CN111340334B - Intelligent work order assignment method, system and medium - Google Patents
Intelligent work order assignment method, system and medium Download PDFInfo
- Publication number
- CN111340334B CN111340334B CN202010088310.9A CN202010088310A CN111340334B CN 111340334 B CN111340334 B CN 111340334B CN 202010088310 A CN202010088310 A CN 202010088310A CN 111340334 B CN111340334 B CN 111340334B
- Authority
- CN
- China
- Prior art keywords
- employee
- work order
- keyword
- vocabulary
- employees
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000009826 distribution Methods 0.000 claims abstract description 9
- 238000013507 mapping Methods 0.000 claims abstract description 5
- 239000013598 vector Substances 0.000 claims description 54
- 239000011159 matrix material Substances 0.000 claims description 42
- 238000004422 calculation algorithm Methods 0.000 claims description 31
- 238000012549 training Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 11
- 238000005065 mining Methods 0.000 claims description 7
- 238000012546 transfer Methods 0.000 claims description 7
- 238000004590 computer program Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- 238000013016 damping Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 5
- 238000004458 analytical method Methods 0.000 claims description 4
- 238000003860 storage Methods 0.000 claims description 4
- 125000004432 carbon atom Chemical group C* 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 31
- 238000012423 maintenance Methods 0.000 abstract description 6
- 206010063385 Intellectualisation Diseases 0.000 abstract description 4
- 239000000284 extract Substances 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 9
- 230000005611 electricity Effects 0.000 description 7
- 230000008901 benefit Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000015572 biosynthetic process Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004904 shortening Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000011158 quantitative evaluation Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
- 230000007306 turnover Effects 0.000 description 1
Images
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
- G06Q10/063114—Status monitoring or status determination for a person or group
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Health & Medical Sciences (AREA)
- Marketing (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- Public Health (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an intelligent work order assignment method, system and medium, the invention analyzes the content of the work order i to be assigned and extracts the keyword to obtain the extracted keyword set Bi(ii) a According to the extracted keyword set BiMapping to a to-be-selected employee list in the enterprise employee set H; determining target staff from the staff list to be selected, and assigning a work order BiThe invention is distributed to the target staff, is carried out in a closed loop around the generation and receiving processing of the work order, carries out intellectualization on the distribution of the previously manually processed work order, can greatly shorten the processing time of the intermediate link of the work order, improves the automation efficiency of the work order processing and the satisfaction degree of the client to the operation and maintenance of the company business, and can be used for processing the submission of the client business requirement and the receiving and processing of the business work order by the staff.
Description
Technical Field
The invention relates to a digitization, informatization and intelligent processing technology of an electric power system, in particular to an intelligent work order allocation method, system and medium, which are used for processing client business requirement submission and receiving and processing business work orders by staff.
Background
With the coming of the information intelligence era, the digitalization, informatization and intellectualization of the power system are continuously developed, and the quantity of the generated power consumption demand worksheets also shows the increase of geometric progression. Due to the characteristics of large data volume, complex content form and the like, the current environmental requirements cannot be met by using the traditional work order service mode. At present, the processing of the work order with the power demand still stays in the traditional receiving, transferring and replying process, namely, the work order is mainly processed through means such as a local business outlet, a 95598 service hotline, a staff telephone and the like, the work order has a longer turnover period, the customer demand cannot be fed back in time, and the phenomenon of disjointed power consumption and power supply links to a certain extent exists. Therefore, how to implement automatic and intelligent assignment of intelligent work orders has become a key technical problem to be solved urgently.
In the current work order system, a large amount of human intervention and intermediate links are needed for transmitting most work orders, so that a large amount of human resources are consumed, the work order processing efficiency is low, and the business development success of a company is influenced.
Disclosure of Invention
The technical problems to be solved by the invention are as follows: the invention is carried out in a closed loop around the generation and receiving processing of the work order, and carries out intellectualization on the work order assignment processed manually in the past, thereby greatly shortening the processing time of the intermediate link of the work order, improving the automation efficiency of work order processing and the satisfaction degree of customers on company business operation and maintenance, and being used for processing the business requirement submission of the customers and the business work order received and processed by the staff.
In order to solve the technical problems, the invention adopts the technical scheme that:
an intelligent work order assignment method comprises the following implementation steps:
1) performing content analysis on the work order i to be dispatched and extracting keywords to obtain an extracted keyword set Bi;
2) According to the extracted keyword set BiMapping to a list of employees to be selected in the enterprise employee set H;
3) determining target staff from the staff list to be selected, and assigning a work order B iAssigned to the target employee.
Optionally, the detailed steps of step 1) include:
1.1) dividing work order content of a work order i to be assigned into sentences by using a sentence separator, segmenting the sentences by using a segmenter to obtain text vocabularies, and filtering the text vocabularies;
1.2) carrying out association rule mining on text vocabularies, defining any vocabularies A and B contained in the same sentence as association relations { A, B }, carrying out association rule joint iteration on the whole text, and generating a maximum frequent vocabulary item set Q, wherein the maximum frequent vocabulary item set Q comprises one or more frequent item sets, and each frequent item set represents a group of keyword sets with association relations generated after the association rule mining;
1.3) generating a corresponding sliding window parameter K according to the frequent item sets with different sizes in the maximum frequent vocabulary item set Q, and generating an adjusted sliding window parameter K ' ═ L '/L) K, wherein L is the sentence length, and L ' is the adjusted sentence length;
1.4) calculating the respective words siThe contained information is associated with an entropy value;
1.5) taking the adjusted sliding window parameter K' as the sliding window of the TextRank algorithm, and calculating each vocabulary siThe included information is associated with entropy as vocabulary s iIterative damping coefficient in the TextRank algorithm, and iterating and calculating each vocabulary s through traversing the maximum frequent vocabulary item set Q by the TextRank algorithmiTo all words s to be converged to a confidence threshold intervaliThe T(s) value sequence of (1) is sorted, and N vocabularies s with top T(s) ranking are selectediAs keywords for text passages.
