CN114742366A - Intelligent work order distribution method based on big data algorithm - Google Patents

Intelligent work order distribution method based on big data algorithm Download PDF

Info

Publication number
CN114742366A
CN114742366A CN202210271066.9A CN202210271066A CN114742366A CN 114742366 A CN114742366 A CN 114742366A CN 202210271066 A CN202210271066 A CN 202210271066A CN 114742366 A CN114742366 A CN 114742366A
Authority
CN
China
Prior art keywords
work order
undertaking
key
phrase
content
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
Application number
CN202210271066.9A
Other languages
Chinese (zh)
Inventor
周洁琴
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing Inspector Intelligent Technology Co ltd
Original Assignee
Nanjing Inspector Intelligent Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing Inspector Intelligent Technology Co ltd filed Critical Nanjing Inspector Intelligent Technology Co ltd
Priority to CN202210271066.9A priority Critical patent/CN114742366A/en
Publication of CN114742366A publication Critical patent/CN114742366A/en
Pending legal-status Critical Current

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
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06311—Scheduling, planning or task assignment for a person or group
    • 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
    • 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)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Strategic Management (AREA)
  • Tourism & Hospitality (AREA)
  • Data Mining & Analysis (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Game Theory and Decision Science (AREA)
  • Educational Administration (AREA)
  • Quality & Reliability (AREA)
  • Databases & Information Systems (AREA)
  • Development Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a work order intelligent distribution method based on a big data algorithm, which comprises the following steps of 1, acquiring all processed historical work order data of n undertaking units, constructing an initial keyword library of the n undertaking units, 2, generating key phrases according to the initial keyword library, and constructing a key phrase library; step 3, training a BERT classification model according to the historical work order: and 4, receiving the citizen appeal content by the hotline, and distributing the appeal content to the undertaking unit according to a distribution rule. The method is based on the intelligent work order dispatching process of the big data, the work order dispatching speed and the work order dispatching accuracy can be rapidly improved, and by giving out key words and key phrases, the work order dispatching method is beneficial for undertaking units and hotline workers to rapidly know the appeal content of citizens, and the work order processing speed is improved.

