Date Extractor Using NER Vector Pattern Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current approaches are inefficient and costly in extracting dates and date ranges from natural language documents due to the variability in date formats and the need to enumerate numerous patterns, as the order of year, month, day is not fixed and some parts may be omitted, making it difficult to achieve precision.
Innovation Solution
A date extractor system utilizing natural language processing (NLP) with named entity recognition (NER) to generate named entity vectors, compare tokens to predetermined date and date range patterns, and extract dates and date ranges, reducing the need for extensive pattern enumeration and coded rules.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If current approaches use extensive pattern enumeration and coded rules to extract dates, then they can handle various date formats, but the implementation becomes time-prohibitive and expensive
Solution Approach 1:
The patent changes the approach from enumerating numerous specific date patterns to using a parametric model that represents date structures generically. Instead of coding rules for each format variation, the system uses parameters to define date components (day, month, year) and their relationships, allowing flexible matching of various formats through parameter substitution rather than pattern enumeration.
Solution Approach 2:
The patent creates a universal date extraction framework that handles multiple date formats through a single unified approach. The named entity recognition machine and pattern comparison mechanism serve multiple functions: identifying date tokens, determining their semantic roles, and matching them against template patterns, eliminating the need for separate specialized rules for each format type.
2Measurement precision
If current approaches enumerate all date patterns with precision, then they can accurately extract dates, but the implementation becomes time-prohibitive and expensive
Solution Approach 1:
The patent performs preliminary actions by using named entity recognition to identify and label potential date tokens before pattern matching. This pre-processing step filters out non-date tokens and organizes candidate tokens with their semantic roles, reducing the search space for subsequent pattern comparison and improving both precision and speed by avoiding exhaustive enumeration of all possible patterns.
Solution Approach 2:
The patent segments the date extraction process into distinct stages: token identification, named entity labeling, pattern template matching, and date construction. This segmentation allows each stage to focus on specific tasks with optimized algorithms, improving overall efficiency while maintaining precision through systematic progression through extraction steps.
3Adaptability or versatility
If the system uses a large number of patterns and rules to handle all date formats, then it can achieve high coverage, but the device complexity increases
Solution Approach 1:
The patent reduces system complexity by transforming the approach from rule-based pattern enumeration to parameter-based template matching. Instead of maintaining numerous coded rules for different date formats, the system uses parameterized templates that represent date structures generically, with parameters instantiated during matching. This significantly reduces the number of explicit rules needed while maintaining comprehensive format coverage.
Solution Approach 2:
The patent introduces named entity recognition as an intermediary layer between raw text tokenization and pattern matching. This intermediary step labels tokens with semantic roles and filters candidates, reducing the complexity of subsequent pattern matching by pre-organizing data and eliminating the need for complex decision logic to handle all format variations directly in the pattern matching stage.
Data Source
AI summary
A date extractor disclosed herein allows extracting dates and date ranges from documents. An implementation of the date extractor is implemented using various computer process instructions including scanning a document to generate a plurality of tokens, assigning labels to token using named entity recognition machine to generate a named entity vector, extracting dates from the named entity vector by comparing each of the named entities of the named entity vector to predetermined patterns of dates to generate a date vector, generating a plurality of date pairs from the date vector, and extracting date-ranges by comparing the plurality of date pairs to predetermined patterns of date ranges.