CN113538011B - Method for associating non-booked contact information with booked user in electric power system - Google Patents
Method for associating non-booked contact information with booked user in electric power system Download PDFInfo
- Publication number
- CN113538011B CN113538011B CN202110840469.6A CN202110840469A CN113538011B CN 113538011 B CN113538011 B CN 113538011B CN 202110840469 A CN202110840469 A CN 202110840469A CN 113538011 B CN113538011 B CN 113538011B
- Authority
- CN
- China
- Prior art keywords
- album
- contact information
- user
- booked
- contact
- 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 31
- 230000005611 electricity Effects 0.000 claims abstract description 13
- 238000005065 mining Methods 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000004458 analytical method Methods 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 10
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 9
- 230000004927 fusion Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 230000003595 spectral effect Effects 0.000 claims description 4
- 238000012423 maintenance Methods 0.000 abstract description 3
- 230000006399 behavior Effects 0.000 description 36
- 238000000605 extraction Methods 0.000 description 5
- 238000007726 management method Methods 0.000 description 4
- 238000003058 natural language processing Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 238000011084 recovery Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
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
- G06Q30/00—Commerce
- G06Q30/01—Customer relationship services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Health & Medical Sciences (AREA)
- Strategic Management (AREA)
- Finance (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Probability & Statistics with Applications (AREA)
- Evolutionary Biology (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Primary Health Care (AREA)
- Tourism & Hospitality (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention provides a method for associating non-in-album contact information with an in-album user in an electric power system, which comprises the following steps of S1, acquiring contact person information bound by the in-album user, and dividing a work order into an in-album contact mode work order and a non-in-album contact mode work order by comparing the contact person information with the user contact information in work order data; step S2, classifying the contact information in the non-booked contact information worksheet as non-booked contact information; step S3, the non-booked contact information and the electricity consumption behavior characteristics of the booked user are respectively analyzed to obtain a characteristic index data set of the non-booked contact information and a characteristic index data set of the booked user; step S4, determining the comprehensive entity similarity between the non-in-album contact information and the in-album user characteristic index; and S5, determining the association relation between the non-booked contact information and the booked user. The invention expands the association relation related to the existing archive information and improves the management and maintenance capability and the customer service accuracy of the non-in-album contact way.
Description
Technical Field
The invention relates to the technical field of power system automation, in particular to a method for associating non-in-album contact information with in-album users in a power system.
Background
Along with the progress of the reform of the electric power system, the electric power clients have increased the personalized and accurate service demands, and electric power companies also continuously push the construction of service channels, such as establishing service channels of entity business halls, telephone service hotlines, online business halls, palm business halls, weChat, payment treasures and the like, so as to meet the different service demands of users. Meanwhile, a big data platform is built, main data in the service field are collected, and various big data-based applications are developed on the basis of the main data. However, the data storage of each service channel is scattered, the client contact data of each channel is not effectively fused and applied, the sharing efficiency of the client contact characteristic data is low, and the accurate service of the client is difficult to carry out.
Particularly, in the process of maintaining and managing information such as non-booked customer contact information, a large number of consultation and message numbers cannot be embodied in a contact information table of a marketing system, namely the non-booked customer contact information, which is called as non-booked contact information. In the consultation and message content of each channel, a great part of work order information of the users relates to the number of the power utilization user, other work order numbers which can create association relation with the number of the power utilization user and some characteristic information of the historical behavior rules of the users. After the business is finished, the part of information cannot be stored and maintained in time, and subsequent non-booked contact information management is difficult to carry out, namely, a great number of repeated works are needed when a customer carries out business interaction each time, the information is difficult to perceive in advance, and the working efficiency of customer service personnel is difficult to improve.
For the historical consultation and effective work sheet data related to the power consumption user number, other work sheet numbers and the characteristic information of the historical behavior rule of the user, the association relationship between the non-in-album contact mode and the historical behavior characteristic rule of the in-album user can be mined through large data means such as text mining, data association and deep learning, so that the effective association relationship creation of the non-in-album contact mode and the in-album user is realized.
