Citation-analysis with umap & hdbscan

Randall Collins: Sociology of Philosophies

Collins, 1998, p. 96: Network of hellenistic philosophy.

Randall Collins: Sociology of Philosophies

Collins, 1998, p. 294: Network of Ch'an philosophy.

Kieran Healy: Co-Citation in Philosophy

Healy, 2013: Co-Citation Network for Philosophy

Scott Weingart: Looking for HPS

Weingart, 2015: Proof that history and philosophy of science exists as a discipline

A vectorized dataset

Sample of a vectorized dataset.

Linear Dimensionality Reduction: SVD

Basic functionality of PCA (or SVD).

Non-Linear Dimensionality Reduction: Umap

Why to use non linear dimensionality-reduction.

Clustering: hDBSCAN

• Combines clustering intuitions of density based clustering with hierachical cluster-extraction
• Generates a useful clustering tree

Sample: 3 Journals

• Herpetology
• Criminology
• Synthese

Sample: List from Philpapers

• ~1000 Journals
• Resampled from WOS
• Keep only > 3 Citations.

Desiderata (Conceptual)

• Temporal umap
• Complex similarity measures
• Specific Question (Ideas: Continental vs. Analytic, Interdisciplin., Package?)

Desiderata (Material)

• Supervisor
• More data!
• Computation-power

Problems

• Reproduces institutionalized biases
• Only as good as the sample
• Weak interpretability

Literature

• Collins, R. (1998). The sociology of philosophies: a global theory of intellectual change. Cambridge, Mass: Belknap Press of Harvard University Press.
• Healy, Kieran (2013). A Co-Citation Network for Philosophy. https://kieranhealy.org/blog/archives/2013/06/18/a-co-citation-network-for-philosophy/, accessed February 28, 2019.
• McInnes, Leland, John Healy, and James Melville (2018). UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction. ArXiv:1802.03426 [Cs, Stat]. http://arxiv.org/abs/1802.03426, accessed February 27, 2019.
• Weingart, Scott B. (2015).Finding the History and Philosophy of Science. Erkenntnis 80(1): 201–213.