Description

Intelligent work order distribution method based on big data algorithm
Technical Field
The invention relates to the field of social governance and natural language processing research, in particular to a work order intelligent distribution method based on a big data algorithm.
Background
In the traditional hot-line worksheet work, workers are required to determine according to the description content of the worksheet and past processing experience, assign different appeal worksheets to corresponding undertaking units, and solve the appeal reflected by people through the handling of the undertaking units. Along with the development of society, the work order acceptance amount is more and more huge, a large amount of manpower, material resources and financial resources are consumed in the traditional allocation operation, and the accuracy rate is not too high. According to the work order processing flow, if the assigned undertaking unit is not accurate, rollback and reassignment are executed, and the process greatly increases the processing time of the work order.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides the intelligent work order dispatching method based on the big data algorithm. The technical scheme is as follows:
in a first aspect, a work order intelligent distribution method based on a big data algorithm is provided, which mainly comprises the following steps:
step 1, obtaining all processed historical work order data of n undertaking units, wherein the historical work order data comprises appeal content of citizens and processing results of the undertaking units, performing word segmentation and part-of-speech tagging on the historical work order content of the n undertaking units by using an LAC word segmentation tool, selecting words with important part-of-speech, extracting keywords by using a TF-IDF algorithm, and constructing an initial keyword library K of the n undertaking units, wherein the K is { K ═ K { (K) } K { (K-IDF) } K1,K2,..,Kn}。
Step 2, generating key phrases according to the initial key word library K, and constructing a key phrase library K '═ K'1,K′2,..,K′n}; when the phrase is constructed, a binary phrase structure of two key word combinations and a ternary phrase structure of three key word combinations are adopted.
The method specifically comprises the following steps: constructing a binary phrase structure and a ternary phrase structure according to a keyword library K, and calculating point mutual information PMI of two or three words, wherein the point mutual information PMI can reflect whether the two or three words have correlation, if the PMI is higher, the two or three words often appear together and have strong correlation, and the PMI calculation formula is as follows:
Figure BDA0003553277600000021
Figure BDA0003553277600000022
wherein p (a)1,a2) Denotes a1And a2Probability that two keywords appear simultaneously as binary phrases in all work orders handled by the undertaking, p (b)1,b2,b3) Denotes b1、b2And b3The probability that three keyword words are used as ternary phrases to appear in all work orders handled by the undertaking unit at the same time. p (a)1)、p(a2)、p(b1)、p(b1)、p(b1) Respectively represent key words a1、a2、b1、b2、b3Probability of occurrence in the work order of the undertaking unit, if PMI (a)1,a2) If the value is larger than the threshold value 1, the key phrase is used as a binary key phrase; if PMI (b)1,b2,b3) If the value is larger than the threshold value 2, the three-element key phrase is used as a ternary key phrase, and if the ternary key phrase contains a binary key phrase, the binary key phrase is deleted, and a key phrase library K' is constructed.
Step 3, training a BERT classification model according to the historical work order: respectively taking 70-90% of appeal content of the historical work order and processing results of all undertaking units as training sets to carry out model training; and (5) taking the appeal content of the residual (30% -10%) historical work order as a test set to perform model test.
Step 4, receiving the citizen appeal content by the hotline, and distributing the appeal content to the undertaking unit according to a distribution rule, wherein the method specifically comprises the following steps:
(1) after citizens dial hot lines to reflect problems, the hot lines can record voice information, and voice data is converted into text information through a voice recognition method to obtain work order content.
