CN110728537A - Prediction payment method based on power consumer behavior label - Google Patents
Prediction payment method based on power consumer behavior label Download PDFInfo
- Publication number
- CN110728537A CN110728537A CN201910904832.9A CN201910904832A CN110728537A CN 110728537 A CN110728537 A CN 110728537A CN 201910904832 A CN201910904832 A CN 201910904832A CN 110728537 A CN110728537 A CN 110728537A
- Authority
- CN
- China
- Prior art keywords
- user
- payment
- information
- time
- power
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000005611 electricity Effects 0.000 claims abstract description 74
- 238000004891 communication Methods 0.000 claims abstract description 10
- 230000004927 fusion Effects 0.000 claims description 8
- 238000003064 k means clustering Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 5
- 238000007689 inspection Methods 0.000 claims description 5
- 230000000087 stabilizing effect Effects 0.000 claims description 5
- 230000001502 supplementing effect Effects 0.000 claims description 5
- 238000005516 engineering process Methods 0.000 claims description 2
- 230000011218 segmentation Effects 0.000 claims description 2
- 230000006399 behavior Effects 0.000 description 51
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 6
- 238000004364 calculation method Methods 0.000 description 3
- 238000005314 correlation function Methods 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000004888 barrier function Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 239000011541 reaction mixture Substances 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- 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
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Item recommendations
-
- 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
Abstract
The invention discloses a prediction payment method based on a power consumer behavior label, and relates to the technical field of power information management; the method comprises the steps of S1, determining a power user behavior label, obtaining and classifying the user behavior label according to the power consumption, the payment history, the amount, the power consumption habit and the basic information of a user, and displaying the user information at a receiving end when the user initiates communication, wherein the user information comprises characteristics, the belonged classification, the dialing history, the historical consultation problem and corresponding solutions; s2, predicting payment according to the user behavior label, and predicting the electricity consumption, the payment time and the payment amount of the user according to the user behavior label; the power user behavior tag is determined through S1, the payment step is predicted according to the user behavior tag through S2, and the like, so that the payment time prediction is realized.
Description
Technical Field
The invention relates to the technical field of power information management, in particular to a forecasting payment method based on a power consumer behavior label.
Background
At present, with the continuous promotion of the innovation of the power system, a large number of electricity selling companies are established, the market development competitive pressure of power grid enterprises is increased, the user behaviors are analyzed and determined, the electricity consumption and the payment amount of a user are accurately predicted, high-quality service push is provided for the customer, and the satisfaction degree and the dependence viscosity of the user on the power grid enterprises are increased.
Currently, when a user dials 95598, different user characteristics cannot be distinguished in time, and the problem and a related solution scheme presented by the user cannot be predicted; meanwhile, under the condition that the power enterprises cannot accurately predict different characteristics, the power consumption of users and the predicted payment time are realized, and the electric charge risk is avoided; in addition, the information push of the power enterprise mostly adopts the modes of business hall propaganda, short message common sending and the like, so that a large amount of resources are wasted, and different customized service schemes cannot be provided for different customers.
The prior art scheme is mainly manual screening of historical dialing problems of users, cannot accurately predict user quantity and payment time, is commonly sent marketing service short messages and publicized offline, and is single in strategy and poor in effect.
Problems with the prior art and considerations:
how to determine the user behavior label and quickly solve the user problem. The method and the system can accurately predict the power consumption and the payment time of the user based on the user behavior tag, and accurately push the schemes of the power consumption habit of the user, the predicted payment time, the optimal payment and benefit of the user and the like, so that the satisfaction degree and loyalty degree of the user are improved.
Disclosure of Invention
The technical problem to be solved by the invention is to provide a prediction payment method based on a power consumer behavior tag, which realizes payment time prediction by determining the power consumer behavior tag through S1, predicting payment steps according to the user behavior tag through S2 and the like.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows: the method comprises the following steps:
s1 determining power consumer behavior tag
The method comprises the steps that user behavior labels and classifications are obtained according to power consumption, payment history, amount, power consumption habits and basic information of a user, when the user initiates communication, the user information is displayed at a receiving end, and the user information comprises characteristics, belonged classifications, dialing history, historical consultation problems and corresponding solutions;
s2 forecasting payment according to user behavior label
And predicting the electricity consumption, the payment time and the payment amount of the user according to the user behavior label.
The further technical scheme is as follows: the method also comprises the following steps of,
s3 information push
And pushing service information to the user according to the user label, the payment amount, the payment time and the prediction result of the electricity consumption.
The further technical scheme is as follows: the step of S1 determining the power consumer behavior tag specifically includes the following steps,
s101 fusion and reconstruction of power marketing business data and customer service data
Carrying out data table association on the user call information and the payment service data by using a user name as a main key;
s102, establishing user behavior labels
Acquiring the electricity consumption, the payment history, the payment amount, the payment mode and the preferred payment information of the user according to the electricity consumption, the payment history, the amount, the electricity consumption habit and the structured data of the basic information of the user, carrying out k-means clustering, and determining a user behavior label;
s103, classifying user behavior labels
Extracting call information and converting the call information into character information, extracting keywords of a landlord, a tenant, password resetting, an intermediary and property rights in the character information, positioning a user according to keyword matching and classifying user behavior labels;
s104 associated user behavior tag and business support system
And summarizing the historical call making problem and problem solution of the user and related problems and solutions, and displaying the characteristics, the belonged classification, the call making history, the historical consultation problem and the corresponding solution of the user at a receiving end when the user initiates communication.