Disclosure of Invention
The invention aims to provide a method for associating non-in-album contact information with an in-album user in an electric power system, which solves the technical problems that the degree of association between the non-in-album contact information and the in-album user in the existing electric power system is low and accurate service cannot be provided.
In one aspect, a method for associating non-booked contact information with booked subscribers in a power system is provided, including:
step S1, acquiring contact person information bound by an in-album user, and dividing the work order into an in-album contact mode work order and a non-in-album contact mode work order by comparing the contact person information with user contact information in work order data;
step S2, classifying the contact information in the non-booked contact information worksheet as non-booked contact information;
step S3, respectively analyzing the non-album contact information and the electricity consumption behavior characteristics of the album user according to a preset characteristic analysis rule to obtain a characteristic index data set of the non-album contact information and a characteristic index data set of the album user;
step S4, forming a comprehensive customer behavior characteristic index system data set by the characteristic index data set of the non-in-stock contact information and the characteristic index data set of the in-stock user, and determining the comprehensive entity similarity between the non-in-stock contact information and the in-stock user characteristic index according to the calculated customer behavior characteristic index system data set;
and S5, determining the association relationship between the non-in-album contact information and the in-album user according to the comprehensive entity similarity between the non-in-album contact information and the characteristic indexes of the in-album user.
Preferably, the step S3 includes: carrying out data noise processing on the non-booked contact information, and filtering and deleting the contact information generated by the test class work order;
carrying out data noise processing on the non-booked contact information, and filtering and deleting the contact information generated by the non-actual user;
replacing the digit string containing the telephone number format in the non-booked contact information;
for incoming number fields for which the incoming number is empty in the non-booked contact information, the first telephone number referred to in the corresponding work order is filled in.
Preferably, in step S3, the analyzing the non-in-album contact information according to the preset feature analysis rule includes:
and carrying out text mining on the non-in-album contact information by adopting regular matching text mining through a preset text recognition rule library to obtain channel characteristic keywords, power failure event characteristic keywords and power utilization service characteristic keywords, and generating a characteristic index data set of the non-in-album contact information.
Preferably, in step S3, the analyzing the electricity behavior features of the in-album user according to the preset feature analysis rule includes:
according to preset in-album user feature index types, counting the in-album user information to obtain the number value of each feature index type, and generating an in-album user feature index data set; the on-book user includes business expansion work list detail list information, power failure event detail list information and on-line electronic channel customer consultation or message information in the customer characteristic index type.
Preferably, the step S4 includes:
carrying out data fusion on the characteristic index data set of the non-booked contact information and the characteristic index data set of the booked user to obtain a comprehensive customer behavior characteristic index system data set;
and performing dimension reduction on each index in the comprehensive customer behavior characteristic index system data set according to a preset index type importance value to obtain a dimension reduction optimized customer behavior characteristic index system data set.
Preferably, the step S4 further includes:
clustering the data set of the customer behavior characteristic index system with optimized dimension reduction according to a preset clustering rule to obtain sample data spectral clusters, and generating different sample clusters { D } with certain association relation among sample individuals in the clusters 1 ,D 2 ,D 3 …D n N=1, 2,3 …, where n is the number of clusters.
Preferably, step S4 further comprises:
determining the pairs of samples in each cluster (x i ,y j ) I, j=1, 2,3, … m, where x i Representing j-th non-booked contact, y j Representing the jth in-album user;
determining attribute similarity of each sample pair according to a preset similarity calculation rule, and generating an attribute similarity matrixWherein (1)>Representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute 1; />Representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute k.
Preferably, the step S4 further includes:
and calculating the comprehensive entity similarity between the non-booked contact information and the booked user characteristic indexes in each sample pair according to the following formula:
wherein, C is the comprehensive entity similarity; alpha i I=1, 2, …, e is the similarity weight of each attribute; e is the number of attribute similarity sequences;representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute k; x is x i Representing j-th non-booked contact, y j Representing the jth in-album user.
