CN113284007A - Power utilization information processing system based on power insurance package and processing method thereof - Google Patents
Power utilization information processing system based on power insurance package and processing method thereof Download PDFInfo
- Publication number
- CN113284007A CN113284007A CN202110586589.8A CN202110586589A CN113284007A CN 113284007 A CN113284007 A CN 113284007A CN 202110586589 A CN202110586589 A CN 202110586589A CN 113284007 A CN113284007 A CN 113284007A
- Authority
- CN
- China
- Prior art keywords
- power consumption
- feedback
- index
- characteristic
- characteristic index
- 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.)
- Granted
Links
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
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/08—Insurance
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Databases & Information Systems (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Finance (AREA)
- Health & Medical Sciences (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Strategic Management (AREA)
- General Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Primary Health Care (AREA)
- Human Resources & Organizations (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Technology Law (AREA)
- Public Health (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses an electricity consumption information processing system based on an electric power insurance package.A database is used for storing the customer name, the electricity consumption information and the customer feedback information of an electricity consumption customer; the data processing module is used for obtaining a power consumption index according to the power consumption information and obtaining a feedback index according to the single piece of client feedback information; the clustering module is used for obtaining a power consumption grade index according to the power consumption index, and clustering and screening according to a plurality of feedback indexes of a plurality of power consumption customers to obtain target power consumption customers with the same characteristic index. The invention also discloses a processing method of the electricity consumption information processing system based on the electric power insurance package, and the electricity consumption information and the customer feedback information are processed and clustered by the electricity consumption information processing system to obtain the grade index of the electricity consumption and the target electricity consumption customers with the same characteristic index. The method can accurately screen the demand of the power insurance package information based on formulation, and screen the power consumption customer groups of the similar products.
Description
Technical Field
The invention relates to the technical field of power markets, in particular to a power utilization information processing system based on a power insurance package and a processing method thereof.
Background
The traditional service mode in the current electric power market belongs to passive service, the corresponding power supply service problem is processed after failure reporting and complaints of customers, and the electricity utilization characteristics and potential requirements of the customers are not systematically analyzed in advance, so that the electricity utilization customer requirements are easily ignored.
In the power industry, electricity customers of industrial and commercial enterprises, particularly customers of small processing and small shops without property services, are restricted by professions, and once the electricity is interrupted due to failure, the problems of unclear electrical failure reason, difficult maintenance, expensive maintenance and the like are often encountered, so that production stagnation and business interruption are often caused, and loss is caused to the operation of the electricity customers. The complaints and complaints caused by the method are increasingly increased, even cause disputes of fault reasons and disputes of handling, and deteriorate the operator environment between the power grid enterprise and the power utilization customers.
At present, customer portrait research in the power industry mainly focuses on tag library establishment of electric charge risks, customer credit, channel preference, electricity consumption behaviors and the like, diversified electricity consumption habits of electricity consumption customers and demand response of the electricity charges are analyzed, and development of basic business work such as electric charge recycling risks, load prediction, complaint monitoring and the like is effectively promoted. However, the current research lacks consideration of diversification of power supply service contents and differentiation of service modes, an active service image system of different power consumption customer groups is not established in a focused manner, the high-quality power supply service of multiple power consumption customers cannot be well promoted, and an electric power insurance active service business mode is supported.
Disclosure of Invention
The system and the method are based on the information requirement of formulating the power insurance package, and carry out processing and clustering on the power consumption information of different power consumption customers to obtain the same power consumption customer group with the same characteristic index and the power consumption index of each power consumption customer, thereby accurately screening the same power consumption customer group.
In order to achieve the above object, the present invention provides an electricity consumption information processing system based on an electric power insurance package, characterized in that: the system comprises a database, a data processing module and a clustering module;
the data processing module is used for obtaining a power consumption index according to the power consumption information and obtaining a feedback index according to the single piece of client feedback information; the power consumption index comprises a power consumption time period and power consumption in the power consumption time period, and the feedback index comprises a feedback time period and TF-IDF weight values of feedback characteristic words;
the clustering module is used for obtaining a power consumption grade index according to the power consumption index, and clustering and screening according to a plurality of feedback indexes of a plurality of power consumption customers to obtain target power consumption customers with the same characteristic index, wherein the characteristic index is the frequency grade of each feedback characteristic word in the feedback time period.
Further, the database is used for storing the customer name, the electricity consumption information and the customer feedback information of the electricity utilization customer.
Further, the data processing module includes a feedback information processing module, and the feedback information processing module is configured to identify time-type data in the feedback information of the single client, so as to obtain a feedback time period in which the time-type data is located.
Further, the feedback information processing module is further configured to perform word segmentation processing on text-type data in the single piece of client feedback information to obtain a plurality of words, perform semantic recognition on the plurality of words by using the feedback feature words as recognition targets to obtain the number of times that each feedback feature value appears, and convert the number of times into a TF-IDF weight of each feedback feature word.
