Skip to content
GitLab
Select Git revision
  • master default
You can move around the graph by using the arrow keys.
Created with Raphaël 2.2.024Oct22211816420Sep1918111095431Jul2925171615121Jun19181712111097654329May2827232221201615141310976328Mar252221201918118754128Feb26252221201913121124Jan2214319Dec1413121065430Nov2928272623211514139765230Oct29[EUROCODE DATA] fix eurocode level for bottom quantile. improve label for the legend.[EUROCODE DATA] improve display for the eurocode return level graphs legend, and title[EUROCODE DATA] improve display for the eurocode return level graphs. rename linear margin temporal models.[TEMPORAL LINEAR MARGIN][BAYESIAN ESTIMATION] modify fevd_fixed prior to ensure that prior is always applied on the shape parameter. add multiprocessing for eurocode return level uncertainty. add option to specify a last year for eurocode uncertainty computation.[TEMPORAL LINEAR MARGIN][BAYESIAN ESTIMATION] add test for non stationary scale and location[TEMPORAL LINEAR MARGIN][BAYESIAN ESTIMATION] add fevdPriorCustom as prior for the fit. update test accordingly. add non stationary version of fevD both in R and also in python. add test for that in python. modify upper and bottom quantile for eurocode[EUROCODE][DRAWING PER REGION] modify test region eurocode. add test_ gev temporal bayesian. add stationary version for the Bayesian. refactor utils at the root, to root_uils to avoid weird import issues.[EXPERIMENT][EUROCODE] refactor eurocode data.[EXTREME ESTIMATOR] refactor margin_fits folder. extreme estimator folder. and extreme_models folder[EXTREME ESTIMATOR] refactor result from model fit. refactor estimator.[EXTREME ESTIMATOR][ESTIMATOR] refactor estimator[EXTREME ESTIMATOR][MARGIN MODEL] refactor linear margin model folder organization. refactor names in univariate trend analysis[EXTREME ESTIMATOR][GEV MARGIN FIT][BAYESIAN] add fevd.R to use fedv package in the Bayesian mode. Also, fix fevd method, by creating my own method "myfevd" that avoid code crash when using mode use.phi=FALSErefactor a bit. add fevd files (work in progress) to use the GMLE method in a Bayesian mode.[TREND ANALYSIS] add verbose argument when creating hypercube visualizer. add shape analysis from the MLE estimator, and the shapes are indeed between -0.4 and 0.4[TREND ANALYSIS] add test_mean, and test_variance for the GevParams class[TREND ANALYSIS] add boolean that accept scale 0 parameter that enables to compute easily some mean difference where scale1 is zero[TREND ANALYSIS] few refactoring of the code[TREND ANALYSIS] add test for the gev params[TREND ANALYSIS] add likely ratio test between non-stationary models. To produce maps to show that the model with both linearities is significantly better than the simpler ones.[POSTER EVAN] improve visualization of the starting years for the poster. some refactorign for the Alps graph and cmap, from now on, we will only remove the last 20 starting years[POSTER EVAN] add main_C for the orientation visualization of the study[POSTER EVAN] add orientation dataset. handle allslopes file for both Crocus and Safran[POSTER EVAN] add poster part D for the EVAN poster[POSTER EVAN] add maxima visualization for the EAN[POSTER EVAN] fix hatch and fix strength value. add mainB for the poster[POSTER EVAN] add hatch to represent the participation of the mean & variance variation[POSTER EVAN] tight layout for the first part of the poster[DATA EXTRACTION] add all_days & all_daily_series property, that was useful to extract some data of snow depth for taline[EUROCODE DATA] add display function for eurocode data[PAPER1][STARTING YEAR] write main for location & scale with common starting year spatially & altitude[PAPER1][STARTING YEAR] write main for location & scale with common starting year spatially.[NON STATIONARY] fix constant gev params. also some refactoring for the regions, the colors, and functions.[EUROCODE DATA] add eurocode regions, and mapping between massifs and departement, and mapping between departement and eurocode region.[GEV TREND TEST] add gev trend test two parameters (for Location and scale parameters). IMPORTANT: Now the definition of test sign correspond to the sign of the strength (i.e. the increase with respect to some quantile)[GEV TREND TEST] rename and refactor the code for one parameter gev trend test (to prepare the introduction of tests with two parameters)[STRENGTH EVOLUTION] improve display of the text in strength display[STRENGTH EVOLUTION] fix strength computation for the scale test.[HYPERCUBE] add seed in fast hypercube mode. some refactoring.[PAPER1] modify the strength criterion for the hypercube (use the 0.98 quantile and its evolution in 10 years)