short syllabus — one week is enough.
- Every day, all six weeks: 45 minutes of Python — reading code and predicting output, because that is how programming is examined.
Phase 2 — Computing and ML (weeks 7–11)
Goal: cover the remaining four sections, each with past-paper practice
- Weeks 7–8: Programming, Data Structures and Algorithms. Stacks, queues, linked lists, trees, hash tables; linear and binary search; selection, bubble, insertion, merge and quick sort; graph traversals and shortest paths.
- Week 9: DBMS & Warehousing. ER and relational models, relational algebra, tuple calculus, SQL, constraints, normal forms, indexing; data transformation; OLAP schemas, concept hierarchies and measures. It carried 18 marks in 2026 — do not rush it.
- Week 10: Machine Learning. Regression (simple, multiple, ridge, logistic), kNN, naive Bayes, LDA, SVM, decision trees, bias–variance, cross-validation, MLPs; k-means/k-medoid, hierarchical clustering, PCA.
- Week 11: AI. Uninformed, informed and adversarial search; propositional and predicate logic; conditional independence, variable elimination and sampling.
Phase 3 — Full papers (weeks 12–15)
Goal: learn to take the paper
- One timed 3-hour paper a week, using only an on-screen calculator. Take the 2024 official paper first and keep 2026 for later.
- Spend as long analysing each paper as taking it; log every lost mark by cause (did not know, misread, calculation slip, time, negative marking, MSQ partly right).
- Fix the two biggest causes the following week before taking the next paper.
Reading a full paper afterwards
The hour after a mock is where the marks are found. DA papers lose marks in a few characteristic ways, and each needs a different fix — so sort every lost mark before you look at the total.
| How the mark was lost | Typical cause | Fix next week |
|---|---|---|
| Python trace went wrong | Slicing bounds, mutable default, integer vs float division | Ten short traces a day, written out line by line |
| Probability numerical wrong | Conditioned on the wrong event | Write the event in words before the formula |
| SQL output wrong | NULLs, duplicates, GROUP BY with HAVING | Run the query by hand on a 4-row table |
| ML MSQ partly right | One option judged by intuition | Prove or refute each option separately |
| Linear algebra slip | Rank or eigenvalue arithmetic | Check with trace = sum of eigenvalues, det = product |
| Negative marks | Guessed MCQs | Guess only after ruling out at least one option |
Phase 4 — Revision (week 16 to the exam)
Goal: nothing new, everything faster
- Two timed papers a week (the remaining official paper plus mocks), alternating with days built from your mistakes log.
- A two-page formula sheet each for P&S, linear algebra and ML; read them daily in the last ten days.
- Download the admit card as soon as it is out and check the centre and reporting time.
If DA is scheduled in the second or third week of February, add the extra weeks to Phase 4.
The week-by-week table
| Week | Main section | Daily | Practice |
|---|---|---|---|
| 1 (from 28 Sep) | P&S: counting, probability, Bayes | Python output tracing | P&S past questions |
| 2 | P&S: distributions, expectation | Python | P&S past questions |
| 3 | P&S: CLT, intervals, tests | Python | P&S past questions |
| 4 | Linear algebra: spaces, rank, eigen | Python | LA past questions |
| 5 | Linear algebra: projections, SVD | Python | LA past questions |
| 6 | Calculus & optimization | Python | Calculus past questions |
| 7 | Data structures | GA (2×/week) | Programming past questions |
| 8 | Algorithms: search, sort, graphs | GA | Programming past questions |
| 9 | DBMS & warehousing | GA | DBMS past questions |
| 10 | Machine learning | P&S revision | ML past questions |
| 11 | AI | LA revision | AI past questions |
| 12 | Weak sections from the log | Mixed problems | Full paper: 2024 |
| 13 | Weak sections | Mixed problems | Full mock + analysis |
| 14 | Weak sections | Mixed problems | Full paper: 2025 |
| 15 | Formula sheets | Mixed problems | Full mock + analysis |
| 16 | Revision | Sheets | 2 full papers |
| 17 | Revision | Sheets | Full paper: 2026 + a mock |
| 18 | Mistakes log only | Sheets | 1–2 papers |
| 19 (to 6 Feb) | Sheets only | Rest | 1 light paper |
Two rules hold across all four phases. First, never let a week pass without Python: output-tracing questions appeared in every official DA paper, and the skill fades quickly without practice. Second, keep a single mistakes log from week one — by January it is the most valuable document you own, because it lists exactly the errors you personally make.
