Transforming a Business Model Based on Data Matching Solutions (Round 6)
As there is currently no off-the-shelf NLP solution is available for the Hungarian language, the common practice is to use manual tagging for content rather than automated one. This creates subjective tags, tag-overlaps, non-relevant tag groups and mainly word-based tags, rather than relevant topics/themes organised in a structured manner. Moreover, 24.hu is constrained by cookie-data for personalisation as no sign-on solution is deployed. Machine learning-based automatic tagging will improve accuracy, recommendation relevance and archive utilisation. Single sign-on on the 24.hu satellite network can add personalisation skills and unlock better targeting capabilities.