Please enable JavaScript.
Coggle requires JavaScript to display documents.
Chapters 13-15 (Build Candidate Models (Model Selection Process…
Chapters 13-15
Build Candidate Models
DataRobot automatically downsamples for us
Optimizting metrics LogLoss
Starting the Process: Time to select target feature
Accuracy _LogLoss
Model Selection Process
Screening
Startup Process: DataRobot
ODBC (Open database connectivity)
HDFS (Handoop System)
URL
Unique
Relies on information from standard stats
Mean, STdev, median min and max
Only numeric columns
Advanced options for building models
How to run the model?
Partitioning
Cross-validation or Train-validation-holdout
Starting the Analytical Process
Quick
ONLY HAVE
Manual
AutoPilot
Download Data into Datarobot (Local File)
Name the Project and label columns and rows
create tags
Feature Understanding & Selection
Interpret the data contained in each feature
Descriptive Statistics
Index header
Unique
Listed as a Var Type
CLick on any data and more is dislayed
Missing Values
Need to avoid so the ML tool doesnt miss value the null or question mark
Bracket denotes an inclusive range and parenthesis denote an exclusive range