CN108830699A - A kind of intelligence accounting technology - Google Patents
A kind of intelligence accounting technology Download PDFInfo
- Publication number
- CN108830699A CN108830699A CN201810519927.4A CN201810519927A CN108830699A CN 108830699 A CN108830699 A CN 108830699A CN 201810519927 A CN201810519927 A CN 201810519927A CN 108830699 A CN108830699 A CN 108830699A
- Authority
- CN
- China
- Prior art keywords
- subject
- accounting
- keyword
- decision
- item
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/12—Accounting
- G06Q40/125—Finance or payroll
-
- 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
- G06F40/284—Lexical analysis, e.g. tokenisation or collocates
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Artificial Intelligence (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- General Business, Economics & Management (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a kind of intelligence to calculate technology, and the accounting method is as follows:The keyword (word for as facilitating subject judgement) in the Electronic Data Processing information and crowdsourcing platform in accounting system is extracted using Ansj participle technique,Construct decision-tree model,The keyword of each history reimbursement document is successively passed through to three layers of decision-tree model of previous step building,According to model-naive Bayesian,It analyzes these keywords and generates a certain section's purpose probability,By the crucial phrase in every single history document at one group of random art without sequencing,Structure of the invention is scientific and reasonable,Enterprise's account accounting item is automatically generated using artificial intelligence technology,Utilize artificial intelligence technology,Simulate logical thinking of the accounting personnel when judging subject,Realize that subject judges automatically by machine,Reduce the generation of mistake,Reduce the time-consuming of calculation work,It improves work efficiency,There are vast potential for future development and promotional value.
Description
Technical field
The invention belongs to calculate technical field, and in particular to a kind of intelligence accounting technology.
Background technique
On the one hand, universal in enterprise with financial Shared Service Center, the accounting processing work of enterprise all concentrates on
Sharing Center's processing, this carries out accounting data analysis for enterprise and provides data basis, and another aspect is continuous with information system
Develop and perfect, enterprise can obtain the more and more and relevant data of accounting processing from system, finally, with calculating
The development of machine, artificial intelligence technology, people can carry out network analysis and utilization to accounting data, create unprecedented valence
Value.
However some unreasonable factors, existing accounting enterprise account are remained when existing accounting enterprise's account
When expend the time it is longer, waste a large amount of manpower, reduce working efficiency;Occurs account often when existing accounting enterprise's account
The puzzlement of error brings very big puzzlement to operator, in order to automatically generate enterprise's account accounting using artificial intelligence technology
Subject, the present invention provide a kind of intelligent accounting technology.
Summary of the invention
The purpose of the present invention is to provide a kind of intelligence to calculate technology, mentioned above in the background art existing to solve
Some unreasonable factors are remained when calculating enterprise's account, and it is longer that when existing accounting enterprise's account expends the time, unrestrained
Take a large amount of manpower, reduces working efficiency;Often occurs the puzzlement of account error when existing accounting enterprise's account, to operation
Personnel bring the problem of very big puzzlement.
To achieve the above object, the present invention provides the following technical solutions:A kind of intelligence accounting technology, the accounting method is such as
Under:
Step 1:Using Ansj participle technique by the key in the Electronic Data Processing information and crowdsourcing platform in accounting system
Word (word for as facilitating subject judgement) extracts;
Step 2:Decision-tree model is constructed, includes three layers:(crowdsourcing quotient extraction+OCR knows for Electronic Data Processing information, invoice terms
Not), type of attachment;Each layer all includes corresponding keyword;
Step 3:The keyword of each history reimbursement document is successively passed through three layers of previous step building by interior circulation
Decision-tree model, according to model-naive Bayesian, by the data for needing to verify substitute into following formula f (x)=argmaxykP (yk |
X)=argmaxykP (yk) ∏ ni=1P (xi | yk) ∑ kP (yk) ∏ ni=1P (xi | yk), raw by analyzing these keywords
At a certain section's purpose probability;
Step 4:It include the odd numbers and subject of each history reimbursement document in supervised learning machine, when in decision-tree model
When consistent in the subject and supervised learning machine of generation, the keyword and its shared specific gravity that each layer in the decision tree includes are i.e.
