Chp3 Exploratory Data Analysis

数据挖掘导论 笔记

1 Summary Statistics

frequency mode percentile mean(trimmed mean) median
range variance standard-deviation
absolute average deviation，$AAD=\frac{1}{m}\sum(|x_i-\bar x|)$
median absolute deviation, $MAD=median(|x_1-\bar x|,...,|x_m-\bar x|)$
interquartile range, $IQR=x_{75\%} -x_{25\%}$

mutivariable：
covariance matrix
correlation matrix

skewness

2 visualization

1.mapping the data to graph
2.arrangement
3.selection

technique
1. with a few variables
stem and leaf plot
histogram
box plot
pie chart
scatter plot
2. with temporary and spacial data
contor plot
surface plot
vector plot
slicing
animation
3. with high demensions
matrix
parallel coordinates
icons

ACCENT Principle

3 OLAP

data cube
pivoting
slicing
dicing
roll up
drill down