‹ Go back to news
Improved Portuguese and Italian parsing – Extract! 2014.2 release
9 October 2014 • Categories: CV parsing, New release, Textkernel
Textkernel’s R&D team has successfully worked on improving its cv parsing, in particular the Portuguese and Italian language models. Check out the full Extract! 2014.2 release notes.
New: Support for parsing LinkedIn/XING JSON input
LinkedIn and Xing allow downloading the profile of an user via their publicly available API. The information is communicated using generic file formats like XML or JSON. Textkernel Extract! was already able to process such XML profiles. This functionality is also powering our “Apply with LinkedIn” and “Apply with Xing” widgets. We have now extended Extract! to also support JSON profiles. If you have a collection of JSON profiles, Extract! can help you import them into your database. Contact Textkernel for more information.
Want to effectively search your data? Check out our semantic search software.
Better parsing by improved handling of column CVs
Most CVs are written as one column and can be easily read line-by-line from left to right. However, some candidates put extra information (such as contact details or skills) on the left or right of this main column. In these cases, reading line by line is no longer an option as it will mix information from different sections. For people, it is very easy to detect the large column and the narrow left/right sidebar. For machines, this is much more difficult. Textkernel’s R&D team successfully trained the software to better recognise and handle such cases, resulting in an overall better parsing for all languages.
Improved Portuguese CV parsing
Textkernel has realised major improvements in the parsing quality of Portuguese CVs for all sections of the CV (such as personal information, education items and working history). This applies to both CVs from Portugal and Brazil.
Improved Italian CV parsing
For Italian parsing, Textkernel has also made large improvements in quality for all sections of the CV.
Would you like to test our improved multilingual parser with Portuguese, Italian and our other 13 languages?
Request a demo.
Improvements in English parsing of South African addresses
Textkernel’s CV parser first recognises the language and country of a CV, in order to better process the CV. For English CVs from South Africa, improvements have been made in parsing of address and phone details.
Contact Textkernel for more information about its multilingual CV parsing.