The further technical scheme is as follows: the step of S102 establishing the user behavior label specifically comprises the steps of calculating the power consumption, the same-month power consumption, the same-quarter power consumption and the payment amount of the user over the years, calculating the average value, the variance and the standard deviation of the power consumption, the payment history, the payment amount, the payment mode and the preferred payment information of the user based on the power consumption, the payment history, the payment amount, the power consumption habits and the structured data of the basic information of the user in the marketing system and the power consumption information acquisition system, and determining the user behavior label.
The further technical scheme is as follows: the step of S103 classifying the user behavior labels specifically comprises the steps of firstly extracting call information based on voice call non-structural data in a customer service system, converting the call information into character information through voice recognition, extracting characteristic hot word keywords of landlords, tenants, password resetting, intermediaries and property rights in the character information through an NLP word segmentation technology, utilizing hot word matching to position users, and performing user division in cooperation with the step S102, wherein the user division is divided into old communities, group rents, vacant rooms, landlords, online electronic payment and offline payment.
The further technical scheme is as follows: the step of S104 associating the user behavior tag with the service support system specifically includes that for the problem and the solution of the problem of the historical call making of the user, the associated problems and the solutions are gathered, the user behavior tag is embedded into the service support system, when the user makes a call, the user characteristics, the belonged classification, the call history, the historical consultation problem and the corresponding solutions are displayed in front of a customer service computer, the customer service specialist is helped to quickly obtain the solutions, and when the user is an offline payment user, suggested online payment information is pushed to the user.
The further technical scheme is as follows: in step S2, the information of the user' S electricity consumption, the historical electricity purchasing amount and the payment time is collated, and a time series is drawn with the time as the horizontal axis and the electricity purchasing amount and the electricity consumption as the vertical axis; and judging whether the time sequence is stable, and if the time sequence is not stable, stabilizing the time sequence and establishing a corresponding time sequence prediction model.
The further technical scheme is as follows: in the step S2, the method specifically includes the following steps:
s201, calculating self-correlation and partial correlation of electricity consumption, historical payment amount and payment time
Calculating self-correlation ACF and partial correlation PACF of the time series of the electricity consumption z1, the historical payment amount z2 and the payment time z 3;
s202 smoothness inspection
If the sequence is not stable, reconstructing the sequence after first-order difference and second-order difference, and performing stability check again until the sequence is stable and stable after n-order difference, wherein d is n;
s203, establishing a time series model
Establishing a corresponding time series model for balancing the complexity of the model and the goodness of fitting data, wherein the smaller the result is, the better the fitting is represented, and the formula is as follows:
AIC ═ 2p + n (log (SSE/n)) formula 3
In the formula 3, p is the number of independent variables in the prediction model, n is the sample size, SSE is the sum of squares of residual errors, the minimum result of AIC is selected, and finally a time series model is established for prediction;
s204, establishing an error compensation function
Construction of daily average air temperature TtThe formula of the incidence relation function r with the electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is as follows:
in the formula 4, x is an average air temperature function and is expressed in centigrade degrees; y is a power purchase amount function, the unit is element, or y is a payment time function, the unit is hour, or y is a power consumption function, the unit is degree; mx is the average of the function x, my is the average of the function y, d is 0,1, 2.. N-1;
for the prediction error etAnd supplementing, wherein the compensation formula is as follows:
etis an error index at time t, rtIs a correlation index, TtIs the change of temperature at time t and time t-1Value, S is TtU is a constant.
The further technical scheme is as follows: the user initiates communication through the mobile terminal.
The further technical scheme is as follows: the user information is displayed on the receiving end through a display of the server.
Adopt the produced beneficial effect of above-mentioned technical scheme to lie in:
the power user behavior tag is determined through S1, the payment step is predicted according to the user behavior tag through S2, and the like, so that the payment time prediction is realized.
The power consumer behavior tag is determined through S1, the power consumer behavior tag is judged according to the predicted payment of the power consumer behavior tag, S2 electricity utilization habits, predicted payment time, payment discount policies, energy-saving schemes and the like are pushed in the forms of short messages, WeChat public numbers and the like, and user satisfaction and loyalty are improved.
See detailed description of the preferred embodiments.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the application, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application, but the present application may be practiced in other ways than those described herein, and it will be apparent to those of ordinary skill in the art that the present application is not limited to the specific embodiments disclosed below.
Example 1:
as shown in fig. 1, the invention discloses a forecasting payment method based on a power consumer behavior tag, which comprises the following steps:
s101 fusion and reconstruction of power marketing business data and customer service data
And associating the call information of the user 95598 with service data such as power acquisition and payment by using a user name as a main key.
S102, establishing user behavior labels
According to the method, structured data of power consumption, payment history, amount, power consumption habits and basic information of a user are used for counting the power consumption, the same-month power consumption, the same-quarter power consumption, payment time and amount of the user over the year, calculating the increase rate, the mean value, the variance and the standard deviation of the power consumption, the overdue payment days, the payment channel, the mean value, the variance and the standard deviation of the payment amount, obtaining information of power consumption, payment preference and the like of the user, carrying out k-means clustering and determining a user behavior label.
