CN107491438A - Business decision elements recognition method and its system based on natural language - Google Patents
Business decision elements recognition method and its system based on natural language Download PDFInfo
- Publication number
- CN107491438A CN107491438A CN201710745216.4A CN201710745216A CN107491438A CN 107491438 A CN107491438 A CN 107491438A CN 201710745216 A CN201710745216 A CN 201710745216A CN 107491438 A CN107491438 A CN 107491438A
- Authority
- CN
- China
- Prior art keywords
- information
- event
- enterprise
- dictionary
- participle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/237—Lexical tools
- G06F40/247—Thesauruses; Synonyms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention relates to the business decision elements recognition method based on natural language and its system, this method includes acquisition enterprise and discloses information;Information is disclosed using enterprise and establishes event and the mapping relations of decision element, forms sample information storehouse;Obtain enterprise dynamic information;Enterprise dynamic information is cleaned, obtains participle;Judge whether participle is synonymous with the word in event dictionary;If so, decision element is then extracted in event sample information storehouse according to participle;If it is not, then carry out multidate information forwarding and simple classification.The present invention realizes the decision element for user that further extraction information may contain, really realize information assistant decision, with the increase of information sample size, dictionary and mapping table can more enrich and accurate, and with reference to the perception to enterprise's situation, information service is more immediately effective and has specific aim, is advantageous to more preferably perceive enterprise's situation.
Description
Technical field
The present invention relates to business decision elements recognition method, more specifically refers to that the business decision based on natural language will
Plain extracting method and its system.
Background technology
Improving level of decision-making is the key of enterprise operation and management, and to improve level of decision-making influences decision-making it is necessary to find
Key element, only realize and break through in the key element for influenceing decision-making, could fundamentally enterprise decision-making capability.Industrial scale decision lures
Because being the input of information, policymaker obtains industry relevant information through various channels, and decision-making is catered to or excited to these information
Person is real or potential investment demand, at this moment just excites decision-making impulsion.It can thus be seen that the outside inducement of industrial scale decision is
Information, Internal Causative Factors are investment demands.
For the above-mentioned input information referred to, existing Data Information Service is all directly provided by information, or right
After current information carries out simple classification, carry out information service, for the information content it is whether useful to user or
Influence is not all judged, also the enterprise management decision-making key element that may contain in information can not be extracted, so that nothing
Method carries out business decision according to information auxiliary user.
Chinese patent 201110236683.7 disclose it is a kind of based on triple user comment summary generation method with
System, this method comprise the following steps:The feature dictionary, mapping vocabulary and emotion dictionary of object are established, and according to feature dictionary
Construction feature tree;Capture user comment webpage;Receive user comment;Each user comment is handled one by one, generation is each
From the comment summary based on comment triple;The comment triple for integrating all user comments is concluded, generates decision-making triple;Meter
Calculate the quantity of feature and emotion word polarity identical decision-making triple;Extract all decision-making triple generation decision-making summaries.Profit
With this method or system, for each user comment generation comment summary, in order to which user checks reference, and by all comments
Triple, which is concluded, to be integrated, decision-making triple of the generation with directive significance, and is extracted the generation of whole decision-making triples and can be reflected always
Body evaluation result, the decision-making summary with decision assistant effect, so as to aid in user rapidly to make correct decisions.
Above-mentioned patent can not real-time update object feature dictionary, mapping vocabulary and emotion dictionary, easily cause user
The decision-making degree of accuracy made is not high.
Therefore, it is necessary to design a kind of business decision elements recognition method based on natural language, realization further carries
The decision element for user that breath of winning the confidence may contain, really realizes information assistant decision, as information sample size increases,
Dictionary and mapping table can more enrich and accurate, and combine the perception to enterprise's situation, information service more immediately effectively and
With specific aim, be advantageous to more preferably perceive enterprise's situation.
The content of the invention
The defects of it is an object of the invention to overcome prior art, there is provided the business decision elements recognition based on natural language
Method and its system.
