Armin Pournaki

That's interesting, intuitively I'd think that you need a certain document length for LDA to reliably capture repeated word co-occurrences as signals for underlying topics... It would be quite useful to do a systematic comparison that doesn't have all the flaws of the above mentioned paper!

09.12.2025 12:21 — 👍 0    🔁 0    💬 1    📌 0

Also, I've never quite seen the value of stemming for topic modeling, but I'd argue that preprocessing in general is an important part of any method. That's why I wouldn't necessarily copy every preprocessing step when comparing methods, especially when the latter are quite different.

09.12.2025 09:01 — 👍 1    🔁 0    💬 1    📌 0

I suspect that a good number of citations stem from the fact that the paper "confirms" a common experience in the CSS community: that embedding-based topic models tend to produce more interpretable topics than LDA _on very short texts_.

09.12.2025 08:56 — 👍 2    🔁 0    💬 3    📌 0

@some4dem.bsky.social researchers @eckolb.bsky.social and @pournaki.bsky.social presenting their work at #ic2s2 on measuring political realignment in Switzerland and extracting conflicting narratives from polarized debates on social media.

22.07.2025 15:33 — 👍 8    🔁 3    💬 0    📌 0

Looking forward to #ic2s2 where I'll present some of our latest work from the @some4dem.bsky.social project:
- Conflicting narratives and polarization (Tue in Pol.Narratives II 2:30pm)
- A political cartography of news sharing (Tue, poster)
- Issue alignment and polarization on Twitter (Thu, poster)

21.07.2025 13:33 — 👍 17    🔁 2    💬 0    📌 0

Thanks for sharing!

14.05.2025 09:39 — 👍 1    🔁 0    💬 0    📌 0