S103, establishing user classification characteristics
And extracting call information and converting the call information into character information, extracting keys of landlords, tenants, password resetting, intermediaries, property rights and the like in the character information, positioning users according to the matching of the keys, and determining classification characteristics of old cells, vacant rooms, group rents and the like.
S104 associated user behavior tag and business support system
And summarizing the historical call making problem, the problem solution and the related problems of the user, and displaying the user characteristics, the belonged classification, the call making history, the historical consultation problem and the corresponding solution at a receiving end when the user initiates communication.
S2 forecasting payment according to user behavior label
And predicting the electricity consumption, the payment time and the payment amount of the user according to the user behavior label.
The information of the user electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is collated, and a time sequence is drawn by taking time as a horizontal axis and taking the electricity purchasing amount and the electricity consumption as vertical axes. Judging whether the time sequence is stable, if not, stabilizing the sequence, and establishing a corresponding time sequence prediction model ARIMA (p, d, q), which comprises the following steps:
s201, calculating self-correlation and partial correlation of electricity consumption, historical payment amount and payment time
And calculating the self-correlation ACF and partial correlation PACF of the time series of the electricity consumption z1, the historical payment amount z2 and the payment time z3, wherein the self-correlation functions ACF and PACF describe the linear correlation between the time series observed value and the past observed value thereof. The calculation formula is as follows:
in formula 1, function ytThe electricity consumption z1 is expressed in unit of degree, or the payment amount is expressed in unit of element; cov (y)t,yt-k) As a function ytAnd a certain time k, yt-kA covariance of (a), which is a constant; var (y)t) Is ytThe variance of (c) is constant.
In the formula 2, rYkIs the autocorrelation of the function Y with a certain time k, rktIs the autocorrelation of Y with the time k, t.
A sequence is not stationary if the autocorrelation and partial correlation maps of the sequence are neither trailing nor truncated.
S202 smoothness inspection
If the sequence is not stable, reconstructing the sequence after first-order difference and second-order difference, and performing stability check again until the sequence is stable and stable after n-order difference, wherein d is n;
s203, establishing a time series model
Establishing a corresponding time series ARIMA (p, d, q) model, setting different combinations of p and q in [1,15], and respectively calculating AIC information entropy of the time series model under different combinations for balancing complexity of the estimated model and goodness of fitting data, wherein the smaller the result is, the better the fitting is represented, wherein the formula is as follows:
AIC ═ 2p + n (log (SSE/n)) formula 3
In the formula 3, p is the number of independent variables in the prediction model, n is the sample size, SSE is the sum of squares of residuals, the minimum result of AIC is selected, and finally, a time series model is established for prediction.
S204, establishing an error compensation function
Construction of daily average air temperature TtThe formula of the incidence relation function r with the electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is as follows:
in the formula 4, x is an average air temperature function and is expressed in centigrade degrees; y is a power purchase amount function, the unit is element, or y is a payment time function, the unit is hour, or y is a power consumption function, the unit is degree; mx is the average of the function x, my is the average of the function y, and d is 0,1, 2.
For the prediction error etAnd supplementing, wherein the compensation formula is as follows:
etis an error index at time t, rtIs a correlation index, TtIs the change value of the temperature at the T moment and the T-1 moment, S is TtThe symbol of (a), u is a constant, and needs to be obtained through a plurality of experiments.
S3 information push
And carrying out information push according to the user label, the electricity consumption, the payment amount and the payment time prediction result. Through the modes of short messages, WeChat public numbers and business halls, the electricity utilization habits are informed, the information of the payment time and the predicted use time of the payment is predicted, the payment preferential policy and the user energy-saving scheme are intelligently pushed, and the user satisfaction is improved.
Example 2:
the invention discloses a prediction payment method based on a power consumer behavior label, which comprises the following steps:
s1 determining power consumer behavior tag
S101 fusion and reconstruction of service data such as power marketing and power utilization information acquisition and 95598 customer service data
And the user call information is associated with service data such as power marketing, power utilization information acquisition and the like by using a user name as a main key to perform data table association, missing data is cleaned, and fusion and reconstruction of power marketing service data and 95598 customer service data are realized.
S102, establishing user behavior labels
The method comprises the steps of counting the power consumption of a user over the years, the power consumption of the same month and season, the payment time and the amount, calculating the increase rate, the mean value, the variance, the standard deviation, the overdue number of payment days, the payment channel and the mean value, the variance and the standard deviation of the payment amount, obtaining the power consumption of the user, the payment preference and other information, carrying out k-means clustering, and determining a user behavior label according to the structured data of the power consumption, the payment history, the payment amount, the payment mode, the power consumption habit and the basic information of the user.
S103 classifying the user classification features
And extracting call information and converting the call information into character information, extracting keys of landlords, tenants, password resetting, intermediaries, property rights and the like in the character information, matching and positioning users according to the keys, and determining classification characteristics of old communities, vacant rooms, group rents and the like, an online electronic payment mode and offline payment characteristics by combining the step S102.
