Machine-learning Recommender Systems Can Inform Collection Development Decisions
Bibliographic record
Abstract
A Review of: Xiao, J., & Gao, W. (2020). Connecting the dots: reader ratings, bibliographic data, and machine-learning algorithms for monograph selection. The Serials Librarian, 78(1-4), 117-122. https://doi.org/10.1080/0361526X.2020.1707599 Objective – To illustrate how machine-learning book recommender systems can help librarians make collection development decisions. Design – Data analysis of publicly available book sales rankings and reader ratings. Setting – The internet. Subjects – 192 New York Times hardcover fiction best seller titles from 2018, and 1,367 Goodreads ratings posted in 2018. Methods – Data were collected using Application Programming Interfaces. The researchers retrieved weekly hardcover fiction best seller rankings published by the New York Times in 2018 in CSV file format. All 52 files, each containing bibliographic data for 15 hardcover fiction titles, were combined and duplicate titles removed, resulting in 192 unique best seller titles. The researchers retrieved reader ratings of the 192 best seller titles from Goodreads. The ratings were limited to those posted in 2018 by the top Goodreads reviewers. A Bayes estimator produced a list of the top ten highest rated New York Times best sellers. The researchers built the recommender system using Python and employed several content-based and collaborative filtering recommender techniques (e.g., cosine similarity, term frequency-inverse document frequency, and matrix factorization algorithms) to identify novels similar to the highest rated best sellers. Main Results – Each recommender technique generated a different list of novels. Conclusion – The main finding from this study is that recommender systems can simplify collection development for librarians and facilitate greater access to relevant library materials for users. Academic libraries can use the same recommender techniques employed in the study to identify titles similar to highly circulated monographs or frequently requested interlibrary loans. There are several limitations to using recommender systems in libraries, including privacy concerns when analyzing user behaviour data and potential biases in machine-learning algorithms.
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 | 0.063 | 0.271 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.013 | 0.013 |
| Science and technology studies | 0.003 | 0.001 |
| Scholarly communication | 0.014 | 0.016 |
| Open science | 0.003 | 0.004 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.020 | 0.016 |
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".