Our cli­ent, the Cen­ter for Docu­men­ta­ti­on and Infor­ma­ti­on (DIZ) of Süd­deut­sche Zei­tung and Süd­west­deut­sche Medi­en­hol­ding, is archi­ve ser­vice pro­vi­der and con­tent mar­ke­ter and the­re­fo­re offers access to con­tents of publi­shing publi­ca­ti­ons and other media. In order to faci­li­ta­te the work and to enhan­ce filing effi­ci­en­cy, we used our Clas­si­fi­ca­tor.


Our goal was to part­ly auto­ma­te the editing pro­cess with the opti­on of fol­lo­wing manu­al revi­si­ons – this way edi­to­ri­al work is get­ting much more effi­ci­ent with the help of Ret­res­cos Seman­ti­cs. Seman­tic enrich­ment and clas­si­fi­ca­ti­on tech­no­lo­gies that are working in the backend are being made admi­nis­tra­ble in your front­end with a well-arran­ged front­end.



All con­tent gets seman­ti­cal­ly ana­ly­sed, which means ent­i­ties like com­pa­nies and per­sons are reco­gnis­ed while the most rele­vant are sug­gested in order to clas­si­fy the text. Based on seman­tic enrich­ment, con­tents auto­ma­ti­cal­ly are allo­ca­ted to cer­tain sub­jects. This clas­si­fi­ca­ti­on enab­les a con­sis­tent archi­ving and mul­ti­ple use of con­tent.

Machi­ne Learning

In addi­ti­on machi­ne learning reco­gni­ses aspects that are not expli­cit­ly men­tio­ned wit­hin the text. Examp­les for that are con­text, back­grounds or cha­rac­te­ris­tics of docu­ments.





Matrix Sur­face

The matrix sur­face of our Clas­si­fi­ca­tor offers a direct over­view about which sub­jects have alre­ady been cove­r­ed. The­re­fo­re you not only gain order wit­hin your archi­ve but it also enab­les an easy inven­to­ry and unco­vers poten­ti­al gaps at the first glance.



Ret­res­cos Seman­ti­cs with its sug­ges­ti­on sys­tem pro­vi­des a cru­ci­al part to effi­ci­ent­ly design edi­to­ri­al work. The spe­cial thing about our DIZ-pro­ject is, that it was the first of this kind – howe­ver it did not stay the only one. Fur­t­her media com­pa­nies like Axel Sprin­ger or DuMont are among our custo­mers.