S104 associated user behavior tag and business support system
And summarizing the historical call-making problem, the problem solution and the related problems of the user, and when the user initiates communication again, displaying the user characteristics, the belonged classification, the call history, the historical consultation problem and the corresponding solution on a receiving end, a large screen and the like. When the user pays off line, electronic channel payment service and electronic channel payment preferential policy recommendation are developed in a targeted manner, user information is updated and perfected in time for the house renting user, and the recommended user is adjusted in time.
S2 forecasting payment according to user behavior label
And predicting the electricity consumption, the payment time and the payment amount of the user according to the user behavior label.
The information of the user electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is collated, and a time sequence is drawn by taking time as a horizontal axis and taking the electricity purchasing amount and the electricity consumption as vertical axes. Judging whether the time sequence is stable, if not, stabilizing the sequence, and establishing a corresponding time sequence prediction model ARIMA (p, d, q), which comprises the following steps:
s201, calculating self-correlation and partial correlation of electricity consumption, historical payment amount and payment time
And calculating the self-correlation ACF and partial correlation PACF of the time series of the electricity consumption z1, the historical payment amount z2 and the payment time z3, wherein the self-correlation functions ACF and PACF describe the linear correlation between the time series observed value and the past observed value thereof. The calculation formula is as follows:
in formula 1, function ytThe electricity consumption z1 is expressed in unit of degree, or the payment amount is expressed in unit of element; cov (y)t,yt-k) As a function ytAnd a certain time k, yt-kA covariance of (a), which is a constant; var (y)t) Is ytThe variance of (c) is constant.
In the formula 2, rYkIs the autocorrelation of the function Y with a certain time k, rktIs the autocorrelation of Y with the time k, t.
A sequence is not stationary if the autocorrelation and partial correlation maps of the sequence are neither trailing nor truncated.
S202 smoothness inspection
And if the sequence is not stable, reconstructing the sequence after first-order difference and second-order difference, and performing stability check again until the sequence is stable and stable after n-order difference, wherein d is equal to n.
S203, establishing a time series model
Establishing a corresponding time sequence ARIMA (p, d, q) model, setting different combinations of p and q in [1,15], respectively calculating AIC information entropies of the time sequence models under different combinations, and using the AIC information entropies to balance the complexity of the models and the goodness of fitting data, wherein the smaller the result is, the better the fitting is, wherein the formula is as follows:
AIC ═ 2p + n (log (SSE/n)) formula 3
In the formula 3, p is the number of independent variables in the prediction model, n is the sample size, SSE is the sum of squares of residuals, the minimum result of AIC is selected, and finally, a time series model is established for prediction.
S204, establishing an error compensation function
Construction of daily average air temperature TtThe formula of the incidence relation function r with the electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is as follows:
in the formula 4, x is an average air temperature function and is expressed in centigrade degrees; y is a power purchase amount function, the unit is element, or y is a payment time function, the unit is hour, or y is a power consumption function, the unit is degree; mx is the average of the function x, my is the average of the function y, and d is 0,1, 2.
For the prediction error etAnd supplementing, wherein the compensation formula is as follows:
etis an error index at time t, rtIs a correlation index, TtIs the change value of the temperature at the T moment and the T-1 moment, S is TtThe symbol of (a), u is a constant, and needs to be obtained through a plurality of experiments.
S3 information push
And carrying out information pushing according to the user label, the electricity consumption, the payment amount, the payment mode and the payment time prediction result, and adjusting pushing personnel in time. Through the modes of short messages, WeChat public numbers and business halls, the electricity utilization habits are informed, the predicted payment time and the predicted service time information of the payment are obtained, the payment preferential policy and the user energy-saving scheme are intelligently pushed, and the user satisfaction is improved.
The invention concept of the application is as follows:
the power consumer behavior label is determined through S1, the power consumer behavior label is judged according to the predicted payment of the power consumer behavior label through S2, the power consumption habit, the predicted payment time, the payment preferential policy, the energy-saving scheme and the like are pushed in the forms of short messages, WeChat public numbers and the like, the satisfaction degree and the loyalty degree of the power consumer are improved, and online electronic payment drainage is realized.
Technical contribution of the present application:
the method associates 95598 appeal with payment service data, and fuses and reconstructs power marketing service data and 95598 service data. Firstly, forming a user label through the attribute of a power user, power utilization information and payment behavior data, finely classifying users, helping customer service specialists to quickly identify client characteristics, and intelligently pushing a user problem solution scheme; and secondly, forecasting is carried out through the electricity purchasing amount and electricity purchasing time, the marketing opportunity of the user is controlled, the payment preferential policy and the energy-saving scheme are intelligently pushed, and the satisfaction degree and loyalty degree of the user are improved.
Description of the method:
s101 fusion and reconstruction of service data such as power marketing and power utilization information acquisition and 95598 customer service data
And the user call information and the service data such as power marketing, power utilization information acquisition and the like are associated by using the user name as a main key to perform data table association, so that the fusion and reconstruction of the power marketing service data and 95598 customer service data are realized.