Preferably, the step S5 includes:
comparing the similarity of the comprehensive entity between the non-in-stock contact information and the in-stock user characteristic index with a preset judging threshold value, and judging that the non-in-stock contact information and the in-stock user characteristic index have an association relationship if the similarity of the comprehensive entity between the non-in-stock contact information and the in-stock user characteristic index is larger than or equal to the preset judging threshold value; if the comprehensive entity similarity between the non-in-stock contact information and the in-stock user characteristic index is smaller than a preset judging threshold value, judging that the non-in-stock contact information and the in-stock user characteristic index have no association relation.
Preferably, the step S5 further includes:
and collecting all the non-registering contact information with the association relationship and the feature indexes of the registering user to obtain a data set of the association relationship between the non-registering contact information and the registering user.
In summary, the embodiment of the invention has the following beneficial effects:
according to the association method of the non-in-album contact information and the in-album user in the electric power system, provided by the invention, natural language processing and text mining technologies are utilized to mine keywords related to various behavior habit characteristic indexes in non-in-album contact information consultation and message text data, and by means of the theory of the attribute similarity and the comprehensive entity similarity calculation method in knowledge fusion, individuals with the behavior characteristic similarity are subjected to soft association from the non-in-album contact information and the related electricity consumption behavior characteristic similarity of the in-album user, so that the association relation related to the existing archive information is expanded, the original method of carrying out direct hard association on the information such as the text mining of the user number, other work order numbers and the like is supplemented and enriched, and the management maintenance capability and customer service accuracy of the non-in-album contact information are greatly assisted.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are required in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that it is within the scope of the invention to one skilled in the art to obtain other drawings from these drawings without inventive faculty.
Fig. 1 is a schematic flow chart of a method for associating non-in-album contact information with an in-album user in an electric power system according to an embodiment of the invention.
Fig. 2 is a logic diagram of a method for associating non-in-album contact information with an in-album user in an electric power system according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings, for the purpose of making the objects, technical solutions and advantages of the present invention more apparent.
Fig. 1 and fig. 2 are schematic diagrams of an embodiment of a method for associating non-in-album contact information with an in-album user in an electric power system according to the present invention. In this embodiment, the method comprises the steps of:
step S1, acquiring contact person information bound by an in-album user, and dividing the work order into an in-album contact mode work order and a non-in-album contact mode work order by comparing the contact person information with user contact information in work order data; it can be understood that the binding contact information of the user in the list is extracted from the marketing system, two kinds of association relations of telephone numbers and user numbers are screened out from the binding contact information, and the association relation is utilized to divide the in-list contact mode worksheet and the out-list contact mode worksheet of the customer service worksheet and other on-line electronic channel consultation or message worksheets.
In a specific embodiment, the correspondence between the user number and the contact information contained in the acquired contact information table; the contact ways are the corresponding contact ways of the user in the album, the contact ways in the online electronic channel consultation or message work orders are matched with the service hot line customer service work order information table, and the work orders of all channels are divided into non-contact way work orders and contact way work orders in the album through the matching results. The method comprises the steps of processing a contact list in a booked mode, wherein the contact list in the booked mode is not processed, and the follow-up natural language processing and text mining and extraction work is mainly conducted on the contact list in the booked mode.
Step S2, classifying the contact information in the non-booked contact information worksheet as non-booked contact information; it can be understood that, by the worksheet division result in step 1, the contact information (contact information) contained in the non-booked contact information worksheet can be classified as non-booked contact information.
Step S3, respectively analyzing the non-album contact information and the electricity consumption behavior characteristics of the album user according to a preset characteristic analysis rule to obtain a characteristic index data set of the non-album contact information and a characteristic index data set of the album user; it can be understood that the non-in-album contact information related data mainly relates to service hotline customer service work list information, specifically includes complaint text and information of consultation or message of various electronic channels on line such as network, palm, micro, etc., and the in-album user related data mainly relates to business expansion work list detail list information, power failure event detail list information and information of consultation or message of various electronic channels on line, as shown in the following table, the method relates to indexes for the non-in-album contact information and the in-album user soft association:
through natural language processing and text mining of the data, all behavior characteristic index values of non-booked contact ways and booked users can be obtained, and a client behavior characteristic index system data set is constructed so as to optimize index system characteristic engineering.