Further, the data processing module further comprises a power consumption information processing module, and the power consumption information processing module is used for acquiring the accumulated power consumption reading of the starting time and the ending time of the power consumption time period, and then acquiring the difference between the accumulated power consumption reading of the ending time and the accumulated power consumption degree of the starting time.
Further, the power usage time period and the feedback time period include up to 30 days, up to 90 days, up to 180 days, and up to one year.
Furthermore, the clustering module comprises a power consumption index clustering module, and the power consumption index clustering module is used for converting the power consumption index into the power consumption grade in the power consumption time period according to the value ranges of different power consumption grades in different power consumption time periods.
Further, the clustering module further comprises a characteristic index clustering module, wherein the characteristic index clustering module is used for constructing feedback time periods in feedback information of a single client and TF-IDF weights of feedback characteristic words into characteristic index row vectors, and constructing a plurality of characteristic index row vectors of a plurality of power utilization clients into a characteristic index matrix X; solving the Euclidean distance between any one characteristic index row vector of the characteristic index matrix X and other characteristic index row vectors to obtain a characteristic index distance row vector D; and selecting a characteristic index row vector corresponding to the minimum value in each characteristic index distance row vector D and the corresponding similar electricity utilization client, and converting the characteristic index row vector corresponding to the minimum value into a characteristic index, thereby obtaining the similar electricity utilization client corresponding to the same characteristic index.
Further, the method for converting the characteristic index row vector into the characteristic index comprises the step of converting the TF-IDF weight in the characteristic index row vector into the frequency grade according to the value range of the TF-IDF weight in different frequency grades of different feedback characteristic words, so that the corresponding characteristic index is obtained.
The invention also provides a processing method of the power utilization information processing system based on the power insurance package demand, which is characterized by comprising the following steps:
obtaining a power consumption index according to the power consumption information and obtaining a feedback index according to single client feedback information; the power consumption index comprises a power consumption time period and power consumption in the power consumption time period, and the feedback index comprises a feedback time period and TF-IDF weight values of feedback characteristic words;
the method comprises the steps of obtaining a power consumption grade index according to the power consumption index, and carrying out clustering and screening according to a plurality of feedback indexes of a plurality of power consumption customers to obtain the same type of power consumption customers with the same characteristic index, wherein the characteristic index is the frequency grade of each feedback characteristic word in a feedback time period.
Further, the clustering and screening method comprises the following steps: establishing a feedback time period in the feedback information of a single client and TF-IDF weight values of feedback feature words as a feature index row vector, and establishing a plurality of feature index row vectors of a plurality of power utilization clients as a feature index matrix X; solving the Euclidean distance between any one characteristic index row vector of the characteristic index matrix X and other characteristic index row vectors to obtain a characteristic index distance row vector D; and selecting a characteristic index row vector corresponding to the minimum value of each characteristic index distance row vector D and the corresponding similar electricity utilization customers, and converting the characteristic index row vector corresponding to the minimum value into the characteristic index according to different frequency grade value ranges of different feedback characteristic values, thereby obtaining the similar electricity utilization customers corresponding to the same characteristic index.
Further, the construction method of the characteristic index row vector comprises the step of forming TF-IDF weight values of all feedback characteristic words into TF-IDF weight value row vector TF _ IDFi.jConverting the feedback time period into a feedback time row vector Ti.jThen the feature index row vector is
Xi(j)=xi,j=(tf_idfi.j Ti.j),1≤j≤ai
Wherein i is the ith electricity utilization client, j is the jth user feedback information of the collusion and electricity utilization client, and aiAnd feeding back the number of information for the ith electricity utilization client in the current year.
Wherein a time row vector T is fed backi.jThe construction method comprises defining Ti.j=(t1,t2,t3,t4),t1,t2,t3,t4Whether the feedback time belongs to 30 days from the present, 90 days from the present, 180 days from the present and one year from the present is respectively represented, values of the feedback time are all 0 or 1, 1 represents that the feedback time is located in the feedback time period, and 0 represents that the feedback time period is not located in the feedback time period.
Further, the method for determining the TF-IDF weight of the feedback feature word comprises,
TF-IDF weight (TF _ IDF) of feedback feature word t in user feedback information dt,dComprises the following steps:
(tf_idf)t,d=tft,d*idft
in the formula, tft,dThe character t representing the feedback is shown in the feedback information d of the userCurrent word frequency, nt,dFor feeding back the times of occurrence of the characteristic word t in the user feedback information d, ndFeeding back the total word number of the information d for the user; idftRepresenting the inverse text frequency of the feedback characteristic word t, N is the total amount of the feedback information of the user, NtAnd feeding back the information quantity of the user containing the feedback characteristic word t.