Performing word segmentation through an LAC word segmentation tool, selecting words with important part of speech, removing stop words, extracting key words through TF-IDF, constructing a binary phrase structure and a ternary phrase structure, judging whether the constructed binary phrase or ternary phrase belongs to the key phrase library K ' according to the key phrase library K ' of all undertaking units, directly taking the binary phrase or ternary phrase as a key phrase of work order content if the constructed binary phrase or ternary phrase belongs to the key phrase library K ', deleting the key words forming the phrase structure to avoid giving repeated information, and if the ternary key phrase in the work order content contains the key phrases in the work orderDeleting the binary key phrase, inputting the work order content according to the trained BERT model, and giving a undertaking unit O with the probability value ranking three above1,O2,O3。
(2) Respectively calculating the keywords, key phrases and undertaking units O of the work order content1,O2,O3Keyword bank K1,K2,K3And Key phrase library K'1,K′2,K′3According to the probability values of the three undertaking units and the corresponding similarity values, different schemes are selected and distributed to the undertaking units.
Work order x and undertaking organization OiThe similarity method is as follows:
Figure BDA0003553277600000031
wherein
Figure BDA0003553277600000032
The representation of the ith key phrase is,
Figure BDA0003553277600000033
represents the jth keyword;
Figure BDA0003553277600000034
representing key phrases
Figure BDA0003553277600000035
Whether the occurrence is in a key phrase library K'iIf the number is 1, otherwise, the number is 0;
Figure BDA0003553277600000036
representing keywords
Figure BDA0003553277600000037
Whether or not to appear in keyword library KiIf the occurrence is marked as 1, otherwise, the occurrence is 0.
Preferably, in step 1, if a certain keyword appears in more than 30% -60% of the undertaking units at the same time, the keyword is deleted as the stop word.
Preferably, the hot wire in step 4 receives the content of the citizen appeal, and the text material can be submitted through a network or a face-to-face text material and the like.
Preferably, in order to further reduce the calculation amount, step 4 (2) is replaced by the following method:
if the undertaking unit O1And undertaking unit O2And if the probability difference is more than 0.2-0.4 (preferably 0.3), directly sending to the undertaking unit with the first probability value, and giving out the keywords and the key phrases.
If the undertaking unit O1And undertaking unit O2The probability difference is less than 0.2-0.4 (preferably 0.3), and the acceptance unit O2And undertaking unit O3If the probability difference is greater than 0.2-0.4 (preferably 0.3), the keywords, key phrases and undertaking units O of the work order content are respectively calculated1And O2Keyword library K1,K2Key phrase library K'1,K′2The similarity of the key phrases is distributed to the undertaking units with higher similarity, and keywords and key phrases are given; if the similarity is the same, the hot-line staff is sent.
If the undertaking unit O1And undertaking unit O2The probability difference is less than 0.2-0.4 (preferably 0.3), and the supporting unit O2And undertaking unit O3The probability difference is less than 0.2-0.4 (preferably 0.3), and the supporting unit O1If the probability of (2) is greater than 0.1, calculating the keywords, key phrases and undertaking units O of the work order content1、O2And O3Keyword library K1,K2,K3Key phrase library K'1,K′2,K′3The text similarity is distributed to the undertaking units with higher similarity, and keywords and key phrases are given; and if the similarity is the same, the hot-line staff is dispatched.
If the condition is other, the information is dispatched to the hotline staff, keywords and key phrases are given, and the information is audited and dispatched manually.
Preferably, the step 4 further comprises directly dispatching the work order to a corresponding department if the work order is returned after dispatching and the returned content contains the 'recommended xx department'; and if the returned content does not contain the recommendation information, the returned content is dispatched to the hotline staff and manually dispatched.
Further, the step 4 further comprises the steps of retraining the BERT model by taking citizen appeal content and return content returned by the work order as training data, and performing iterative optimization on the BERT model.
Compared with the prior art, one of the technical schemes has the following beneficial effects: the work orders are classified and distributed through a big data classification algorithm BERT model and a similarity method, so that the labor consumption is reduced, and the work order distribution speed and distribution accuracy are improved; by constructing a key word library and a key phrase library, the citizen appeal content is subjected to key word and key phrase extraction and is used as key information to be sent to undertaking units or hotline workers, so that the citizen appeal content can be quickly known, and the worksheet processing is accelerated.
Detailed Description