In a specific embodiment, before text mining and extraction are performed on the feature index data, non-in-album contact information and all relevant information of all aspects of all power grid contact behaviors of in-album users, such as channels, services, events and the like, are analyzed, and a behavior feature index system is mainly created from three aspects, namely event class relevant attributes, electricity utilization service class relevant attributes and channel contact class relevant attributes. The event type related attribute mainly comprises power failure events and comprises 5 indexes; the relevant attribute of the power utilization business class comprises 24 indexes (wherein 'the number of times of dealing with business expansion of various businesses' and 'the number of times of dealing with business expansion of various businesses' respectively comprise 14 index fields according to business classification; and 'the number of times of touching channels of dealing with business expansion of business' respectively comprise 5 index fields according to business related channel classification); the channel contact class related attribute mainly comprises 2 indexes; therefore, the client behavior characteristic index system considers three aspects, namely 31 indexes; the subsequent non-booked contact way and booked user characteristic index mining extraction and soft association relation creation are developed based on the index system optimization result. In the customer behavior characteristic index system, indexes of the non-booked contact ways and indexes of the booked users are in one-to-one correspondence, for example: in the related characteristic attribute of the event (power failure), the total number of the 'message power failure event' of the non-in-album contact way corresponds to the total number of the 'power failure event influence' of the in-album user; in the related characteristic attribute of the power utilization business, the 'appeal' of the non-in-album contact mode relates to the total number of business expansion business to the 'total number of business expansion business handling' used for the album user, and the like. Because, only two sample indexes are in one-to-one correspondence, the clustering of different clusters according to the same type of indexes and the soft association relation mining based on different index similarity on opposite samples in different clusters can be realized.
Specifically, consider the pretreatment of service hot line customer service work order information form and electronic channel consultation or message work order data related to the non-in-album contact mode, namely the noise elimination and missing value treatment of the non-in-album contact mode field, and the data noise treatment of the resort content before natural language treatment and text data mining. The analyzing the non-booked contact information according to the preset feature analysis rule comprises the following steps: and carrying out data noise processing on the non-in-album contact information, and filtering and deleting the contact information generated by the test class worksheet, specifically, filtering and deleting test class worksheets such as 'test', 'internal dial test'. And carrying out data noise processing on the non-in-album contact information, and filtering and deleting the contact information generated by the non-actual user, specifically, filtering related contact modes of the non-actual user, such as public security, urban management, internal exchange of a company and the like, relating to worksheets. The method comprises the steps of replacing a digital string character string which does not contain a telephone number format in the album contact information, particularly replacing the digital string character string which possibly contains the telephone number format, such as the number of a charging pile device, the number of a problem, and the like in a text of a service work order information table, and preventing noise from being generated when the telephone number text is mined. For the incoming number field with the empty incoming number in the non-album contact information, the first phone number related to the corresponding work order is used for filling, specifically, the incoming number field with the empty incoming number in the customer service work order information table is filled with the first phone number related to the appeal text, because the records with the empty incoming number field are all the message work orders, the incoming number, namely the first phone number of the message text, is related in the incoming content of the work order, and the message text is in standard format, such as: "mobile phone number, message content: … …).
After the filtering processing, text mining is carried out on the non-in-album contact information by adopting regular matching text mining through a preset text recognition rule base, channel characteristic keywords, power failure event characteristic keywords and power utilization service characteristic keywords are obtained, and a characteristic index data set of the non-in-album contact information is generated. It is understood that the mining is mainly based on the three aspect feature indexes contained in the created behavior feature index system for the non-booked contact ways. The main mining method is to use a regular matching text mining algorithm to carry out text mining on the text subjected to natural language word segmentation processing by constructing a text recognition rule base, wherein the mining object is characteristic index keyword information. Such as: the 'appeal' of the non-booked contact way relates to the index of the total times of the power failure event, and the mining calculation method is that (the 'power failure' keywords in the appeal text are mined through regular fuzzy matching, and statistics is carried out according to the non-booked contact way); the index of 'appeal related to various business expansion times' of the non-booked contact ways mainly comprises the following steps: new installation, capacity increase, capacity reduction recovery, temporary electricity utilization, renaming, passing, selling, changing, temporarily dismantling, suspending business, refunding, changing settlement house, and reporting 14 business related fields, wherein the mining calculation method comprises the following steps: a. counting the historical handling quantity of each business subclass of the business expansion business of the in-album user; b. according to the ranking in a, the quantity of customer service work orders related to business subclasses in the non-in-album contact way appeal content is fuzzy matched according to the keywords of each business subclass; c. screening and marking business subclasses with large consultation quantity and large handling quantity; d. naming the field according to the common frequent terminology of a business subclass recorded in the incoming call appeal text; e. and mining each business keyword in the appeal text through regular fuzzy matching, and carrying out statistics and the like according to the non-booked contact way.