Generate an intelligent accounting criteria, for judging subject, when in the subject and supervised learning machine generated in decision-tree model not
When consistent, supervised learning machine can accordingly adjust keyword and its specific gravity shared in this layer in three layers of decision tree, until raw
At correct subject, an intelligent accounting criteria is then generated;
Step 5:Outer circulation, by the crucial phrase in every single history document at one group of random art without sequencing,
A subject is generated according to words art, and is compared with the correct subject in machine learning, unanimously, then an intelligence is generated and calculates
Rule;It is inconsistent, then assign correct subject to the words art.
As a preferred technical solution of the present invention, in the step 1, accounting system include title, explanation,
Budgetary accounting item, expense item, the accounting item of generation.
It is described Step 1: in two as a preferred technical solution of the present invention, Electronic Data Processing information include title, explanation,
Budgetary accounting item, expense item.
As a preferred technical solution of the present invention, in the step 1, crowdsourcing platform:Each appended original of expense report
The type (abbreviation type of attachment) of beginning voucher, invoice terms.
As a preferred technical solution of the present invention, in the step 2, each layer all includes corresponding keyword,
Such as Electronic Data Processing layer includes title, explanation, budgetary accounting item, expense item;Invoice terms include dinner cost, mineral water etc.;It is attached
Part type includes invoice, contract, inventory etc..
As a preferred technical solution of the present invention, in the step 3, analyzes these keywords and generate a certain subject
Probability, system assert that high that subject of probability is correct subject, during generating subject, the keyword in each layer
This layer of different specific gravity can be accounted for when judging accounting item;With the entrance of a large amount of history reimbursement document, can construct more next
More decision-tree models forms random forest.
As a preferred technical solution of the present invention, the step 5 is pair the supplement Step 3: four, utilizes claimer
Speech habits of the member in reimbursement determine accounting item.
Compared with prior art, the beneficial effects of the invention are as follows:Structure of the invention is scientific and reasonable, safe and convenient to use, benefit
Manually intellectual technology automatically generates enterprise's account accounting item, using artificial intelligence technology, simulates accounting personnel in judgement section
Logical thinking when mesh realizes that subject judges automatically by machine, reduces the generation of mistake, reduce the consumption of calculation work
When, it improves work efficiency, there is vast potential for future development and promotional value.
Specific embodiment
A kind of intelligence accounting technology, includes the following steps:
Step 1:Using Ansj participle technique by the key in the Electronic Data Processing information and crowdsourcing platform in accounting system
Word (word for as facilitating subject judgement) extracts, and accounting system includes title, explanation, budgetary accounting item, fee item
Mesh, generation accounting item, Electronic Data Processing information includes title, explanation, budgetary accounting item, expense item;
Step 2:Decision-tree model is constructed, includes three layers:(crowdsourcing quotient extraction+OCR knows for Electronic Data Processing information, invoice terms
Not), type of attachment;Each layer all includes corresponding keyword, if Electronic Data Processing layer includes title, explanation, budgetary accounting item, expense
Project etc.;Invoice terms include dinner cost, mineral water etc.;Type of attachment includes invoice, contract, inventory etc.;
Step 3:The keyword of each history reimbursement document is successively passed through three layers of previous step building by interior circulation
Decision-tree model, according to model-naive Bayesian, by the data for needing to verify substitute into following formula f (x)=argmaxykP (yk |
X)=argmaxykP (yk) ∏ ni=1P (xi | yk) ∑ kP (yk) ∏ ni=1P (xi | yk), raw by analyzing these keywords
At a certain section's purpose probability, system assert that high subject of probability is correct subject, during generating subject, each layer
In keyword this layer of different specific gravity can be accounted for when judging accounting item;With the entrance of a large amount of history reimbursement document,
More and more decision-tree models can be constructed, random forest is formed;
Step 4:It include the odd numbers and subject of each history reimbursement document in supervised learning machine, when in decision-tree model
When consistent in the subject and supervised learning machine of generation, the keyword and its shared specific gravity that each layer in the decision tree includes are i.e.