In order to clarify the technical solution and the working principle of the present invention, the embodiments of the present disclosure will be described in further detail below. All the above optional technical solutions may be combined arbitrarily to form the optional embodiments of the present disclosure, and are not described herein again.
The terms "step 1," "step 2," "step 3," and the like in the description and claims of this application are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It should be understood that the data so used may be interchanged under appropriate circumstances such that the embodiments of the application described herein may be practiced in sequences other than those described herein.
In a first aspect: the embodiment of the disclosure provides a work order intelligent distribution method based on a big data algorithm, which mainly comprises the following steps:
step 1, acquiring all processed historical work order data of n undertaking units, wherein the historical work order data comprises appeal content of citizens and processing results of the undertaking units, and performing word segmentation and part-of-speech tagging on the historical work order content of the n undertaking units by using an LAC word segmentation toolNote that words with important parts of speech are selected, and words with unimportant parts of speech, such as prepositions, quantifiers and the like, are deleted. Extracting keywords by using TF-IDF algorithm, and constructing an initial keyword library K of n undertaking units1,K2,..,Kn}。
Preferably, in the step 1, if a certain keyword appears in more than 30% -60% of undertaking units at the same time, deleting the keyword as a stop word; if the segmented words appear in a plurality of undertaking units, the segmented words are low in importance to the content of the historical work order and do not help to judge the undertaking units.
Step 2, generating key phrases according to the initial key word library K, and constructing a key phrase library K '═ K'1,K′2,..,K′n}；
Because most phrases have binary and ternary structures, the phrase construction adopts a binary phrase structure of two keyword combinations and a ternary phrase structure of three keyword combinations.
The method specifically comprises the following steps: and constructing a binary phrase structure and a ternary phrase structure according to the keyword library K, and calculating point mutual information PMI of two or three words, wherein the point mutual information PMI can reflect whether the two or three words have correlation or not, if the PMI is higher, the two or three words often appear together and have strong correlation, and the possibility of being used as a phrase is higher. The PMI calculation formula is as follows:
Figure BDA0003553277600000051
Figure BDA0003553277600000052
wherein p (a)1,a2) Denotes a1And a2Probability of two keywords as binary phrases appearing simultaneously in all work orders handled by the undertaking, p (b)1,b2,b3) Is shown by b1、b2And b3The probability that three keyword words as ternary phrases appear in all work orders handled by undertaking units at the same time. p (a)1)、p(a2)、p(b1)、p(b1)、p(b1) Respectively represent keywords a1、a2、b1、b2、b3Probability of occurrence in the work order of the undertaking unit, if PMI (a)1,a2) If the value is larger than the threshold value 1, the key phrase is used as a binary key phrase; if PMI (b)1,b2,b3) If the value is larger than the threshold value 2, the three-element key phrase is used as a ternary key phrase, and if the ternary key phrase contains a binary key phrase, the binary key phrase is deleted, repeated information is avoided being given, and a key phrase library K' is constructed.
Step 3, training a BERT classification model according to the historical work order: respectively taking 70-90% of appeal content of the historical work order and processing results of all undertaking units as training sets to carry out model training; and (4) taking the appeal content of the rest (30% -10%) of the historical work orders as a test set to perform model test. Because the processing result of the undertaking unit is not available in the actual dispatching process, the processing result of the undertaking unit is only used as a training set and not used as a testing set.
And 4, receiving the citizen appeal content by the hotline, and distributing the appeal content to the undertaking unit according to a distribution rule. The method comprises the following specific steps:
(1) after citizens dial hot lines to reflect problems, the hot lines can record voice information, and voice data is converted into text information through a voice recognition method to obtain work order content.