Further, the euclidean distance between two feature index row vectors of the same electricity consumer is as follows:
in the formula, the v1 th and v2 th characteristic index row vectors of the u electricity consumer are xu,v1,xu,v2(1≤u≤N,1≤v1,v2≤au)。
The Euclidean distance between two appeal characteristic elements of different electricity utilization clients is as follows:
in the formula, the v characteristic index row vector of the u electricity consumer is xu,v(1≤u≤N,1≤v≤au) The nth characteristic index row vector of the mth electricity consumer is xm,n(1≤m≤N,1≤n≤am)。
Further, the feature index distance row vector D (u, v) is expressed as:
in the formula, u is the u-th (u is more than or equal to 1 and less than or equal to N) electricity customer, v is the v-th (v is more than or equal to 1 and less than or equal to a)u) Characteristic index row vector xu,vAnd the corresponding distance matrix is denoted as D (u, v).
The invention has the beneficial effects that: and precisely screening the same type of electricity utilization client groups. The method comprises the steps of obtaining power consumption information and power consumption client feedback information of power consumption clients, obtaining power consumption grades in different power consumption time periods according to the power consumption information, respectively processing according to time type data and text type data in the single power consumption client feedback information to obtain feedback time periods and TF-IDF weights of feedback characteristic words, constructing a plurality of characteristic index row vectors of a plurality of feedback indexes of the power consumption clients, solving the Euclidean distance between any two characteristic index row vectors to obtain characteristic index distance row vectors, determining corresponding characteristic indexes and similar power consumption clients according to the minimum value of each row of characteristic index distance row vectors to obtain similar power consumption clients with the same characteristic indexes, and screening out similar power consumption client groups with similar tea products based on the information requirement of making power insurance packages.
Drawings
FIG. 1 is a schematic diagram of an electrical information processing system according to the present invention.
The components in the figures are numbered as follows: the system comprises a database 100, a data processing module 200, a feedback information processing module 210, a power consumption information processing module 220, a clustering module 300, a power consumption index clustering module 310 and a characteristic index clustering module 320.
Detailed Description
The following detailed description is provided to further explain the claimed embodiments of the present invention in order to make it clear for those skilled in the art to understand the claims. The scope of the invention is not limited to the following specific examples. It is intended that the scope of the invention be determined by those skilled in the art from the following detailed description, which includes claims that are directed to this invention.
As shown in fig. 1, the present invention provides an electricity consumption information processing system based on an electric power insurance package, which includes a database 100, a data processing module 200, and a clustering module 300;
the database 100 is used for storing the customer name, the power consumption information and the customer feedback information of the power consumption customer, wherein the power consumption information is power consumption readings at a plurality of time points with the same time interval and comprises time type data and text type data, and the customer feedback information is information of events such as power failure, equipment damage, fire and the like fed back by the power consumption customer through a plurality of ways and comprises the time type data and the text type data.
The data processing module 200 includes a power consumption information processing module 220, the power consumption information processing module 220 is configured to obtain cumulative power consumption readings of a starting time and a terminating time of a power consumption time period, and then obtain a difference between the cumulative power consumption readings of the terminating time and the cumulative power consumption degrees of the starting time, so as to obtain a power consumption index, where the power consumption index refers to a power consumption amount in a certain power consumption time period, and the power consumption time period in this embodiment includes 30 days up to the present, 90 days up to the present, 180 days up to the present, and one year up to the present.
The data processing module 200 further includes a feedback information processing module 210, and the feedback information processing module 210 is configured to identify time-type data in a single piece of client feedback information, and obtain a feedback time period where the time-type data is located, where in this embodiment, the feedback time period includes 30 days up to the present, 90 days up to the present, 180 days up to the present, and a year up to the present. The feedback information processing module 210 is further configured to perform effective word segmentation on text type data in a single piece of client feedback information to obtain a plurality of effective words, perform semantic recognition on the plurality of effective words with the feedback feature words as recognition targets, that is, power failure, equipment damage, and fire, to obtain the number of times that each feedback feature value appears in the piece of client feedback information, and convert the number of times into a TF-IDF weight of each feedback feature word.
The TF-IDF weight of the feedback characteristic word is determined by the TF-IDF weight (TF _ IDF) of the feedback characteristic word t in the user feedback information dt,dComprises the following steps:
(tf_idf)t,d=tft,d*idft
in the formula, tft,dRepresenting the word frequency, n of the feedback characteristic word t appearing in the user feedback information dt,dFor feeding back the times of occurrence of the characteristic word t in the user feedback information d, ndFeeding back the total word number of the information d for the user; idftRepresenting the inverse text frequency of the feedback characteristic word t, N is the total amount of the feedback information of the user, NtAnd feeding back the information quantity of the user containing the feedback characteristic word t.
The clustering module 300 includes a power consumption index clustering module 310, and the power consumption index clustering module 310 is configured to convert the power consumption index into the power consumption grade in the power consumption time period according to the value ranges of different power consumption grades in different power consumption time periods. In this embodiment, the power consumption levels in each power consumption time period are divided into three levels, i.e., high, medium, and low, and therefore the power consumption levels include high/medium/low power consumption for 30 days up to now, high/medium/low power consumption for 90 days up to now, high/medium/low power consumption for 180 days up to now, and high/medium/low power consumption for one year up to now.