Generate an intelligent accounting criteria, for judging subject, when in the subject and supervised learning machine generated in decision-tree model not
When consistent, supervised learning machine can accordingly adjust keyword and its specific gravity shared in this layer in three layers of decision tree, until raw
At correct subject, an intelligent accounting criteria is then generated;
Step 5:Outer circulation, by the crucial phrase in every single history document at one group of random art without sequencing,
A subject is generated according to words art, and is compared with the correct subject in machine learning, unanimously, then an intelligence is generated and calculates
Rule;It is inconsistent, then assign correct subject to the words art, step 5 is pair the supplement Step 3: four, is being reported using reimbursement personnel
Speech habits when pin determine accounting item.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (7)
1. a kind of intelligence calculates technology, it is characterised in that:The accounting method is as follows:
Step 1:Using Ansj participle technique by the keyword in the Electronic Data Processing information and crowdsourcing platform in accounting system
(word for as facilitating subject judgement) extracts;
Step 2:Decision-tree model is constructed, includes three layers:It is Electronic Data Processing information, invoice terms (crowdsourcing quotient extraction+OCR identification), attached
Part type;Each layer all includes corresponding keyword;
Step 3:The keyword of each history reimbursement document is successively passed through three layers of decision of previous step building by interior circulation
Tree-model, according to model-naive Bayesian, by the data for needing to verify substitute into following formula f (x)=argmaxykP (yk | x)=
ArgmaxykP (yk) ∏ ni=1P (xi | yk) ∑ kP (yk) ∏ ni=1P (xi | yk), certain is generated by analyzing these keywords
One section's purpose probability;
Step 4:It include the odd numbers and subject of each history reimbursement document in supervised learning machine, when being generated in decision-tree model
Subject and supervised learning machine in it is consistent when, the keyword and its shared specific gravity that each layer in the decision tree includes generate
One intelligent accounting criteria, for judging subject, when the subject generated in decision-tree model with it is inconsistent in supervised learning machine
When, supervised learning machine can accordingly adjust keyword and its specific gravity shared in this layer in three layers of decision tree, until generating just
Then true subject generates an intelligent accounting criteria;
Step 5:Outer circulation, by the crucial phrase in every single history document at one group of random art without sequencing, according to
It talks about art and generates a subject, and be compared with the correct subject in machine learning, unanimously, then generate an intelligence and calculate rule
Then;It is inconsistent, then assign correct subject to the words art.
2. a kind of intelligence according to claim 1 calculates technology, it is characterised in that:In the step 1, accounting system
System includes title, explanation, budgetary accounting item, expense item, the accounting item of generation.
3. a kind of intelligence according to claim 1 or 2 calculates technology, it is characterised in that:It is described Step 1: in two, zooming
Changing information includes title, explanation, budgetary accounting item, expense item.
4. a kind of intelligence according to claim 1 calculates technology, it is characterised in that:In the step 1, crowdsourcing platform:Often
The type (abbreviation type of attachment) of the appended original certificate of one expense report, invoice terms.
5. a kind of intelligence according to claim 1 calculates technology, it is characterised in that:In the step 2, each layer is all wrapped
Containing corresponding keyword, such as Electronic Data Processing layer includes title, explanation, budgetary accounting item, expense item;Invoice terms include meal
Take, mineral water etc.;Type of attachment includes invoice, contract, inventory etc..
6. a kind of intelligence according to claim 1 calculates technology, it is characterised in that:In the step 3, these passes are analyzed
Keyword generates a certain section's purpose probability, and system assert that high subject of probability is correct subject, during generating subject,
Keyword in each layer can account for this layer of different specific gravity when judging accounting item;With a large amount of history reimbursement document
Into can construct more and more decision-tree models, form random forest.
7. a kind of intelligence according to claim 1 calculates technology, it is characterised in that:The step 5 is pair Step 3: four
Supplement, using reimbursement personnel reimbursement when speech habits determine accounting item.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810519927.4A CN108830699A (en) | 2018-05-28 | 2018-05-28 | A kind of intelligence accounting technology |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810519927.4A CN108830699A (en) | 2018-05-28 | 2018-05-28 | A kind of intelligence accounting technology |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108830699A true CN108830699A (en) | 2018-11-16 |
Family
ID=64145739