To achieve the above object, the present invention uses following technical scheme:Business decision elements recognition based on natural language
Method, methods described include:
Obtain enterprise and disclose information;
Information is disclosed using enterprise and establishes event and the mapping relations of decision element, forms sample information storehouse;
Obtain enterprise dynamic information;
The enterprise dynamic information is cleaned, obtains participle;
Judge whether the participle is synonymous with the word in event dictionary;
If so, decision element is then extracted in event sample information storehouse according to participle;
If it is not, then carry out multidate information forwarding and simple classification.
Its further technical scheme is:Information is disclosed using enterprise and establishes event and the mapping relations of decision element, is formed
The step of sample information storehouse, including step in detail below:
Extraction enterprise discloses the mapping of entity, event, keyword, decision element and event and decision element in information
Relation;
Preserve information original text, summary, entity, event, keyword and the event and the mapping relations of decision element, shape
Into sample information storehouse.
Its further technical scheme is:Before obtaining the step of enterprise discloses information, in addition to:
Establish event dictionary and entity dictionary.
Its further technical scheme is:Information is disclosed using enterprise and establishes event and the mapping relations of decision element, is formed
After the step of sample information storehouse, in addition to:
Root in event dictionary is subjected to synonym merging according to naturally semantic;
Root in entity dictionary is subjected to synonym merging according to naturally semantic.
Its further technical scheme is:The enterprise dynamic information is cleaned, obtains the step of segmenting, including it is following
Specific steps:
Segment processing and subordinate sentence processing are carried out to the content of text of the enterprise dynamic information, obtain every section of text and
Every text;
Word segmentation processing is carried out to every section of text and every text, obtains the participle in every section and every.
Its further technical scheme is:The participle is judged whether after the step synonymous with the word in event dictionary, also
Including:
Judge whether event dictionary or sample information storehouse can cover multidate information;
If can, into end step;
If can not, existing event dictionary and sample information storehouse are expanded using machine learning mode.
Present invention also offers the business decision elements recognition system based on natural language, including information acquiring unit, sample
This information storehouse forms unit, multidate information acquiring unit, participle acquiring unit, judging unit, extraction decision element and simple
Processing unit;
The information acquiring unit, information is disclosed for obtaining enterprise;
The sample information storehouse forms unit, and establishing the mapping of event and decision element for disclosing information using enterprise closes
System, form sample information storehouse;
The multidate information acquiring unit, for obtaining enterprise dynamic information;
The participle acquiring unit, for being cleaned to the enterprise dynamic information, obtain participle;
The judging unit, for judging whether the participle is synonymous with the word in event dictionary;
The extraction decision element, for if so, then extracting decision element in event sample information storehouse according to participle;
The simple process unit, for if it is not, then carrying out multidate information forwarding and simple classification.
Its further technical scheme is:The sample information storehouse, which forms unit, includes extraction module and preserving module;
The extraction module, entity, event, keyword, decision element and thing in information are disclosed for extracting enterprise
The mapping relations of part and decision element;
The preserving module, for preserving information original text, summary, entity, event, keyword and the event and decision-making
The mapping relations of key element, form sample information storehouse.
Its further technical scheme is:The system also establishes unit including dictionary;
The dictionary establishes unit, for establishing event dictionary and entity dictionary.
Its further technical scheme is:The system also includes the first combining unit and the second combining unit;
First combining unit, for root in event dictionary to be carried out into synonym merging according to naturally semantic;
Second combining unit, for root in entity dictionary to be carried out into synonym merging according to naturally semantic.
Compared with the prior art, the invention has the advantages that:The business decision key element based on natural language of the present invention carries
Method is taken, information is disclosed by obtaining enterprise, utilizes its mapping pass with decision element of the event establishment screened in information
It is table, after enterprise dynamic information is got, multidate information is cleaned, obtain participle, according in participle and event dictionary
Word contrast, if synonymous, decision element is extracted in event sample information storehouse according to participle, realizes further extraction information
The decision element for user that may contain, really realizes information assistant decision, with information sample size increase, dictionary and
Mapping table can more enrich and accurate, and combine the perception to enterprise's situation, and information service more effectively and has a pin immediately
To property, be advantageous to more preferably perceive enterprise's situation.