In the above technical solution, the clustering module 300 further includes a characteristic index clustering module 320, where the characteristic index clustering module 320 is configured to construct characteristic index row vectors from feedback time periods in feedback information of a single client and TF-IDF weights of feedback characteristic words, and construct multiple characteristic index row vectors of multiple electricity consumers as a characteristic index matrix X.
The construction method of the characteristic index row vector comprises the steps of forming TF-IDF weight values of all feedback characteristic words into TF-IDF weight value row vector TF _ IDFi.jConverting the feedback time period into a feedback time row vector Ti.jThen the feature index row vector is
Xi(j)=xi,j=(tf_idfi.j Ti.j),1≤j≤ai
Wherein i is the ith electricity utilization client, j is the jth user feedback information of the collusion and electricity utilization client, and aiAnd feeding back the number of information for the ith electricity utilization client in the current year.
Wherein a time row vector T is fed backi.jThe construction method comprises defining Ti.j=(t1,t2,t3,t4),t1,t2,t3,t4Respectively indicating whether the feedback time belongs to the distanceThe values of 30 days, 90 days, 180 days and one year are 0 or 1, 1 represents that the feedback time period is within, and 0 represents that the feedback time period is not within.
Assuming that the number of the electricity consumers is N, the user feedback information of each electricity consumer within 1 year up to now has a1,a2,…,ai,…,aNStrip (i is more than or equal to 1 and less than or equal to N, j is more than or equal to 1 and less than or equal to ai) And summarizing the user feedback information of all the electricity customers, and forming a characteristic index matrix X to be expressed as follows:
and solving the Euclidean distance between any one characteristic index row vector of the characteristic index matrix X and other characteristic index row vectors to obtain a characteristic index distance row vector D.
In the above technical solution, the euclidean distance between two feature index row vectors of the same electricity consumer is:
in the formula, the v1 th and v2 th characteristic index row vectors of the u electricity consumer are xu,v1,xu,v2(1≤u≤N,1≤v1,v2≤au)。
The Euclidean distance between two appeal characteristic elements of different electricity utilization clients is as follows:
in the formula, the v characteristic index row vector of the u electricity consumer is xu,v(1≤u≤N,1≤v≤au) The nth characteristic index row vector of the mth electricity consumer is xm,n(1≤m≤N,1≤n≤am)。
In the above technical solution, the feature index distance row vector D (u, v) is expressed as:
in the formula, u is the u-th (u is more than or equal to 1 and less than or equal to N) electricity customer, v is the v-th (v is more than or equal to 1 and less than or equal to a)u) Characteristic index row vector xu,vThe number of the feature index distance row vectors is
And selecting a characteristic index row vector corresponding to the minimum value in each characteristic index distance row vector D and the corresponding similar electricity utilization client, and converting the characteristic index row vector corresponding to the minimum value into a characteristic index, thereby obtaining the similar electricity utilization client corresponding to the same characteristic index.
I.e. in D (u, v), if D (x)u,v,xm,n) When D is min (D (u, v)), two feature index row vectors x are obtainedu,v,xm,nViewed as the same class, wherein u is 1. ltoreq. u, m. ltoreq. N, v. ltoreq. 1. ltoreq. v. ltoreq. au,1≤n≤am. As can be seen, all in allClass I, formAnd (4) the same type of electricity utilization client groups.
In the above technical solution, the method for converting the characteristic index row vector into the characteristic index includes converting the TF-IDF weight in the characteristic index row vector into a frequency class according to the value ranges of the TF-IDF weights in different frequency classes of different feedback characteristic words, thereby obtaining the corresponding characteristic index. In this embodiment, the frequency level of each feedback feature word is divided into three levels, i.e., high, medium, and low, and the frequency level of the feedback feature word in the feature index includes high/medium/low equipment damage frequency, high/medium/low power failure frequency, and high/medium/low fire frequency.
The invention also provides a processing method of the power utilization information processing system based on the power insurance package demand, which comprises the following steps:
step 1: the power consumption information processing module 220 obtains the accumulated power consumption readings of the starting time and the ending time of the power consumption time period, and then obtains the difference between the accumulated power consumption readings of the ending time and the accumulated power consumption degrees of the starting time to obtain the power consumption in different power consumption time periods;
the feedback information processing module 210 identifies time-type data in a single piece of client feedback information to obtain a feedback time period in which the time-type data is located, and obtains a TF-IDF weight of each feedback feature word according to text-type data in the client feedback information;
step 2: the power consumption index clustering module 310 converts the power consumption index into the power consumption grade in the power consumption time period according to the value ranges of different power consumption grades in different power consumption time periods;
the characteristic index clustering module 320 constructs feedback time periods in the feedback information of the single client and TF-IDF weights of the feedback characteristic words into characteristic index row vectors, and constructs a plurality of characteristic index row vectors of a plurality of power utilization clients into a characteristic index matrix X; solving the Euclidean distance between any one characteristic index row vector of the characteristic index matrix X and other characteristic index row vectors to obtain a characteristic index distance row vector D; and selecting a characteristic index row vector corresponding to the minimum value of each characteristic index distance row vector D and the corresponding similar electricity utilization customers, and converting the characteristic index row vector corresponding to the minimum value into the characteristic index according to different frequency grade value ranges of different feedback characteristic values, thereby obtaining the similar electricity utilization customers corresponding to the same characteristic index.
