Skip to content
This repository was archived by the owner on Jan 5, 2026. It is now read-only.

Releases: mab-plus/fanti-mechanical-dating-analysis

Figures renamed; README_figures; env equation product; results prose; reproducible exports

15 Sep 00:43

Choose a tag to compare

  • Rename figures to match manuscript captions & order (Fig1..Fig10)
  • Add figures/README_figures.txt (captions, CC BY for Fanti 2015)
  • Fix environment equation (product of exponentials)
  • Results §3.2: prose + uncertainty (1σ & 95% HDI)
  • Scripts export figures with new filenames (reproducible)

Figures & scripts: vector PDFs, VIF thresholds, 14C refs

12 Sep 20:32

Choose a tag to compare

Plotting: add TS (σᵣ = 243.22 MPa) and ¹⁴C 1325±65 AD refs across figures (line, band, point+xerr)

New vector PDF exports for figures (journal-ready)

VIF chart: thresholds at 4 and 10, labels + values on bars

Observed vs predicted: identity line, RMSE/MAE box

Cleanup: remove obsolete fanti_experiment.py; refresh comparison to Fanti (2015) with proper CC-BY/DOI attribution

Repro: python scripts/modeling/modelisation.py rebuilds figures

Plots (VIF/R²), ^14C markers, SI cleanup

11 Sep 15:22

Choose a tag to compare

Summary

This release polishes analysis figures and notation for the paper and repo:

  • clearer VIF diagnostics (threshold lines at 4 and 10 + labels),
  • Age–strength figure now includes ^14C 1325±65 AD vertical band/line and the TS target (horizontal line + point with x-error),
  • tidy R² Monte Carlo and Observed vs Predicted plots,
  • SI and terminology cleanup (T/RH, spaces before units, °C, %).

Added

  • ^14C reference overlay on the Age vs Strength plot: vertical line + ±65 AD band, and TS point with xerr; horizontal TS target line for σ_r = 243.22 MPa.

Changed

  • VIF bar chart: variable names (ln(σ_r), ln(E_f), η_i), dashed thresholds at 4 (investigate) and 10 (serious), numeric labels atop bars.
  • Observed vs Predicted: identity line y=x; axis labels include units (AD); optional RMSE/MAE textbox.
  • R² distribution (Monte Carlo): tidied layout; optional median + IQR overlay.
  • LaTeX/fig captions: isotope as (^{14})C (not “C14”); T/RH for temperature/relative humidity; SI typography (20~°C, 5~\%, 243.22~MPa).

Fixed

  • Minor wording: AIC phrasing → “did not perform significantly better (ΔAIC < 2)”.

How to reproduce

pip install -r requirements.txt
# regenerate main figures
python scripts/modeling/modelisation.py

Plot references: TS target + 14C (line, band, point+xerr)

10 Sep 00:25

Choose a tag to compare

Adds clear visual references to figures:

TS horizontal reference line at σᵣ = 243.22 MPa

^14C vertical reference at 1325 AD with ±65 yr shaded band

^14C point at (1325 AD, 243.22 MPa) with horizontal error bar

Legend refresh & consistent z-order to prevent occlusion

pip install -r requirements.txt
python scripts/modeling/modelisation.py

Fanti Mechanical Dating v2.2.3 — Finalized metadata

21 Jun 20:34

Choose a tag to compare

This release finalizes metadata consistency

Updating README CITATION and scripts/modeling/figure6.png

21 Jun 18:16

Choose a tag to compare

Updating README, CITATION and scripts/modeling/figure6.png

Fixes the layout of the MCMC trace graph

21 Jun 16:19

Choose a tag to compare

Fixes the layout of the MCMC trace graph for better readability.

Version 2.2.0 – Fanti data integration

11 Jun 22:56

Choose a tag to compare

Version 2.2.0: Fanti data integration

Major additions:

  • Integrated Fanti et al. (2015) mechanical data (9 samples)
  • Comparative analysis script showing physics-based vs empirical approaches
  • Documentation updates for reproducibility
  • Strengthens manuscript argument for viscoelastic modeling"

v2.1.1 - English translations

10 Jun 22:38

Choose a tag to compare

Changes in v2.1.1

  • Translated all figure labels from French to English
  • Translated all code comments to English
  • Improved international accessibility
  • No changes to calculations or results

v2.1 — Amélioration validation croisée

28 Mar 18:57

Choose a tag to compare

v2.1 — Clarification des sorties et stabilité accrue

  • Suppression des UndefinedMetricWarning liés à r2_score
  • R² affiché uniquement s'il est défini
  • Message explicite : "non défini" ou "basé sur un seul fold, résultat fragile"
  • Nettoyage de analysis.py et fanti_crossed_analysis.py