Dividing words through an LAC word dividing tool, selecting words with important word characteristics, stopping using words, extracting keywords through TF-IDF, constructing a binary phrase structure and a ternary phrase structure, judging whether the constructed binary phrase or ternary phrase belongs to the key phrase library K ' according to the key phrase library K ' of all undertaking units, directly taking the binary phrase or ternary phrase as the key phrase of work order content if the constructed binary phrase or ternary phrase belongs to the key phrase library K ', deleting the keywords forming the phrase structure to avoid giving repeated information, and deleting the binary key phrase to avoid giving the repeated information if the ternary key phrase in the work order content contains the binary key phrase in the work order content. After the processing, the keywords and the key phrases in the appeal content are obtained. According to the trainingThe trained BERT model inputs the work order content and gives a undertaking unit O with the probability value ranked in the top three1,O2,O3。
(2) Respectively calculating the keywords, key phrases and undertaking units O of the work order content1,O2,O3Keyword library K1,K2,K3And Key phrase library K'1,K′2,K′3According to the probability values of the three undertaking units and the corresponding similarity values, different schemes are selected and distributed to the undertaking units.
Work order x and undertaking organization OiThe similarity method is as follows:
Figure BDA0003553277600000061
wherein
Figure BDA0003553277600000062
The (i) th key phrase is represented,
Figure BDA0003553277600000063
represents the jth keyword;
Figure BDA0003553277600000064
representing key phrases
Figure BDA0003553277600000065
Whether the occurrence is in a key phrase library K'iIf the number is 1, otherwise, the number is 0;
Figure BDA0003553277600000066
representing keywords
Figure BDA0003553277600000067
Whether it appears in the keyword library KiIf the occurrence is marked as 1, otherwise, the occurrence is 0.
Preferably, the hot line in step 4 receives the content of the citizen appeal, and the modes of submitting the text material through the network, interviewing the text material and the like can be replaced.
Preferably, in order to further reduce the amount of calculation, step 4 (2) is replaced by the following method:
if the undertaking unit O1And undertaking unit O2And if the probability difference is more than 0.2-0.4 (preferably 0.3), directly sending to the undertaking unit with the first probability value, and giving out the keywords and the key phrases.
If the undertaking unit O1And undertaking unit O2The probability difference is less than 0.2-0.4 (preferably 0.3), and the supporting unit O2And undertaking unit O3If the probability difference is greater than 0.2-0.4 (preferably 0.3), the keywords, key phrases and undertaking units O of the work order content are calculated respectively1And O2Keyword library K1,K2Key phrase library K'1,K′2The similarity of the keywords is distributed to the undertaking units with higher similarity, and the keywords and the key phrases are given; if the similarity is the same, the hot-line staff is sent.
If the undertaking unit O1And undertaking unit O2The probability difference is less than 0.2-0.4 (preferably 0.3), and the supporting unit O2And undertaking unit O3The probability difference is less than 0.2-0.4 (preferably 0.3), and the acceptance unit O1If the probability of (2) is greater than 0.1, calculating the keywords, key phrases and undertaking units O of the work order content1、O2And O3Keyword library K1,K2,K3Key phrase library K'1,K′2,K′3The text similarity is distributed to the undertaking units with higher similarity, and keywords and key phrases are given; if the similarity is the same, the information is distributed to the hotline staff.
If the condition is other, the information is dispatched to the hotline staff, keywords and key phrases are given, and the information is audited and dispatched manually.
Preferably, the step 4 further comprises directly dispatching the work order to a corresponding department if the work order is returned after dispatching and the returned content contains the 'recommended xx department'; and if the returned content does not contain the recommendation information, the returned content is dispatched to the hotline staff and manually dispatched.
Further, the step 4 further comprises the steps of retraining the BERT model by taking citizen appeal content and return content returned by the work order as training data, and performing iterative optimization on the BERT model.
The invention has been described above by way of example, it is obvious that the specific implementation of the invention is not limited by the above-described manner, and that various insubstantial modifications are possible using the method concepts and technical solutions of the invention; or directly apply the conception and the technical scheme of the invention to other occasions without improvement and equivalent replacement, and the invention is within the protection scope of the invention.

