|
| 1 | +# Test the vignette |
| 2 | + |
| 3 | + Code |
| 4 | + print(looss_1) |
| 5 | + Output |
| 6 | + |
| 7 | + Computed from 4000 by 100 subsampled log-likelihood |
| 8 | + values from 3020 total observations. |
| 9 | + |
| 10 | + Estimate SE subsampling SE |
| 11 | + elpd_loo -1968.5 15.6 0.3 |
| 12 | + p_loo 3.1 0.1 0.4 |
| 13 | + looic 3936.9 31.2 0.6 |
| 14 | + ------ |
| 15 | + MCSE of elpd_loo is 0.0. |
| 16 | + MCSE and ESS estimates assume independent draws (r_eff=1). |
| 17 | + |
| 18 | + All Pareto k estimates are good (k < 0.7). |
| 19 | + See help('pareto-k-diagnostic') for details. |
| 20 | + |
| 21 | +--- |
| 22 | + |
| 23 | + Code |
| 24 | + print(looss_1b) |
| 25 | + Output |
| 26 | + |
| 27 | + Computed from 4000 by 200 subsampled log-likelihood |
| 28 | + values from 3020 total observations. |
| 29 | + |
| 30 | + Estimate SE subsampling SE |
| 31 | + elpd_loo -1968.3 15.6 0.2 |
| 32 | + p_loo 3.2 0.1 0.4 |
| 33 | + looic 3936.7 31.2 0.5 |
| 34 | + ------ |
| 35 | + MCSE of elpd_loo is 0.0. |
| 36 | + MCSE and ESS estimates assume independent draws (r_eff=1). |
| 37 | + |
| 38 | + All Pareto k estimates are good (k < 0.7). |
| 39 | + See help('pareto-k-diagnostic') for details. |
| 40 | + |
| 41 | +--- |
| 42 | + |
| 43 | + Code |
| 44 | + print(aploo_1) |
| 45 | + Output |
| 46 | + |
| 47 | + Computed from 2000 by 3020 log-likelihood matrix. |
| 48 | + |
| 49 | + Estimate SE |
| 50 | + elpd_loo -1968.4 15.6 |
| 51 | + p_loo 3.2 0.2 |
| 52 | + looic 3936.8 31.2 |
| 53 | + ------ |
| 54 | + Posterior approximation correction used. |
| 55 | + MCSE of elpd_loo is 0.0. |
| 56 | + MCSE and ESS estimates assume independent draws (r_eff=1). |
| 57 | + |
| 58 | + All Pareto k estimates are good (k < 0.7). |
| 59 | + See help('pareto-k-diagnostic') for details. |
| 60 | + |
| 61 | +--- |
| 62 | + |
| 63 | + Code |
| 64 | + print(looapss_1) |
| 65 | + Output |
| 66 | + |
| 67 | + Computed from 2000 by 100 subsampled log-likelihood |
| 68 | + values from 3020 total observations. |
| 69 | + |
| 70 | + Estimate SE subsampling SE |
| 71 | + elpd_loo -1968.2 15.6 0.4 |
| 72 | + p_loo 2.9 0.1 0.5 |
| 73 | + looic 3936.4 31.1 0.8 |
| 74 | + ------ |
| 75 | + Posterior approximation correction used. |
| 76 | + MCSE of elpd_loo is 0.0. |
| 77 | + MCSE and ESS estimates assume independent draws (r_eff=1). |
| 78 | + |
| 79 | + All Pareto k estimates are good (k < 0.7). |
| 80 | + See help('pareto-k-diagnostic') for details. |
| 81 | + |
| 82 | +--- |
| 83 | + |
| 84 | + Code |
| 85 | + print(looss_2) |
| 86 | + Output |
| 87 | + |
| 88 | + Computed from 4000 by 100 subsampled log-likelihood |
| 89 | + values from 3020 total observations. |
| 90 | + |
| 91 | + Estimate SE subsampling SE |
| 92 | + elpd_loo -1952.0 16.2 0.2 |
| 93 | + p_loo 2.6 0.1 0.3 |
| 94 | + looic 3903.9 32.4 0.4 |
| 95 | + ------ |
| 96 | + MCSE of elpd_loo is 0.0. |
| 97 | + MCSE and ESS estimates assume independent draws (r_eff=1). |
| 98 | + |
| 99 | + All Pareto k estimates are good (k < 0.7). |
| 100 | + See help('pareto-k-diagnostic') for details. |
| 101 | + |
| 102 | +--- |
| 103 | + |
| 104 | + Code |
| 105 | + print(comp) |
| 106 | + Output |
| 107 | + elpd_diff se_diff subsampling_se_diff |
| 108 | + model2 0.0 0.0 0.0 |
| 109 | + model1 16.5 22.5 0.4 |
| 110 | + |
| 111 | +--- |
| 112 | + |
| 113 | + Code |
| 114 | + print(comp) |
| 115 | + Output |
| 116 | + elpd_diff se_diff subsampling_se_diff |
| 117 | + model2 0.0 0.0 0.0 |
| 118 | + model1 16.1 4.4 0.1 |
| 119 | + |
| 120 | +--- |
| 121 | + |
| 122 | + Code |
| 123 | + print(comp2) |
| 124 | + Output |
| 125 | + elpd_diff se_diff subsampling_se_diff |
| 126 | + model2 0.0 0.0 0.0 |
| 127 | + model1 16.3 4.4 0.1 |
| 128 | + |
| 129 | +--- |
| 130 | + |
| 131 | + Code |
| 132 | + print(comp3) |
| 133 | + Output |
| 134 | + elpd_diff se_diff subsampling_se_diff |
| 135 | + model2 0.0 0.0 0.0 |
| 136 | + model1 16.5 4.4 0.3 |
| 137 | + |
0 commit comments