Post-Processing Methods
gdtm.models.cstb module
Cross-Source Topic Blending (CSTB)
- gdtm.models.cstb.cstb(topic_sets, source_thresholds=(3,), word_threshold=3, topn=5, words_per_topic=10)
Cross-Source Topic Blending: Blends topics by finding them in other data sources.
- Parameters
topic_sets – list of topic sets, one for each data source.
source_thresholds – list, required: minimum number of sources that a topic must appear in to be integrated (ell in paper)
word_threshold – int, required: minimum number of words overlapping to match a topic between sources (chi in paper)
topn – int, optional: number of words per topic to consider for blending (psi in paper)
words_per_topic – int, optional: number of words per topic to put into final topic