The electricity consumption information processing process of the present invention is illustrated below with reference to specific examples: in the following table, 4 business clients (N ═ 4) and 11 client feedback information samples are taken as examples. By 11, 16 days in 2020, 4 pieces of customer feedback information, 1 piece of customer feedback information of the electricity customer 2, 5 pieces of customer feedback information of the electricity customer 3, and 1 piece of customer feedback information of the electricity customer 4, which are initiated by the electricity customer 1 in the past year, are collected.
Table 1 customer feedback information summary table
The user feedback information processing process is as follows:
1. constructing a feature index matrix X
According to the processing method, the characteristic index row vector corresponding to each feedback record of each power consumption client is obtained, and the TF-IDF weight row vector TF _ IDF is used for calculating the power consumption index row vectori.jAnd a feedback time row vector Ti.jTwo parts are formed.
For the electricity consumption client 1, the total number of the characteristic index row vectors is 4 within 1 year, and the characteristic index row vector is marked as a1I.e. a14, the characteristic index matrix of the electricity customer is marked as X1And is represented by
Similarly, for the electricity consumer 2, a2=1,
X2=[x2,1]=[tf_idf2.1 T2.1]=[0.17488 0.08333 0.52465 0 0 1 0]。
For the electricity consumer 3, a3=5,
For the electricity consumers 4, a4=1,
X4=[x4,1]=[tf_idf4.1 T4.1]=[0 0.1035 0.07237 1 0 0 0]。
Summarizing the characteristic index matrixes of the electricity customers to form a total characteristic index matrix X which is expressed as:
2. And constructing a feature index distance row vector matrix D.
And calculating the Euclidean distance between the self characteristic index row vectors of the same power consumption client.
For example, the self characteristic index row vector of the electricity client 1 comprises x1,1,x1,2,x1,3,x1,4Then, the distance between two is:
similarly, d (x) can also be calculated1,1,x1,3)=0.1286,d(x1,1,x1,4)=0.2797,d(x1,2,x1,3)=0.1027,d(x1,2,x1,4)=0.1818,d(x1,3,x1,4)=0.2307。
And calculating Euclidean distances between different characteristic index row vectors of different electricity customers.
If the 1 st characteristic index row vector of the electricity customer 1 is x1,1And the 1 st characteristic index row vector of the 3 rd electricity consumer is x3,1Then, the euclidean distance between two feature index row vector elements is:
similarly, the Euclidean distance between different characteristic index row vectors of the same power customer and different characteristic indexes of other power customers are calculatedAnd summarizing Euclidean distances between the line vectors to form a characteristic index distance line vector D. For example, the v (1 ≦ v ≦ a) of the u-th (1 ≦ u ≦ 4) electricity consumer in the feature index matrix Xu) Characteristic index row vector xu,vAnd the corresponding characteristic index distance matrix is marked as D (u, v) and expressed as:
D(u,v)=[d(xu,v,x1,1),…,d(xu,v,x1,4),…,d(xu,v,xm,n),…,d(xu,v,x3,1),…,d(xu,v,x4,1)]1×11
and (3) longitudinally arranging the characteristic index distance row vectors in sequence to form a characteristic index distance matrix W with the size of 11 multiplied by 11.
3. Cluster screening
And clustering each row vector D (u, v) in the characteristic index distance matrix W, wherein the distance method between the clusters is nearest neighbor. The judgment conditions are as follows: in D (u, v), if D (x)u,v,xm,n) When D is min (D (u, v)), two feature index row vectors x are obtainedu,v,xm,nViewed as a class, xu,v,xm,nThe row vectors are indexed for two different features. Wherein u is more than or equal to 1, m is more than or equal to 4, v is more than or equal to 1 and less than or equal to au,1≤n≤am。
Using the feature index row vector x3,1For example, the corresponding feature index distance row vector is represented as:
D(3,1)=[d(x3,1,x1,1),…,d(x3,1,x1,4),d(x3,1,x2,1),d(x3,1,x3,1),…,d(x3,1,x3,5),d(x3,1,x4,1)]
=[0.03333,0.1228,0.09524,0.2618,1.5208,0,0,0,0,0.1191,1.4175]。
dividing D (3,1) by D (x)3,1,x3,1) The minimum value outside (which is the Euclidean distance from itself and is not taken into account) is denoted by min (D (3,1)), and the feature index row vector set corresponding to the minimum value is determined to be x3,1For the same kind of feature index row vector, specifically, min (D (3,1)) takes a value of 0, and the corresponding feature data point set is { x }3,2,x3,3,x3,4Are x3,1And { x3,2,x3,3,x3,4Are grouped into one class, denoted as { x }3,1,x3,2,x3,3,x3,4And indicates that the 4 characteristic index row vectors inside the electricity client 3 have similar characteristics.