The invention will be further described with specific embodiment below in conjunction with the accompanying drawings.
Brief description of the drawings
Fig. 1 is the flow for the business decision elements recognition method based on natural language that the specific embodiment of the invention provides
Figure;
Fig. 2 is the flow chart in the formation event sample information storehouse that the specific embodiment of the invention provides;
Fig. 3 is being cleaned to the enterprise dynamic information and obtaining the flow of participle for specific embodiment of the invention offer
Figure;
Fig. 4 is the structural frames for the business decision elements recognition system based on natural language that the specific embodiment of the invention provides
Figure;
Fig. 5 is the structured flowchart that the sample information storehouse that the specific embodiment of the invention provides forms unit;
Fig. 6 is the structured flowchart for the participle acquiring unit that the specific embodiment of the invention provides.
Embodiment
In order to more fully understand the technology contents of the present invention, technical scheme is entered with reference to specific embodiment
One step introduction and explanation, but it is not limited to this.
Specific embodiment as shown in figs. 1 to 6, the business decision elements recognition based on natural language that the present embodiment provides
It method, can be used in the assorting process to the magnanimity company information on internet, realize that further extraction information may
The decision element for user contained, really realizes information assistant decision, with the increase of information sample size, dictionary and mapping
Table can more enrich and accurate, and combine the perception to enterprise's situation, and information service more effectively and has a specific aim immediately,
Be advantageous to more preferably perceive enterprise's situation.
As shown in figure 1, the business decision elements recognition method based on natural language that the present embodiment provides, this method bag
Include:
S1, acquisition enterprise disclose information;
S2, disclose using enterprise information and establish event and the mapping relations of decision element, form sample information storehouse;
S3, obtain enterprise dynamic information;
S4, the enterprise dynamic information is cleaned, obtain participle;
S5, judge whether the participle is synonymous with the word in event dictionary;
S6, if so, then in event sample information storehouse extracting decision element according to participle;
S7, if it is not, then carrying out multidate information forwarding and simple classification.
For above-mentioned S1 steps, specifically using crawler technology, the open money related to enterprise is crawled on the internet
News, the disclosure information include but is not limited to all kinds of news report, company's bulletin etc..
In other embodiment, above-mentioned S1 steps, before obtaining the step of enterprise discloses information, in addition to:
Establish event dictionary and entity dictionary.
Further, above-mentioned S2 steps, disclose information using enterprise and establish event and the mapping relations of decision element,
The step of forming sample information storehouse, including step in detail below:
S21, extraction enterprise disclose entity, event, keyword, decision element and event in information and decision element
Mapping relations;
S22, the mapping pass for preserving information original text, summary, entity, event, keyword and the event and decision element
System, form sample information storehouse.
Above-mentioned S21 steps, information is specifically disclosed by Machine automated processing enterprise, information can be disclosed to enterprise and entered
Row extraction word segmentation processing, contrasted using participle with the event dictionary and entity dictionary originally established, participle is positioned, had
Body, according to natural semantic understanding, the participle in every section or every is contrasted with the event dictionary and entity dictionary originally established,
Same event will be classified as with word synonymous in event dictionary, the word synonymous with entity dictionary is classified as identical entity, so as to obtain reality
The mapping relations of body, event, keyword, decision element and event and decision element.
Above-mentioned S22 steps, specifically by information original text, summary, entity, event, keyword and the event with determining
The mapping relations of plan key element are stored in database, form sample information storehouse.
In the present embodiment, above-mentioned S2 steps, disclose information using enterprise and establish the mapping of event and decision element and close
System, formed sample information storehouse the step of after, in addition to:
S20, the root in event dictionary is subjected to synonym merging according to naturally semantic;
S30, the root in entity dictionary is subjected to synonym merging according to naturally semantic.
