CN110096593A - A method of the outer paging system of building intelligence - Google Patents
A method of the outer paging system of building intelligence Download PDFInfo
- Publication number
- CN110096593A CN110096593A CN201910323073.7A CN201910323073A CN110096593A CN 110096593 A CN110096593 A CN 110096593A CN 201910323073 A CN201910323073 A CN 201910323073A CN 110096593 A CN110096593 A CN 110096593A
- Authority
- CN
- China
- Prior art keywords
- paging system
- outer paging
- keyword
- intelligence
- outgoing call
- 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
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000013135 deep learning Methods 0.000 claims abstract description 7
- 239000000284 extract Substances 0.000 claims description 2
- 238000009825 accumulation Methods 0.000 abstract description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000003466 anti-cipated effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a kind of methods of the outer paging system of building intelligence, its key points of the technical solution are that extracting the keyword that user speaks first, construct keyword identification module, then have outer paging system H based on keyword identification modulei, i is positive integer；Use outer paging system HiIt carries out outgoing call accumulation corpus manually to be marked with keyword search corpus, obtains the intention classification that user speaks；Based on the intention classification, deep learning text classifier training intention assessment disaggregated model y is usedi, outgoing call system H is obtained after the completion of trainingi+1, this process of iterative cycles, so that it may obtain outer paging system Hi+n, n=2,3 ....Whole process is easy to operate, and can constantly recycle and constantly upgrade outer paging system, while using the method manually marked, so that intent classifier is more accurate, the outer paging system of building also can more and more mature and intelligence.
Description
Technical field
This disclosure relates to interactive system field, more particularly to a kind of method of the outer paging system of building intelligence.
Background technique
Artificial intelligence has the tendency that growing in intensity in recent years, is on the one hand to have breakthrough progress because of core algorithm,
Meanwhile the refinement of requirement and the division of labor of the society to working efficiency has been pushed artificial intelligence among the tide of development.Many enterprises
Industry such as real estate, finance, insurance and bank etc. improves working efficiency to reduce enterprises recruit persons for jobs cost, gradually begins to use machine
Device people replaces people to go to complete to sell or pay a return visit task.
When the phone that we are connected to outgoing call robot is more and more, major electric business platform nearly all has intelligent visitor
The problem of clothes, but be clear and be, either outgoing call robot or intelligent customer service, they can solve, is all very limited, in addition to
Outside fixed template, it is difficult to understand for really being intended to and solve the problems, such as user, especially Chinese is of extensive knowledge and profound scholarship, same in short different
The tone, the meaning be exactly it is completely different, to the outer paging system of intelligence be one it is great challenge, thus how outside training smart
Paging system is particularly important.
Summary of the invention
In order to solve the above technical problems, present disclose provides a kind of methods of the outer paging system of building intelligence, with building one
More intelligence can flexibly handle the outer paging system of customer problem, and specific technical solution is as follows:
A method of the outer paging system of building intelligence, comprising:
Step 1, extracts the keyword that user speaks, establishes keyword identification module, constructs outgoing call based on keyword identification module
System Hi, i is positive integer；
Step 2 uses the outer paging system HiOutgoing call is carried out, corpus is accumulated；
Then step 3 is manually marked using corpus described in the keyword search；
Step 4 is based on deep learning text classifier, training intention assessment disaggregated model yi；
Step 5, the intention assessment disaggregated model yiBy outer paging system HiIt is upgraded to outer paging system Hi+1。
As specific embodiment, by the outer paging system Hi+1It is input to described Step 2: being recycled in three, four and five, it can
Obtain Hi+n, n=2,3 ....
As specific embodiment, the intention classification for manually being labeled as speaking to user is labeled classification.
As specific embodiment, the deep learning text classifier include fastText, TextCNN, TextRNN,
RCNN and sequence with attention mechanism are to series model.
As specific embodiment, the outer paging system HiIncluding talking about art template, words art template is multiple words art templates
Set.
To sum up, the beneficial effect of the disclosure is: extracting the keyword that user speaks first, building keyword identifies mould
Block then has outer paging system H based on keyword identification modulei, i is positive integer；Use outer paging system HiIt carries out outgoing call and accumulates corpus,
It with keyword search corpus, is manually marked, obtains the intention classification that user speaks；Based on the intention classification, depth is used
Spend learning text classifier training intention assessment disaggregated model yi, outgoing call system H is obtained after the completion of trainingi+1, iterative cycles this
A process, so that it may obtain outer paging system Hi+n, n=2,3 ....Whole process is easy to operate, and can constantly recycle outgoing call system
System constantly upgrading, while using the method manually marked, so that intent classifier is more accurate, the outer paging system of building also can be more next
It is more mature and intelligent.
Detailed description of the invention
Fig. 1 is disclosure flow diagram.