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810519927.4A Pending CN108830699A (en) | 2018-05-28 | 2018-05-28 | A kind of intelligence accounting technology |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108830699A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111078921A (en) * | 2019-03-18 | 2020-04-28 | 广东小天才科技有限公司 | Subject identification method and electronic equipment |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101025818A (en) * | 2006-02-21 | 2007-08-29 | 高新生 | Algorithm for computer automatially processing accounting entry |
CN102508860A (en) * | 2011-09-29 | 2012-06-20 | 广州中浩控制技术有限公司 | Data mining method based on XBRL (extensible business reporting language) embodiment document |
KR20140015685A (en) * | 2012-07-10 | 2014-02-07 | 김형남 | Advertise match system and service thereof |
CN104715300A (en) * | 2013-12-12 | 2015-06-17 | 上海莞东拿信息科技有限公司 | Network broker information managing and processing system adopting novel program slicing technology and orienting commercial real estate industry |
CN107209750A (en) * | 2014-11-19 | 2017-09-26 | 里德爱思唯尔股份有限公司雷克萨斯尼克萨斯分公司 | System and method for automatically identifying the potential material facts in document |
-
2018
- 2018-05-28 CN CN201810519927.4A patent/CN108830699A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101025818A (en) * | 2006-02-21 | 2007-08-29 | 高新生 | Algorithm for computer automatially processing accounting entry |
CN102508860A (en) * | 2011-09-29 | 2012-06-20 | 广州中浩控制技术有限公司 | Data mining method based on XBRL (extensible business reporting language) embodiment document |
KR20140015685A (en) * | 2012-07-10 | 2014-02-07 | 김형남 | Advertise match system and service thereof |
CN104715300A (en) * | 2013-12-12 | 2015-06-17 | 上海莞东拿信息科技有限公司 | Network broker information managing and processing system adopting novel program slicing technology and orienting commercial real estate industry |
CN107209750A (en) * | 2014-11-19 | 2017-09-26 | 里德爱思唯尔股份有限公司雷克萨斯尼克萨斯分公司 | System and method for automatically identifying the potential material facts in document |
Non-Patent Citations (3)
Title |
---|
令才汇: "令才人工智能数据助手", 《令才汇》 * |
崇凌,等: "当会计核算遇上人工智能", 《中国会计报》 * |
翁崇凌,等: "会计创新三大法宝", 《智慧财务官》 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111078921A (en) * | 2019-03-18 | 2020-04-28 | 广东小天才科技有限公司 | Subject identification method and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Shen et al. | Bi-directional block self-attention for fast and memory-efficient sequence modeling | |
Clark et al. | Improving coreference resolution by learning entity-level distributed representations | |
CN104572709B (en) | Data digging system for enterprise innovation system | |
CN110968701A (en) | Relationship map establishing method, device and equipment for graph neural network | |
CN105095190B (en) | A kind of sentiment analysis method combined based on Chinese semantic structure and subdivision dictionary | |
CN107704512A (en) | Financial product based on social data recommends method, electronic installation and medium | |
CN113255321B (en) | Financial field chapter-level event extraction method based on article entity word dependency relationship | |
CN101299256A (en) | Method and system for recognizing electric contract based on grid computing and data integration | |
CN109740160A (en) | A kind of task dissemination method based on artificial intelligence semantic analysis | |
Chi et al. | Establish a patent risk prediction model for emerging technologies using deep learning and data augmentation | |
CN114997263B (en) | Method, device, equipment and storage medium for analyzing training rate based on machine learning | |
Zaidi et al. | Learning graph representation of bug reports to triage bugs using graph convolution network | |
Wang et al. | Cross-domain contract element extraction with a bi-directional feedback clause-element relation network | |
Cong et al. | A small sample data-driven method: User needs elicitation from online reviews in new product iteration | |
CN108830699A (en) | A kind of intelligence accounting technology | |
Gong | Network data mining based on artificial intelligence inference engine | |
Tannert et al. | FlowchartQA: the first large-scale benchmark for reasoning over flowcharts | |
CN106934024A (en) | A kind of data processing method and device | |
CN114708073B (en) | Intelligent detection method and device for surrounding mark and serial mark, electronic equipment and storage medium | |
Wen et al. | Few-shot named entity recognition with joint token and sentence awareness | |
CN114003270A (en) | Deep learning-based software security vulnerability repairer recommendation method | |
CN113742495A (en) | Rating characteristic weight determination method and device based on prediction model and electronic equipment | |
Zhou et al. | Intelligent detection on construction project contract missing clauses based on deep learning and NLP | |
Hao et al. | A Novel Public Sentiment Analysis Method Based on an Isomerism Learning Model via Multiphase Processing | |
Salamanos et al. | HyperGraphDis: Leveraging Hypergraphs for Contextual and Social-Based Disinformation Detection |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 601-022, floor 6, No. 15, wanquanzhuang Road, Haidian District, Beijing 100089 Applicant after: BEIJING LINGCAI TECHNOLOGY CO.,LTD. Address before: No. 507, floor 5, building 2, yard 77, Shuangqing Road, Haidian District, Beijing 100080 Applicant before: BEIJING LINGCAI TECHNOLOGY CO.,LTD. |
|
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181116 |