A typical LKT workflow begins by loading a dataset, standardizing key columns, creating a response variable, and organizing observations by student and time. Cross-validation folds may also be created during preprocessing so model evaluation can be handled consistently later.
Step 1
Initial loading of data
Embedded CRAN section
Step 2
computeSpacingPredictors
computeSpacingPredictors is used to derive time-based features from learner interaction histories. In general terms, the function works from timestamps and knowledge components to calculate intervals between interactions, prior spacing, and related time-sensitive features that can then be used in the model.
- Inputs: interaction data with timestamps and knowledge-component structure
- Process: identify event timing, calculate elapsed intervals, and derive spacing-oriented predictors
- Purpose: make temporal learning history available to the model in a usable form