Extraction of time-related expressions using text mining with application to Hebrew

Area of Science:

• Computational Linguistics
• Digital Humanities
• Religious Studies

Background:

• Rabbinic texts contain complex temporal expressions often linked to citations.
• Identifying these time-related expressions is crucial for accurate historical and linguistic analysis.
• Previous methods lacked efficiency and precision in extracting temporal data from such specialized corpora.

Purpose of the Study:

• To develop a semi-automatic method for extracting time-related expressions from rabbinic texts.
• To improve the efficiency and accuracy of temporal data identification in historical religious documents.
• To create and test novel statistical functions and heuristic methods for this specific linguistic task.

Main Methods:

• A semi-automatic approach was employed, initially identifying all expressions near rabbinic references.
• Two novel statistical functions were formulated to analyze and filter potential time expressions.
• Grammatical screenings and heuristic methods were applied to refine the extracted data.

Main Results:

• The methods successfully filtered potential time-related expressions, reducing 484,681 initial phrases to 575.
• A significant reduction of 99.88% in candidate expressions was achieved.
• The approach demonstrated high efficacy on a corpus of responsa documents with marked rabbinic citations.

Conclusions:

• The developed semi-automatic method, utilizing statistical functions and screening, is effective for extracting time expressions from rabbinic literature.
• This approach offers a substantial improvement in precision and efficiency for temporal data analysis in specialized corpora.
• The findings facilitate more accurate computational analysis of historical religious texts.