Optionally, the filtering the text vocabulary in step 1.1) includes: eliminating text vocabularies without relevant semantics, marking vocabulary attributes and keeping the text vocabularies including verbs, nouns and adjectives.
Optionally, step 1.4) calculating the vocabulary siThe functional expression of the associated entropy values involved is shown as follows:
in the above-mentioned formula, the compound has the following structure,is a word siThe contained information is related to entropy, X is a vocabulary,is a word siThe probability of occurrence in the jth frequent item set in the maximum frequent vocabulary item set Q,n is the number of the frequent itemsets of the maximum frequent vocabulary item set Q.
Optionally, the detailed steps of step 2) include:
2.1) according to the extracted keyword set BiAnd generating a keyword vector P of the work order i to be dispatched from a total keyword set B extracted from all the work ordersiThe keyword vector PiThe number of the elements in (B) is the same as that of the keywords in the total keyword set B if the keywords in the total keyword set B are contained in the keyword set B iMiddle key word vector PiThe value of the corresponding element in (1), otherwise the keyword vector PiThe value of the corresponding element in (1) is 0;
2.2) searching for a regular function meeting the following formula based on the label variables corresponding to all the employees in the enterprise employee set H and the adjustment label vectors thereofAdding the staff corresponding to the maximum value into a list of the staff to be selected;
in the above formula, PiFor work order B to be assignediS is the employee in the historical worksheet data training set, l is the total number of employees in the enterprise employee set H, Qj TCentralizing employee H for enterprise employee HjTag variable Q ofjTransposed matrix of (2), Qs TCentralizing employee H for enterprise employee HsTag variable Q ofsThe transpose matrix of (a) is,centralizing employee H for enterprise employee HjAdjusted label vector QjThe transpose matrix of' is then,centralizing employee H for enterprise employee HsAdjusted label vector QsThe transposed matrix of.
Optionally, before the step 2.2), a step of generating a tag variable corresponding to any employee in the enterprise employee set H and an adjustment tag vector thereof is further included:
s1) selecting part or all of the work orders from all the work orders to establish a training set, and establishing a relation matrix R according to the distribution relation between the work orders and the employees in the training set:
in the above formula, with rijRepresenting elements in a relation matrix R, wherein i ranges from 1 to k, j ranges from 1 to l, k is the number of work orders, l is the number of employees, and if the element R is ijA value equal to 1 indicates that the corresponding work order i is assigned to employee j, if element rijIf the work order is equal to 0, the corresponding work order i is not distributed to the employee j;
s2) calculating an arbitrary employee H according to the following equationjTag variable Q ofj;
In the above formula, Hj TTo represent employee HjVector of (a), r1j~rkjRespectively indicate whether the 1 st to k th worksheets are allocated to the employee HjA result of (1) or (0) and a value of 1 indicates assignment to employee Hj(ii) a P denotes a set of keyword vectors, P1~PkA keyword vector representing 1 to k work orders; k is the total number of all work orders.
S3) calculating any employee H in the enterprise employee set H according to P (i) ═ i/njThe keyword occurrence frequencies f (1) -f (k) of all k work orders in the training set, wherein i represents the occurrence frequency of the keyword, and n represents the total number of all the keywords; for the tag variable Q according tojCorrecting to obtain an adjusted label vector Qj’;
In the above formula, f (1) to f (k) are employee HjKeyword frequency of occurrence, Q, for all k work ordersjFor employee HjThe tag variable of (1).
Optionally, the step 3) of determining the target employee from the list of employees to be selected specifically means: and calculating the task quantity of each employee according to the complex coefficient and the quantity of the distributed work orders of each employee in the list of the employees to be selected, and then selecting the employee with the minimum task quantity as the target employee.
Optionally, the work order B to be dispatched in the step 3)iThe method also comprises a step of automatic or manual task transfer distribution after the target employees are assigned, wherein the step of task transfer distribution comprises the steps of finding out the employees with the best affinity relationship of the target employees from an affinity relationship matrix D between the employees in a preset enterprise employee set H as new target employees, and assigning a work order B to be assignediTo a new target employee; the generating step of the affinity relationship matrix D comprises the following steps: firstly, respectively calculating semantic distances between labels of any two employees, and then generating an affinity relationship matrix D according to the following formula;
in the above formula, with dijRepresenting elements in the intimacy relationship matrix D, wherein the value range of i and j is 1-l, l is the number of employees, and the element DijRepresenting the semantic distance between the tags of employee i, employee j; the tags of the staff comprise one or more of post work types, service objects, situations and regions.
In addition, the invention also provides an intelligent work order dispatching system, which comprises a computer device, wherein the computer device is programmed or configured to execute the steps of the intelligent work order dispatching method, or a computer program which is programmed or configured to execute the intelligent work order dispatching method is stored on a memory of the computer device.
Furthermore, the present invention also provides a computer readable storage medium having stored thereon a computer program programmed or configured to execute the intelligent work order dispatching method.
Compared with the prior art, the invention has the following advantages: the invention analyzes the content of the work order i to be dispatched and extracts the keywords to obtain an extracted keyword set Bi(ii) a According to the extracted keyword set BiMapping to a list of employees to be selected in the enterprise employee set H; determining target staff from the staff list to be selected, and assigning a work order BiThe invention is distributed to the target staff, is carried out in a closed loop around the generation and receiving processing of the work order, carries out intellectualization on the distribution of the previously manually processed work order, can greatly shorten the processing time of the intermediate link of the work order, improves the automation efficiency of the work order processing and the satisfaction degree of the client to the operation and maintenance of the company business, and can be used for processing the submission of the client business requirement and the receiving and processing of the business work order by the staff.
