Quick answer. The RBI Grade B DSIM syllabus has nine units, and both Paper I and Paper II draw from the same pool. They run from classical probability and sampling theory through linear models, inference, stochastic processes, multivariate analysis, econometrics and time series, optimization and statistical computing, and then into machine learning and database management. The Board states that the standard is that of a postgraduate degree examination.
Read unit 8 and unit 9 before you read anything else on this page. Neural networks, gradient boosting, natural language processing, SQL joins and star schemas are now formally examined by a central bank.
The Nine Units
| Unit | Title | Familiar to a Statistics postgraduate? |
|---|---|---|
| 1 | Theory of Probability, Probability Distributions and Sampling Theory | Yes |
| 2 | Linear Models and Economic Statistics | Mostly |
| 3 | Statistical Inference: Estimation, Testing and Non Parametric | Yes |
| 4 | Stochastic Processes | Yes |
| 5 | Multivariate Analysis | Mostly |
| 6 | Econometrics and Time Series | Partly |
| 7 | Optimization and Statistical Computing | Partly, and the computing half is often new |
| 8 | Data Science, AI and Machine Learning | Usually not |
| 9 | Database and Data Warehouse Management | Almost never |
Unit 1: Theory of Probability, Probability Distributions and Sampling Theory
Probability foundations. Classical and axiomatic approaches to probability and their properties. Bayes theorem and its applications. Strong and weak laws of large numbers. Characteristic functions. Central limit theorem. Probability inequalities.
Standard distributions. Binomial, Poisson, Geometric, Negative binomial, Uniform, Normal, Exponential, Logistic, Log normal, Beta, Gamma, Weibull, and Bivariate normal.
Exact sampling distributions. Chi square, Student’s t, F and Z distributions and their applications. Asymptotic sampling distributions and large sample tests. Association and analysis of contingency tables.
Sampling methods. Simple random sampling, stratified random sampling, systematic sampling, cluster sampling, two stage sampling, probability proportional to size. Ratio estimation, regression estimation, non sampling errors and the problem of non response.
Unit 2: Linear Models and Economic Statistics
Linear algebra. Vectors, matrices, spanning of vector space, matrix algebra, inverse of partitioned matrices, g inverse, orthogonal matrices, properties of idempotent matrices, characteristic roots and vectors, Cayley Hamilton theorem, quadratic forms, definite, semi definite and indefinite forms, simultaneous reduction of two quadratic forms, properties of similar matrices.
Simple linear regression. Assumptions, estimation and inference, diagnostic checks, polynomial regression, transformations on Y or X including Box Cox, square root and log, method of weighted least squares, inverse regression.
Multiple regression. Standard Gauss Markov setup, least squares estimation and related properties, regression analysis with correlated observations, simultaneous estimation of linear parametric functions, testing of hypotheses, confidence intervals and regions.
Modern regression topics. Multicollinearity and shrinkage models: ridge regression, LASSO, Elastic Net. Model selection criteria, residual diagnostics, categorical data analysis using dummy variables, outlier detection and treatment.
Economic statistics. Definition and construction of index numbers, standard index numbers, conversion of chain base to fixed base and back, base shifting, splicing and deflating. Measurement of economic inequality: Gini coefficient, Lorenz curves. Basics of macroeconomics and national accounts.
Note that LASSO and Elastic Net appear here in Unit 2 and again in Unit 8. Shrinkage methods are examined from both a linear models angle and a machine learning angle.
Unit 3: Statistical Inference
Estimation. Unbiasedness, sufficiency, consistency and efficiency. Factorization theorem. Complete statistic. Minimum variance unbiased estimator. Rao Blackwell and Lehmann Scheffe theorems and their applications. Cramer Rao inequality.
Methods of estimation. Method of moments, maximum likelihood, least squares, minimum chi square, and basic idea of Bayes estimators.
Testing of significance. Type I and Type II errors, critical region, level of significance, size, p value and its interpretation, power, best critical region, most powerful test, uniformly most powerful test, Neyman Pearson theory. Likelihood ratio tests. Tests of goodness of fit. Bartlett’s test for homogeneity of variances.