Specifically, the analyzing the electricity behavior features of the in-album user according to the preset feature analysis rule includes: according to preset in-album user feature index types, counting the in-album user information to obtain the number value of each feature index type, and generating an in-album user feature index data set; the on-book user includes business expansion work list detail list information, power failure event detail list information and on-line electronic channel customer consultation or message information in the customer characteristic index type. It can be understood that the user behavior characteristic index considered in the album mainly relates to three data sources, namely business expansion work list detail list information, power failure event detail list information and on-line electronic channel customer consultation or message information; and acquiring the user characteristic index data table in the album by a corresponding index value statistical calculation method for the data source. Such as: the index of the total number of times of influence of the power failure event of the user is recorded, and the extraction and calculation method comprises the following steps: counting and calculating the total times of the user affected by the power failure event in a certain period of time; the index of the business expansion times of the business expansion of the business of the user in the book mainly comprises the following steps: new installation, capacity increase, capacity reduction recovery, temporary electricity utilization, renaming, passing, selling, changing, temporarily dismantling, suspending business, refunding, changing settlement house, and reporting 14 business related fields, wherein the extraction and calculation method comprises the following steps: and according to the field names, the fuzzy matching users accept business expansion business subclass names, and calculate the number of each field according to the business subclass names, wherein the names of the business related fields are consistent with those of the non-booked contact information related fields, and the like.
Step S4, forming a comprehensive customer behavior characteristic index system data set by the characteristic index data set of the non-in-stock contact information and the characteristic index data set of the in-stock user, and determining the comprehensive entity similarity between the non-in-stock contact information and the in-stock user characteristic index according to the calculated customer behavior characteristic index system data set; it will be appreciated that.
In a specific embodiment, a characteristic index data set of non-in-album contact information and a characteristic index data set of an in-album user are subjected to data fusion to obtain a comprehensive customer behavior characteristic index system data set; the method comprises the steps of performing dimension reduction on each index in a comprehensive customer behavior characteristic index system data set according to a preset index type importance value to obtain a dimension reduction optimized customer behavior characteristic index system data set, wherein it is understood that feature engineering optimization is performed mainly by means of feature principal component analysis, and the optimized customer behavior characteristic index system data set of the secondary key feature index is obtained according to the original index importance dimension reduction.
Specifically, clustering is carried out on the data set of the customer behavior characteristic index system with optimized dimension reduction according to a preset clustering rule, sample data spectral clusters are obtained, and different sample clusters { D } with certain association relation among sample individuals in the clusters are generated 1 ,D 2 ,D 3 …D n N=1, 2,3 …, where n is the number of clusters. It can be understood that, based on the customer behavior characteristic index system data set after dimension reduction optimization, all indexes of the relevant characteristic attributes of the electricity service, all indexes of the relevant attribute of the channel contact and the index of the total number of outage events under the relevant characteristic attributes of the event (outage) are selected as clustering input indexes, namely, samples are selectedThe group characteristic attribute is used as the model input of group clustering division, and the similarity matrix, the degree matrix, the Laplace matrix and the characteristic matrix of the sample data are calculated in sequence based on the group characteristic attribute index. And then, carrying out min-max standardization on the feature matrix, and carrying out K-means clustering based on the standardized feature matrix, thereby completing sample data spectral clustering and generating different sample clusters with certain association relations among sample individuals in the clusters.
