File size: 1,256 Bytes
17391b0 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 | {
"n_images": 8125,
"accuracy_a": 0.9911384615384615,
"accuracy_b": 0.9907692307692307,
"mcnemar": {
"both_correct": 8004,
"only_a_correct": 49,
"only_b_correct": 46,
"both_wrong": 26,
"statistic": 0.042105263157894736,
"p_value": 0.8374192519835879,
"method": "chi-squared (continuity corrected)"
},
"bootstrap": {
"observed_difference": -0.00036923076923078835,
"ci_low": -0.0027076923076923,
"ci_high": 0.001969230769230834,
"n_resamples": 10000,
"confidence": 0.95,
"p_value": 0.8012
},
"per_class": {
"mean_difference": 0.0037079681416117925,
"t_statistic": 1.23458202602619,
"p_value": 0.22476960764208556,
"df": 37,
"ci_low": -0.0023775384859782067,
"ci_high": 0.009793474769201792,
"cohens_d": 0.2002756507644646,
"effect_size_label": "small"
},
"power": {
"effect_size": 0.2002756507644646,
"n": 38,
"alpha": 0.05,
"power": 0.22522047555120145,
"n_for_80_percent_power": 198
},
"notes": [
"NOTE: the per-class test is underpowered \u2014 a null result here is inconclusive, not evidence of equivalence."
],
"baseline_file": "artifacts/preds_baseline.npz",
"challenger_file": "artifacts/preds_challenger.npz"
} |