For above-mentioned S20 steps and S30 steps, the entity in event participle and entity dictionary in event dictionary
Participle, it can increase because of the information content increase of information, for the characteristics of Chinese, in event dictionary and entity dictionary
There is synonym in word, therefore, it is necessary to which synonym is merged, to improve mapping efficiency, for example investment is synonym with providing funds, hair
Cloth product and product are reached the standard grade a synonym;As long as that is synonymous, then an event word is merged into, is stored in event dictionary
It is interior, also it is processed as entity dictionary.
After above-mentioned S1, S2, S20, S30 step, " event-decision element " mapping table, that is, sample are formd
The relation of event and decision element in this information storehouse, in this, as the benchmark of judgement, the enterprise dynamic information of acquisition is carried out
Judge, in order to further extract the decision element for user that information may contain, really realize that information auxiliary is determined
Plan.
For above-mentioned S3 steps, particular by crawler technology, the enterprise dynamic information on internet, the enterprise are crawled
Multidate information includes but is not limited to all kinds of news report, company's bulletin etc..Using enterprise dynamic information, can combine to enterprise's state
The perception of gesture, information service is more immediately effective and has specific aim, helps more preferably to perceive enterprise's situation.
For above-mentioned S4 steps, the enterprise dynamic information is cleaned, obtains the step of segmenting, including following tool
Body step:
S41, the content of text to the enterprise dynamic information carry out segment processing and subordinate sentence is handled, and obtains every section of text
And every text;
S42, word segmentation processing is carried out to every section of text and every text, obtain the participle in every section and every.
In other embodiment, above-mentioned S3 steps, S4 steps and carry out before being placed on S1 steps.
For above-mentioned participle, specifically, according to natural semantic understanding, the participle in every section or every is established with original
Event dictionary and entity dictionary contrast, same event will be classified as with word synonymous in event dictionary, it is synonymous with entity dictionary
Word is classified as identical entity.
For above-mentioned S5 steps, judge whether participle is synonymous with the word in event dictionary, is mainly based upon natural language
Treatment technology, if participle is synonymous with the word in event dictionary, containing corresponding with event dictionary semanteme in enterprise dynamic information
Participle, can enter S6 steps, extract decision element, if participle is not synonymous with the word in event dictionary, enter normal letter
Breath forwarding and simple classification.
In other embodiment, above-mentioned S5 steps, judge the participle whether the step synonymous with the word in event dictionary
Afterwards, in addition to:
Judge whether event dictionary or sample information storehouse can cover multidate information;If can, into end step;If no
Can, then existing event dictionary and sample information storehouse are expanded using machine learning mode.
Existing event dictionary and sample information storehouse are expanded using machine learning mode, event dictionary can be caused and reflected
Firing table can be more abundant and accurate with the sample size increase of information.
For example, the extraction of the business decision key element based on natural language is carried out to Beijing Mo Bai Science and Technology Ltd.s, entirely
Workflow is as follows:Information 1 --- L'Oreal jointly rub visit bicycle this transboundary market you to ratherSource page:http:// www.jiemian.com/article/1227353.html.The content of information 1 is cleaned, cleaned again after obtaining participle;Extraction
Keyword ---【Entity:Rub and visit bicycle, L'Oreal】【Event:Jointly, transboundary market】【Attribute:L'Oreal's FSI, across
Market on boundary】;According to NLP, event " uniting " semantic word identical with database is merged into " strategic cooperation ";Or by event
" transboundary marketing " semantic word identical with database is merged into " marketing ";Due to event " strategic cooperation " with decision-making
Key element " operation-cooperative relationship-strategic cooperation " maps, event " marketing " with decision element " operation-marketing-
Promotion method " maps, and so far, this information has been converted into the pass of " entity-event-attribute " via common news report
Keyword triple, and event therein directly corresponds to specific decision element.Information 2:Rub visit bicycle announce enter Linfen, Shanxi,
Accelerate to move into three, four line cities, source pagehttp://www.cyzone.cn/a/20170505/310580.html.Work
Flow is as follows:The content of information 2 is cleaned, cleaned again after obtaining participle;Extract keyword ---【Entity:Linfen county of Shanxi Province, rubs
Visit bicycle】【Event:Move into】【Attribute:Linfen, Shanxi, three, four line cities】;According to NLP, by event " garrison " and database
Identical semantic word is merged into " market expansion ";Due to event " market expansion " with decision element " operation-market-market
Strategy " maps, and so far, this information has been converted into the keyword of " entity-event-attribute " via common news report
Triple, and event therein directly corresponds to specific decision element.