Specifically, based on the behavior feature index system data set omega which is subjected to feature principal component analysis, screening and filtering, calculating the attribute similarity of each sample pair by means of different types of index attribute similarity calculation methods related to a knowledge fusion theory to obtain an attribute similarity matrix. Determining the pairs of samples in each cluster (x i ,y j ) I, j=1, 2,3, … m, where x i Representing j-th non-booked contact, y j Representing the jth in-album user;
determining attribute similarity of each sample pair according to a preset similarity calculation rule, and generating an attribute similarity matrixWherein (1)>Representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute 1; />Representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute k; it can be understood that different attribute similarities of each sample are calculated by adopting different methods according to the attribute value types, for example, numerical attribute similarity is calculated by cosine distance, enumeration attribute similarity is calculated by Jaccard similarity coefficient, time attribute similarity is calculated by time interpolation method, and text attribute similarity is calculated by edit distance.
More specifically, the comprehensive entity similarity between the non-in-album contact information and the in-album user characteristic index in each sample pair is calculated according to the following formula:
wherein, C is the comprehensive entity similarity; alpha i I=1, 2, …, e is the similarity weight of each attribute; e is the number of attribute similarity sequences;representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute k; x is x i Representing j-th non-booked contact, y j Representing the jth in-album user; specifically, the subjective weight part is obtained by combining subjective and objective weighting methods, and the subjective weight part adopts a hierarchical analysis method, and is divided according to the judgment of business personnel, if a non-in-album contact way can generate a relatively strong association relation with a group of in-album users through a power failure event, so that more weight is given to the related characteristic indexes of the power failure event relative to the characteristic indexes with smaller individual differences such as 'inquiring electric charge'. And the objective weight part is calculated by using a CRITIC method based on the calculation result of the index similarity among users.
And S5, determining the association relationship between the non-in-album contact information and the in-album user according to the comprehensive entity similarity between the non-in-album contact information and the characteristic indexes of the in-album user. It can be understood that the threshold value beta is set based on a combination mode of the objective data distribution characteristic analysis conclusion and subjective expert experience judgment, and the relevance judgment is carried out on the sample pair.
In a specific embodiment, comparing the similarity of the comprehensive entity between the non-in-stock contact information and the in-stock user characteristic index with a preset judging threshold value, and judging that the non-in-stock contact information and the in-stock user characteristic index have an association relationship if the similarity of the comprehensive entity between the non-in-stock contact information and the in-stock user characteristic index is greater than or equal to the preset judging threshold value; and collecting all the non-in-album contact information with the association relationship with the feature indexes of the in-album user to obtain a data set of the association relationship between the non-in-album contact information and the in-album user, wherein it can be understood that ranking judgment can be carried out on the non-in-album contact manner and all the associated in-album user affinity and affinity by integrating the size of the entity similarity C. If the comprehensive entity similarity between the non-in-stock contact information and the in-stock user characteristic index is smaller than a preset judging threshold value, judging that the non-in-stock contact information and the in-stock user characteristic index have no association relation.
In summary, the embodiment of the invention has the following beneficial effects:
according to the association method of the non-in-album contact information and the in-album user in the electric power system, provided by the invention, natural language processing and text mining technologies are utilized to mine keywords related to various behavior habit characteristic indexes in non-in-album contact information consultation and message text data, and by means of the theory of the attribute similarity and the comprehensive entity similarity calculation method in knowledge fusion, individuals with the behavior characteristic similarity are subjected to soft association from the non-in-album contact information and the related electricity consumption behavior characteristic similarity of the in-album user, so that the association relation related to the existing archive information is expanded, the original method of carrying out direct hard association on the information such as the text mining of the user number, other work order numbers and the like is supplemented and enriched, and the management maintenance capability and customer service accuracy of the non-in-album contact information are greatly assisted.
The foregoing disclosure is illustrative of the present invention and is not to be construed as limiting the scope of the invention, which is defined by the appended claims.