And (3) sequentially judging the minimum value of each row vector D (u, v) in the characteristic index distance matrix W and the corresponding characteristic index row vector by taking the minimum distance as a principle, and summarizing to obtain the following similar characteristic index row vectors:
x3,5,{x1,1,x3,1,x3,2,x3,3,x3,4},x1,3,x1,2,x1,4,x2,1,x4,1。
and converting the characteristic index row vectors of the same type into characteristic indexes of the same type, wherein the characteristic indexes of the same type of each power consumption client are shown in a table 2, and the representative characteristic index is the characteristic index with the longest feedback time period and the highest frequency grade in a certain feedback characteristic word of a certain power consumption client.
TABLE 2 summary of similar characteristic indexes
4. Electric power insurance package
Firstly, determining the period of the electric power insurance according to the feedback time period in the characteristic indexes, for example, if the feedback time period in a certain characteristic index is 1 year from the present, the period of the electric power insurance is 1 year, and then selecting the power consumption grade of the power consumption index from the present year; and then determining the premium grade according to the feedback characteristic words in the characteristic indexes and the frequency grade of the feedback characteristic words:
(1) when a "fire" occurs in the characteristic indicator, it is a first premium, and then three levels of high/medium/low may be classified in the first premium according to the frequency level of the fire.
(2) And when the characteristic indexes have 'equipment damage' and/or 'power failure' and no 'fire disaster', judging the premium grade according to the power consumption grade index.
(2.1) when the grade of the electricity consumption in the electricity application time period is a high grade, the second-grade premium is paid, and then three grades of high/medium/low are divided in the second-grade premium according to the frequency grade of 'equipment damage' and/or 'power failure'.
(2.2) when the grade of the electricity consumption in the electricity application time period is the middle/low grade, the third-grade premium is paid, and then the third-grade premium is divided into a high grade, a middle grade and a low grade according to the frequency grade of 'equipment damage' and/or 'power failure'.
The first, second and third premium are decreased in size in turn, and the high/medium/low premium in the same premium is decreased in size in turn.
And according to the specified rule, sequentially judging the characteristic indexes of the four electricity utilization users, thereby making the most appropriate power insurance package for each electricity utilization client. And the similar electric insurance packages can be formulated according to the characteristic indexes corresponding to the similar characteristic index row vectors, so that the targeted active service is performed on the similar electricity utilization customer groups.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Claims (10)
1. The utility model provides an electricity consumption information processing system based on electric power insurance package which characterized in that: comprises a database (100), a data processing module (200) and a clustering module (300);
the data processing module (200) is used for obtaining a power consumption index according to the power consumption information and obtaining a feedback index according to the single piece of client feedback information; the power consumption index comprises a power consumption time period and power consumption in the power consumption time period, and the feedback index comprises a feedback time period and TF-IDF weight values of feedback characteristic words;
the clustering module (300) is used for obtaining a power consumption grade index according to the power consumption index, and clustering and screening according to a plurality of feedback indexes of a plurality of power consumption customers to obtain the same type of power consumption customers with the same characteristic index, wherein the characteristic index is the frequency grade of each feedback characteristic word in the feedback time period.
2. The power insurance package-based electricity information processing system according to claim 1, wherein: the data processing module (200) further comprises: the feedback information processing module (210) is used for identifying the time type data in the single piece of client feedback information to obtain the feedback time period of the time type data.
3. The power insurance package-based electricity information processing system according to claim 2, wherein: the feedback information processing module (210) is further configured to perform word segmentation processing on text-type data in the single piece of client feedback information to obtain a plurality of words, perform semantic recognition on the plurality of words by using the feedback feature words as recognition targets to obtain the number of times that each feedback feature value appears, and convert the number of times into a TF-IDF weight of each feedback feature word.
4. The power insurance package-based electricity information processing system according to claim 3, wherein: the data processing module (200) further comprises a power consumption information processing module (220), wherein the power consumption information processing module (220) is used for acquiring accumulated power consumption readings of the starting time and the ending time of the power consumption time period and then acquiring the difference between the accumulated power consumption readings of the ending time and the accumulated power consumption degrees of the starting time.
5. The power insurance package-based electricity information processing system according to claim 4, wherein: the power usage time period and the feedback time period include up to 30 days, up to 90 days, up to 180 days, and up to one year.
6. The power insurance package-based electricity information processing system according to claim 1, wherein: the clustering module (300) comprises a power consumption index clustering module (310), wherein the power consumption index clustering module (310) is used for converting power consumption indexes into power consumption grades in the power consumption time periods according to the value ranges of different power consumption grades in different power consumption time periods.