The above-mentioned business decision elements recognition method based on natural language, discloses information by obtaining enterprise, utilizes money
Its mapping table with decision element of the event establishment screened in news, after enterprise dynamic information is got, to dynamic
Information is cleaned, and obtains participle, is contrasted according to participle and the word in event dictionary, if synonymous, according to participle in event sample
Decision element is extracted in this information storehouse, realizes the decision element for user that further extraction information may contain, very
Just realize information assistant decision, with the increase of information sample size, dictionary and mapping table can more enrich and accurate, and with reference to pair
The perception of enterprise's situation, information service is more immediately effective and has specific aim, is advantageous to more preferably perceive enterprise's situation.
As shown in figure 4, present embodiments providing the business decision elements recognition system based on natural language, it includes information
Acquiring unit 1, sample information storehouse form unit 2, multidate information acquiring unit 5, participle acquiring unit 6, judging unit 7, extraction
Decision element unit 8 and simple process unit 9.
Information acquiring unit 1, information is disclosed for obtaining enterprise.
Sample information storehouse forms unit 2, and event and the mapping relations of decision element are established for disclosing information using enterprise,
Form sample information storehouse.
Multidate information acquiring unit 5, for obtaining enterprise dynamic information.
Acquiring unit 6 is segmented, for being cleaned to the enterprise dynamic information, obtains participle.
Judging unit 7, for judging whether the participle is synonymous with the word in event dictionary.
Decision element unit 8 is extracted, for if so, then extracting decision element in event sample information storehouse according to participle.
Simple process unit 9, for if it is not, then carrying out multidate information forwarding and simple classification.
Above-mentioned information acquiring unit 1 is specifically to use crawler technology, crawls the disclosure related to enterprise on the internet
Information, the disclosure information include but is not limited to all kinds of news report, company's bulletin etc..
In other embodiment, the above-mentioned system also establishes unit including dictionary;Above-mentioned dictionary establishes unit, is used for
Establish event dictionary and entity dictionary.
Further, above-mentioned sample information storehouse, which forms unit 2, includes extraction module 21 and preserving module 22.
Extraction module 21, entity, event, keyword, decision element and event in information are disclosed for extracting enterprise
With the mapping relations of decision element.
Preserving module 22, will for preserving information original text, summary, entity, event, keyword and the event and decision-making
The mapping relations of element, form sample information storehouse.
Above-mentioned extraction module 21 is specifically to disclose information by Machine automated processing enterprise, can disclose information to enterprise
Extraction word segmentation processing is carried out, is contrasted using participle with the event dictionary and entity dictionary originally established, participle is positioned,
Specifically, according to natural semantic understanding, by the participle in every section or every and the event dictionary and entity dictionary pair originally established
Than same event will be classified as with word synonymous in event dictionary, the word synonymous with entity dictionary is classified as identical entity, so as to obtain
The mapping relations of entity, event, keyword, decision element and event and decision element.
Above-mentioned preserving module 22 be specifically by information original text, summary, entity, event, keyword and the event with
The mapping relations of decision element are stored in database, form sample information storehouse.
Further, above-mentioned system also includes the first combining unit 3 and the second combining unit 4.
First combining unit 3, for root in event dictionary to be carried out into synonym merging according to naturally semantic.
Second combining unit 4, for root in entity dictionary to be carried out into synonym merging according to naturally semantic.
Specifically, in event dictionary event participle and entity dictionary in entity participle, can because information information
Amount increases and increased, and for the characteristics of Chinese, the word in event dictionary and entity dictionary has synonym, therefore,
Need to merge synonym, to improve mapping efficiency, for example investment is synonym with providing funds, and release product and product position of reaching the standard grade are same
Adopted word;As long as that is synonymous, then an event word is merged into, is stored in event dictionary, it is also such for entity dictionary
Processing.
