TREX supports several other languages for which restrictions currently apply as regards TREX functionality. TREX currently supports the following additional languages with restrictions (May 2006/external software version 3.7.3):
Greek
Hebrew
Polish
Romanian
Thai
Turkish
Hungarian
You can find the most up-to-date information about the languages that TREX supports on SAP Service Marketplace in the
Platform Availability Matrix (PAM). For more information about activating these additional languages, see Configuring Language Recognition Manually.
Certain restrictions apply because the linguistic processing development for these additional languages is still at a relatively early stage. TREX functions such as search, attribute query, query-based classification, and other functions that use text-mining functions rarely or never work at the same level of quality as for fully supported languages.
However, the linguistic text-mining functions sometimes delivery results of less quality that in the case for the fully supported languages. Results of poor quality can occur in the following areas:
Feature extraction
Automatically calculated document and/or class features may be of poor quality.
Example-based classification
When using this classification method the elements may be classified less precisely.
Linguistic search
Incomplete or unexpected grammatical variations of a search term may be returned in the search results list.