7. The power insurance package-based electricity information processing system according to claim 1, wherein: the clustering module (300) further comprises a characteristic index clustering module (320), wherein the characteristic index clustering module (320) is used for constructing a characteristic index row vector x by the feedback time period in the feedback information of the single client and the TF-IDF weight of each feedback characteristic wordi,jConstructing a plurality of characteristic index row vectors of a plurality of electricity utilization customers into a characteristic index matrix X; solving the Euclidean distance between any one characteristic index row vector and other characteristic index row vectors in the characteristic index matrix X to obtain a characteristic index distance row vector D; and selecting a characteristic index row vector corresponding to the minimum value in each characteristic index distance row vector D and a corresponding similar electricity utilization client, and converting the characteristic index row vector corresponding to the minimum value into a characteristic index to obtain the similar electricity utilization client corresponding to the same characteristic index.
8. The power insurance package-based electricity information processing system according to claim 1, wherein: the method for converting the characteristic index row vector into the characteristic index comprises the step of converting the TF-IDF weight in the characteristic index row vector into the frequency grade according to the value range of the TF-IDF weight in different frequency grades of different feedback characteristic words, so as to obtain the corresponding characteristic index.
9. A processing method of a power utilization information processing system based on power insurance package requirements is characterized by comprising the following steps:
obtaining a power consumption index according to the power consumption information and obtaining a feedback index according to single client feedback information; the power consumption index comprises a power consumption time period and power consumption in the power consumption time period, and the feedback index comprises a feedback time period and TF-IDF weight values of feedback characteristic words;
the method comprises the steps of obtaining a power consumption grade index according to the power consumption index, and carrying out clustering and screening according to a plurality of feedback indexes of a plurality of power consumption customers to obtain the same type of power consumption customers with the same characteristic index, wherein the characteristic index is the frequency grade of each feedback characteristic word in a feedback time period.
10. The processing method of the electricity consumption information processing system based on the demand of the power insurance package according to claim 9, wherein: the clustering and screening method comprises the following steps: establishing a feedback time period in the feedback information of a single client and TF-IDF weight values of feedback feature words as a feature index row vector, and establishing a plurality of feature index row vectors of a plurality of power utilization clients as a feature index matrix X; solving the Euclidean distance between any one characteristic index row vector of the characteristic index matrix X and other characteristic index row vectors to obtain a characteristic index distance row vector D; and selecting a characteristic index row vector corresponding to the minimum value of each characteristic index distance row vector D and the corresponding similar electricity utilization customers, and converting the characteristic index row vector corresponding to the minimum value into the characteristic index according to different frequency grade value ranges of different feedback characteristic values, thereby obtaining the similar electricity utilization customers corresponding to the same characteristic index.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110586589.8A CN113284007B (en) | 2021-05-27 | 2021-05-27 | Power consumption information processing system based on electric insurance package and processing method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110586589.8A CN113284007B (en) | 2021-05-27 | 2021-05-27 | Power consumption information processing system based on electric insurance package and processing method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113284007A true CN113284007A (en) | 2021-08-20 |
CN113284007B CN113284007B (en) | 2023-07-04 |
Family
ID=77282222
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110586589.8A Active CN113284007B (en) | 2021-05-27 | 2021-05-27 | Power consumption information processing system based on electric insurance package and processing method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113284007B (en) |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107506475A (en) * | 2017-09-08 | 2017-12-22 | 国网辽宁省电力有限公司 | A kind of magnanimity electric power customer service file classification method based on Spark |
US20180165302A1 (en) * | 2016-12-12 | 2018-06-14 | Facebook, Inc. | Systems and methods to provide local suggestions based on spectral clustering |
CN109063217A (en) * | 2018-10-29 | 2018-12-21 | 广州供电局有限公司 | Work order classification method, device and its relevant device in Electric Power Marketing System |
WO2019214133A1 (en) * | 2018-05-08 | 2019-11-14 | 华南理工大学 | Method for automatically categorizing large-scale customer complaint data |
US20190392493A1 (en) * | 2018-06-21 | 2019-12-26 | Lisa Therese Miller | Direct-To-Business Feedback Communication And Database Management System |
