| !\"#! \"-../0,!\" \"-../0,\"#$%& \"-../0,\"'%( \"-../0,\"$)* \"-../0,$+ TABLEAU 3 -Matrice de confusion pour that dans le Penn Treebank. !\" \"#$%& \"'%( \"$)* $+, 12 3 4 1 3 3 12 1 1 3, 2 11 156 7 3, 3 1 14 28 3, 3 3 3 3 3 Pour ce qui concerne l'annotation de l'\u00e9chantillon du corpus d'apprenants, on obtient une pr\u00e9cision globale de 91 %. Ce chiffre est \u00e0 comparer avec les 96 % obtenus lors de DT TPRON RB Tagged DT 15 4 0 Tagged TPRON 1 2 0 Tagged RB 0 0 0 TABLEAU 4 -Matrice de confusion pour this dans le corpus Diderot-Longdale. that DT TPRON TCOM TREL RB Tagged DT 0 1 0 0 0 Tagged TPRON 0 0 0 0 0 Tagged TCOM 2 9 21 1 3 l'annotation this Tagged TREL 0 1 7 4 0 |