> 0.00 SISUS: Stable Isotope Sourcing using Sampling http://StatAcumen.com/sisus, by Erik Barry Erhardt Starting Tue May 29 21:48:48 2007 Version: v.0.08-010 > 0.01 Reading workbook: SISUS_v0_08-010_Getting_Started.xls reading up to 11 sheets, as requested: 1r 2r 3r 4r 5r 6 7r 8r 9r 10r 11 > 24.08 Process input data and assign variables ### SISUS Model Settings ### analysis.name = SISUS Getting Started filename.prefix = SISUS_GS_v0_08 mixtures.group.name = Mixtures Group Name m = 3 Mixtures n = 4 Sources k = 2 Isotopes Include SIMPLEX (sum of proportions is 1) constraint -- for Mixing Model Correcting source isotopic signatures for discrimination Using SOURCE CONCENTRATIONS NOT using ASSIMILATION EFFICIENCIES Including additional LINEAR CONSTRAINTS Importance resampling using PRIOR weighting of uniform samples, if specified, in addition to uniform samples from solution polytope Including ISOTOPE STDDEV on Convex Hull plots NOT using BIOMASS PER INDIVIDUAL NOT using NUMBER OF INDIVIDUALS ### SISUS Model Assumptions (under development) ### Mixing Model using Simplex Discrimination values are accurately specified. Elemental concentrations are accurately specified. Assimilation efficiencies are the same for all sources and elements. Additional linear constraints are accurately specified. Prior (expert) knowledge properly specified as Dirichlet distribution over the simplex. NOT using BIOMASS PER INDIVIDUAL NOT using NUMBER OF INDIVIDUALS ### SISUS Output Settings ### Mixtures are not a time series > 24.24 Set random seed = 76543 > 24.25 Dirichlet prior on p vector Dirichlet prior on p vector Input priors: 0.2, 0.1, 0.2, 0.5 Input precision: 1 Alphas: 0.8, 0.4, 0.8, 2 Expected: 0.2, 0.1, 0.2, 0.5 Variance: 0.032, 0.018, 0.032, 0.05 StdDev: 0.178885438199983, 0.134164078649987, 0.178885438199983, 0.223606797749979 > 24.26 Dirichlet prior for p is not uniform -- resample > 24.26 Draw 0 samples from MVN isotope distribution > 24.36 Plot Convex Hull ==> process Mixture: Mixture 1 > 33.66 Draw uniform samples from convex solution polytope Drawing 1 sample(s) of size 10000 for a total of 10000 samples: 1 (v) Solution Polytope with 2 vertices > 63.43 1 of 1 = 100 % of mvn samples lead to feasible solutions > 63.43 Resample based on Dirichlet prior for p, with parameters alpha= 0.8 0.4 0.8 2 > 70.00 Back-solve from biomass to isotopes contributions > 90.27 Plots and Numerical Output > 90.28 Uniform Sample > 90.28 Biomass Marginal Histogram > 91.42 Biomass Scatterplot Matrix (long time for large samples) > 135.06 Biomass Numerical Summaries > 135.30 Isotope 1 Marginal Histogram > 136.25 Isotope 1 Scatterplot Matrix (long time for large samples) > 177.23 Isotope 1 Numerical Summaries > 177.33 Isotope 2 Marginal Histogram > 178.19 Isotope 2 Scatterplot Matrix (long time for large samples) > 218.28 Isotope 2 Numerical Summaries > 218.38 Write out samples > 219.04 Dirichlet Resample > 219.05 Biomass Marginal Histogram > 219.76 Biomass Scatterplot Matrix (long time for large samples) > 271.22 Biomass Numerical Summaries > 271.36 Isotope 1 Marginal Histogram > 272.01 Isotope 1 Scatterplot Matrix (long time for large samples) > 314.61 Isotope 1 Numerical Summaries > 314.71 Isotope 2 Marginal Histogram > 315.36 Isotope 2 Scatterplot Matrix (long time for large samples) > 361.46 Isotope 2 Numerical Summaries > 361.56 Write out samples ==> process Mixture: Mixture 2 > 362.25 Draw uniform samples from convex solution polytope Drawing 1 sample(s) of size 10000 for a total of 10000 samples: 1 (v) WARNING: No vertices in solution polytope, Unfeasible Solution -- check Convex Hull plots. Solution Polytope with 0 vertices > 362.27 0 of 1 = 0 % of mvn samples lead to feasible solutions Information: No feasible solutions. Check Convex Hull plots. ==> process Mixture: Mixture 3 > 362.28 Draw uniform samples from convex solution polytope Drawing 1 sample(s) of size 10000 for a total of 10000 samples: 1 (v) Solution Polytope with 2 vertices > 392.95 1 of 1 = 100 % of mvn samples lead to feasible solutions > 392.95 Resample based on Dirichlet prior for p, with parameters alpha= 0.8 0.4 0.8 2 > 400.23 Back-solve from biomass to isotopes contributions > 420.77 Plots and Numerical Output > 420.77 Uniform Sample > 420.77 Biomass Marginal Histogram > 421.50 Biomass Scatterplot Matrix (long time for large samples) > 471.74 Biomass Numerical Summaries > 471.88 Isotope 1 Marginal Histogram > 472.54 Isotope 1 Scatterplot Matrix (long time for large samples) > 512.87 Isotope 1 Numerical Summaries > 512.96 Isotope 2 Marginal Histogram > 513.88 Isotope 2 Scatterplot Matrix (long time for large samples) > 556.70 Isotope 2 Numerical Summaries > 557.01 Write out samples > 557.70 Dirichlet Resample > 557.70 Biomass Marginal Histogram > 558.40 Biomass Scatterplot Matrix (long time for large samples) > 603.40 Biomass Numerical Summaries > 603.54 Isotope 1 Marginal Histogram > 604.17 Isotope 1 Scatterplot Matrix (long time for large samples) > 648.08 Isotope 1 Numerical Summaries > 648.18 Isotope 2 Marginal Histogram > 648.80 Isotope 2 Scatterplot Matrix (long time for large samples) > 689.75 Isotope 2 Numerical Summaries > 689.84 Write out samples > 690.52 SISUS Complete