CN105991298B - A kind of method and apparatus for reforming ticket - Google Patents
A kind of method and apparatus for reforming ticket Download PDFInfo
- Publication number
- CN105991298B CN105991298B CN201510040759.7A CN201510040759A CN105991298B CN 105991298 B CN105991298 B CN 105991298B CN 201510040759 A CN201510040759 A CN 201510040759A CN 105991298 B CN105991298 B CN 105991298B
- Authority
- CN
- China
- Prior art keywords
- ticket
- feature
- sift
- system operation
- reforming
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The embodiment of the present invention provides a kind of method and apparatus for reforming ticket, and method includes: that starting ticket reforms mechanism；It is filtered out from error message and needs to reform reforming user and reforming business for ticket, the error message is to calculate information corresponding to the vicious daily ticket that generates in expense mechanism in ticket；According to it is described reform user and reform business carry out ticket reform to obtain correct ticket, the correct ticket eliminates the mistake in the vicious daily ticket.Due to regular ticket parallel in advance before the process that ticket is reformed, therefore the efficiency that ticket is reformed is greatly improved, it is done again since provided technology can voluntarily be chosen to reform user and reform business jargon substance of going forward side by side, therefore the fast quick-recovery of mistake that charge system can be likely to occur, it can be commonly used in the charge system of field of telecommunications, promote the level of user service.
Description
Technical field
The present invention relates to communication charge technologies, particularly relate to a kind of method and apparatus for reforming ticket.
Background technique
Charge system is business operation support system (BOSS, Business&Operation Support System)
Core, is responsible for every communication bill record expense of user, and provides query function.It is especially heavy that it calculates the accuracy taken
It wants, how customer perception clearly, therefore while guaranteeing that charge system is accurately run, quickly finds the problem simultaneously
Timely repair system is calculated expense mistake and is very important.
Since the configuration control of charge system office data is inadequate, network side ticket issues the factors such as exception and program exception,
Calculation takes possibility of the result there is mistake, once discovery, which is calculated, takes mistake, it is necessary to the calculation expense for reforming to correct mistake by ticket
As a result.Existing mode of reforming is extracted by system maintenance personnel to impacted user group after finding the problem, and is determined
User scope is reformed, ticket reforms operation and carries out reforming for the whole month generally be directed to a major class user, a major class business, reforming
When need to stop the charging of all users of such business.
It includes: to generate former entrance CDR file according to detailed monofile that charging bill, which is reformed, exports CDR file；According to accumulation
Amount operational order file is set as zero to relevant cumulant file relative recording；The detailed list that need to be analyzed is defined by configuration file
Data and the call bill data that need to be exported.In practical business, it should be noted that configuration file, log between multiple program entities
The problem of file, input directory, output directory, input directory and output file are with the presence or absence of repeating.
Existing ticket reform system there is a problem of it is more: can not automatic identification when initiate to reform operation, determine words
Substance does object and needs artificial participate in；Ticket is reformed to be carried out using mode identical with usually charge system process, in ticket weight
Need to stop charge system, traffic affecting continuity when doing；When a large amount of tickets are reformed, efficiency is also relatively low, is unable to satisfy the moon
Bottom a large number of users occurs calculating the demand that charging error reforms ticket, and in other words, ticket, which calculates expense and goes wrong, needs artificially to judge occur
Determine that ticket object is also required to think to participate in after problem, higher to personnel requirement, the response time is long, cannot accomplish to automate；Words
Substance does operation and uses existing charging flow, does not talk with substance and does independent processing, the inefficiency when ticket is reformed；For existing
Ticket reform mechanism to certain business carry out ticket reform when, need to stop the accounting routine of the business, non-ticket reforms object
The processing of ticket also needs to stop, thus traffic affecting continuity.
Summary of the invention
The technical problem to be solved in the present invention is to provide a kind of method and apparatus for reforming ticket, solve in the prior art,
Ticket, which is reformed, needs maintenance personnel to find to calculate expense mistake, takes the user of mistake to calculation and business carries out artificial screening, and in original
There is completion ticket in charging flow to reform work, results in the need for the defect for the accounting routine for stopping the business.