Specific embodiment
The disclosure is described in further detail below in conjunction with attached drawing.
As shown in Figure 1, extracting the keyword that user speaks, keyword identification module is constructed, keyword identification module is based on
It is then outer paging system Hi, i is positive integer.
Use HiOutgoing call is carried out, is passed through using words art template and the dialogue of user largely accumulates corpus, to the corpus of accumulation
Manually marked.There are many kinds of modes for mark, can be labeled by training pattern, the method being labeled using model
Rapid and convenient, but precision cannot be very high.Using the method heavy workload manually marked, the manpower for needing to expend is more, but this
It is open to be labeled using manual method, corpus can be subjected to keyword search first, then the keyword searched is carried out
Artificial mark, what is not only marked is high-efficient, and speaking to user for pin-point accuracy of energy carries out intent classifier, so that subsequent training
Effect is more preferable.
The classification for being intended to classification is obtained after the completion of artificial mark, is input to deep learning text classifier, training is anticipated
Figure identification model yi, it is based on yiConstruct outer paging system Hi+1.Process before iterative cycles is constantly trained upgrading, will
Obtain more and more intelligent outer paging system Hi+n, n=2,3 ....The intelligent outgoing call systematic comparison obtained in this way is stablized, and
Very precisely.For different industries, different keywords is obtained, constructs the outer paging system of intelligence, energy using disclosed method
Enough outer paging systems for accurately customizing different industries, making accurately intelligent outgoing call is no longer problem.
Such as real estate industry, it is clear that keyword just has house, location, average price, area etc., is based on these keywords and use
Family, which is spoken, constructs keyword identification model, there has been outer paging system Hi, outer paging system is constructed according still further to disclosed method.
If financial industry, the keyword being related to can also be segmented, for example recommend loan, and keyword just has loan, benefit
Breath, the terms of loan, make loans time and mode of repayment etc., if financing keyword be usually financing, year income, the financing that breaks even,
Income height etc., it is again different keyword that collection, which is refunded then, and the final outer paging system of intelligence is constructed based on keyword.
The deep learning text classifier that the disclosure uses includes fastText, TextCNN, TextRNN, RCNN and has
The sequence of attention mechanism arrives series model etc..Equally, outer paging system HiIn include words art template, words art template be also through depth
Neural metwork training forms.
The above are disclosure exemplary embodiment, the protection scope of the disclosure is limited by claims and its equivalent.
Claims (5)
1. a kind of method of the outer paging system of building intelligence characterized by comprising
Step 1, extracts the keyword that user speaks, establishes keyword identification module, constructs outgoing call based on keyword identification module
System Hi, i is positive integer；
Step 2 uses the outer paging system HiOutgoing call is carried out, corpus is accumulated；
Then step 3 is manually marked using corpus described in the keyword search；
Step 4 is based on deep learning text classifier, training intention assessment disaggregated model yi；
Step 5, the intention assessment disaggregated model yiBy outer paging system HiIt is upgraded to outer paging system Hi+1。
2. a kind of method of the outer paging system of building intelligence as described in claim 1, which is characterized in that by the outer paging system
Hi+1It is input to described Step 2: being recycled in three, four and five, H can be obtainedi+n, n=2,3 ....
3. the method for the outer paging system of a kind of building intelligence as described in claim 1, which is characterized in that it is described it is artificial be labeled as to
The intention classification that family is spoken is labeled classification.
4. a kind of method of the outer paging system of building intelligence as described in claim 1, which is characterized in that the deep learning text
Classifier includes fastText, TextCNN, TextRNN, RCNN and sequence with attention mechanism to series model.