Non parametric tests. Kolmogorov Smirnov test, sign test, Wilcoxon signed rank test, Wilcoxon rank sum test, Mann Whitney U test, Kruskal Wallis one way ANOVA, Friedman’s test, Kendall’s tau, Spearman rank correlation.
Also included. Distribution of order statistics, distribution fitting, and kernel density estimation.
Unit 4: Stochastic Processes
Poisson process. Arrival, interarrival and conditional arrival distributions. Non homogeneous processes. Law of rare events and the Poisson process. Compound Poisson processes.
Markov chains. Transition probability matrix, Chapman Kolmogorov equations, regular chains and stationary distributions, periodicity, limit theorems, patterns for recurrent events.
Brownian motion. Limit of random walk, defining characteristics and peculiarities. Martingales.
Unit 5: Multivariate Analysis
Multivariate normal distribution, its properties and characterization.
Logit and Probit models.
Mahalanobis D squared statistic. Linear discriminant analysis. Canonical correlation analysis. Principal components analysis. Factor analysis. Cluster analysis.
This unit is shorter than the multivariate syllabus of most Statistics examinations, but it goes further into applied classification. Logit and probit models, LDA and clustering sit alongside the classical theory.
Unit 6: Econometrics and Time Series
General linear model. Extensions, ordinary least squares and generalized least squares estimation and prediction, heteroscedastic disturbances, pure and mixed estimation. Autocorrelation, its consequences and related tests. Theil BLUS procedure. Multicollinearity, its implications and tools for handling it. Ridge regression.
Regression variants. Linear and stochastic regression, instrumental variable regression, panel regression, autoregressive linear regression, distributed lag models, estimation of lags by OLS.
Simultaneous equations. Simultaneous linear equations model and its generalization, identification problem, restrictions on structural parameters, rank and order conditions, different estimation methods, prediction and simultaneous confidence intervals.
Time series. Exploratory analysis. Weak and strong stationarity. AR, MA and ARMA processes and their properties. Model identification based on ACF and PACF. Model estimation and diagnostic tests. Box Jenkins models. ARCH and GARCH models.
Non stationary models. ARIMA and SARIMA, determination of the order of integration, trend stationary and difference stationary processes, tests of non stationarity.
Unit 7: Optimization and Statistical Computing
Optimization. Unconstrained optimization using calculus: Taylor’s theorem, convex functions, coercive functions. Unconstrained optimization via iterative methods: Newton’s method, gradient and conjugate gradient methods, quasi Newton methods. Constrained optimization: penalty methods, Lagrange multipliers. Convex sets, convex hull, formulation of a linear programming problem, theorems on vertices of feasible regions and optimality, graphical solution, simplex method.
Statistical computing. Simulation techniques for various probability models. Resampling: jack knife, bootstrap, cross validation. Robust linear regression, nonlinear and generalized linear regression, tree structured regression and classification. Analysis of incomplete data: EM algorithm, single and multiple imputation. Bayesian modelling and estimation. Markov Chain Monte Carlo and annealing techniques, Gibbs sampling, Metropolis Hastings algorithm. Neural networks, association rules and learning algorithms.
This unit is where most candidates underestimate the paper. The optimization half is standard postgraduate material. The computing half, with MCMC, Gibbs sampling and Metropolis Hastings, is a research level topic that many Statistics Master’s programmes touch only briefly.
Unit 8: Data Science, Artificial Intelligence and Machine Learning
Foundations. Introduction to supervised and unsupervised pattern classification. Unsupervised and reinforcement learning. Basics of optimization. Model accuracy measures.
Core algorithms. Linear regression, logistic regression, penalized regression, naive Bayes, nearest neighbour, decision tree, support vector machine. Kernel density estimation and kernel discriminant analysis. Classification under a regression framework. Neural networks. Kernel regression, and tree and random forests.
Clustering. Hierarchical and non hierarchical methods: k means, k medoids, linkage methods. Cluster validation indices: Dunn index, Gap statistics.
Ensemble methods. Bagging with random forest. Boosting with adaptive boosting and gradient boosting.
Deep learning. Recurrent Neural Network. Convolutional Neural Network. Natural Language Processing.
