The intra-corpus evaluation scores obtained by retraining tmChem (see "Cross-text-genre to cross-corpus evaluation" section) show that the precision (F-measure) values on abstracts are at least 12% (6%) higher than those on full texts.
In this paper, we take this idea one step further and perform a cross-text-genre evaluation by assessing the performance of chemical NER tools trained on scientific articles a problem much better researched on patent corpora.
In this paper, we performed a cross-text-genre evaluation by measuring the tagging quality of the two NER baselines trained on the abstract of scientific articles when evaluated on patent corpora.
We also contrast our cross-text-genre results to those obtained after retraining a chemical NER tool on patent corpora, showing that taking away the text-genre difference significantly boosts results, i.e., that the different characteristics of patent versus scientific texts strongly impact chemical NER performance.
The performance values attained in cross-text-genre evaluations show that the F-measure values of the models trained on the abstracts of scientific articles decrease by around 10% when tested on patent abstracts and by nearly 18% when applied to patent full texts.
Experiments were performed on cross-lingual text classification and cross-domain digit image recognition tasks.
Draw or put stamps on the screen, highlight or cross out text, and so on.
Experiments were performed for cross-domain object recognition and cross-lingual text categorization.
Prettenhofer and Stein [78] proposed a new HTL method called Cross-Language Structural Correspondence Learning (CL-SCL) for cross-language text classification that is built upon structural correspondence learning (SCL) [79].
The new price controls would limit the retail charge for sending a cross-border text message while traveling in Europe to 11 cents.
