Skip to content

Latest commit

 

History

History

analysis

README.md

Estimation of the likelihood of police notification on NCVS

Estimate the likelihood of police notification on NCVS data via logistic regression with survey weights using

Rscript analysis/ncvs_estimate_weights_logistic.R '0110';
Rscript analysis/ncvs_estimate_weights_logistic.R '0010';

and via the SuperLearner with

Rscript analysis/ncvs_estimate_weights_superlearner.R '0110' 50;
Rscript analysis/ncvs_estimate_weights_superlearner.R '0010' 50;

For this analysis, you will need to have TensorFlow installed. The code to install it is commented out in the R file. The last number on each line indicates the number of cores the process is parallelized onto.

Estimation of the likelihood of police notification on NIBRS

Obtain the predictions of the likelihood of police notification on NIBRS with

Rscript analysis/nibrs_get_weights.R;
Rscript analysis/nibrs_get_weights.R mult;

The code in the second line generates the predictions for the incidents with one or more offenders.

Analysis of NIBRS

Conduct an exploratory data analysis of NIBRS with

Rscript analysis/nibrs_eda;
Rscript analysis/nibrs_eda_mult;

Fit the regression models with

Rscript analysis/nibrs_fit_regression.R;
Rscript analysis/nibrs_fit_regression_multiple.R;

Run the sensitivty analysis and the model diagnostics using

Rscript analysis/sensitivity.R;
Rscript analysis/nibrs_modeldiagnostics;

utils.R and utils_regression.R include functions used throughout the analysis.