In order to solve the above technical problems, the embodiment of the present invention provides a kind of method for reforming ticket, it is applied to default clothes
Business device, method include: that starting ticket reforms mechanism；It is filtered out from error message and needs to reform reforming user and reforming for ticket
Business, the error message are to calculate information corresponding to the vicious daily ticket that generates in expense mechanism in ticket；According to institute
It states to reform user and reform business progress ticket and reforms to obtain correct ticket, the correct ticket eliminates the vicious day
Mistake in normal ticket.
In the method, before starting ticket reforms mechanism further include: will be calculated in ticket the day generated in expense mechanism
Normal ticket is transmitted to predetermined server, daily ticket is carried out in predetermined server it is regular in real time with sequence, it is regular in real time with
Merger ticket after sequence can be extracted after ticket reforms mechanism starting to be reformed for ticket.
In the method, is filtered out from error message and need to reform reforming user and reforming business and include: for ticket
The scale invariant feature for obtaining error message converts SIFT feature；Every two SIFT is constructed based on the relationship between SIFT feature
Structure feature between feature；System operation situation I1 to be detected is determined according to the structure feature set that the structure feature is formed
Similarity between tag system operation conditions I2；It determines to make mistakes when the similarity is greater than predetermined similarity threshold values and lead
Reforming for causing and reforms business at user.
In the method, the scale invariant feature conversion SIFT feature for obtaining error message includes: extraction error message
Scale invariant feature convert SIFT feature set, include scale invariant feature conversion SIFT special in the SIFT feature set
It levies, includes user information and business information in the error message；Relationship between SIFT feature constructs every two SIFT
Structure feature between feature includes: that each structure feature is constructed based on the relationship between SIFT feature, and structure feature includes
Relative distance, relative angle and relative scalar between two SIFT features；Determine the distance between two described structure features
It is the weighted sum of relative distance, relative angle and relative scalar；It is determined according to the structure feature set that the structure feature is formed
Similarity between system operation situation I1 to be detected and tag system operation conditions I2 includes: to run shape for examining system to be checked
Condition I1 and tag system operation conditions I2 determines the distance between the structure feature set of two kinds of system operation situations, thus really
Determine the similarity between system operation situation I1 to be detected and tag system operation conditions I2, includes in the structure feature set
The structure feature；It is determined when the similarity is greater than predetermined similarity threshold values and reforms user caused by making mistakes and reform industry
Business includes: to be compared the similarity with predetermined similarity threshold, when the similarity is greater than predetermined similarity threshold values
It determines and needs to reform, so that it is determined that reforming user caused by making mistakes and reforming business.
In the method, the scale invariant feature conversion SIFT feature set for extracting error message is specifically included: extracting
The SIFT feature set of error messagedik,θik,sikIndicate i-th kind of system operation shape
K-th of SIFT feature of condition is in the position of corresponding error message, angle and scale.
In the method, each structure feature is constructed based on the relationship between SIFT feature, structure feature includes two
Relative distance, relative angle and relative scalar between a SIFT feature specifically include: based on the relationship structure between SIFT feature
Build structure featureWithWherein, dij, θijAnd sij, and, dKl,θklAnd sklRespectively
Indicate relative distance, relative angle and the relative scalar in operating condition between two SIFT features.
In the method, before the scale invariant feature conversion SIFT feature set for extracting error message further include: right
Whole SIFT features are clustered, and the difference of classification is clustered according to belonging to each SIFT feature, uses vkIndicate i-th kind of operation shape
K-th of SIFT feature of conditionCorresponding error message.
In the method, determine that the distance between two structure features are relative distance, relative angle and relative scalar
Weighted sum comprise determining that dl, dθ, dsRespectively indicate relative distance, relative angle and the opposite ruler between two structure features
It spends, then dl=| dij-dkl|, dθ=| θij-θkl| and ds=| sij-skl|；The distance between two structure featuresThen
It is the weighted sum of the relative distance, relative angle and relative scalarAnd vi=
vk, vj=vl；Alternatively, when v is not presenti=vk, vj=vlWhen,The weight can be adjusted to be suitable for
Different applicable cases.