Above-mentioned information acquiring unit 1, sample information storehouse form unit 2, the first combining unit 3 and the second combining unit 4
" event-decision element " mapping table, that is, event and the relation of decision element in sample information storehouse are formd, in this, as
The benchmark of judgement, the enterprise dynamic information of acquisition is judged, in order to further extract that information may contain for
The decision element of user, really realizes information assistant decision.
Above-mentioned multidate information acquiring unit 5 crawls the enterprise dynamic information on internet particular by crawler technology,
The enterprise dynamic information includes but is not limited to all kinds of news report, company's bulletin etc..Using enterprise dynamic information, can combine pair
The perception of enterprise's situation, information service is more immediately effective and has specific aim, helps more preferably to perceive enterprise's situation.
Further, above-mentioned participle acquiring unit 6 includes segmentation subordinate sentence module 61 and word segmentation processing module 62.
Subordinate sentence module 61 is segmented, for being carried out to the content of text of the enterprise dynamic information at segment processing and subordinate sentence
Reason, obtain every section of text and every text.
Word segmentation processing module 62, for carrying out word segmentation processing to every section of text and every text, obtain every section and every
Participle in sentence.
Above-mentioned participle acquiring unit 6 is particularly according to natural semantic understanding, by the participle in every section or every and original
Event dictionary and entity the dictionary contrast first established, will be classified as same event, with entity dictionary with word synonymous in event dictionary
Synonymous word is classified as identical entity.
Above-mentioned judging unit 7 judges whether participle is synonymous with the word in event dictionary, is mainly based upon at natural language
Reason technology, contain if participle is synonymous with the word in event dictionary, in enterprise dynamic information semantic with event dictionary corresponding point
Word, decision element is can extract out, if participle is not synonymous with the word in event dictionary, enter the forwarding of normal information and simple point
Class.
In other embodiment, above-mentioned system also includes level of coverage judging unit, and it is used to judge event dictionary or sample
Whether this information storehouse can cover multidate information;If can, terminate;If can not, existing thing is expanded using machine learning mode
Part dictionary and sample information storehouse.
Existing event dictionary and sample information storehouse are expanded using machine learning mode, event dictionary can be caused and reflected
Firing table can be more abundant and accurate with the sample size increase of information.
The above-mentioned business decision elements recognition system based on natural language, discloses information by obtaining enterprise, utilizes money
Its mapping table with decision element of the event establishment screened in news, after enterprise dynamic information is got, to dynamic
Information is cleaned, and obtains participle, is contrasted according to participle and the word in event dictionary, if synonymous, according to participle in event sample
Decision element is extracted in this information storehouse, realizes the decision element for user that further extraction information may contain, very
Just realize information assistant decision, with the increase of information sample size, dictionary and mapping table can more enrich and accurate, and with reference to pair
The perception of enterprise's situation, information service is more immediately effective and has specific aim, is advantageous to more preferably perceive enterprise's situation.
The above-mentioned technology contents that the present invention is only further illustrated with embodiment, in order to which reader is easier to understand, but not
Represent embodiments of the present invention and be only limitted to this, any technology done according to the present invention extends or recreation, by the present invention's
Protection.Protection scope of the present invention is defined by claims.
Claims (10)
1. the business decision elements recognition method based on natural language, it is characterised in that methods described includes:
Obtain enterprise and disclose information;
Information is disclosed using enterprise and establishes event and the mapping relations of decision element, forms sample information storehouse;
Obtain enterprise dynamic information;
The enterprise dynamic information is cleaned, obtains participle;
Judge whether the participle is synonymous with the word in event dictionary;
If so, decision element is then extracted in event sample information storehouse according to participle;
If it is not, then carry out multidate information forwarding and simple classification.
2. the business decision elements recognition method according to claim 1 based on natural language, it is characterised in that utilize enterprise
Industry discloses information and establishes event and the mapping relations of decision element, the step of forming sample information storehouse, including step in detail below:
The mapping that extraction enterprise discloses entity, event, keyword, decision element and event and decision element in information is closed
System;
Information original text, summary, entity, event, keyword and the event and the mapping relations of decision element are preserved, forms sample
This information storehouse.