Practical machine learning. Recursive Feature Elimination. Variance Inflation Factor. Ensembling and stacking methods. Elastic Net regularization. Hyperparameter tuning via grid search. Feature importance interpretation. Cross validation strategies.
Unit 9: Database and Data Warehouse Management
Fundamentals. Data structures. Relational Database Management Systems and non traditional NoSQL databases. Principles of database normalization, data redundancy elimination, consistency maintenance.
SQL. Structured Query Language for querying, updating, aggregating and managing relational data. Database joins: inner join, left join, right join, outer join, with applications to data merging and integration.
NoSQL. Document based, key value, wide column stores and graph databases.
Data warehousing. Concepts, star and snowflake schemas, ETL processes, OLAP versus OLTP. Database indexing and optimization. Basics of big data frameworks and storage systems.
Paper III: English
The official syllabus for Paper III is one sentence:
The paper on English shall be framed in a manner to assess the writing skills including expression and understanding of the topic.
That is all the Board specifies. Past papers have typically asked for a precis, an essay and a reading comprehension. The paper is typed on a keyboard and must be answered in English only.
What This Syllabus Tells You About the Job
Read the nine units together and the picture is clear. This is not a classical Statistics examination with a data science section bolted on.
Units 1 to 5 are the classical core, and they are what a Statistics postgraduate already has.
Units 6 and 7 are where research level techniques enter: GARCH, Bayesian estimation, MCMC, simulation.
Units 8 and 9 are a working data scientist’s toolkit: model selection, ensembles, deep learning and the databases you would query to get the data in the first place.
The Board’s own job description confirms it. The role names classical and Bayesian inference, high dimensional regression, time series models, simulation methods, complex survey design, and the management and querying of large structured and unstructured datasets.
Prepare for the job the syllabus describes, not the one the cadre name suggests.
Where Most Candidates Lose
| Unit | Why it is missed |
|---|---|
| 7, computing half | MCMC, Gibbs sampling and Metropolis Hastings are rarely taught in depth at Master’s level |
| 8 | Almost no Statistics postgraduate has studied RNN, CNN or NLP formally |
| 9 | Almost nobody with a Statistics background has written SQL joins or seen a star schema |
| 6, non stationary | ARCH and GARCH sit beyond most syllabi |
Four blind spots, all in the second half of the list. Study the list backwards from Unit 9 and you will find your gaps faster than working forwards from Unit 1.
Frequently Asked Questions
How many units are in the RBI DSIM syllabus?
Nine. Both Paper I and Paper II draw from the same nine units.
Is machine learning in the RBI DSIM syllabus?
Yes. Unit 8 covers supervised and unsupervised learning, logistic and penalized regression, naive Bayes, nearest neighbour, decision trees, support vector machines, neural networks, random forests, boosting, recurrent and convolutional neural networks, and natural language processing.
Is SQL in the RBI DSIM syllabus?
Yes. Unit 9 covers SQL querying, updating and aggregating, all four join types, NoSQL databases, star and snowflake schemas, ETL processes and OLAP versus OLTP.
What is the standard of the RBI DSIM papers?
The Board states that the standard of the papers is that of a postgraduate degree examination of a recognised Indian institution or university.
Do Paper I and Paper II have different syllabi?
No. Both papers draw from the same nine units. Paper I is objective and Paper II is descriptive.
Is Bayesian statistics in the RBI DSIM syllabus?
Yes, in two places. Unit 3 includes the basic idea of Bayes estimators, and Unit 7 includes Bayesian modelling and estimation, Markov Chain Monte Carlo, Gibbs sampling and the Metropolis Hastings algorithm.
What is in the RBI DSIM Paper III English syllabus?
The Board specifies only that the paper assesses writing skills including expression and understanding of the topic. It is typed on a keyboard and must be answered in English.
Related Pages
- RBI Grade B DSIM: Complete Guide
- RBI Grade B Exam Pattern
- Best Books for RBI DSIM
- RBI DSIM Preparation Roadmap
This syllabus is reproduced from RBI Advertisement RBISB/DA/01/2026-27, Appendix III. The official RBI website is the final authority and the syllabus may be revised in a future cycle.