Docstoc

eda

Document Sample
eda Powered By Docstoc
					Principios de
 Estad´ıstica



Graphical
Tools
                                            ıstica
                         Principios de Estad´

                                              ıa    e
                Leonardo Collado Torres y Mar´ Guti´rrez Arcelus
                                               o
                   Licenciatura en Ciencias Gen´micas, UNAM
                   www.lcg.unam.mx/~lcollado/index.php
                   www.lcg.unam.mx/~mgutierr/index.php



                                             e
                              Cuernavaca, M´xico
                             Febrero - Junio, 2009


                                                                   1/6
                Exploratory Data Analysis with R

Principios de
 Estad´ıstica



Graphical
Tools




                 1 Graphical Tools




                                                   2/6
                Diaps. de Jim

Principios de
 Estad´ıstica



Graphical
Tools




                    Estas diapositivas corresponden a las 28, 29 y 30 de la
                              o
                    presentaci´n EDA de Jim.




                                                                              3/6
                boxplots

Principios de
 Estad´ıstica       boxplot : A plotting method for generating Tukey’s
                    boxplots.
Graphical           Excellent for comparing location shifts of k distributions of
Tools
                    varying size.
                    Assess skewness and spread of either of one or more
                    distributions.
                    Boxplots are often a much better summary of a
                    distribution than are histograms as they do not suffer from
                    either bandwidth choice or the need to have large data
                    sets.
                Example
                What does skewness look like on a boxplot, spread? can we
                generate some data to exemplify these things? (hint:
                remember all of the random number generators which we
                talked about in the first lecture)                               4/6
                Anatomy of a boxplot

Principios de
 Estad´ıstica



Graphical
Tools


                    A, B : lower/upper adjacent values:

                                 r      |q75 − q25 |             (1)
                                A = inf{xi : xi > q25 − 1.5r }   (2)
                                B = sup{xi : xi < q75 + 1.5r }   (3)




                                                                   5/6
                Anatomy of a boxplot

Principios de
 Estad´ıstica                             The anatomy of a Boxplot

                                outlier              q
                   3
Graphical
Tools                           outlier              q

                         B
                   2
                   1




                        q0.75
                   0




                        q0.5
                        q0.25
                   −1
                   −2




                         A

                                outlier              q




                                                                     6/6

				
DOCUMENT INFO