Vorträge & Konferenzen

LitSOM — Automatic Mapping of Russian Authors using SOM

Mas­ter the­sis in Applied Com­put­er Sci­ence at U of Salzburg


The Lit­SOM pro­posed in this Mas­ter the­sis applies Self-Orga­niz­ing Maps (SOM) and Learn­ing Vec­tor Quan­ti­za­tion (LVQ) to Russ­ian lit­er­a­ture. The SOM is used to cre­ate a map of con­tem­po­rary Russ­ian nov­els, which enables lit­er­ary schol­ars to study inter­tex­tu­al rela­tion­ships. Thus, Lit­SOM can be regard­ed as an exam­ple for the emerg­ing sci­en­tif­ic dis­ci­pline of dig­i­tal human­i­ties. The qual­i­ty of the result­ing maps of Russ­ian lit­er­a­ture is assessed sub­jec­tive­ly, i.e., from a philol­o­gist’s point of view, and objec­tive­ly, i.e., in a “clas­sic” prob­lem of text min­ing, name­ly, author clas­si­fi­ca­tion. For that pur­pose the SOM can be eas­i­ly con­vert­ed to imple­ment LVQ.

Down­load Lit­SOM pre­sen­ta­tion.