S102, establishing user behavior labels
And carrying out data mining on structured data such as user electricity consumption, user payment history, amount, electricity utilization habits, user basic information and the like in a marketing system, electricity utilization information acquisition system and other systems of a national network company. And combing 1800 outgoing tenants and 5000 non-renters of one district as analysis samples. The method comprises the steps of counting the power consumption, the same-month power consumption, the same-quarter power consumption, the payment time and the amount of the user over the years, calculating the increase rate, the mean value, the variance, the standard deviation of the power consumption, the overdue days of payment, the payment channel and the mean value, the variance and the standard deviation of the payment amount, obtaining the power consumption, the payment preference and other information of the user, carrying out k-means clustering, and determining a user behavior label.
S103, establishing user classification characteristics
And extracting call information and converting the call information into character information, extracting keys of landlords, tenants, password resetting, intermediaries, property rights and the like in the character information, positioning users according to the matching of the keys, and determining classification characteristics of old cells, vacant rooms, group rents and the like.
S104 associated user behavior tag and business support system
And summarizing the historical call making problem, the problem solution and the related problems of the user, and displaying the user characteristics, the belonged classification, the call making history, the historical consultation problem and the corresponding solution at a receiving end when the user initiates communication.
S2 forecasting payment according to user behavior label
And predicting the electricity consumption, the payment time and the payment amount of the user according to the user behavior label.
The information of the user electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is collated, and a time sequence is drawn by taking time as a horizontal axis and taking the electricity purchasing amount and the electricity consumption as vertical axes. Judging whether the time sequence is stable, if not, stabilizing the sequence, and establishing a corresponding time sequence prediction model ARIMA (p, d, q), which comprises the following steps:
s201, calculating self-correlation and partial correlation of electricity consumption, historical payment amount and payment time
And calculating the self-correlation ACF and partial correlation PACF of the time series of the electricity consumption z1, the historical payment amount z2 and the payment time z3, wherein the self-correlation functions ACF and PACF describe the linear correlation between the time series observed value and the past observed value thereof. The calculation formula is as follows:
in the formula 1, the reaction mixture is,function ytThe electricity consumption z1 is expressed in unit of degree, or the payment amount is expressed in unit of element; cov (y)t,yt-k) As a function ytAnd a certain time k, yt-kA covariance of (a), which is a constant; var (y)t) Is ytThe variance of (c) is constant.
In the formula 2, rYkIs the autocorrelation of the function Y with a certain time k, rktIs the autocorrelation of Y with the time k, t.
A sequence is not stationary if the autocorrelation and partial correlation maps of the sequence are neither trailing nor truncated.
S202 smoothness inspection
And if the sequence is not stable, reconstructing the sequence after first-order difference and second-order difference, and performing stability check again until the sequence is stable and stable after n-order difference, wherein d is equal to n.
S203, establishing a time series model
Establishing a corresponding time sequence ARIMA (p, d, q) model, setting different combinations of p and q in [1,15], respectively calculating AIC information entropies of the time sequence models under different combinations, and using the AIC information entropies to balance the complexity of the models and the goodness of fitting data, wherein the smaller the result is, the better the fitting is, wherein the formula is as follows:
AIC ═ 2p + n (log (SSE/n)) formula 3
In the formula 3, p is the number of independent variables in the prediction model, n is the sample size, SSE is the sum of squares of residuals, the minimum result of AIC is selected, and finally, a time series model is established for prediction.
S204, establishing an error compensation function
Construction of daily average air temperature TtThe formula of the incidence relation function r with the electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is as follows:
in the formula 4, x is an average air temperature function and is expressed in centigrade degrees; y is a power purchase amount function, the unit is element, or y is a payment time function, the unit is hour, or y is a power consumption function, the unit is degree; mx is the average of the function x, my is the average of the function y, and d is 0,1, 2.
For the prediction error etAnd supplementing, wherein the compensation formula is as follows:
etis an error index at time t, rtIs a correlation index, TtIs the change value of the temperature at the T moment and the T-1 moment, S is TtThe symbol of (a), u is a constant, and needs to be obtained through a plurality of experiments.
S3 information push
And carrying out information push according to the user label, the electricity consumption, the payment amount and the payment time prediction result. Through the modes of short messages, WeChat public numbers and business halls, the electricity utilization habits are informed, the information of the payment time and the predicted use time of the payment is predicted, the payment preferential policy and the user energy-saving scheme are intelligently pushed, and the user satisfaction is improved.
Description of technical effects:
data fusion and reconstruction are realized, data barriers are broken through associating 95598 data with payment and electricity utilization information acquisition service data, and cross-unit and cross-professional data sharing is realized.
The user labels are formed by applying a k-means clustering algorithm, fine classification is carried out on the users, the user labels are embedded into the 95598 business system, customer service professionals are helped to quickly identify the characteristics of the customers, a user-related problem solution scheme is intelligently pushed, customer service problem solution efficiency is improved, and user experience is improved.
The method is determined by the user behavior label, the electricity consumption, the payment time and the amount of money of the user are predicted, the marketing opportunity can be controlled in time, the electricity consumption habit, the payment predicted time, the payment preferential policy, the energy-saving scheme and the like can be intelligently pushed, and the satisfaction degree and loyalty degree of the user are improved.