Drawings
FIG. 1 is a schematic diagram of a basic flow of a method according to an embodiment of the present invention.
FIG. 2 is an undirected graph of a TextRank algorithm model in an embodiment of the invention.
FIG. 3 is a directed graph of a TextRank algorithm model in an embodiment of the invention.
FIG. 4 is a comparison graph of PR curves for the keyword "document" in the embodiment of the present invention.
FIG. 5 is a comparison graph of PR curves for the keyword "pay" in accordance with the present invention.
FIG. 6 is a diagram illustrating a ternary relationship between client, middle factor and employee in an embodiment of the present invention.
FIG. 7 is a service model constructed based on a ternary relationship according to the present embodiment.
Detailed Description
As shown in fig. 1, the implementation steps of the intelligent work order assignment method of the embodiment include:
1) performing content analysis on the work order i to be dispatched and extracting keywords to obtain an extracted keyword set Bi;
2) According to the extracted keyword set BiMapping to a list of employees to be selected in the enterprise employee set H;
3) determining target staff from the staff list to be selected, and assigning a work order BiAssigned to the target employee.
In this embodiment, the detailed steps of step 1) include:
1.1) dividing work order content of a work order i to be assigned into sentences by using a sentence separator, segmenting the sentences by using a segmenter to obtain text vocabularies, and filtering the text vocabularies;
1.2) carrying out association rule mining on text vocabularies, defining any vocabularies A and B contained in the same sentence as association relations { A, B }, carrying out association rule joint iteration on the whole text, and generating a maximum frequent vocabulary item set Q, wherein the maximum frequent vocabulary item set Q comprises one or more frequent item sets, and each frequent item set represents a group of keyword sets with association relations generated after the association rule mining;
1.3) generating a corresponding sliding window parameter K according to frequent item sets with different sizes in the maximum frequent vocabulary item set Q, and generating an adjusted sliding window parameter K ═ L '/L (where L is the sentence length and L' is the adjusted sentence length);
1.4) calculating the respective vocabulary siThe contained information is associated with an entropy value;
1.5) taking the adjusted sliding window parameter K' as the sliding window of the TextRank algorithm, and calculating each vocabulary siThe included information is associated with entropy as vocabulary siIterative damping coefficient in the TextRank algorithm, and iterating and calculating each vocabulary s through traversing the maximum frequent vocabulary item set Q by the TextRank algorithmiT(s) value sequence of (a) for all words s to converge to a confidence threshold intervaliThe value sequence of T(s) is sorted, and N vocabularies s with top rank of T(s) are selectediAs keywords for the text passage.
The extraction of keywords in text data mining is the basic work for relevant research, and the quality of the extraction effect is directly related to the reading of text content information. Because the TextRank algorithm has simple and effective correlation with weak languageAnd the like, which have become widely used keyword extraction methods at present, but are greatly influenced by word frequency, so that the method has no great advantages compared with other keyword extraction algorithms. The embodiment introduces the incidence relation between the information entropy and the vocabulary on the basis of the existing TextRank algorithm, solves the single characteristic of the vocabulary word frequency, better considers the global and context semantic characteristics of the vocabulary, and can obtain better extraction effect. The TextRank algorithm firstly utilizes the participle to segment the text, eliminates the vocabulary with useless habits by taking 'sentence' as a unit, and reserves verbs, nouns and the like in the vocabulary as a candidate keyword option set Q i={Si1,Si2,…,SinWhere i represents the sentence number. Through setting the value of the parameter K of the sliding window, the words with the distance less than K in the sentence are generalized to the same window for marking, namely { s0,s1,…,sk}{s1,s2,…,sk+1},…,{sn-k,s1,…,snAnd establishing a link relation between the words appearing in the same window to form a corresponding keyword link matrix. The TextRank algorithm model mainly comprises an undirected graph and a directed graph, which are specifically shown in FIG. 2 and FIG. 3.
Wherein, the matrix expression of the undirected graph is as follows:
the formula for iteratively calculating the t(s) value sequence for an undirected graph is as follows:
wherein, the matrix expression of the directed graph is as follows:
the formula for iteratively computing the t(s) value sequence for a directed graph is as follows:
in the formulae (2) and (4), T (S)i) Is a word siD is a damping coefficient, In (S)i) Representation and vocabulary SiLinked vocabulary set, out (S)j) Representing a directional vocabulary SjSet of words of (1), mjkIs a word Sk,SjWeight of edges between, mjiIs a word Si,SjWeight of edges between, SkFor directing to the vocabulary SjThe vocabulary set out (S)j) The words in (1). In the embodiment, the document is traversed, iterated and converged to a certain interval range, so that the T(s) value sequences of all the words are obtained, and the words with the value ranking at the top are selected as keywords of the text.
In this embodiment, the filtering the text vocabulary in step 1.1) includes: eliminating text vocabularies without relevant semantics, marking vocabulary attributes and keeping the text vocabularies comprising verbs, nouns and adjectives.
