Hierarchical Clustering with Contiguity Constraint in <i>R</i>
Bibliographic record
Abstract
This article presents a new implementation of hierarchical clustering for the R language that allows one to apply spatial or temporal contiguity constraints during the clustering process. The need for contiguity constraint arises, for instance, when one wants to partition a map into different domains of similar physical conditions, identify discontinuities in time series, group regional administrative units with respect to their performance, and so on. To increase computation efficiency, we programmed the core functions in plain C. The result is a new R function, constr.hclust, which is distributed in package adespatial. The program implements the general agglomerative hierarchical clustering algorithm described by Lance and Williams (1966; 1967), with the particularity of allowing only clusters that are contiguous in geographic space or along time to fuse at any given step. Contiguity can be defined with respect to space or time. Information about spatial contiguity is provided by a connection network among sites, with edges describing the links between connected sites. Clustering with a temporal contiguity constraint is also known as chronological clustering. Information on temporal contiguity can be implicitly provided as the rank positions of observations in the time series. The implementation was mirrored on that found in the hierarchical clustering function hclust of the standard R package stats (R Core Team 2022). We transcribed that function from Fortran to C and added the functionality to apply constraints when running the function. The implementation is efficient. It is limited mainly by input/output access as massive amounts of memory are potentially needed to store copies of the dissimilarity matrix and update its elements when analyzing large problems. We provided R computer code for plotting results for numbers of clusters.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.011 | 0.054 |
| Meta-epidemiology (narrow) | 0.004 | 0.003 |
| Meta-epidemiology (broad) | 0.003 | 0.005 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.005 | 0.003 |
| Open science | 0.005 | 0.005 |
| Research integrity | 0.002 | 0.005 |
| Insufficient payload (model declined to judge) | 0.044 | 0.038 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".