Please enable JavaScript.
Coggle requires JavaScript to display documents.
Feature Understanding and Selection (feature understanding: interpret data…
Feature Understanding and Selection
feature understanding: interpret data contained in each feature
descriptive statistics
data types
feature content
missing data
descriptive statistics
index number specifies which feature is being discussed
can expand feature to show more details
hover over blue bars to see how many values are missing
unique notes how many unique values exist for each specific feature
each row/feature can have different number of unique values
split values into different bins represented by bars
brackets vs parenthesis
brackets denote an inclusive range - range includes number next to a bracket
parenthesis denotes an exclusive range - range does not include the number next to parenthesis
data types
categorical data - minary or multi=class
numeric - any type of number including integers and decimals (some categorical can be coded as numeric)
boolean - true or false
text type -
algorithms struggling with missing values - regression, neural networks, support vector machines
nulls - values never entered or retrieved