CN110826886A (en) * | 2019-10-29 | 2020-02-21 | 南京华盾电力信息安全测评有限公司 | Electric power customer portrait construction method based on clustering algorithm and principal component analysis |
CN110874532A (en) * | 2018-08-30 | 2020-03-10 | 北京京东尚科信息技术有限公司 | Method and device for extracting keywords of feedback information |
CN111177389A (en) * | 2019-12-30 | 2020-05-19 | 佰聆数据股份有限公司 | NLP technology-based classification method, system and storage medium for power charge notification and customer appeal collection |
CN111612228A (en) * | 2020-05-12 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | User electricity consumption behavior analysis method based on electricity consumption information |
CN111652757A (en) * | 2020-04-22 | 2020-09-11 | 国网江苏省电力有限公司营销服务中心 | Method and device for analyzing customer behaviors in electric power business hall |
CN111709791A (en) * | 2020-06-19 | 2020-09-25 | 四川中电启明星信息技术有限公司 | Power supply marketing service method based on improved feature word weight algorithm |
US20210133848A1 (en) * | 2019-11-05 | 2021-05-06 | Shopify Inc. | Systems and methods for using keywords extracted from reviews |
-
2021
- 2021-05-27 CN CN202110586589.8A patent/CN113284007B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180165302A1 (en) * | 2016-12-12 | 2018-06-14 | Facebook, Inc. | Systems and methods to provide local suggestions based on spectral clustering |
CN107506475A (en) * | 2017-09-08 | 2017-12-22 | 国网辽宁省电力有限公司 | A kind of magnanimity electric power customer service file classification method based on Spark |
WO2019214133A1 (en) * | 2018-05-08 | 2019-11-14 | 华南理工大学 | Method for automatically categorizing large-scale customer complaint data |
US20190392493A1 (en) * | 2018-06-21 | 2019-12-26 | Lisa Therese Miller | Direct-To-Business Feedback Communication And Database Management System |
CN110874532A (en) * | 2018-08-30 | 2020-03-10 | 北京京东尚科信息技术有限公司 | Method and device for extracting keywords of feedback information |
CN109063217A (en) * | 2018-10-29 | 2018-12-21 | 广州供电局有限公司 | Work order classification method, device and its relevant device in Electric Power Marketing System |
CN110826886A (en) * | 2019-10-29 | 2020-02-21 | 南京华盾电力信息安全测评有限公司 | Electric power customer portrait construction method based on clustering algorithm and principal component analysis |
US20210133848A1 (en) * | 2019-11-05 | 2021-05-06 | Shopify Inc. | Systems and methods for using keywords extracted from reviews |
CN111177389A (en) * | 2019-12-30 | 2020-05-19 | 佰聆数据股份有限公司 | NLP technology-based classification method, system and storage medium for power charge notification and customer appeal collection |
CN111652757A (en) * | 2020-04-22 | 2020-09-11 | 国网江苏省电力有限公司营销服务中心 | Method and device for analyzing customer behaviors in electric power business hall |
CN111612228A (en) * | 2020-05-12 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | User electricity consumption behavior analysis method based on electricity consumption information |
CN111709791A (en) * | 2020-06-19 | 2020-09-25 | 四川中电启明星信息技术有限公司 | Power supply marketing service method based on improved feature word weight algorithm |
Non-Patent Citations (1)
Title |
---|
孙毅 等: "面向售电侧改革的用户分层聚类与套餐推荐方法", 《电网技术》, pages 447 - 454 * |
Also Published As
Publication number | Publication date |
---|---|
CN113284007B (en) | 2023-07-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US10504194B2 (en) | Correlating consumption and voltage data to identify line loss in an electric grid | |
CN111401777B (en) | Enterprise risk assessment method, enterprise risk assessment device, terminal equipment and storage medium | |
CN109597936B (en) | New user screening system and method | |
CN106570778A (en) | Big data-based data integration and line loss analysis and calculation method | |
CN102542474A (en) | Method for sorting inquiry results and device | |
KR101868729B1 (en) | Resource portfolio processing method, device, apparatus and computer strorage medium | |
CN111881221A (en) | Method, device and equipment for customer portrait in logistics service | |
US20210125272A1 (en) | Using Inferred Attributes as an Insight into Banking Customer Behavior | |
CN108304990B (en) | Power failure sensitivity pre-judging method and system | |
CN112184489A (en) | Power consumer grouping management system and method | |
CN110866698A (en) | Device for assessing service score of service provider | |
CN105359172A (en) | Calculating a probability of a business being delinquent | |
CN115204881A (en) | Data processing method, device, equipment and storage medium | |
Triantis | Fuzzy non-radial data envelopment analysis (DEA) measures of technical efficiency in support of an integrated performance measurement system | |
CN113284007A (en) | Power utilization information processing system based on power insurance package and processing method thereof | |
CN109697203A (en) | Index unusual fluctuation analysis method and equipment, computer storage medium, computer equipment | |
CN110472827B (en) | System policy early warning method, device, server and readable storage medium | |
CN108109002B (en) | Data processing method and device | |
KR101997613B1 (en) | Matching system and method for corporate and event using artificial intelligence | |
CN111061615A (en) | Data monitoring method and device for data warehouse, server and storage medium | |
CN106021326A (en) | A stream-computing event processing method and device | |
TWI700596B (en) | Information integrating system and information integrating method | |
WO2023144949A1 (en) | Risk countermeasure assistance device, learning device, risk countermeasure assistance method, learning method, and program | |
CN115409381A (en) | Line loss cause determination method and device, electronic equipment and storage medium | |
WO2013190627A1 (en) | Correlation analyzing device and correlation analyzing method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |