
Course unit
COMPUTATIONAL STATISTICS
SCP4063598, A.A. 2018/19
Information concerning the students who enrolled in A.Y. 2017/18
ECTS: details
Type 
ScientificDisciplinary Sector 
Credits allocated 
Core courses 
SECSS/01 
Statistics 
9.0 
Course unit organization
Period 
Second semester 
Year 
2nd Year 
Teaching method 
frontal 
Type of hours 
Credits 
Teaching hours 
Hours of Individual study 
Shifts 
Laboratory 
3.0 
22 
53.0 
No turn 
Lecture 
6.0 
42 
108.0 
No turn 
Examination board
Board 
From 
To 
Members of the board 
4 Commissione a.a.2018/19 
01/10/2018 
30/09/2019 
GRIGOLETTO
MATTEO
(Presidente)
CANALE
ANTONIO
(Membro Effettivo)
SCARPA
BRUNO
(Membro Effettivo)

Prerequisites:

The following previous courses are required: Mathematics, Statistics I and II, Linear algebra, Probability. 
Target skills and knowledge:

Knowledge of computational tools useful for inferential purposes. Programming abilities that allow the implementation, with the software R, of functions that apply the required algorithms. 
Examination methods:

Written exam in which the student is required to write and comment programs in R, with the objective to solve specific inferential problems. 
Assessment criteria:

Evaluation of the understanding of theoretical and practical computational tools useful for solving inferential problems. 
Course unit contents:

Simulation techniques and statistical applications.
Introduction to simulation: generation from uniform random variables, generation by inversion, generation by acceptancerejection, importance sampling, RaoBlackwell, antithetic variables. Applications: multidimensional integrals, evaluation of efficiency and robustness of inferential methods, hypotheses testing in nonstandard settings.
Inference with Bootstrap. Introduction to Bootstrap, parametric and nonparametric Bootstrap, application examples (quantiles, linear models).
Nonparametric estimation. Density function: the kernel method, the choice of the smoothing parameter, automatic criteria (cross validation, SheatherJones). Regression function: local polynomial regression, splines, equivalent degrees of freedom, AICc and GCV, using the Bootstrap for evaluating precision. Applications to real data.
Numerical exploration of the likelihood function. Introduction to numerical optimization and differentiation algorithms in R. Use of these algorithms for computing maximum likelihood estimators. Confidence regions based on the profile likelihood or on the Fisher information matrix. 
Planned learning activities and teaching methods:

Lectures and laboratories, all based on the software R. Teaching is always interactive, with questions and presentation of case studies that provoke critical discussion. 
Additional notes about suggested reading:

The lectures and laboratories are all available in the Moodle platform. In the same platform past exams, data sets and more teaching materials are also available. 
Textbooks (and optional supplementary readings) 

Innovative teaching methods: Teaching and learning strategies
 Laboratory
 Interactive lecturing
 Problem solving
Innovative teaching methods: Software or applications used
 Moodle (files, quizzes, workshops, ...)
 R

