Please enable JavaScript.
Coggle requires JavaScript to display documents.
Feature engineering, Steps in feature engineering - Coggle Diagram
Feature engineering
variable characteristics
missing data
Missing data imputation
Mean median mode
Random & arbitrary
missing indicators
MICE
Regression
cardinality
Category frequency
variable distribution
Outliers
feature magnitude
Variable type
continuous
discreet
categorical
Nominal
ordinal
dates
mixed variable
Steps in feature engineering
Missing data
categorical variable imputation
Variable tranformation
Handling outliers
Scaling the features
Method of feature selection
Wrapper method
step forward
Step backward
Exhaustive search
filter method
constant
quazi - constant
Duplicated
statistical measure
Fischer score
univariate method
Mutual information
correlation methods
embedded method
LASSO
Decision Tree derived
Regression coefecient
Recursive feature elimination
Steps in feature engineering
Creating features from dates and time
Extracting features from transactions and time series
Extracting features from text