In the method, for system operation situation I1 to be detected and tag system operation conditions I2, two germlines are determined
The distance between the structure feature set for operation conditions of uniting, so that it is determined that system operation situation I1 to be detected and tag system operation
Similarity between situation I2 includes: to extract structure feature from system operation situation I1 to be detected to form first structure feature set
GOF1 is closed to indicate operation conditions, and, the second structure feature, which is extracted, from tag system operation conditions I2 forms structure feature
Set GOF2 indicates operation conditions；The distance between the structure feature set for determining two kinds of system operation situations is first structure
The sum of characteristic set GOF1 minimum value and the second structure feature set GOF2 minimum value and average value
So that it is determined that system operation situation I1 to be detected and mark
Label system operation situation I2 between similarity be
In the method, calculates and described reform user's cumulant and account for reforming ticket corresponding with business of reforming
Negative increment, checked and write off in billing and accounting system it is described reform expense caused by ticket and increase the amount of money, complete the reduction of billing and accounting system.
A kind of device for reforming ticket, comprising: ticket reforms start unit, reforms mechanism for starting ticket；Ticket weight
Selection unit is done, needs to reform reforming user and reforming business for ticket, the mistake letter for filtering out from error message
Breath is to calculate information corresponding to the vicious daily ticket that generates in expense mechanism in ticket；Ticket reforms unit, is used for basis
It is described reform user and reform business carry out ticket reform to obtain correct ticket, the correct ticket eliminates described vicious
Mistake in daily ticket.
In the device, it includes: characteristic extracting module that ticket, which reforms selection unit, for extracting the scale of error message
Invariant features convert SIFT feature set, include SIFT feature in the SIFT feature set, include use in the error message
The information at family and business；Structure feature module, for constructing each structure feature, structure based on the relationship between SIFT feature
Feature includes relative distance, relative angle and the relative scalar between two SIFT features；Structure feature spacing module, for true
The distance between fixed two structure features are the weighted sums of relative distance, relative angle and relative scalar；Similarity module, is used for
For system operation situation I1 to be detected and tag system operation conditions I2, the structure characteristic collection of two kinds of system operation situations is determined
The distance between close, so that it is determined that the similarity between system operation situation I1 to be detected and tag system operation conditions I2, institute
Stating includes structure feature in structure feature set；Module is chosen, for comparing the similarity and predetermined similarity threshold
Compared with, it is determining when the similarity is greater than predetermined similarity threshold values to need to reform, so that it is determined that reforming user caused by making mistakes
With reform business.
The advantageous effects of the above technical solutions of the present invention are as follows: due to the preparatory and professional etiquette before the process that ticket is reformed
Whole ticket, therefore the efficiency that ticket is reformed is greatly improved, and reform user since provided technology can voluntarily be chosen
It goes forward side by side the fast quick-recovery of mistake that jargon substance does, therefore can be likely to occur to charge system with the business of reforming, can be universal
Applied in the charge system of field of telecommunications, the level of user service is promoted.
Detailed description of the invention
Fig. 1 shows a kind of method flow schematic diagrams for reforming ticket；
Fig. 2 indicates a kind of apparatus structure schematic diagram for reforming ticket.
Specific embodiment
To keep the technical problem to be solved in the present invention, technical solution and advantage clearer, below in conjunction with attached drawing and tool
Body embodiment is described in detail.
In the embodiment of the present invention, ticket is reformed into judgement and ticket reforms object select automation, it would be possible to if appearance
Substance is done can talked about so that substance is prepared to be likely to occur with expectation with daily charging flow parallelization
The time that ticket is reformed is reduced when substance is done.
The reason of causing ticket mistake includes that office data configuration error, network abnormal call ticket, charge system program etc. are multi-party
Face, by dimensions such as this part system reason combination customer complaints, is built after classifying to previous generation ticket error issues
Vertical self-adapting estimation model finds that ticket mistake and automatic screening reform user and reform business automatically.
The embodiment of the present invention provides a kind of method for reforming ticket, as shown in Figure 1, being applied to predetermined server, comprising:
Step 101, starting ticket reforms mechanism；
Step 102, it is filtered out from error message and needs to reform reforming user and reforming business for ticket, the mistake letter
Breath is to calculate information corresponding to the vicious daily ticket that generates in expense mechanism in ticket；
Step 103, user and the business of reforming reformed is reformed to obtain correct ticket, the correct ticket is eliminated
Mistake in the vicious daily ticket.
