Special Topics in Statistics - DEs/UEM

Works developed for the course of Special Topics in Statistics in the second semester of 2017 taught by professor Josmar Mazucheli at DEs/UEM.

Application of Association Rules and Classification Methods

In this work the techniques of association rules and classification methods were applied in two real problems.

1. Determine the relationship between climatological variables and rainfall using data from climatological station at Maringá. The source of the data is Instituo Nacional de Meteorologia (INMET).

2. Prediction of Tennis Matches from the data available by Francesco. The following classification methods were used: Classification Based on Associations, k-Nearest Neighbour (KNN), Linear Discriminant Analysis (LDA) and Naye Bayes.

---

Report

Presentation

R script

---

Permutation Test for Comparison of Multivarite Distributions

This work presents a theoretical overview of permutation tests with focus on comparison of multivariate distributions. Specially, I consider the Hotelling T2 test and the Nearest neighbor test.

---

Report

Presentation

R script

---

Bayesian Inference for the parameters of Marshall-Olkin extended Weibull distribution

In this work the issue of estimation the parameters of Marshall-Olkin extended Weibull distribution is studied under the Bayesian paradigmh. An real application considering interval censoring is presented.

---

Report

Presentation

R and SAS scripts

---

Bootstrap methods applied to Bivariate Exponential distributions

This work discussed the Bootstrap methods under the theoretical framework to estimation the parameters and functions of the parameters of the Gumbel Bivariate Exponential (GBE) and Hougaard Bivariate Exponential (HBE) distribuitions. Simulation examples are presented using R and SAS software.

---

Report

R scripts:

• GBE
• HBE

SAS scripts:

• GBE
• HBE

---