Go Back

Correlation Results

Current Conditions

  • Human Assessment Type: Adequacy, 5-point scale
  • Target Language: French
  • Correlation Level: system
  • Track: Single Reference Track (remove)

Subdivisions

By source language: By data genre:

Ranking

RankMetric NameSpearman's RhoKendall's TauPearson's RGraphs
Value95% confidence intervalValue95% confidence intervalValue95% confidence interval
1NIST (case-sensitive)0.7393(0.3654, 0.9078)0.6190(0.1564, 0.8589)0.8456(0.5880, 0.9474)graph_scatterplot
2TERp-0.7775(-0.9224, -0.4408)-0.6124(-0.8561, -0.1460)-0.8572(-0.9516, -0.6150)graph_scatterplot
3MT-mNCD0.7107(0.3121, 0.8966)0.5619(0.0697, 0.8341)0.9031(0.7272, 0.9677)graph_scatterplot
4Bkars0.7286(0.3451, 0.9036)0.5810(0.0978, 0.8425)0.9179(0.7658, 0.9728)graph_scatterplot
5i_letter_recall0.6786(0.2549, 0.8837)0.5048(-0.0101, 0.8081)0.7462(0.3786, 0.9104)graph_scatterplot
6SVM_rank0.7429(0.3722, 0.9092)0.5810(0.0978, 0.8425)0.7891(0.4647, 0.9268)graph_scatterplot
7MT-NCD0.7286(0.3451, 0.9036)0.5810(0.0978, 0.8425)0.9043(0.7304, 0.9681)graph_scatterplot
8badger_2.0_lite0.7750(0.4357, 0.9214)0.6190(0.1564, 0.8589)0.8933(0.7024, 0.9643)graph_scatterplot
9TESLA-M0.7357(0.3586, 0.9064)0.5429(0.0424, 0.8256)0.7715(0.4287, 0.9201)graph_scatterplot
10TESLA0.6286(0.1715, 0.8630)0.5048(-0.0101, 0.8081)0.8097(0.5081, 0.9344)graph_scatterplot
11i_letter_BLEU0.7429(0.3722, 0.9092)0.5810(0.0978, 0.8425)0.8103(0.5096, 0.9346)graph_scatterplot
12meteor-next-rank0.7250(0.3384, 0.9022)0.5619(0.0697, 0.8341)0.8587(0.6186, 0.9522)graph_scatterplot
13Stanford0.2179(-0.3314, 0.6568)0.1238(-0.4148, 0.5981)0.4063(-0.1339, 0.7603)graph_scatterplot
14badger_2.0_full0.7750(0.4357, 0.9214)0.6190(0.1564, 0.8589)0.8989(0.7167, 0.9663)graph_scatterplot
15ATEC_2.10.7786(0.4430, 0.9228)0.6190(0.1564, 0.8589)0.9040(0.7297, 0.9680)graph_scatterplot
16BLEU-4-mteval-v13a (case-sensitive)0.8321(0.5575, 0.9426)0.6762(0.2508, 0.8827)0.7995(0.4865, 0.9306)graph_scatterplot

16 metrics (including 2 baseline metrics)
15 data points (total number of systems used)