Using provided technology, reforms user since provided technology can voluntarily be chosen and reform business and carry out
Ticket is reformed, therefore the fast quick-recovery of mistake that can be likely to occur to charge system, can be commonly used for field of telecommunications
In charge system, the level of user service is promoted.
Since the configuration control of charge system office data is inadequate, network side ticket issues the factors such as exception and program exception,
Therefore it calculates and takes possibility of the result there is mistake, at this moment it needs to be determined that impacted business and impacted business were related to
Therefore user usually extracts impacted business, nonoculture is error message if the user that is related to is corresponding.
In order to obtain impacted business, the user that error message is related to, i.e., needs are filtered out from error message and reformed
Reforming for ticket and reforms business at user, and the feature extraction algorithm of customer complaint is docked with customer service Request System, to user
It complains and carries out careful classification.If it was found that certain classification customer complaint increase sharply after, using feature extraction algorithm to customer complaint from
Various dimensions (feature) extract feature vector, it is contemplated that test object (customer complaint) may be due to factors such as the bursts of mistake
Influence fractional mutations can occur, therefore take SIFT operator to identify AD HOC to obtain SIFT feature.
In a preferred embodiment, before step 101 further include:
It is transmitted to predetermined server by the daily ticket generated in expense mechanism is calculated in ticket, to daily in predetermined server
Ticket carry out in real time it is regular with sequence, in real time it is regular with sequence after merger ticket can ticket reform mechanism starting after
It is extracted and is reformed for ticket.It is regular in real time by being carried out to ticket, complete the regular work for reforming ticket.Ticket is reformed in realization
The server of method be predetermined server, realize that server regular in real time is predetermined server, predetermined server can be
The property server different from predetermined server, the two can also be located in identical property server.It is regular in real time to refer to, it will
Daily ticket is transmitted to predetermined server, to daily ticket carry out in real time it is regular with sequence, in real time it is regular with sort after returning
And ticket can be extracted when carrying out and reforming ticket.
It is regular in real time to be completed at the same time parallel with ticket calculation expense mechanism, when needing ticket to reform, it is no longer necessary to return
The error message of target user is extracted when rolling, but the regular of daily ticket and row are had been completed before needing ticket to reform
Sequence substantially increases so as to reform the ticket chosen from merger ticket in mechanism and need to reform in ticket and reforms ticket
Efficiency.
In a preferred embodiment, comprising: the daily ticket for taking portal generation for calculation is talked about according to setting rule
It is single to merge, it is ensured that generate cumulant ticket according to setting rule for each user, the setting rule can be online ticket, language
The type of call of sound ticket, charging policy, period etc..Daily ticket is merged according to these setting rules, it is ensured that each
User generates cumulant ticket by regular (day, week, the moon and other conditions) are set.This process real-time perfoming, as words
One of the daily preparation that substance is done.
Generating cumulant ticket according to setting rule for each user can be component part regular in real time, be also possible to
Independently of process regular in real time.In order to facilitate the understanding of those skilled in the art, it in the present embodiment, will be generated according to setting rule tired
For accumulated amount ticket as a part regular in real time, the cumulant ticket of generation is also a kind of daily ticket, can be reformed in ticket
It is reformed after being extracted in mechanism.
In a preferred embodiment, it is filtered out from error message and needs to reform reforming for ticket and user and reform business
Include:
The scale invariant feature for obtaining error message converts SIFT feature；
The structure feature between every two SIFT feature is constructed based on the relationship between SIFT feature；
System operation situation I1 and tag system to be detected are determined according to the structure feature set that the structure feature is formed
Similarity between operation conditions I2；
It is determined when the similarity is greater than predetermined similarity threshold values and reforms user caused by making mistakes and reform business.