Claims (6)
1. A method for associating non-booked contact information with booked subscribers in an electric power system, comprising the steps of:
step S1, acquiring contact person information bound by an in-album user, and dividing the work order into an in-album contact mode work order and a non-in-album contact mode work order by comparing the contact person information with user contact information in work order data;
step S2, classifying the contact information in the non-booked contact information worksheet as non-booked contact information;
step S3, respectively analyzing the non-album contact information and the electricity consumption behavior characteristics of the album user according to a preset characteristic analysis rule to obtain a characteristic index data set of the non-album contact information and a characteristic index data set of the album user;
step S4, forming a comprehensive customer behavior characteristic index system data set by the characteristic index data set of the non-in-stock contact information and the characteristic index data set of the in-stock user, and determining the comprehensive entity similarity between the non-in-stock contact information and the in-stock user characteristic index according to the calculated customer behavior characteristic index system data set;
the method comprises the steps of carrying out data fusion on a characteristic index data set of non-in-album contact information and a characteristic index data set of an in-album user to obtain a comprehensive customer behavior characteristic index system data set;
performing dimension reduction on each index in the comprehensive customer behavior characteristic index system data set according to a preset index type importance value to obtain a dimension reduction optimized customer behavior characteristic index system data set;
clustering the data set of the customer behavior characteristic index system with optimized dimension reduction according to a preset clustering rule to obtain sample data spectral clusters, and generating different sample clusters { D } with certain association relation among sample individuals in the clusters 1 ,D 2 ,D 3 …D n N=1, 2,3 …, where n is the number of clusters;
determining the pairs of samples in each cluster (x i ,y j ) I, j=1, 2,3, … m, where x i Representing j-th non-booked contact, y j Representing the jth in-album user;
determining attribute similarity of each sample pair according to a preset similarity calculation rule, and generating an attribute similarity matrixWherein (1)>Representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute 1; />Representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute k;
and calculating the comprehensive entity similarity between the non-in-album contact information and the in-album user characteristic index in each sample pair according to the following formula:
wherein, C is the comprehensive entity similarity; alpha i I=1, 2, …, e is the similarity weight of each attribute; e is the number of attribute similarity sequences;representing pairs of samples (x i ,y j ) Attribute similarity with respect to attribute k; x is x i Representing j-th non-booked contact, y j Representing the jth in-album user;
and S5, determining the association relationship between the non-in-album contact information and the in-album user according to the comprehensive entity similarity between the non-in-album contact information and the characteristic indexes of the in-album user.
2. The method according to claim 1, wherein the step S3 includes:
carrying out data noise processing on the non-booked contact information, and filtering and deleting the contact information generated by the test class work order;
carrying out data noise processing on the non-booked contact information, and filtering and deleting the contact information generated by the non-actual user;
replacing the digit string containing the telephone number format in the non-booked contact information;
for incoming number fields for which the incoming number is empty in the non-booked contact information, the first telephone number referred to in the corresponding work order is filled in.
3. The method of claim 2, wherein in step S3, the analyzing the non-in-album contact information according to the preset feature analysis rule includes:
and carrying out text mining on the non-in-album contact information by adopting regular matching text mining through a preset text recognition rule library to obtain channel characteristic keywords, power failure event characteristic keywords and power utilization service characteristic keywords, and generating a characteristic index data set of the non-in-album contact information.
4. A method according to claim 3, wherein in step S3, the analyzing the electricity behavior characteristics of the in-album user according to the preset characteristic analysis rule includes:
according to preset in-album user feature index types, counting the in-album user information to obtain the number value of each feature index type, and generating an in-album user feature index data set; the on-book user includes business expansion work list detail list information, power failure event detail list information and on-line electronic channel customer consultation or message information in the customer characteristic index type.
5. The method of claim 4, wherein the step S5 includes:
comparing the similarity of the comprehensive entity between the non-in-stock contact information and the in-stock user characteristic index with a preset judging threshold value, and judging that the non-in-stock contact information and the in-stock user characteristic index have an association relationship if the similarity of the comprehensive entity between the non-in-stock contact information and the in-stock user characteristic index is larger than or equal to the preset judging threshold value; if the comprehensive entity similarity between the non-in-stock contact information and the in-stock user characteristic index is smaller than a preset judging threshold value, judging that the non-in-stock contact information and the in-stock user characteristic index have no association relation.