Starting late?
| If you start in | Keep | Trim | First full paper |
|---|---|---|---|
| Early November (≈13 weeks) | P&S, Programming, DBMS in full; ML core; AI search and logic | Calculus to maxima/minima; data-warehousing detail | Week 8 |
| Early December (≈9 weeks) | P&S, Programming, DBMS; past-paper questions for the rest | ML to the recurring themes; skip deep theory | Week 5 |
| January (≈5 weeks) | The three official papers by section, twice | Anything not already studied once | Week 1 |
The three sections kept in every row — Probability & Statistics, Programming and DBMS — together carried 42 of 85 subject marks in 2024, 44 in 2025 and 53 in 2026 by our count. If time is short, they are where it goes furthest.
Practise it, don't just read it
Topic quizzes fit Phase 1 and Phase 2: ten questions on one topic, marked instantly, with a worked explanation for every option.
The ProSyllabus GATE DA library currently has three topic quizzes: Matrix Rank and Inverse (Section 2, Linear Algebra), Backpropagation in Neural Networks (Section 6, multi-layer perceptrons) and Gradient Descent Variants (optimisation practice for Machine Learning — the syllabus names no gradient methods, though a 2026 question applied an SGD update). Each has ten questions and a worked explanation for every option. Probability, programming and DBMS have no quizzes yet — drill those from the section index of the official papers.
Are four months enough for GATE DA?
Usually yes, for a candidate with an engineering, mathematics or statistics background. About nineteen weeks remain from late September to 6 February 2027 — enough to cover all seven sections by early December and spend the rest on full papers and revision.
Which GATE DA section should I start with?
Probability and Statistics. It carried the most marks in the official papers (15, 19 and 21 of 85 in 2024 to 2026) and machine learning and AI questions build on it. Linear algebra and calculus should follow before machine learning.
How many hours a day for GATE DA?
This plan assumes about 20 hours a week — roughly three hours on weekdays and five on one weekend day. The planner on this page recalculates the section split for any number of hours and any February exam date.
When is the GATE DA 2027 exam?
GATE 2027 is on 6, 7, 13, 14, 20 and 21 February 2027. The day of the DA paper had not been announced when this was checked.
How should I practise Python for GATE DA?
By reading code and predicting its output. Python output-tracing questions appeared in all three official DA papers, so daily practice tracing loops, list operations, slicing, recursion and simple data structures is more useful than writing large programs.
More in this series
- GATE DA topic-wise weightage, counted from 2024–2026 papers
- GATE DA marks vs score vs rank, with calculator
- GATE DA previous year papers 2024–2026, with answer keys
- GATE DA or GATE CS — which paper, and writing both
- GATE DA without a CS degree
- Probability & Statistics for GATE DA — the heaviest section
- GATE DA cutoffs, rank vs IIT and placements
- GATE DA notice board — official updates
Sources
- official — GATE 2027 important dates
- official — GATE 2027 question paper pattern
- official — GATE 2027 Information Brochure, revised 17 Sep 2026
- official — GATE 2027 DA syllabus (IIT Madras)
- official — GATE 2024 DA question paper (IISc)
- official — GATE 2025 DA question paper (IIT Roorkee)
- official — GATE 2026 DA question paper (IIT Guwahati)
official = a document published by the conducting body. reported = a news or coaching site we could not check against an original. Where sources disagree this page says so rather than picking one.
