This repository contains the data, code, and supporting files for interpolating functions for longevity outome variables based on a set of feature variables that were inputs to the simulation process using the OG-USA macroeconomic model.
We have included the Jupyter notebook SeparateCSVs.ipynb in the /code directory of this repository. This notebook uses the master data file for the surrogate model from all_sources_together.csv and creates 6 separate .csv files, one for each interpolation model. We have 6 outcome variables of interest, listed down the left column of the following table, which we are predicting from the feature variables in the next 7 columns of the table. Only model 6 that predicts the NPV variable is a function of all 7 feature variables. Model 6 is the only model that includes the discount_rate variable as a feature variable. All the other models (models 1 - 5) are only functions of the first 6 feature variables.
| Model num |
Response variable |
Feature variable age_effect |
Feature variable initial_effect |
Feature variable final_effect |
Feature variable mort_effect |
Feature variable prod_effect |
Feature variable fert_effect |
Feature variable discount_rate |
Observations | CSV filename |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | pop_diffs_2045_2065 |
Yes | Yes | Yes | Yes | Yes | Yes | No | 472 | model1_pop2045.csv |
| 2 | pop_diffs_2025_2100 |
Yes | Yes | Yes | Yes | Yes | Yes | No | 472 | model2_pop2025.csv |
| 3 | pop_diffs_2050 |
Yes | Yes | Yes | Yes | Yes | Yes | No | 472 | model3_pop2050.csv |
| 4 | avg_diff |
Yes | Yes | Yes | Yes | Yes | Yes | No | 472 | model4_avgdiff.csv |
| 5 | avg_gdp_pc_diff |
Yes | Yes | Yes | Yes | Yes | Yes | No | 472 | model5_avggdppc.csv |
| 6 | NPV |
Yes | Yes | Yes | Yes | Yes | Yes | Yes | 2,832 | model6_npv.csv |
In the .csv data files listed in the last column of the table above, which are created in the cells below, each row is an observation and each column is a variable. The first column of the .csv file is always the predicted variable for the model, one of the variables from the first column of the table above. Then the following columns of the .csv file are the feature variables. The first row of the .csv file contains the variable names. All data .csv files in the last column of the table above are stored in the /data/ folder of the GitHub repository for this project at https://github.com/OpenRG/LongevityInterpolant.