Replication material for the paper "Efficient Sampling for Realized Variance Estimation in Time-Changed Diffusion Models" (https://arxiv.org/abs/2212.11833)
by Timo Dimitriadis, Roxana Halbleib, Jeannine Polivka, Jasper Rennspies, Sina Streicher and Axel Friedrich Wolter
-
Replication material for the simulations of the article are contained in the folders 'SCS-FUCHS/R_sim' and 'simulations'. To run these, first run the 'run_FUCHS_sim_script.sh' file (that runs the files within the sub-folder 'R_sim') on a cluster. The file 'simulations/sim_BiasApprox_RR2.R' produces the additional bias approximations of Figure 7 and the file 'simulations/sim_eval_RR2.R' produces the evaluation plots shown in the paper.
-
Replication material for the applications of the article is given in the folders 'SCS-FUCHS/R_appl' and 'application'. For the application, the necessary data files are not included in the repository as they cannot be made publicly available and require 28GB of storage space. Based on the data, the file 'prepare_ticks.R' prepares the data for processing and the loop in the file 'SCS-FUCHS/run_FUCHS_resample_script.sh' resamples the data (on a cluster) according to the different sampling schemes. The file 'application/main_application_RR2_RVest.sh' runs the RV estimator for each stock, and the file 'application/application_RVPAVG_est_RR3.R' computes the pre-averaging RV estimator for each stock. The files 'application/application_eval_RR2.R' (Fig. 9 and 10), 'application/application_eval_RR3_SqReturn.R' (Fig. G.6 and G.7) and 'application/application_eval_RVPAVG_RR3.R' (Fig. 11) produce the tables and plots for the estimation application of the paper. The results for the forecasting application are generated by calling the file 'application/application_HAR_forecasting_RR3.R'.