In a preferred embodiment,
The scale invariant feature conversion SIFT feature for obtaining error message includes: to extract the scale invariant feature of error message
SIFT feature set is converted, includes that scale invariant feature converts SIFT feature, the error message in the SIFT feature set
In include user information and business information；
It includes: based on SIFT that the structure feature between every two SIFT feature is constructed based on the relationship between SIFT feature
Relationship between feature constructs each structure feature, and structure feature includes relative distance between two SIFT features, opposite
Angle and relative scalar；
Determine that the distance between two described structure features are the weighted sums of relative distance, relative angle and relative scalar；
System operation situation I1 and tag system to be detected are determined according to the structure feature set that the structure feature is formed
Similarity between operation conditions I2 includes: that system operation situation I1 to be detected and tag system operation conditions I2 are determined
The distance between the structure feature set of two kinds of system operation situations, so that it is determined that system operation situation I1 to be detected and label system
The similarity united between operation conditions I2, includes the structure feature in the structure feature set；
It is determined when the similarity is greater than predetermined similarity threshold values and reforms user caused by making mistakes and reform business packet
It includes: the similarity is compared with predetermined similarity threshold, determined when the similarity is greater than predetermined similarity threshold values
It needs to reform, so that it is determined that reforming user caused by making mistakes and reforming business.
After the SIFT feature set for extracting error message further include: carried out with K-means algorithm to whole SIFT features
Cluster, and belong to the difference for clustering classification according to each SIFT feature, corresponding SIFT feature is indicated using the label of word,
Such as k-th of SIFT feature of i-th kind of operation conditionsCorresponding error message vkIt indicates.
In a preferred embodiment, the SIFT feature set for extracting error message specifically includes:
Extract the SIFT feature set of error messageWherein, k indicates that SIFT is special
The number of sign, (dik,θik,sik) indicate k-th of SIFT feature of i-th kind of system operation situation in the position of corresponding error message
It sets, angle and scale.
In a preferred embodiment, each structure feature, structure feature are constructed based on the relationship between SIFT feature
It is specifically included including relative distance, relative angle and the relative scalar between two SIFT features:
Structure feature is constructed based on the relationship between SIFT featureWithIts
In, dij, θijAnd sij, and, dKl,θklAnd sklRespectively indicate relative distance, the phase in operating condition between two SIFT features
To angle and relative scalar.dij, θijAnd sijIt respectively indicates in operating condition between i-th of SIFT feature and j-th of SIFT feature
Relative distance, relative angle and relative scalar, dkl, θklAnd sklRespectively indicate k-th of SIFT feature and l in operating condition
Relative distance, relative angle and relative scalar between a SIFT feature.
In a preferred embodiment, before the SIFT feature set for extracting error message further include:
Whole SIFT features is clustered, the difference of classification is clustered according to belonging to each SIFT feature, uses vkIt indicates
K-th of SIFT feature of i-th kind of operation conditionsCorresponding error message.
In a preferred embodiment, determine that the distance between two structure features are relative distance, relative angle and phase
Weighted sum to scale includes:
Determine dl, dθ, dsRelative distance, relative angle and the relative scalar between two structure features are respectively indicated, then dl
=| dij-dkl|, dθ=| θij-θkl| and ds=| sij-skl|；
The distance between two structure featuresIt is then the relative distance, relative angle and relative scalar
Weighted sumAnd error message vi=vk, vj=vl；Alternatively, when there is no mistakes to believe
Cease vi=vk, vj=vlWhen,The weight can be adjusted to be suitable for different applicable cases.Wherein,
dij, θijAnd sijRespectively indicate relative distance in operating condition between i-th of SIFT feature and j-th of SIFT feature, relative angle
Degree and relative scalar, dkl, θklAnd sklRespectively indicate the phase in operating condition between k-th of SIFT feature and first of SIFT feature
It adjusts the distance, relative angle and relative scalar, wl、wθAnd wsRespectively indicate relative distance, relative angle and the corresponding power of relative scalar
Value.
In different application, to the relative distance in structure feature, relative angle and relative scalar have different requirements respectively.
For example, corresponding weight can be adjusted to 0 if the variation on scale will not occur for system operation situation to be detected.
In a preferred embodiment, it for system operation situation I1 to be detected and tag system operation conditions I2, determines
The distance between the structure feature set of two kinds of system operation situations, so that it is determined that system operation situation I1 to be detected and label system
System operation conditions I2 between similarity include:
Structure feature formation first structure characteristic set GOF1 is extracted from system operation situation I1 to be detected to indicate to transport
Row situation, and, the second structure feature formation structure feature set GOF2 is extracted from tag system operation conditions I2 to indicate
Operation conditions；
The distance between the structure feature set for determining two kinds of system operation situations be first structure characteristic set GOF1 most
The sum of small value and the second structure feature set GOF2 minimum value and average value
So that it is determined that the similarity between system operation situation I1 to be detected and tag system operation conditions I2 is
Similarity is compared with predetermined similarity threshold, that is, can determine in system to be detected actual operation conditions and
Error User is chosen, to realize that automatic selection Error User and self-adapting start ticket are reformed.