Claims (6)

1. An intelligent work order dispatching method based on a big data algorithm is characterized by mainly comprising the following steps:
step 1, obtaining all processed historical work order data of n undertaking units, wherein the historical work order data comprises appeal content of citizens and processing results of the undertaking units, performing word segmentation and part-of-speech tagging on the historical work order content of the n undertaking units by using an LAC word segmentation tool, selecting words with important parts-of-speech, extracting keywords by using a TF-IDF algorithm, and constructing an initial keyword bank K ═ { K ═ K { K } of the n undertaking units1,K2,..,Kn}；
Step 2, generating key phrases according to the initial key word library K, and constructing a key phrase library K '═ K'1,K′2,..,K′n}; when the phrase is constructed, a binary phrase structure of two key word combinations and a ternary phrase structure of three key word combinations are adopted;
the method specifically comprises the following steps: constructing a binary phrase structure and a ternary phrase structure according to a keyword library K, and calculating point mutual information PMI of two or three words, wherein the point mutual information PMI can reflect whether the two or three words have correlation, if the PMI is higher, the two or three words often appear together and have strong correlation, and the PMI calculation formula is as follows:
Figure FDA0003553277590000011
Figure FDA0003553277590000012
wherein p (a)1,a2) Denotes a1And a2Probability of two keywords as binary phrases appearing simultaneously in all work orders handled by the undertaking, p (b)1,b2,b3) Is shown by b1、b2And b3Probability that three keyword words appear simultaneously in all work orders handled by the undertaking unit as a ternary phrase, p (a)1)、p(a2)、p(b1)、p(b1)、p(b1) Respectively represent keywords a1、a2、b1、b2、b3Probability of occurrence in the work order of the undertaking unit, if PMI (a)1,a2) If the value is larger than the threshold value 1, the key phrase is used as a binary key phrase; if PMI (b)1,b2,b3) If the value is larger than the threshold value 2, the three-element key phrase is used as a ternary key phrase, and if the ternary key phrase contains a binary key phrase, the binary key phrase is deleted, and a key phrase library K' is constructed;
step 3, training a BERT classification model according to the historical work order: respectively taking 70-90% of appeal content of the historical work order and processing results of all undertaking units as training sets to carry out model training; taking the appeal content of the residual (30% -10%) historical work order as a test set, and carrying out model test;
step 4, receiving the citizen appeal content by the hotline, and distributing the appeal content to the undertaking unit according to a distribution rule, wherein the method specifically comprises the following steps:
(1) after citizens dial hot lines to reflect problems, the hot lines can record voice information, and voice data is converted into text information through a voice recognition method to obtain work order content;
performing word segmentation through an LAC word segmentation tool, selecting words with important parts of speech, removing stop words, extracting key words through TF-IDF, constructing a binary phrase structure and a ternary phrase structure, and judging whether the constructed binary phrase or ternary phrase belongs to a key phrase library K ' according to key phrase libraries K ' of all undertaking units 'If the work order belongs to the work order, the binary phrase or the ternary phrase is directly used as a key phrase of the work order content, keywords forming a phrase structure are deleted, repeated information is avoided being given, if the ternary key phrase in the work order content contains the binary key phrase in the work order content, the binary key phrase is deleted, the work order content is input according to a trained BERT model, and a undertaking unit O with the probability value of being ranked first three is given1,O2,O3；
(2) Respectively calculating the keywords, key phrases and undertaking units O of the work order content1,O2,O3Keyword bank K1,K2,K3And Key phrase library K'1,K′2,K′3According to the probability values of the three undertaking units and the corresponding similarity values, different schemes are selected and distributed to the undertaking units;
work order x and undertaking order OiThe similarity method is as follows:
Figure FDA0003553277590000021
wherein
Figure FDA0003553277590000022
The (i) th key phrase is represented,
Figure FDA0003553277590000023
represents the jth keyword;
Figure FDA0003553277590000024
representing key phrases
Figure FDA0003553277590000025
Whether the occurrence is in a key phrase library K'iIf the number is 1, otherwise, the number is 0;
Figure FDA0003553277590000026
indicating the criticalityWord
Figure FDA0003553277590000027
Whether it appears in the keyword library KiIf the occurrence is marked as 1, otherwise, the occurrence is 0.
2. The intelligent work order dispatching method based on big data algorithm as claimed in claim 1, wherein in step 1, if a certain keyword occurs in more than 30% -60% of the undertaking units at the same time, the keyword is deleted as a stop word.
3. The intelligent work order distribution method based on the big data algorithm as claimed in claim 1, wherein the hot wire in step 4 receives the content of the citizen's appeal, and can be replaced by submitting the text material through the network or by way of surface-handing the text material.
4. The intelligent work order distribution method based on the big data algorithm as claimed in any one of claims 1-3, wherein step 4 (2) is replaced by the following method:
if the undertaking unit O1And undertaking unit O2If the probability difference is more than 0.2-0.4 (preferably 0.3), directly sending to the undertaking unit with the first probability value, and giving out keywords and key phrases;
if the undertaking unit O1And undertaking unit O2The probability difference is less than 0.2-0.4 (preferably 0.3), and the supporting unit O2And undertaking unit O3If the probability difference is greater than 0.2-0.4 (preferably 0.3), the keywords, key phrases and undertaking units O of the work order content are calculated respectively1And O2Keyword library K1,K2Key phrase library K'1,K′2The similarity of the key phrases is distributed to the undertaking units with higher similarity, and keywords and key phrases are given; if the similarity is the same, the hot-line staff is dispatched;
if the undertaking unit O1And undertaking unit O2The probability difference is less than 0.2-0.4 (preferably 0.3), and the supporting unit O2And undertaking unit O3The probability difference is less than 0.2-0.4 (excellent)0.3) is selected, and the undertaking unit O1If the probability of (2) is greater than 0.1, calculating the keywords, key phrases and undertaking units O of the work order content1、O2And O3Keyword library K1,K2,K3Key phrase library K'1,K′2,K′3The text similarity is distributed to the undertaking units with higher similarity, and keywords and key phrases are given; if the similarity is the same, the hot-line staff is dispatched;
if the condition is other, the information is dispatched to the hotline staff, keywords and key phrases are given, and the information is audited and dispatched manually.
5. The intelligent work order dispatching method based on the big data algorithm as claimed in claim 4, wherein step 4 further comprises directly dispatching the work order to a corresponding department if the work order is returned after dispatching and the returned content contains the "recommended xx department"; and if the returned content does not contain the recommendation information, the returned content is dispatched to the hotline staff and manually dispatched.
6. The intelligent work order distribution method based on big data algorithm as claimed in claim 5, wherein step 4 further comprises retraining the BERT model by using citizen appeal content and return content of work order return as training data, and performing iterative optimization on the BERT model.
CN202210271066.9A 2022-03-18 2022-03-18 Intelligent work order distribution method based on big data algorithm Pending CN114742366A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210271066.9A CN114742366A (en) 2022-03-18 2022-03-18 Intelligent work order distribution method based on big data algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210271066.9A CN114742366A (en) 2022-03-18 2022-03-18 Intelligent work order distribution method based on big data algorithm