3. the business decision elements recognition method according to claim 1 or 2 based on natural language, it is characterised in that obtain
Before the step of taking enterprise to disclose information, in addition to:
Establish event dictionary and entity dictionary.
4. the business decision elements recognition method according to claim 3 based on natural language, it is characterised in that utilize enterprise
Industry discloses information and establishes event and the mapping relations of decision element, after the step of forming sample information storehouse, in addition to:
Root in event dictionary is subjected to synonym merging according to naturally semantic;
Root in entity dictionary is subjected to synonym merging according to naturally semantic.
5. the business decision elements recognition method according to claim 4 based on natural language, it is characterised in that to described
Enterprise dynamic information is cleaned, and obtains the step of segmenting, including step in detail below:
Segment processing and subordinate sentence processing are carried out to the content of text of the enterprise dynamic information, obtain every section of text and every
Text;
Word segmentation processing is carried out to every section of text and every text, obtains the participle in every section and every.
6. the business decision elements recognition method according to claim 5 based on natural language, it is characterised in that judge institute
Participle is stated whether after the step synonymous with the word in event dictionary, in addition to:
Judge whether event dictionary or sample information storehouse can cover multidate information;
If can, into end step;
If can not, existing event dictionary and sample information storehouse are expanded using machine learning mode.
7. the business decision elements recognition system based on natural language, it is characterised in that including information acquiring unit, sample information
Storehouse forms unit, multidate information acquiring unit, participle acquiring unit, judging unit, extraction decision element unit and simple place
Manage unit;
The information acquiring unit, information is disclosed for obtaining enterprise;
The sample information storehouse forms unit, and event and the mapping relations of decision element are established for disclosing information using enterprise,
Form sample information storehouse;
The multidate information acquiring unit, for obtaining enterprise dynamic information;
The participle acquiring unit, for being cleaned to the enterprise dynamic information, obtain participle;
The judging unit, for judging whether the participle is synonymous with the word in event dictionary;
The extraction decision element unit, for if so, then extracting decision element in event sample information storehouse according to participle;
The simple process unit, for if it is not, then carrying out multidate information forwarding and simple classification.
8. the business decision elements recognition system according to claim 7 based on natural language, it is characterised in that the sample
This information storehouse, which forms unit, includes extraction module and preserving module;
The extraction module, for extract enterprise disclose entity, event, keyword, decision element and event in information with
The mapping relations of decision element;
The preserving module, for preserving information original text, summary, entity, event, keyword and the event and decision element
Mapping relations, form sample information storehouse.
9. the business decision elements recognition system according to claim 8 based on natural language, it is characterised in that the system
System also establishes unit including dictionary;
The dictionary establishes unit, for establishing event dictionary and entity dictionary.
