| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| en:cnk:syn2010 [2016/12/11 11:34] – old revision restored (2015/10/22 21:01) veronikapojarova | en:cnk:syn2010 [2016/12/11 16:27] (current) – veronikapojarova |
|---|
| |
| SYN2010 is a synchronic representative corpus of written Czech comprising 100 million tokens. It is a sequel to the corpora [[en:cnk:SYN2000]] and [[en:cnk:SYN2005]] and together with them forms a series of synchronic representative corpora that cover three successive periods. | SYN2010 is a synchronic representative corpus of written Czech comprising 100 million tokens. It is a sequel to the corpora [[en:cnk:SYN2000]] and [[en:cnk:SYN2005]] and together with them forms a series of synchronic representative corpora that cover three successive periods. |
| **All corpora contain different texts and are therefore disjoint**. The basic characteristic freatures of the SYN2010 are identical to those of the corpus [[en:SYN2005|SYN2005]], which is predominantly related to the same conception of [[en:pojmy:reprezentativnost|representativeness]] based on the reception of written language and the resulting composition of the corpus. The SYN2010 corpus is [[en:pojmy:lemma|lemmatized]] and [[en:pojmy:tag|morphologically tagged]]. | **All corpora contain different texts and are therefore disjunctive**. The basic characteristic features of the SYN2010 are identical to those of the corpus [[en:cnk:SYN2005|SYN2005]], which is predominantly related to the same conception of [[en:pojmy:reprezentativnost|representativeness]] based on the reception of written language and the resulting composition of the corpus. The SYN2010 corpus is [[en:pojmy:lemma|lemmatized]] and [[en:pojmy:tag|morphologically tagged]]. |
| |
| |
| ====== Changes compared to the SYN2005 corpus ====== | ====== Changes compared to the SYN2005 corpus ====== |
| |
| Compared to the corpus [[en:SYN2005|SYN2005]], the SYN2010 corpus saw **significant improvements in lemmatization** and **[[en:pojmy:tag|morphological tagging]]**; both basically identical to the processing of the [[en:SYN2009PUB|SYN2009PUB]] corpus. Therefore, although [[en:SYN2005|SYN2005]] and SYN2010 do not differ in their understanding of [[en:pojmy:reprezentativnost|representativeness]], **these differences should be taken into account** when comparing their lexical frequencies. | Compared to the corpus [[en:cnk:SYN2005|SYN2005]], the SYN2010 corpus saw **significant improvements in lemmatization** and **[[en:pojmy:tag|morphological tagging]]**; both basically identical to the processing of the [[en:cnk:SYN2009PUB|SYN2009PUB]] corpus. Therefore, although [[en:cnk:SYN2005|SYN2005]] and SYN2010 do not differ in their understanding of [[en:pojmy:reprezentativnost|representativeness]], **these differences should be taken into account** when comparing their lexical frequencies. |
| |
| ====== Composition of SYN2010 ====== | ====== Composition of SYN2010 ====== |
| Some of the fiction texts may have been published earlier, but there is a general rule that the corpus consists mainly of newer texts, whereas the proportion of older texts is decreasing. Compared to the SYN2005 corpus, the lemmatization and morphological tagging of the SYN2010 corpus have been significantly improved; both of them correspond with the processing of the [[en:cnk:SYN2009PUB]]. | Some of the fiction texts may have been published earlier, but there is a general rule that the corpus consists mainly of newer texts, whereas the proportion of older texts is decreasing. Compared to the SYN2005 corpus, the lemmatization and morphological tagging of the SYN2010 corpus have been significantly improved; both of them correspond with the processing of the [[en:cnk:SYN2009PUB]]. |
| |
| <WRAP clear></WRAP> | ===== The general composition of SYN2010 ===== |
| |
| [{{:en:cnk:syn2010-slozeni.gif|Structure of corpus SYN2010: <fc #ffcc00>40 % fiction</fc>, <fc #cc0000>27 % technical literature</fc>, <fc #3333ff>33 % journalism</fc>}}] | [{{:en:cnk:syn2010-slozeni.gif|Structure of corpus SYN2010: <fc #ffcc00>40 % fiction</fc>, <fc #cc0000>27 % technical literature</fc>, <fc #3333ff>33 % journalism</fc>}}] |