In this embodiment, step 1.4) calculates the vocabulary siThe functional expression of the associated entropy values contained is shown below:
in the above-mentioned formula, the compound has the following structure,is a word siThe contained information is related to entropy value, X is vocabulary,is a word siAnd the probability of the occurrence in the jth frequent item set in the maximum frequent vocabulary item set Q is n, and the number of the frequent item sets in the maximum frequent vocabulary item set Q is n. In massive text contents, the information entropy can carry out quantitative evaluation on the text, and vocabularies are expressedThe expected value of the information contained in the whole text can be obtained according to the amount of the carried information. Generally, for a certain vocabulary, if the information entropy value of the vocabulary is larger, the distribution of the vocabulary in the whole text is more even, which indicates that the probability of the occurrence of the vocabulary is larger, and the vocabulary comprises three characteristics of monotonicity, nonnegativity and accumulation. In the embodiment, the text keyword information is obtained by calculation by adopting the associated entropy concept, so that the importance of vocabulary frequency is weakened, and a more accurate keyword extraction effect can be achieved for the vocabulary context information. According to the obtained associated entropy information of the vocabulary, the information of the degree of linkage between the vocabulary and other vocabularies can be obtained, namely the degree of damping between the vocabulary and other associated vocabularies. If the vocabulary appears less frequently in the text, but the vocabulary often appears in more important positions in the text structure (such as higher degree of information associated with high frequency vocabulary), it indicates that the importance of the low frequency vocabulary in the text should be improved, i.e. the weight of the vocabulary associated entropy information is enhanced, so that the low frequency and important vocabulary can be better processed. In the embodiment, by extracting the texts of the categories such as the text set of the dog hunting, the text set of the university of the double denier and the like, such as the history, law, education, art, computer and the like as the experiment samples, the number of the samples in each category is kept about 1000. Before processing the experimental text, firstly, a Chinese word segmentation device is used for preprocessing the experimental text, stop words and useless symbols are removed, and the part of speech of the words is labeled. Selecting an experiment operating environment: the operating system is win 7, kernel Intel Core i7 CPU M540, and memory 4G. For these documents, the number of keywords is basically kept around 5, and therefore the number of extracted keywords is set to 5 and 6, and the specific experimental effects are shown in tables 1 and 2 below.
Table 1: five keywords are extracted.
Table 2: six keywords are extracted.
Experimental data show that, under the condition of a large amount of text data information, compared with the TFIDF algorithm, the TR algorithm and the Tag-TR algorithm, the keyword extraction algorithm for improving the associated entropy provided by the embodiment can achieve a better effect. For the explanation of text content, the context meaning is not only reflected in the frequency of the occurrence of the words, but also the positions of the occurrence of the words in the context and the association between the words have great significance, so that the association information of the words in the text is calculated for the selection of the keywords. In order to further observe the characteristics of the algorithm in the keyword extraction process, the cutoff values of the keywords in different types of documents (such as the keyword "document" and the keyword "pay") are extracted, and PR curves represented by the cutoff values are shown in fig. 4 and fig. 5, wherein a curve "TFIDF" represents a TFIDF algorithm for comparison, a curve "TR" represents a TR algorithm for comparison, a curve "Tag-TR" represents a Tag-TR algorithm for comparison, and a curve "this paper" represents the method of the embodiment. As can be seen from fig. 4 and 5, the area of the curve and the coordinate axis in this embodiment is the largest, which indicates that the keyword extraction effect in this embodiment has certain advantages compared with the TFIDF algorithm, the TR algorithm, and the Tag-TR algorithm, and a better keyword extraction effect can be obtained by mining the association relationship among the words, performing quantitative analysis on the association entropy, and combining TR iterative computation to select k words ranked earlier as keyword candidates.
In this embodiment, the detailed steps of step 2) include:
2.1) according to the extracted keyword set BiAnd generating a keyword vector P of the work order i to be dispatched from a total keyword set B extracted from all the work ordersiThe keyword vector PiThe number of the elements in (B) is the same as that of the keywords in the total keyword set B if the keywords in the total keyword set B are contained in the keyword set BiMiddle key word vector PiThe value of the corresponding element in (1), otherwise the keyword vector PiThe value of the corresponding element in (1) is 0;
2.2) searching for a regular function meeting the following formula based on the label variables corresponding to all the employees in the enterprise employee set H and the adjustment label vectors thereofAdding the staff corresponding to the maximum value into a list of the staff to be selected;
in the above formula, PiFor work order B to be assignediS is the employee in the historical worksheet data training set, l is the total number of employees in the enterprise employee set H, Qj TCentralizing employee H for enterprise employee HjTag variable Q ofjTransposed matrix of (2), Qs TCentralizing employee H for enterprise employee HsTag variable Q ofsThe transpose matrix of (a) is,centralizing employee H for enterprise employee HjAdjusted label vector QjThe transpose matrix of' is then,centralizing employee H for enterprise employee HsAdjusted label vector QsThe transposed matrix of.
In this embodiment, before the step 2.2), the method further includes the steps of generating a tag variable corresponding to any employee in the enterprise employee set H and adjusting a tag vector thereof:
s1) selecting part or all of the work orders from all the work orders to establish a training set, and establishing a relation matrix R according to the distribution relation between the work orders and the employees in the training set:
in the above formula, with rijRepresenting elements in a relational matrix R, where i takes on a valueThe range is 1-k, the value range of j is 1-l, k is the number of work orders, l is the number of employees, if the element rijA value equal to 1 indicates that the corresponding work order i is assigned to employee j, if element rijIf the work order is equal to 0, the corresponding work order i is not distributed to the employee j;
s2) calculating an arbitrary employee H according to the following formulajTag variable Q ofj;
In the above formula, Hj TTo represent employee HjVector of (a), r1j~rkjRespectively indicate whether the 1 st to k th worksheets are allocated to the employee HjA result of (1) or (0) and a value of 1 indicates assignment to employee Hj(ii) a P denotes a set of keyword vectors, P1~PkA keyword vector representing 1 to k work orders; k is the total number of all work orders.
S3) calculating any employee H in the enterprise employee set H according to P (i) ═ i/njThe keyword occurrence frequencies f (1) -f (k) of all k work orders in the training set, wherein i represents the occurrence frequency of the keyword, and n represents the total number of all the keywords; for the tag variable Q according to jCorrecting to obtain an adjusted label vector Qj’;
In the above formula, f (1) to f (k) are employee HjKeyword frequency of occurrence, Q, for all k work ordersjFor employee HjThe tag variable of (1).