Claims (10)
1. A prediction payment method based on a power consumer behavior label is characterized in that: the method comprises the following steps:
s1 determining power consumer behavior tag
The method comprises the steps that user behavior labels and classifications are obtained according to power consumption, payment history, amount, power consumption habits and basic information of a user, when the user initiates communication, the user information is displayed at a receiving end, and the user information comprises characteristics, belonged classifications, dialing history, historical consultation problems and corresponding solutions;
s2 forecasting payment according to user behavior label
And predicting the electricity consumption, the payment time and the payment amount of the user according to the user behavior label.
2. The predictive payment method based on the power consumer behavior tag as claimed in claim 1, wherein: the method also comprises the following steps of,
s3 information push
And pushing service information to the user according to the user label, the payment amount, the payment time and the prediction result of the electricity consumption.
3. The predictive payment method based on the power consumer behavior tag as claimed in claim 1, wherein: the step of S1 determining the power consumer behavior tag specifically includes the following steps,
s101 fusion and reconstruction of power marketing business data and customer service data
Carrying out data table association on the user call information and the payment service data by using a user name as a main key;
s102, establishing user behavior labels
Acquiring the electricity consumption, the payment history, the payment amount, the payment mode and the preferred payment information of the user according to the electricity consumption, the payment history, the amount, the electricity consumption habit and the structured data of the basic information of the user, carrying out k-means clustering, and determining a user behavior label;
s103, classifying user behavior labels
Extracting call information and converting the call information into character information, extracting keywords of a landlord, a tenant, password resetting, an intermediary and property rights in the character information, positioning a user according to keyword matching and classifying user behavior labels;
s104 associated user behavior tag and business support system
And summarizing the historical call making problem and problem solution of the user and related problems and solutions, and displaying the characteristics, the belonged classification, the call making history, the historical consultation problem and the corresponding solution of the user at a receiving end when the user initiates communication.
4. The predictive payment method based on the power consumer behavior tag as claimed in claim 3, wherein: the step of S102 establishing the user behavior label specifically comprises the steps of calculating the power consumption, the same-month power consumption, the same-quarter power consumption and the payment amount of the user over the years, calculating the average value, the variance and the standard deviation of the power consumption, the payment history, the payment amount, the payment mode and the preferred payment information of the user based on the power consumption, the payment history, the payment amount, the power consumption habits and the structured data of the basic information of the user in the marketing system and the power consumption information acquisition system, and determining the user behavior label.
5. The predictive payment method based on the power consumer behavior tag as claimed in claim 3, wherein: the step of S103 classifying the user behavior labels specifically comprises the steps of firstly extracting call information based on voice call non-structural data in a customer service system, converting the call information into character information through voice recognition, extracting characteristic hot word keywords of landlords, tenants, password resetting, intermediaries and property rights in the character information through an NLP word segmentation technology, utilizing hot word matching to position users, and performing user division in cooperation with the step S102, wherein the user division is divided into old communities, group rents, vacant rooms, landlords, online electronic payment and offline payment.
6. The predictive payment method based on the power consumer behavior tag as claimed in claim 3, wherein: the step of S104 associating the user behavior tag with the service support system specifically includes that for the problem and the solution of the problem of the historical call making of the user, the associated problems and the solutions are gathered, the user behavior tag is embedded into the service support system, when the user makes a call, the user characteristics, the belonged classification, the call history, the historical consultation problem and the corresponding solutions are displayed in front of a customer service computer, the customer service specialist is helped to quickly obtain the solutions, and when the user is an offline payment user, suggested online payment information is pushed to the user.
7. The predictive payment method based on the power consumer behavior tag as claimed in claim 1, wherein: in step S2, the information of the user' S electricity consumption, the historical electricity purchasing amount and the payment time is collated, and a time series is drawn with the time as the horizontal axis and the electricity purchasing amount and the electricity consumption as the vertical axis; and judging whether the time sequence is stable, and if the time sequence is not stable, stabilizing the time sequence and establishing a corresponding time sequence prediction model.
8. The predictive payment method based on the power consumer behavior tag as claimed in claim 1, wherein: in the step S2, the method specifically includes the following steps:
s201, calculating self-correlation and partial correlation of electricity consumption, historical payment amount and payment time
Calculating self-correlation ACF and partial correlation PACF of the time series of the electricity consumption z1, the historical payment amount z2 and the payment time z 3;
s202 smoothness inspection
If the sequence is not stable, reconstructing the sequence after first-order difference and second-order difference, and performing stability check again until the sequence is stable and stable after n-order difference, wherein d is n;
s203, establishing a time series model
Establishing a corresponding time series model for balancing the complexity of the estimated model and the goodness of fitting data, wherein the smaller the result is, the better the fitting is represented, wherein the formula is as follows:
AIC ═ 2p + n (log (SSE/n)) formula 3
In the formula 3, p is the number of independent variables in the prediction model, n is the sample size, SSE is the sum of squares of residual errors, the minimum result of AIC is selected, and finally a time series model is established for prediction;
s204, establishing an error compensation function
Construction of daily average air temperature TtThe formula of the incidence relation function r with the electricity consumption z1, the historical electricity purchasing amount z2 and the payment time z3 is as follows:
in the formula 4, x is an average air temperature function and is expressed in centigrade degrees; y is a power purchase amount function, the unit is element, or y is a payment time function, the unit is hour, or y is a power consumption function, the unit is degree; mx is the average of the function x, my is the average of the function y, d is 0,1, 2.. N-1;
for the prediction error etAnd supplementing, wherein the compensation formula is as follows:
etis an error index at time t, rtIs a correlation index, TtIs the change value of the temperature at the T moment and the T-1 moment, S is TtU is a constant.
