How to Do Lexical Quality Estimation of a Large OCRed Historical Finnish Newspaper Collection with Scarce Resources
Bibliographic record
Abstract
The National Library of Finland has digitized and made available the historical newspapers published in Finland between 1771 and 1910 (Bremer-Laamanen 2014; Kettunen et al. 2014). This collection contains approximately 1.95 million pages in Finnish and Swedish. The Finnish part of the collection consists of about 2.40 billion words. The National Library’s Digital Collections are offered via the digi.kansalliskirjasto.fi web service, also known as Digi. An open data package of the whole collection was released in early 2017 (Pääkkönen et al. 2016). Quality of OCRed collections is an important topic in digital humanities, as it affects general usability and searchability of collections. There is no single available method to assess quality of large collections, but different methods can be used to approximate quality. This paper discusses different corpus analysis style methods to approximate overall lexical quality of the Finnish part of the Digi collection. Methods include usage of parallel samples and word error rates, usage of morphological analysers, frequency analysis of words and comparisons to comparable edited lexical data. Our aim in the quality analysis is twofold: firstly to analyse the present state of the lexical data and secondly, to establish a set of assessment methods that build up a compact procedure for quality assessment after e.g. re-OCRing or post-correction of the material. Résumé La Bibliothèque nationale de Finlande a numérisé et rendu disponible les journaux historiques publiés en Finlande entre 1771 et 1910 (Bremer-Laamanen 2014 ; Kettunen et al. 2014). Cette collection contient environ 1,95 million pages en finnois et suédois. La partie finnoise de la collection compte environ 2,40 milliards de mots. Les Collections numérisées de la Bibliothèque Nationale sont offertes sur le service web digi.kansalliskirjasto.fi, également appelé Digi. Un ensemble de données disponibles de la collection entière est sorti début 2017 (Pääkkönen et al. 2016). La qualité de collections en OCR est un thème important pour les humanités numériques, puisqu’elle concerne l’utilité et la facilité de recherche de collections. Il n’y a pas qu’une seule méthode pour évaluer la qualité de grandes collections, mais des méthodes différentes peuvent être employées pour en estimer la qualité. Cet article discute de méthodes différentes d’analyses de corpus visant à estimer la qualité lexicale totale de la partie finnoise de la collection Digi. Les méthodes comprennent l’usage d’échantillons parallèles et de fréquences d’erreur de mot, l’usage d’analyseurs morphologiques, l’analyse de fréquence de mots et les comparaisons à des données lexicales rédigées comparables. Notre objectif dans l’analyse de qualité est double : premièrement, analyser l’état actuel des données lexicales et, deuxièmement, établir un ensemble de méthodes d’évaluation qui constituent une procédure compacte pour l’évaluation de la qualité après, par exemple, la retransformation en OCR ou après les après corrections du matériel. Mots-clés: qualité d’OCR; estimation de qualité lexicale; collection de journaux finnois du 19e siècle
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 | 0.007 | 0.054 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.011 | 0.007 |
| Science and technology studies | 0.002 | 0.001 |
| Scholarly communication | 0.005 | 0.005 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.004 |
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".