In the power industry, the electricity utilization behavior of customers is similar to the purchasing of ' commodities ', and the good quality of the commodities ' is related to the practical product experience of the customers. The "electricity" of a commodity has not only the general inherent commodity attribute but also the characteristics of intangibility, dependence on the use of a third party, and the like due to the particularity of the commodity. There are two main bodies in the production and use of electric power commodities, one is a customer who uses electric power, and the other is an enterprise who provides and maintains electric power. In the whole circulation process of the 'electric power commodity', the actual contact between a client and an enterprise is the most, and the nature and the characteristics of the 'electric power commodity' are not concerned much. Thus, the main part of the recommendation algorithm is mainly oriented between the client and the enterprise when being introduced. The connection between the customer power utilization and the enterprise power supply can be called as power supply service, and the requirement of the customer on the electric power commodity is mainly transferred to the enterprise. For an enterprise, the basic elements of the enterprise are employees in the enterprise, and the quality of an employee team is directly related to the business benefits of the enterprise. For the enterprise staff, because the post work types, the service objects and the like are different, the enterprise staff can carry out labeling according to the difference of the post work types, the service objects and the like. The label can reflect the evaluation of the client on the opinion of the enterprise staff, can also be a word which directly reflects the working attribute of the enterprise staff, and of course, any kind of label is a summary of the information related to the stock of the enterprise staff, and is not limited to a single working post of the enterprise staff, so that an association channel can be established with the electricity client through the label. Factors such as situation, region and the like can be introduced when describing the electricity demand of the client and the tag characteristics of the staff, and the effect that the recommendation service is influenced due to insufficient information can be compensated to a certain extent. The labeling processing of the enterprise employees can not only calculate the similarity of the clients by utilizing the semantic distance between the tag sets fed back by the clients to the electricity utilization service, but also recommend the clients and serve the employees with high degree of fitness by the relationship between the tags and the work order content. In traditional research content, students usually combine tags with a recommendation algorithm based on content to solve the sparsity problem frequently occurring in the recommendation algorithm to a certain extent, so that the advantages of the recommendation algorithm based on the tags are proved, but the recommendation accuracy is not enough, so that the customer satisfaction is low, and the power supply service quality is influenced. In order to recommend enterprise employees to clients more accurately, an intermediate transition matrix is added between the clients and the employees, and an original client-employee relationship matrix R is decomposed into a client-intermediate factor relationship matrix P and an intermediate factor-employee relationship matrix Q, so that a ternary relationship of the clients, the intermediate factors and the employees shown in FIG. 6 is formed, and a plurality of clients form associations with a plurality of enterprise employees through k intermediate factors.
For the relation matrix of the client and the intermediate factor, the main body of the relation matrix is mainly composed of a keyword vector, and the relation matrix of the intermediate factor and the enterprise employee is composed of a label vector. Compared with the previous single client-employee relationship, the client-intermediate factor-employee ternary relationship can find a more accurate recommendation relationship, and can accurately meet the requirements of a real environment. The embodiment sets up a service model from three major relations of the client and the intermediate factor, the interior of the intermediate factor, the intermediate factor and the employee, and mainly comprises a client-intermediate factor model, an intermediate factor recommendation model and an intermediate factor-employee model. Fig. 7 shows a service model constructed in the present embodiment based on three major relationships, namely, client-middle factor model, enterprise employee and tag vector, and middle factor-employee model, in which the recommendation relationship between the keyword vector and the tag vector constitutes a middle factor recommendation model.
The content described in the step 2.1) is the client-intermediary factor model, and the specific analysis meaning of the work order formed according to the power consumption requirement of the client is different due to the different content, so that the corresponding keyword vectors are different. The total keyword set extracted from all the customer electricity demand worksheet training sets is B ═ B 1,b2,...,bnAnd a keyword set B of a single power demand worksheetiThen a subset of data set B, namely B, should be presentiE.g. B. According to the keywords contained in the single customer work order and the corresponding relationship in the total keyword set B, the corresponding keyword vector PiAny element P of (a) is:
if P ═ 0, it means that the customer demand work order does not contain the keyword, and if P ═ 1, it means that the customer demand work order contains the keyword.
The steps S1) to S3) are described as an intermediate factor-employee model, and H is known from the relationship matrix R between the dispatch worksheets and the employees in the training setjThe corresponding vectors are:
where k is the basis for the quantity of work orders. Due to the work order BiCorresponding key word vector Pi=[pi1,pi2,...,pin]And n is the number of keywords. The employee H is obtained by calculationjTag variable Q ofj. In order to achieve a better practical effect, according to the frequency f (i) of the keywords appearing in the training work order set, the importance degree of the keywords with high appearance probability is higher than that of the keywords with low appearance probability (after the invalid phrases are removed). Variable Q in tagjBased on the keyword appearance frequency f (i) as an adjustment factor, calculating to obtain an adjustment label vector Q'j。
The step 2.2) is described as an intermediate factor recommendation model. The recommendation of the intermediate factor is particularly important in order to be able to correspond the work order set with the employee set. And a certain relation exists between the keyword vector and the label vector, the new keyword vector is recommended in return from the label vector obtained by the learning in the training set, and the employee with the highest conformity degree is found, so that the work order is distributed to the employee. For work order B iIn other words, the key vector is Pi=[pi1,pi2,...,pin]Finding the most suitable staff HjServing, i.e. presence tag vector QjAnd adjust tag vector Q'jThe maximum value of the regular function delta phi is satisfied. That is, employee H is presentjThe condition of the power utilization service work for processing the work order content is met. Staff meeting conditions of electricity utilization service work for processing work order contentThe number of the employees can be one or more, so that the employees need to be taken as a list of the employees to be selected so as to determine the target employees from the list of the employees to be selected.