6. The method of claim 5, wherein said step S5 further comprises:
and collecting all the non-registering contact information with the association relationship and the feature indexes of the registering user to obtain a data set of the association relationship between the non-registering contact information and the registering user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110840469.6A CN113538011B (en) | 2021-07-24 | 2021-07-24 | Method for associating non-booked contact information with booked user in electric power system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110840469.6A CN113538011B (en) | 2021-07-24 | 2021-07-24 | Method for associating non-booked contact information with booked user in electric power system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113538011A CN113538011A (en) | 2021-10-22 |
| CN113538011B true CN113538011B (en) | 2024-02-09 |
Family
ID=78088927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110840469.6A Active CN113538011B (en) | 2021-07-24 | 2021-07-24 | Method for associating non-booked contact information with booked user in electric power system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113538011B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117290689A (en) * | 2023-09-21 | 2023-12-26 | 湖北太昇科技有限公司 | Smart home-based user binding method and system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108924371A (en) * | 2018-08-20 | 2018-11-30 | 国家电网有限公司客户服务中心 | The method that family number is identified by incoming number during electric power customer service |
| CN111861240A (en) * | 2020-07-27 | 2020-10-30 | 深圳前海微众银行股份有限公司 | Suspicious user identification method, device, equipment and readable storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11200581B2 (en) * | 2018-05-10 | 2021-12-14 | Hubspot, Inc. | Multi-client service system platform |
-
2021
- 2021-07-24 CN CN202110840469.6A patent/CN113538011B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108924371A (en) * | 2018-08-20 | 2018-11-30 | 国家电网有限公司客户服务中心 | The method that family number is identified by incoming number during electric power customer service |
| CN111861240A (en) * | 2020-07-27 | 2020-10-30 | 深圳前海微众银行股份有限公司 | Suspicious user identification method, device, equipment and readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113538011A (en) | 2021-10-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106651424B (en) | Power user portrait establishing and analyzing method based on big data technology | |
| CN106447346A (en) | Method and system for construction of intelligent electric power customer service system | |
| CN112632405B (en) | Recommendation method, recommendation device, recommendation equipment and storage medium | |
| CN111552813A (en) | Power knowledge graph construction method based on power grid full-service data | |
| CN106528850B (en) | Gate inhibition's data exception detection method based on machine learning clustering algorithm | |
| CN113010578B (en) | Community data analysis method and device, community intelligent interaction platform and storage medium | |
| CN116384889A (en) | Intelligent analysis method for information big data based on natural language processing technology | |
| CN110544035A (en) | internal control detection method, system and computer readable storage medium | |
| CN111723256A (en) | Government affair user portrait construction method and system based on information resource library | |
| CN107977855B (en) | Method and device for managing user information | |
| CN113538011B (en) | Method for associating non-booked contact information with booked user in electric power system | |
| CN112199488B (en) | Incremental knowledge graph entity extraction method and system for power customer service question and answer | |
| CN113688870A (en) | Group renting house identification method based on user electricity utilization behavior by adopting hybrid algorithm | |
| CN112488736A (en) | Method and system for analyzing government affair hotline work order data in field of residential construction | |
| Shi et al. | Human resources balanced allocation method based on deep learning algorithm | |
| Yanti et al. | Application of named entity recognition via Twitter on SpaCy in Indonesian (case study: Power failure in the Special Region of Yogyakarta) | |
| Yu et al. | Research on situational perception of power grid business based on user portrait | |
| CN115759253A (en) | Power grid operation and maintenance knowledge map construction method and system | |
| Zhu et al. | Telecom customer churn prediction based on classification algorithm | |
| CN113971213A (en) | Smart city management public information sharing system | |
| Zhang et al. | A text mining based method for policy recommendation | |
| CN111666378A (en) | Chinese yearbook title classification method based on word vectors | |
| Zhang | Intelligent Mining Method of Massive Digital Archives Based on Artificial Intelligence | |
| CN113537734B (en) | Energy data application catalog extraction method based on maximum correlation minimum redundancy | |
| Yao et al. | [Retracted] Analysis on the Establishment and Management of Library Resource Base Based on Modern Information Technology |
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 |