Replication data and code for "Local Linear Estimation of Functional Regression and its Derivative for Functional Time Series under Long-Range Dependence"

Replication code and data accompanying the paper "Local Linear Estimation of Functional Regression and its Derivative for Functional Time Series under Long-Range Dependence" by Abdelhak Chouaf (Laboratory of Statistics and Stochastic Processes, Djillali Liabès University, Sidi Bel Abbès, Algeria). CONTENTS rv_dataset.csv: Daily realized variance (5-minute intraday returns) for 8 stock market indices (.SPX, .GDAXI, .FCHI, .FTSE, .OMXSPI, .N225, .KS11, .HSI), 2615 trading days. Originally distributed by the Oxford-Man Institute Realized Library (now discontinued); retrieved from the public replication archive of Son et al. (2023, Journal of Forecasting, 42(7), 1539-1559). simulation.py: Monte Carlo simulation (Section 5 of the paper) implementing the functional local linear estimator under Hermite-rank long-range dependent errors, reproducing the functional double bandwidth dichotomy (Table 1, Figure 3). make_figures.py: Generates Figures 3 and 4 (skewness/kurtosis vs. bandwidth; histograms) from the simulation output. real_data.py: Prepares the S&P 500 functional time series (Section 6): computes the GPH long-memory diagnostic and builds the 20-day realized-volatility curves. real_fit.py: Fits the functional local linear and local constant estimators to the real data via leave-one-curve-out cross-validation, reproducing Table 2. make_real_figures.py: Generates Figure 5 (example volatility curves and full daily series). REQUIREMENTS Python 3.10+, numpy, scipy, pandas, matplotlib. USAGE python simulation.py        (produces summary.json, res_m.npy, res_b1.npy)python make_figures.py      (produces fig_dichotomy_moments.pdf, fig_dichotomy_hist.pdf)python real_data.py         (produces real_data_curves.npz; requires rv_dataset.csv)python real_fit.py          (produces real_fit_results.json; requires real_data_curves.npz)python make_real_figures.py (produces fig_real_curves.pdf) LICENSE Code: MIT License. Data: see original source (Son et al., 2023) for terms of reuse.