10. the business decision elements recognition system according to claim 9 based on natural language, it is characterised in that described
System also includes the first combining unit and the second combining unit;
First combining unit, for root in event dictionary to be carried out into synonym merging according to naturally semantic;
Second combining unit, for root in entity dictionary to be carried out into synonym merging according to naturally semantic.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710745216.4A CN107491438A (en) | 2017-08-25 | 2017-08-25 | Business decision elements recognition method and its system based on natural language |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710745216.4A CN107491438A (en) | 2017-08-25 | 2017-08-25 | Business decision elements recognition method and its system based on natural language |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107491438A true CN107491438A (en) | 2017-12-19 |
Family
ID=60646739
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710745216.4A Pending CN107491438A (en) | 2017-08-25 | 2017-08-25 | Business decision elements recognition method and its system based on natural language |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107491438A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108763324A (en) * | 2018-05-03 | 2018-11-06 | 苏州朗动网络科技有限公司 | Recognition methods, device, storage medium and the computer equipment of business data |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120030206A1 (en) * | 2010-07-29 | 2012-02-02 | Microsoft Corporation | Employing Topic Models for Semantic Class Mining |
CN103049532A (en) * | 2012-12-21 | 2013-04-17 | 东莞中国科学院云计算产业技术创新与育成中心 | Method for creating knowledge base engine on basis of sudden event emergency management and method for inquiring knowledge base engine |
CN103544255A (en) * | 2013-10-15 | 2014-01-29 | 常州大学 | Text semantic relativity based network public opinion information analysis method |
CN102708096B (en) * | 2012-05-29 | 2014-10-15 | 代松 | Network intelligence public sentiment monitoring system based on semantics and work method thereof |
CN104820629A (en) * | 2015-05-14 | 2015-08-05 | 中国电子科技集团公司第五十四研究所 | Intelligent system and method for emergently processing public sentiment emergency |
-
2017
- 2017-08-25 CN CN201710745216.4A patent/CN107491438A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120030206A1 (en) * | 2010-07-29 | 2012-02-02 | Microsoft Corporation | Employing Topic Models for Semantic Class Mining |
CN102708096B (en) * | 2012-05-29 | 2014-10-15 | 代松 | Network intelligence public sentiment monitoring system based on semantics and work method thereof |
CN103049532A (en) * | 2012-12-21 | 2013-04-17 | 东莞中国科学院云计算产业技术创新与育成中心 | Method for creating knowledge base engine on basis of sudden event emergency management and method for inquiring knowledge base engine |
CN103544255A (en) * | 2013-10-15 | 2014-01-29 | 常州大学 | Text semantic relativity based network public opinion information analysis method |
CN104820629A (en) * | 2015-05-14 | 2015-08-05 | 中国电子科技集团公司第五十四研究所 | Intelligent system and method for emergently processing public sentiment emergency |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108763324A (en) * | 2018-05-03 | 2018-11-06 | 苏州朗动网络科技有限公司 | Recognition methods, device, storage medium and the computer equipment of business data |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110516208B (en) | System and method for extracting PDF document form | |
CN103544255B (en) | Text semantic relativity based network public opinion information analysis method | |
CN106296195A (en) | A kind of Risk Identification Method and device | |
CN103309862B (en) | Webpage type recognition method and system | |
CN106372132A (en) | Artificial intelligence-based query intention prediction method and apparatus | |
CN104182412B (en) | A kind of web page crawl method and system | |
CN105912625A (en) | Linked data oriented entity classification method and system | |
CN106021410A (en) | Source code annotation quality evaluation method based on machine learning | |
CN103064971A (en) | Scoring and Chinese sentiment analysis based review spam detection method | |
CN105045847B (en) | A kind of method that Chinese institutional units title is extracted from text message | |
CN105528422A (en) | Focused crawler processing method and apparatus | |
CN106021383A (en) | Method and device for computing similarity of webpages | |
CN113722490B (en) | Visual rich document information extraction method based on key value matching relation | |
CN105912645A (en) | Intelligent question and answer method and apparatus | |
CN106126502A (en) | A kind of emotional semantic classification system and method based on support vector machine | |
CN107943514A (en) | The method for digging and system of core code element in a kind of software document | |
CN110472209A (en) | Table generation method, device and computer equipment based on deep learning | |
CN111967970B (en) | Bank product recommendation method and device based on spark platform | |
Catalano et al. | Phylogenetic inference based on landmark data in 41 empirical data sets | |
CN113609892A (en) | Handwritten poetry recognition method integrating deep learning with scenic spot knowledge map | |
CN109408643B (en) | Fund similarity calculation method, system, computer equipment and storage medium | |
CN107491438A (en) | Business decision elements recognition method and its system based on natural language | |
CN110889274B (en) | Information quality evaluation method, device, equipment and computer readable storage medium | |
RU2003118580A (en) | METHOD FOR RECOGNIZING GRAPHIC OBJECTS USING THE PRINCIPLE OF INTEGRITY | |
CN110362673A (en) | Computer vision class papers contents method of discrimination and system based on abstract semantic analysis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171219 |
|
RJ01 | Rejection of invention patent application after publication |