9. The predictive payment method based on the power consumer behavior tag as claimed in claim 1, wherein: the user initiates communication through the mobile terminal.
10. The predictive payment method based on the power consumer behavior tag as claimed in claim 1, wherein: the user information is displayed on the receiving end through a display of the server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910904832.9A CN110728537A (en) | 2019-09-24 | 2019-09-24 | Prediction payment method based on power consumer behavior label |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910904832.9A CN110728537A (en) | 2019-09-24 | 2019-09-24 | Prediction payment method based on power consumer behavior label |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110728537A true CN110728537A (en) | 2020-01-24 |
Family
ID=69218341
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910904832.9A Pending CN110728537A (en) | 2019-09-24 | 2019-09-24 | Prediction payment method based on power consumer behavior label |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110728537A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401431A (en) * | 2020-03-12 | 2020-07-10 | 成都小步创想慧联科技有限公司 | Group renting house identification method and system and storage medium |
| CN111582600A (en) * | 2020-05-15 | 2020-08-25 | 中国银行股份有限公司 | Behavior period determination method and device |
| CN111612230A (en) * | 2020-05-13 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | Client appeal trend early warning analysis method |
| CN112035715A (en) * | 2020-07-10 | 2020-12-04 | 广西电网有限责任公司 | User label design method and device |
| CN112184193A (en) * | 2020-10-26 | 2021-01-05 | 支付宝(杭州)信息技术有限公司 | Payment processing method and device |
| CN112215513A (en) * | 2020-10-22 | 2021-01-12 | 国网辽宁省电力有限公司营销服务中心 | Offline analysis method and system for user behavior events of power system |
| CN112381295A (en) * | 2020-11-13 | 2021-02-19 | 深圳供电局有限公司 | Resident electricity utilization reminding method and system based on electricity utilization behavior preference |
| CN112488421A (en) * | 2020-12-15 | 2021-03-12 | 国网雄安金融科技集团有限公司 | Tracking and predicting method and device for electric charge account receivable |
| CN112529620A (en) * | 2020-12-07 | 2021-03-19 | 北京来也网络科技有限公司 | RPA and AI-based generation method and device of electric power receivable report |
| CN113554454A (en) * | 2021-06-30 | 2021-10-26 | 西安图迹信息科技有限公司 | Big data is sold system with electric power and equipment thereof |
| CN113592140A (en) * | 2021-06-22 | 2021-11-02 | 国网宁夏电力有限公司吴忠供电公司 | Electric charge payment prediction model training system and electric charge payment prediction model |
| CN114429341A (en) * | 2022-01-24 | 2022-05-03 | 吉林银行股份有限公司 | Grouped payment method, device and equipment |
| CN115660663A (en) * | 2022-12-29 | 2023-01-31 | 北京易思汇商务服务有限公司 | Intelligent paying reminding method and system for study reservation |
| CN116433403A (en) * | 2023-06-14 | 2023-07-14 | 国网安徽省电力有限公司营销服务中心 | Account tracking-based electric enterprise accounts receivable early warning method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150088783A1 (en) * | 2009-02-11 | 2015-03-26 | Johnathan Mun | System and method for modeling and quantifying regulatory capital, key risk indicators, probability of default, exposure at default, loss given default, liquidity ratios, and value at risk, within the areas of asset liability management, credit risk, market risk, operational risk, and liquidity risk for banks |
| CN106651424A (en) * | 2016-09-28 | 2017-05-10 | 国网山东省电力公司电力科学研究院 | Electric power user figure establishment and analysis method based on big data technology |
| CN106776879A (en) * | 2016-11-29 | 2017-05-31 | 国网山东省电力公司电力科学研究院 | A kind of client's paying service information-pushing method |
| CN107423859A (en) * | 2017-08-07 | 2017-12-01 | 国家电网公司客户服务中心 | A kind of built-up pattern modeling method and system |
-
2019
- 2019-09-24 CN CN201910904832.9A patent/CN110728537A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150088783A1 (en) * | 2009-02-11 | 2015-03-26 | Johnathan Mun | System and method for modeling and quantifying regulatory capital, key risk indicators, probability of default, exposure at default, loss given default, liquidity ratios, and value at risk, within the areas of asset liability management, credit risk, market risk, operational risk, and liquidity risk for banks |
| CN106651424A (en) * | 2016-09-28 | 2017-05-10 | 国网山东省电力公司电力科学研究院 | Electric power user figure establishment and analysis method based on big data technology |
| CN106776879A (en) * | 2016-11-29 | 2017-05-31 | 国网山东省电力公司电力科学研究院 | A kind of client's paying service information-pushing method |
| CN107423859A (en) * | 2017-08-07 | 2017-12-01 | 国家电网公司客户服务中心 | A kind of built-up pattern modeling method and system |
Non-Patent Citations (1)
| Title |
|---|
| 章维维: "基于ARIMA修正模型的电力市场价格预测研究" * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401431A (en) * | 2020-03-12 | 2020-07-10 | 成都小步创想慧联科技有限公司 | Group renting house identification method and system and storage medium |