In this embodiment, the step 3) of determining the target employee from the to-be-selected employee list specifically means: and calculating the task quantity of each employee according to the complex coefficient and the quantity of the distributed work orders of each employee in the list of the employees to be selected, and then selecting the employee with the minimum task quantity as the target employee. The complexity coefficient is set according to the complexity of the work order, for example, the complexity coefficient can be calculated according to the length of the solution time of the staff with general proficiency, one hour is taken as a standard unit, the complexity degree of the long time is high, and the complexity degree of the short time is low. Thus employee workload ═ number of work orders ═ complexity. According to the existing workload of the staff, the assignment weight of the staff work order is adjusted (the relation which is inversely related to the workload is formed, namely the probability of reassigning the work order with large workload is smaller, and the probability of reassigning the work order with small workload is larger), so that the balance of work order distribution can be realized.
Because the label vector set of the staff is obtained by learning from the training set, the characteristics of sample localization, semantic diversity and the like still exist, and certain uncertain factors exist in the staff, the accuracy of the recommended service cannot be completely ensured. In order to enhance the flexible adjustment of the service work and provide the staff with the assistance recommendation service based on the recommendation service, the work order B to be assigned in step 3) of the embodimentiAfter the assignment to the target employees, the method also comprises the step of automatic or manual task transfer assignment, wherein the step of task transfer assignment comprises the steps of finding out the employee with the best intimacy relationship of the target employees from an intimacy relationship matrix D between the employees in a preset enterprise employee set H as a new target employee, and assigning a work order B to be assignediTo a new target employee; the generating step of the affinity relationship matrix D comprises the following steps: firstly, respectively calculating semantic distances between labels of any two employees, and then generating an affinity relationship matrix D according to the following formula;
in the above formula, with dijRepresenting elements in the intimacy relationship matrix D, wherein the value range of i and j is 1-l, l is the number of employees, and the element DijRepresenting the semantic distance between the tags of employee i, employee j; the tags of the staff comprise one or more of post work types, service objects, situations and regions. If the employee state or on-duty condition matched with the recommendation model does not meet the power demand of the work order, the work order can be automatically transferred according to the requirement, and the label of the employee can be adjusted in time according to the intimacy matrix; staff or administrators can also perform manual transfers as needed. Therefore, the propagation mode of the work order is changed from the conventional central regulation mode to the multipoint decentralized regulation mode, the staff operation resources are greatly utilized, the work order processing efficiency is improved, and the service quality is improved.
To sum up, in order to better solve the inconsistency of the linking steps between the client and the enterprise power supply service, the intelligent work order dispatching method of the embodiment starts with the client, the staff and the connection mechanism between the client and the staff, designs a processing system capable of intelligently processing the work order, receives and analyzes the requirements of the client in the form of the work order through a uniform simulation window, adopts an efficient dispatching method, finds the staff with the highest fitness through continuous recommendation and simulation for service, greatly reduces the transmission and circulation of intermediate links, ensures that each work order can be dispatched in time, accurately conveys the requirements to the corresponding service staff, reduces the manual intervention, improves the working efficiency of the staff, saves the human resources of the staff of the power supply enterprise, improves the handling efficiency of the power utilization service to the greatest extent, and provides work order reply, The functions of reminding, forwarding and the like can better solve the optimization problem of the mode of 'last kilometer'. The invention establishes an intelligent maintenance work order dispatching management technology based on a dynamic planning technology around the formation mode of a customer work order, the recommendation matching problem between employees and the work order, the feedback of the customer to the employee processing work order and the like, adopts digital storage, automatic data matching, system implementation and other technical methods to expand standardized management and digital conversion, completes transverse coordination in the operation and maintenance process in the form of a work list, and combines various types of work list operation and maintenance personnel with the dynamic programming technology on the basis of longitudinal data analysis, thereby realizing the association of a cooperative work list, personnel and work instructions. The method can greatly improve the processing efficiency of the work order, build a direct bridge between the client and the staff and promote the improvement of the human resource efficiency of the company. The invention can realize direct contact between the client and the processing staff, greatly reduces unnecessary intermediate transmission links from the formation of the client work order to the processing of the client work order by the staff, and the intermediate links almost do not need manual telephone notification, transmission and the like, thereby shortening the time limit of the whole work order processing, realizing zero-distance contact between the client and the first-line staff, knowing the most urgent business requirements of the client and forming the most effective staff solution.
In addition, the present embodiment also provides an intelligent work order dispatching system, which includes a computer device programmed or configured to execute the steps of the intelligent work order dispatching method, or a computer program programmed or configured to execute the intelligent work order dispatching method is stored in a memory of the computer device.
Furthermore, the present embodiment also provides a computer-readable storage medium having stored thereon a computer program programmed or configured to execute the aforementioned intelligent work order assignment method.
The above description is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may occur to those skilled in the art without departing from the principle of the invention, and are considered to be within the scope of the invention.
Claims (9)
1. An intelligent work order assignment method is characterized by comprising the following implementation steps:
1) content analysis and keyword extraction are carried out on the work order i to be dispatchedObtaining an extracted keyword set Bi;
2) According to the extracted keyword set B iMapping to a list of employees to be selected in the enterprise employee set H;
3) determining target staff from the staff list to be selected, and assigning a work order BiAssigning to a target employee;
the detailed steps of the step 1) comprise:
1.1) dividing work order content of a work order i to be assigned into sentences by using a sentence separator, segmenting the sentences by using a segmenter to obtain text vocabularies, and filtering the text vocabularies;
1.2) carrying out association rule mining on text vocabularies, defining any vocabularies A and B contained in the same sentence as association relations { A, B }, carrying out association rule joint iteration on the whole text, and generating a maximum frequent vocabulary item set Q, wherein the maximum frequent vocabulary item set Q comprises one or more frequent item sets, and each frequent item set represents a group of keyword sets with association relations generated after the association rule mining;
1.3) generating a corresponding sliding window parameter K according to the frequent item sets with different sizes in the maximum frequent vocabulary item set Q, and generating an adjusted sliding window parameter K ' ═ L '/L) K, wherein L is the sentence length, and L ' is the adjusted sentence length;
1.4) calculating the respective words siThe contained information is associated with an entropy value;
1.5) taking the adjusted sliding window parameter K' as the sliding window of the TextRank algorithm, and calculating each vocabulary s iThe information associated entropy contained is taken as a vocabulary siIterative damping coefficient in the TextRank algorithm, and iterating and calculating each vocabulary s through traversing the maximum frequent vocabulary item set Q by the TextRank algorithmiTo all words s to be converged to a confidence threshold intervaliThe T(s) value sequence of (1) is sorted, and N vocabularies s with top T(s) ranking are selectediAs keywords for text passages.