It reforms according to the object-of reforming chosen automatically and user and reforms business, in the merger ticket run parallel with production
Middle fast selecting target ticket, in a preferred embodiment, ticket are reformed after mechanism starting, further includes: according to described heavy
It is user and reforms business, target ticket is chosen in merger ticket, carry out ticket and reform to obtain correct ticket, according to correct words
Corresponding cumulant and account negative increment are singly calculated, the reduction of billing and accounting system is completed.
In an application scenarios, mistake occurs for the configuration of gprs service office data, using skill provided in an embodiment of the present invention
Art, comprising:
Adaptive ticket reforms starter and plays a game data configuration situation using the calculating of SIFT algorithm progress feature vector, obtains
Space and range information to office data configuring condition.
After corresponding space in advance clusters correct office data using K-means clustering algorithm with citing information feeding
To pattern recognition model in calculated, obtain corresponding bias；
When bias is more than threshold range, it is confirmed as meeting the entry condition that ticket is reformed, adaptive ticket, which is reformed, to be opened
Dynamic device starting ticket reforms mechanism.
After process is reformed in starting, according to the situation of office data mistake, it is automatically positioned the user group and the scope of business of influence, it is complete
It at the selection work of target customer, obtains reforming user and reforms business, reform unit into ticket and start ticket and reform work.
In this application, client is reformed according to target and the gprs service of mistake occurs, utilizes the merger handled in advance
Ticket is rapidly completed ticket and reforms work.
Due to regular ticket parallel in advance before the process that ticket is reformed, the effect that ticket is reformed is greatly improved
Rate, and done since provided technology can voluntarily be chosen to reform user and reform business jargon substance of going forward side by side, it can be right
The fast quick-recovery of the mistake that charge system is likely to occur, can be commonly used in the charge system of field of telecommunications, promote user
The level of service.
As shown in Fig. 2, the whole system for reforming ticket includes: menu manager, words if the dress setting and counting expense entrance for reforming ticket
Single calculate takes mechanism, billing and accounting system and ticket storage etc..
Before the function and process that starting ticket is reformed, further includes: the daily words generated in expense mechanism will be calculated in ticket
Patrilineal line of descent with only one son in each generation is delivered to predetermined server, regular in real time to the progress of daily ticket in predetermined server and sequence, regular in real time and sequence
Merger ticket later can be reformed after ticket is reformed and is extracted in mechanism.It is regular in real time by being carried out to ticket, it is complete
At the regular work for reforming ticket.The server for realizing the method for reforming ticket is predetermined server, realizes clothes regular in real time
Business device is predetermined server, and predetermined server can be the property server different from predetermined server, and the two can also be located at
In identical property server.It is regular in real time to refer to, daily ticket is transmitted to predetermined server, daily ticket is carried out real-time
Regular and sequence, the merger ticket with after sequence regular in real time can be extracted when carrying out and reforming ticket.
It is regular in real time to be completed parallel in charge system with ticket calculation expense mechanism, when needing ticket to reform, it is no longer necessary to
The error message of target user is extracted in rollback, but the regular of daily ticket is had been completed before needing ticket to reform
With sequence, the efficiency for reforming ticket is substantially increased.Since the configuration control of charge system office data is inadequate, network side ticket issues
The factors such as exception and program exception, calculation take possibility of the result there is mistake, at this moment it needs to be determined that impacted business, and
Therefore the user that impacted business is related to usually extracts impacted business, nonoculture is poor if the user that is related to is corresponding
Wrong information.
The embodiment of the present invention provides a kind of device for reforming ticket, as shown in Figure 2, comprising:
Ticket reforms start unit 201, reforms mechanism for starting ticket；
Ticket reforms selection unit 202, for filtered out from error message need to reform ticket reform user and again
Do business, the error message is to calculate information corresponding to the vicious daily ticket that generates in expense mechanism in ticket；
Ticket reforms unit 203, for doing in mechanism in the substance independently of ticket calculation expense mechanism, to described heavy
It is user and reforms business and reformed to obtain correct ticket, the correct ticket eliminates in the vicious daily ticket
Mistake.
