Information extraction method for administrative penalty decision
A technology of information extraction and text information, applied in the fields of natural language processing and legal artificial intelligence, can solve the problems of low efficiency and low accuracy of information extraction, solve the dependency of similar texts, improve accuracy and efficiency, and prevent information loss Effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0046] An information extraction method for an administrative penalty decision, such as figure 1 shown, including:
[0047] Step 1: Crawl from the administrative penalty document website to obtain the administrative penalty decision letters of each province; it will be used to build a data set later.
[0048] Step 2: Extract the text content of the administrative penalty decision obtained in step 1 in the html tag, construct the original data set, and obtain the .csv file.
[0049] Step 3: According to the normative rules for writing administrative punishment decisions, use regular expressions to perform data preprocessing on administrative punishment decisions to be processed, construct data sets, and obtain .csv files.
[0050] Step 4: Input the data set constructed in step 3 into the information extraction module trained with the original data set constructed in step 2, and output the information extraction results of administrative punishment documents. Information extra...
Embodiment 2
[0052] According to the information extraction method of an administrative penalty decision document described in Embodiment 1, the difference is that:
[0053] In step 2, use the strip() function in python to remove the html label and label, obtain the text content of the administrative punishment decision letter, the text content of the administrative penalty decision letter includes the decision letter number, parties, subject qualification certificate name, unified social credit code, domicile (address), legally responsible person (responsible person, operator) ), identity card number, source of the case and investigation process, case facts, proof of evidence (notification of administrative punishment, statements, defenses, hearing opinions of the parties, review and acceptance and reasons), qualitative nature of illegal acts, basis for punishment, discretionary Facts and reasons, implementation methods and deadlines of administrative penalties, remedies and deadlines. ...
Embodiment 3
[0062] According to the information extraction method of an administrative penalty decision document described in Embodiment 2, the difference is that:
[0063] In step four, the steps are as follows:
[0064] Input the data set constructed in step 3 into the pre-training language module, and according to the text characteristics of the administrative penalty decision, obtain the short text information sequence through the sliding window self-attention mechanism, including the decision document number, party, subject qualification certificate name, unified society Credit code, domicile (address), legally responsible person (responsible person, operator), ID number; through the combination of sliding window self-attention mechanism and global attention mechanism to obtain the source of the case, investigation process, case facts, evidence (administrative Notification of punishment, parties’ statement, defense, hearing opinions, review and adoption and reasons), qualitativ...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap