Editing Automated Resolution of Noisy Bibliographic References

Title: Automated Resolution of Noisy Bibliographic References

Research Question: How can we develop an efficient method to resolve noisy bibliographic references obtained from OCR methods and link them to records in a bibliographic database?

Methodology:

1. Identify the problem: The researchers focused on the issue of resolving noisy bibliographic references, which are often riddled with errors due to OCR methods. They used the NASA Astrophysics Data System (ADS), which has gathered over three million references from scanned astronomical literature.

2. Propose a solution: The researchers developed a method that allows a controlled merging of correction, parsing, and matching, inspired by dependency grammars. They also employed various heuristics to improve recall.

3. Implement the solution: The researchers used a three-step procedure to correct the OCR results, parse the corrected string, and match it against the database. They then introduced a heuristic approach to improve recall, which involved techniques like lemmatization, stemming, and string similarity.

Results:

1. Evaluation of the method: The researchers found that their method was effective in resolving noisy references and linking them to the bibliographic database. They reported improvements in recall and precision rates.

Implications:

1. Significance of the research: The automated resolution of noisy bibliographic references is a crucial problem for linking scholarly publications, especially for historical literature. The researchers' method has the potential to improve the accuracy and efficiency of such processes.

2. Future directions: The researchers suggest further exploration of the heuristics they employed and the possibility of incorporating machine learning techniques to enhance the system's performance. They also encourage the application of their method to other domains facing similar challenges.

In conclusion, the researchers developed an efficient method for resolving noisy bibliographic references, which has significant implications for the field of scholarly linkage and the automated processing of historical literature. Their approach involves a controlled merging of correction, parsing, and matching, along with heuristic techniques to improve recall. The method's effectiveness and potential for further improvement make it a valuable contribution to the field.

Link to Article: https://arxiv.org/abs/0401028v1
Authors: 
arXiv ID: 0401028v1

[[Category:Computer Science]]
[[Category:Method]]
[[Category:Bibliographic]]
[[Category:References]]
[[Category:Researchers]]
[[Category:Noisy]]