Please enable JavaScript.
Coggle requires JavaScript to display documents.
Feature Understanding and Selection (Data types (Categorical (Boolean…
Feature Understanding and Selection
Descriptive statistics
Features listed in rows under with names under Feature Name
Expand feature details by clicking on feature name
Bracket
Denotes inclusive range
Parenthesis
Exclusive range
Data types
Nature of data inside a feature
Categorical
Example: gender
Binary categorical (2 categories), multi categorical
Boolean
True/false
Numeric :
Most common
Integers and decimals
Text
Data Robot is capable of detecting and tagging currencies
And measurements
And dates
Care about missing values
Data Robot coverts to 1 consistent value
Avoids them being categorized as multiple unique values
Algorithms that struggle with missing values: regression, neutral networks, support vector machines
Data Robot converts some missing values to nulls
Null = values never entered or received
When a cell is not applicable
No response or it may have been coded with a character value that means missing or null