Regression model to predict runtime for point-buffer ejamit analyses

Weighted runtime model for point-buffer ejamit analyses, fit from Analysis_timing_results*.csv files with extra emphasis on small point-count runs.

Usage

modelEjamit

Format

An object of class lm of length 13.

Details

Regression model to predict runtime for point-buffer ejamit analyses

The model is trained from point-buffer rows in all Analysis_timing_results*.csv files in data-raw/. Small runs such as 1, 2, and 10 points are up-weighted so predictions are more accurate for small analyses. ejamit runtime is modeled from input_number and radius using weighted least squares.