2. The intelligent work order assignment method as claimed in claim 1, wherein the filtering of text vocabulary in step 1.1) comprises: eliminating text vocabularies without relevant semantics, marking vocabulary attributes and keeping the text vocabularies comprising verbs, nouns and adjectives.
3. The intelligent work order assignment method as claimed in claim 1, wherein step 1.4) calculates vocabulary siThe functional expression of the associated entropy values contained is shown below:
in the above formula, the first and second carbon atoms are,is a word siThe contained information is related to entropy, X is a vocabulary,is a word siAnd the probability of the occurrence in the jth frequent item set in the maximum frequent vocabulary item set Q is n, and the number of the frequent item sets in the maximum frequent vocabulary item set Q is n.
4. The intelligent work order assignment method as claimed in claim 1, wherein the detailed steps of step 2) include:
2.1) according to the extracted keyword set BiAnd generating a keyword vector P of the work order i to be dispatched from a total keyword set B extracted from all the work ordersiThe keyword vector PiThe number of the elements in (B) is the same as that of the keywords in the total keyword set B if the keywords in the total keyword set B are contained in the keyword set BiMiddle key word vector PiThe value of the corresponding element in (1), otherwise the keyword vector PiThe value of the corresponding element in (1) is 0;
2.2) searching for a regular function meeting the following formula based on the label variables corresponding to all the employees in the enterprise employee set H and the adjustment label vectors thereofAdding the staff corresponding to the maximum value into a list of the staff to be selected;
in the above formula, PiFor work order B to be assignediS is the employee in the historical worksheet data training set, l is the total number of employees in the enterprise employee set H, Qj TCentralizing employee H for enterprise employee HjTag variable Q ofjTransposed matrix of (2), Qs TCentralizing employee H for enterprise employee HsTag variable Q ofsTransposed matrix of, Q'j TCentralizing employee H for enterprise employee HjAdjusted label vector Qj'transpose matrix, Q's TCentralizing employee H for enterprise employee HsAdjusted label vector QsThe transposed matrix of.
5. The intelligent work order assignment method according to claim 4, wherein step 2.2) is preceded by the step of generating a label variable corresponding to any employee in the enterprise employee set H and an adjustment label vector thereof:
S1) selecting part or all of the work orders from all the work orders to establish a training set, and establishing a relation matrix R according to the distribution relation between the work orders and the employees in the training set:
in the above formula, with rijRepresenting elements in a relation matrix R, wherein i ranges from 1 to k, j ranges from 1 to l, k is the number of work orders, l is the number of employees, and if the element R isijA value equal to 1 indicates that the corresponding work order i is assigned to employee j, if element rijEqual to 0 indicates that the corresponding work order i is not dividedAn assignment employee j;
s2) calculating an arbitrary employee H according to the following formulajTag variable Q ofj;
In the above formula, Hj TTo represent employee HjVector of (a), r1j~rkjRespectively indicate whether the 1 st to k th worksheets are allocated to the employee HjA result of (1) or (0) and a value of 1 indicates assignment to employee Hj(ii) a P denotes a set of keyword vectors, P1~PkA keyword vector representing 1 to k work orders; k is the total number of all work orders;
s3) calculating any employee H in the enterprise employee set H according to P (i) ═ i/njThe keyword occurrence frequencies f (1) -f (k) of all k work orders in the training set, wherein i represents the occurrence frequency of the keyword, and n represents the total number of all the keywords; for the tag variable Q according tojCorrecting to obtain an adjusted label vector Q j’;
In the above formula, f (1) to f (k) are employee HjKeyword frequency of occurrence, Q, for all k work ordersjFor employee HjThe tag variable of (1).
6. The intelligent work order assignment method according to claim 1, wherein the step 3) of determining the target employee from the list of employees to be selected specifically means: and calculating the task quantity of each employee according to the complex coefficient and the quantity of the distributed work orders of each employee in the list of the employees to be selected, and then selecting the employee with the minimum task quantity as the target employee.
7. The intelligent work order assignment method as claimed in claim 1, whichCharacterized in that the work order B to be dispatched in the step 3)iAfter the assignment to the target employees, the method also comprises the step of automatic or manual task transfer assignment, wherein the step of task transfer assignment comprises the steps of finding out the employee with the best intimacy relationship of the target employees from an intimacy relationship matrix D between the employees in a preset enterprise employee set H as a new target employee, and assigning a work order B to be assignediAssigning to a new target employee; the generating step of the affinity relationship matrix D comprises the following steps: firstly, respectively calculating semantic distances between labels of any two employees, and then generating an affinity relationship matrix D according to the following formula;
In the above formula, with dijRepresenting elements in the intimacy relationship matrix D, wherein the value range of i and j is 1-l, l is the number of employees, and the element DijRepresenting the semantic distance between the tags of employee i, employee j; the tags of the staff comprise one or more of post work types, service objects, situations and regions.