| CN111401431B (en) * | 2020-03-12 | 2023-07-25 | 成都小步创想慧联科技有限公司 | Group renting room identification method and system and storage medium |
| CN111612230A (en) * | 2020-05-13 | 2020-09-01 | 国网河北省电力有限公司电力科学研究院 | Client appeal trend early warning analysis method |
| CN111582600B (en) * | 2020-05-15 | 2022-06-03 | 中国银行股份有限公司 | Behavior period determination method and device |
| CN111582600A (en) * | 2020-05-15 | 2020-08-25 | 中国银行股份有限公司 | Behavior period determination method and device |
| CN112035715A (en) * | 2020-07-10 | 2020-12-04 | 广西电网有限责任公司 | User label design method and device |
| CN112035715B (en) * | 2020-07-10 | 2023-04-14 | 广西电网有限责任公司 | User label design method and device |
| CN112215513A (en) * | 2020-10-22 | 2021-01-12 | 国网辽宁省电力有限公司营销服务中心 | Offline analysis method and system for user behavior events of power system |
| CN112184193A (en) * | 2020-10-26 | 2021-01-05 | 支付宝(杭州)信息技术有限公司 | Payment processing method and device |
| CN112381295A (en) * | 2020-11-13 | 2021-02-19 | 深圳供电局有限公司 | Resident electricity utilization reminding method and system based on electricity utilization behavior preference |
| CN112529620A (en) * | 2020-12-07 | 2021-03-19 | 北京来也网络科技有限公司 | RPA and AI-based generation method and device of electric power receivable report |
| CN112488421A (en) * | 2020-12-15 | 2021-03-12 | 国网雄安金融科技集团有限公司 | Tracking and predicting method and device for electric charge account receivable |
| CN112488421B (en) * | 2020-12-15 | 2023-04-28 | 国网雄安金融科技集团有限公司 | Tracking and predicting method and device for accounts receivable of electric charge |
| CN113592140A (en) * | 2021-06-22 | 2021-11-02 | 国网宁夏电力有限公司吴忠供电公司 | Electric charge payment prediction model training system and electric charge payment prediction model |
| CN113554454A (en) * | 2021-06-30 | 2021-10-26 | 西安图迹信息科技有限公司 | Big data is sold system with electric power and equipment thereof |
| CN114429341A (en) * | 2022-01-24 | 2022-05-03 | 吉林银行股份有限公司 | Grouped payment method, device and equipment |
| CN114429341B (en) * | 2022-01-24 | 2022-12-02 | 吉林银行股份有限公司 | Grouped payment method, device and equipment |
| CN115660663A (en) * | 2022-12-29 | 2023-01-31 | 北京易思汇商务服务有限公司 | Intelligent paying reminding method and system for study reservation |
| CN116433403A (en) * | 2023-06-14 | 2023-07-14 | 国网安徽省电力有限公司营销服务中心 | Account tracking-based electric enterprise accounts receivable early warning method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110728537A (en) | Prediction payment method based on power consumer behavior label | |
| US8375068B1 (en) | Extensible framework and graphical user interface for sharing, comparing, and displaying resource usage data | |
| US20110022429A1 (en) | Resource reporting | |
| US20110023045A1 (en) | Targeted communication to resource consumers | |
| Sullivan et al. | Estimating power system interruption costs: A guidebook for electric utilities | |
| Smith et al. | Bicycle commuting in Melbourne during the 2000s energy crisis: A semiparametric analysis of intraday volumes | |
| Bijwaard et al. | Early mover advantages: An empirical analysis of European mobile phone markets | |
| US20200118223A1 (en) | Using artificial intelligence to process data extracted from utility bills | |
| CN112184313A (en) | Power marketing business application system based on client portrait data | |
| Worden et al. | Willingness to pay and pricing for broadband across the rural/urban divide in Canada | |
| Kim et al. | An intelligent product recommendation model to reflect the recent purchasing patterns of customers | |
| CN114445138A (en) | Hotel room type pricing method, device, equipment and storage medium | |
| Lyons | Timing and determinants of local residential broadband adoption: evidence from Ireland | |
| Eekhout et al. | Entrepreneurs' mobile phone appropriation and technical efficiency of informal firms in Dakar (Senegal) | |
| CN109544271A (en) | A kind of trade managing system | |
| US20210027332A1 (en) | Method of operating short-term rental restaurant and persuading users to visit short-term rental restaurant based on profile | |
| KR100799627B1 (en) | Customer analysis service method inside the business sector which uses the real-time integration of customer information | |
| Beko et al. | Demand models for direct mail and periodicals delivery services: Results for a transition economy | |
| Su et al. | The analysis on the determinants of mobile VIP customer churn: A logistic regression approach | |
| CN112767079A (en) | Operation service platform | |
| Gaudin | Using water bills to reinforce price signals: Evidence from the USA | |
| Kim et al. | A Multiattribute Model of the Timing of Buyer's Upgrading to Improved Versions of High Technology Products | |
| Milovanovic | Potentials of electronic business development in Serbia | |
| Fikri et al. | Technology Readiness of e-Government in the Use of Poverty Data for Social Assistance in Indonesia | |
| CN102799593A (en) | Personalized searching and sequencing method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200124 |