Learned listening groups span country boundaries in a LastFM sample
Source check: Source checked. Not independent scientific confirmation.
Finding: A fitted spectral-plus-k-means partition of 511 LastFM users identified groups spanning seven countries. Posthoc comparisons associated the learned groups with different listening breadth, activity and artist profiles. These are descriptive patterns in a selected historical sample, not universal listener types.
Technology: Sequential BiFLICKER spectral estimation; fitted k-means unsupervised clustering
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- K-means fits latent assignments in a federated spectral embedding, enabling group-specific listening comparisons.
- Test and result
- Seven learned groups from binary user–artist incidence; posthoc event-count summaries and leave-one-user-out intervals characterized the same sample.
- Limitation / still unknown
- Filtering removed low-degree users and very popular artists. Group number and genre interpretations are analysis-dependent. Breadth/artist profiles reuse clustering inputs; jackknife intervals and centralized agreement are not independent cohort validation. No causal country or preference claim.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/html/2610.25997v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Actual primary Methods3, LastFM Results6 and implementation/posthoc summariesS1; exact current v1 history..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 1/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.