5. a kind of method of the outer paging system of building intelligence as described in benefit requires 1, which is characterized in that the outer paging system HiIncluding
Talk about art template.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910323073.7A CN110096593A (en) | 2019-04-22 | 2019-04-22 | A method of the outer paging system of building intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910323073.7A CN110096593A (en) | 2019-04-22 | 2019-04-22 | A method of the outer paging system of building intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110096593A true CN110096593A (en) | 2019-08-06 |
Family
ID=67445480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910323073.7A Pending CN110096593A (en) | 2019-04-22 | 2019-04-22 | A method of the outer paging system of building intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110096593A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111897589A (en) * | 2020-06-27 | 2020-11-06 | 中国建设银行股份有限公司 | Intelligent outbound system configuration method and device and electronic equipment |
| CN113159901A (en) * | 2021-04-29 | 2021-07-23 | 天津狮拓信息技术有限公司 | Method and device for realizing financing lease service session |
| CN113239154B (en) * | 2021-06-01 | 2023-08-01 | 平安科技（深圳）有限公司 | Quality inspection method and device for seat conversation operation, computer equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7853451B1 (en) * | 2003-12-18 | 2010-12-14 | At&T Intellectual Property Ii, L.P. | System and method of exploiting human-human data for spoken language understanding systems |
| CN108363690A (en) * | 2018-02-08 | 2018-08-03 | 北京十三科技有限公司 | Dialog semantics Intention Anticipation method based on neural network and learning training method |
| CN109104534A (en) * | 2018-10-22 | 2018-12-28 | 北京智合大方科技有限公司 | A kind of system for improving outgoing call robot and being intended to Detection accuracy, recall rate |
| CN109446306A (en) * | 2018-10-16 | 2019-03-08 | 浪潮软件股份有限公司 | Task-driven multi-turn dialogue-based intelligent question and answer method |
| CN109493166A (en) * | 2018-10-23 | 2019-03-19 | 深圳智能思创科技有限公司 | A kind of construction method for e-commerce shopping guide's scene Task conversational system |
-
2019
- 2019-04-22 CN CN201910323073.7A patent/CN110096593A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7853451B1 (en) * | 2003-12-18 | 2010-12-14 | At&T Intellectual Property Ii, L.P. | System and method of exploiting human-human data for spoken language understanding systems |
| CN108363690A (en) * | 2018-02-08 | 2018-08-03 | 北京十三科技有限公司 | Dialog semantics Intention Anticipation method based on neural network and learning training method |
| CN109446306A (en) * | 2018-10-16 | 2019-03-08 | 浪潮软件股份有限公司 | Task-driven multi-turn dialogue-based intelligent question and answer method |
| CN109104534A (en) * | 2018-10-22 | 2018-12-28 | 北京智合大方科技有限公司 | A kind of system for improving outgoing call robot and being intended to Detection accuracy, recall rate |
| CN109493166A (en) * | 2018-10-23 | 2019-03-19 | 深圳智能思创科技有限公司 | A kind of construction method for e-commerce shopping guide's scene Task conversational system |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111897589A (en) * | 2020-06-27 | 2020-11-06 | 中国建设银行股份有限公司 | Intelligent outbound system configuration method and device and electronic equipment |
| CN111897589B (en) * | 2020-06-27 | 2023-06-09 | 建信金融科技有限责任公司 | Intelligent outbound system configuration method and device and electronic equipment |
| CN113159901A (en) * | 2021-04-29 | 2021-07-23 | 天津狮拓信息技术有限公司 | Method and device for realizing financing lease service session |
| CN113159901B (en) * | 2021-04-29 | 2024-06-04 | 天津狮拓信息技术有限公司 | Method and device for realizing financing lease business session |
| CN113239154B (en) * | 2021-06-01 | 2023-08-01 | 平安科技（深圳）有限公司 | Quality inspection method and device for seat conversation operation, computer equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109241255B (en) | Intention identification method based on deep learning | |
| CN110232114A (en) | Sentence intension recognizing method, device and computer readable storage medium | |
| CN107818164A (en) | A kind of intelligent answer method and its system | |
| CN110096593A (en) | A method of the outer paging system of building intelligence | |
| CN104978587B (en) | A kind of Entity recognition cooperative learning algorithm based on Doctype | |
| CN108038208B (en) | Training method and device of context information recognition model and storage medium | |
| CN105095444A (en) | Information acquisition method and device | |
| CN111476023A (en) | Method and device for identifying entity relationship | |
| CN102314417A (en) | Method for identifying Web named entity based on statistical model | |
| CN104462600A (en) | Method and device for achieving automatic classification of calling reasons | |
| CN107844558A (en) | The determination method and relevant apparatus of a kind of classification information | |
| CN106529525A (en) | Chinese and Japanese handwritten character recognition method | |
| CN111695338A (en) | Interview content refining method, device, equipment and medium based on artificial intelligence | |
| CN107886231A (en) | The QoS evaluating method and system of customer service | |
| CN110321564A (en) | A kind of more wheel dialogue intension recognizing methods | |
| CN111428012B (en) | Intelligent question-answering method, device, equipment and storage medium based on attention mechanism | |
| CN115994317A (en) | Incomplete multi-view multi-label classification method and system based on depth contrast learning | |
| CN111462752A (en) | Client intention identification method based on attention mechanism, feature embedding and BI-L STM | |
| CN112989023A (en) | Label recommendation method, device, equipment, storage medium and computer program product | |
| CN116010581A (en) | Knowledge graph question-answering method and system based on power grid hidden trouble shooting scene | |
| CN112417132A (en) | New intention recognition method for screening negative samples by utilizing predicate guest information | |
| CN116071150A (en) | Data processing method, bank product popularization, wind control system, server and medium | |
| CN117520503A (en) | Financial customer service dialogue generation method, device, equipment and medium based on LLM model | |
| CN116860953B (en) | Question-answer matching method and system based on question-answer system | |
| CN110782221A (en) | Intelligent interview evaluation system and method |
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: 20190806 |