Publications (1)

Publication Number Publication Date
CN114742366A true CN114742366A (en) 2022-07-12

Family

ID=82277856

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210271066.9A Pending CN114742366A (en) 2022-03-18 2022-03-18 Intelligent work order distribution method based on big data algorithm

Country Status (1)

Country Link
CN (1) CN114742366A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116029492A (en) * 2022-12-01 2023-04-28 广州云趣信息科技有限公司 Order sending method and device
CN117273314A (en) * 2023-09-11 2023-12-22 中通服网络信息技术有限公司 Rule base-based method for matching and consignment of underwriting units and automatic dispatching

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116029492A (en) * 2022-12-01 2023-04-28 广州云趣信息科技有限公司 Order sending method and device
CN116029492B (en) * 2022-12-01 2023-12-01 广州云趣信息科技有限公司 Order sending method and device
CN117273314A (en) * 2023-09-11 2023-12-22 中通服网络信息技术有限公司 Rule base-based method for matching and consignment of underwriting units and automatic dispatching

Similar Documents

Publication Publication Date Title
CN114742366A (en) Intelligent work order distribution method based on big data algorithm
CN107886231B (en) Service quality evaluation method and system for customer service
CN112419096B (en) Automatic circulation method of user electricity appeal work order based on NLP information extraction and few-sample self-learning
CN112163424A (en) Data labeling method, device, equipment and medium
CN110188047A (en) A kind of repeated defects report detection method based on binary channels convolutional neural networks
CN107833059B (en) Service quality evaluation method and system for customer service
WO2021036439A1 (en) Method for responding to complaint, and device
CN111429184A (en) User portrait extraction method based on text information
WO2023071120A1 (en) Method for recognizing proportion of green assets in digital assets and related product
CN113590807A (en) Scientific and technological enterprise credit evaluation method based on big data mining
CN115878778A (en) Natural language understanding method facing business field
CN110196897A (en) A kind of case recognition methods based on question and answer template
CN117172508B (en) Automatic dispatch method and system based on city complaint worksheet recognition
CN110347812A (en) A kind of search ordering method and system towards judicial style
CN114418327A (en) Automatic order recording and intelligent order dispatching method for customer service system
CN114330318A (en) Method and device for recognizing Chinese fine-grained entities in financial field
CN112036179B (en) Electric power plan information extraction method based on text classification and semantic frame
CN113240562A (en) Method and system for recommending and matching obstetrical and academic research projects based on nlp
CN107480126A (en) A kind of engineering material classification intelligent identification Method
CN115357718B (en) Method, system, device and storage medium for discovering repeated materials of theme integration service
CN107886233B (en) Service quality evaluation method and system for customer service
WO2023071129A1 (en) Method for identifying proportion of green assets and related product
WO2022227196A1 (en) Data analysis method and apparatus, computer device, and storage medium
CN115952282A (en) Intelligent bank customer complaint diversion handling method and system based on NLP technology
CN110414819B (en) Work order scoring 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