8. An intelligent work order dispatching system comprising a computer device, wherein the computer device is programmed or configured to perform the steps of the intelligent work order dispatching method of any one of claims 1-7, or wherein a memory of the computer device has stored thereon a computer program programmed or configured to perform the intelligent work order dispatching method of any one of claims 1-7.
9. A computer readable storage medium having stored thereon a computer program programmed or configured to perform the intelligent work order dispatch method of any one of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010088310.9A CN111340334B (en) | 2020-02-12 | 2020-02-12 | Intelligent work order assignment method, system and medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010088310.9A CN111340334B (en) | 2020-02-12 | 2020-02-12 | Intelligent work order assignment method, system and medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111340334A CN111340334A (en) | 2020-06-26 |
CN111340334B true CN111340334B (en) | 2022-06-28 |
Family
ID=71181526
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010088310.9A Active CN111340334B (en) | 2020-02-12 | 2020-02-12 | Intelligent work order assignment method, system and medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111340334B (en) |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111881684A (en) * | 2020-07-02 | 2020-11-03 | 中国地质大学(武汉) | Intelligent field data acquisition method based on mobile terminal |
CN111950929A (en) * | 2020-08-25 | 2020-11-17 | 上海逸迅信息科技有限公司 | Workload balanced distribution method and device for project type tasks |
CN112001572B (en) * | 2020-10-27 | 2021-01-05 | 绿漫科技有限公司 | Work order intelligent allocation method |
CN112541351B (en) * | 2020-12-08 | 2023-09-05 | 济南市城乡建设数字化中心 | Method and system for dispatching government service heat line work orders in living construction field |
CN113723778A (en) * | 2021-08-16 | 2021-11-30 | 杭州智果科技有限公司 | Intelligent order dispatching method applied to after-sale work order system |
CN114168819B (en) * | 2022-02-14 | 2022-07-12 | 北京大学 | Post matching method and device based on graph neural network |
CN115438995B (en) * | 2022-09-21 | 2023-10-10 | 青岛酷特智能股份有限公司 | Business processing method and equipment for clothing customization enterprise based on knowledge graph |
CN116777148B (en) * | 2023-05-31 | 2023-12-05 | 江苏瑞德信息产业有限公司 | Intelligent distribution processing system for service work orders based on data analysis |
CN117094539B (en) * | 2023-10-20 | 2024-01-16 | 国网浙江省电力有限公司杭州供电公司 | Power-preserving intelligent industrial personal management control method, system, equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103473262A (en) * | 2013-07-17 | 2013-12-25 | 北京航空航天大学 | Automatic classification system and automatic classification method for Web comment viewpoint on the basis of association rule |
CN107301229A (en) * | 2017-06-22 | 2017-10-27 | 美味不用等(上海)信息科技股份有限公司 | Feedback assigning method and system based on semantic analysis |
CN109993542A (en) * | 2017-12-28 | 2019-07-09 | 青岛日日顺电器服务有限公司 | A kind of method, apparatus, server and storage medium for assigning work order |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11042820B2 (en) * | 2018-03-06 | 2021-06-22 | Ricoh Company, Ltd. | Asset management and work order application synchronization for intelligent lockers |
-
2020
- 2020-02-12 CN CN202010088310.9A patent/CN111340334B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103473262A (en) * | 2013-07-17 | 2013-12-25 | 北京航空航天大学 | Automatic classification system and automatic classification method for Web comment viewpoint on the basis of association rule |
CN107301229A (en) * | 2017-06-22 | 2017-10-27 | 美味不用等(上海)信息科技股份有限公司 | Feedback assigning method and system based on semantic analysis |
CN109993542A (en) * | 2017-12-28 | 2019-07-09 | 青岛日日顺电器服务有限公司 | A kind of method, apparatus, server and storage medium for assigning work order |
Also Published As
Publication number | Publication date |
---|---|
CN111340334A (en) | 2020-06-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111340334B (en) | Intelligent work order assignment method, system and medium | |
CN106649434B (en) | Cross-domain knowledge migration label embedding method and device | |
Gautam et al. | Sentiment analysis of twitter data using machine learning approaches and semantic analysis | |
CN109165294B (en) | Short text classification method based on Bayesian classification | |
Zhang et al. | Semi-supervised learning combining co-training with active learning | |
CN111143549A (en) | Method for public sentiment emotion evolution based on theme | |
US11023503B2 (en) | Suggesting text in an electronic document | |
CN110532390B (en) | News keyword extraction method based on NER and complex network characteristics | |
CN112883193A (en) | Training method, device and equipment of text classification model and readable medium | |
CN113627797B (en) | Method, device, computer equipment and storage medium for generating staff member portrait | |
CN113360647B (en) | 5G mobile service complaint source-tracing analysis method based on clustering | |
CN116562265B (en) | Information intelligent analysis method, system and storage medium | |
CN110222192A (en) | Corpus method for building up and device | |
CN112818091A (en) | Object query method, device, medium and equipment based on keyword extraction | |
CN114722810A (en) | Real estate customer portrait method and system based on information extraction and multi-attribute decision | |
Abdullahi et al. | Development of machine learning models for classification of tenders based on UNSPSC standard procurement taxonomy | |
JP2009053743A (en) | Document similarity derivation apparatus, document similarity derivation method and document similarity derivation program | |
Xie et al. | Construction of unsupervised sentiment classifier on idioms resources | |
CN108694165B (en) | Cross-domain dual emotion analysis method for product comments | |
CN110633468A (en) | Information processing method and device for object feature extraction | |
Sun | Research on product attribute extraction and classification method for online review | |
Anuradha et al. | Fuzzy based summarization of product reviews for better analysis | |
CN114969371A (en) | Heat sorting method and device of combined knowledge graph | |
Gündoğan et al. | Creating special issues automatically for papers accepted in journals | |
Messaoudi et al. | Topic extraction in social networks |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |