# US 2019/0244116 A1 · Method and system to mine rule intents from documents

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0244116 A1
- Document ID: US20190244116A1
- Status: Published application
- Kind: A1
- Application number: 16/265,965
- Filed: 2019-02-01
- Published: 2019-08-08
- Earliest claimed: 2018-02-02
- Assignees: Tata Consultancy Services Limited
- Applicants: Tata Consultancy Services Limited
- Inventors: Pavan Kumar CHITTIMALLI; Ravindra NAIK
- Web page: https://patentgrep.com/patent/US20190244116A1

## CPC

- G06N 5/025: Extracting rules from data
- G06F 40/205: Parsing
- G06F 40/289: Phrasal analysis, e.g. finite state techniques or chunking
- G06F 40/30: Semantic analysis
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system and method for method and a system for mining rule intents from documents is provided, wherein the rule intents are basic atomic facts present in a sentence. The proposed method and system for mining of rule intents from documents is performed in multiple stages that include extraction of rule sentences from input documents that are further parsed to create dependency trees. The created dependency trees are further analyzing based on a set of plurality of heuristic rules to mine rule intents. The mined rule intents are saved and further displayed in Semantics of Business Vocabulary and Rules (SBVR) format. The mined rule intents that are displayed automatically in the SBVR format can be easily analyzed as they SBVR is a comprehensive standard for business rule representation by Object Management Group (OMG) in accordance with set of a standard pre-defined vocabularies.

## Claims

### Claim 1

A processor-implemented method for mining rule intents from documents, the method comprising:
extracting a plurality of rule sentence from the documents as an input;
creating a dependency tree for each of the extracted plurality of rule sentence;
extracting a plurality of arguments from the dependency tree based on a plurality of heuristic rules;
analyzing the extracted plurality of arguments to mine rule intents; and
saving the mined rule intents in a rule-base database.

### Claim 2 (depends on 1)

The method of claim 1, wherein the documents comprises of the plurality of input sentences are structured or unstructured documents that is written in natural language and contains noise.

### Claim 3 (depends on 1)

The method of claim 1, wherein the plurality of rule sentence is extracted by segregating rule sentence from noise based on learning models across different domains.

### Claim 4 (depends on 1)

The method of claim 1, wherein the dependency tree is created for each of the extracted plurality of sentences based on parser techniques that includes Stanford Dependency Parser.

### Claim 5 (depends on 1)

The method of claim 1, wherein the plurality of arguments comprise of multiple attributes including value and role attributes.

### Claim 6 (depends on 4)

The method of claim 4, wherein the role attributes hold multiple values that include subject, object, preposition, split and negation.

### Claim 7 (depends on 1)

The method of claim 1, wherein the a plurality of heuristic rules are performed individually for each role attribute based on nodes(n) and edges of dependency tree.

### Claim 8 (depends on 1)

The method of claim 1, wherein the rule intents are mined using extracted plurality of arguments (A) along with multiple parameters that include string (S) and headword (H) parameters.

### Claim 9 (depends on 1)

The method of claim 1, wherein the mined rule intents are saved as vocabularies in Semantics of Business Vocabulary and Rules (SBVR) format.

### Claim 10

A system comprising:
a rule intent mining unit (102) for mining rule intents from documents;
a memory (104) for storing instructions;
one or more communication interfaces(106);
one or more hardware processors(108) communicatively coupled to the memory using the one or more communication interfaces, wherein the one or more hardware processors are configured by the instructions to execute the one or more modules of the mining unit(102) comprising:
a rule sentence extractor(202) for extracting a plurality of sentences from the documents as an input;
a natural language parser(204) for creating a dependency tree for each of the extracted plurality of rule sentence;
a rule intent extractor(206) for extracting a plurality of arguments from the dependency tree based on a plurality of heuristic rules and analyzing the extracted plurality of arguments to mine rule intents;
a rule-base database(208) saving the mined rule intents; and
a display module(210) for displaying saved mined rule intents.

### Claim 11 (depends on 10)

The system of claim 10, wherein the saved mined rule intents are displayed on the display module (210) in saved SBVR format.

### Claim 12

One or more non-transitory machine readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors cause:
extracting a plurality of rule sentence from the documents as an input;
creating a dependency tree for each of the extracted plurality of rule sentence;
extracting a plurality of arguments from the dependency tree based on a plurality of heuristic rules;
analyzing the extracted plurality of arguments to mine rule intents; and
saving the mined rule intents in a rule-base database.