Expense mistake and user's screening operation after generation mistake, which will be calculated, transfers to independent adaptive ticket to reform judgment means
It carries out, it includes: that ticket reforms start unit and ticket reforms selection unit that adaptive ticket, which reforms judgment means, and ticket reforms work
Independent ticket is transferred to reform unit completion.
In a preferred embodiment, further includes:
Ticket reforms start unit, for starting the function and process that ticket is reformed.
In a preferred embodiment, ticket reforms selection unit and includes:
Characteristic extracting module, the scale invariant feature for extracting error message convert SIFT feature set, the SIFT
Include SIFT feature in characteristic set, includes the information of user and business in the error message；
Structure feature module, for constructing each structure feature, structure feature packet based on the relationship between SIFT feature
Include relative distance, relative angle and the relative scalar between two SIFT features；
Structure feature spacing module, for determine the distance between two structure features be relative distance, relative angle and
The weighted sum of relative scalar；
Similarity module, for determining two kinds for system operation situation I1 to be detected and tag system operation conditions I2
The distance between structure feature set of system operation situation, so that it is determined that system operation situation I1 to be detected and tag system fortune
Similarity between row situation I2 includes structure feature in the structure feature set；
Module is chosen, for the similarity to be compared with predetermined similarity threshold, is greater than in the similarity pre-
It is determined when determining similarity threshold values and needs to reform, so that it is determined that reforming user caused by making mistakes and reforming business.
In a preferred embodiment, characteristic extracting module includes:
SIFT feature collection modules, for extracting the SIFT feature set of error message
dik,θik,sikIndicate k-th of SIFT feature of i-th kind of system operation situation in the position of corresponding error message, angle and ruler
Degree.
In a preferred embodiment, structure feature module includes:
Structure feature constructs module, for constructing structure feature based on the relationship between SIFT feature
WithWherein, dij, θijAnd sij, and, dKl,θklAnd sklRespectively indicate two SIFT in operating condition
Relative distance, relative angle and relative scalar between feature.
In a preferred embodiment, similarity module includes:
Structure feature extraction module forms structure feature for extracting structure feature from system operation situation I1 to be detected
Gather to indicate operation conditions, and, structure feature formation structure feature set, which is extracted, from tag system operation conditions I2 comes
Indicate operation conditions；
Structure feature set spacing module, between the structure feature set GOF for determining two kinds of system operation situations
Distance be fisrt feature set GOF1 minimum value and with second feature set GOF2 minimum value and average value
So that it is determined that the similarity between system operation situation I1 to be detected and tag system operation conditions I2 is
As shown in Fig. 2, the line that ticket reforms unit 203 to financial system indicates: calculating corresponding cumulant and account
After negative increment of being engaged in, the reduction of billing and accounting system is completed.
Be using the advantage after this programme: ticket reforms unit and reforms object according to what is chosen automatically, calculates with ticket
Substance is done in mechanism if expense mechanism is run parallel, and basis reforms business and reforms user's fast selecting target from merger ticket
Ticket, and corresponding cumulant and account negative increment are calculated, complete the reduction work of billing and accounting system.Ticket reforms unit can
Designated user's specified services carry out ticket and reform work, after charging goes wrong, guarantee do not influencing the non-premise for reforming object
Under, opponent influences user's progress ticket and reforms work.
The above is a preferred embodiment of the present invention, it is noted that for those skilled in the art
For, without departing from the principles of the present invention, it can also make several improvements and retouch, these improvements and modifications
It should be regarded as protection scope of the present invention.
Claims (11)
1. a kind of method for reforming ticket, it is applied to predetermined server, which is characterized in that method includes:
Starting ticket reforms mechanism；
It is filtered out from error message and needs to reform reforming user and reforming business for ticket, the error message is calculated in ticket
Information corresponding to the vicious daily ticket generated in expense mechanism, comprising: the scale invariant feature for obtaining error message turns
Change SIFT feature；The structure feature between every two SIFT feature is constructed based on the relationship between SIFT feature；According to described
The structure feature set that structure feature is formed determines between system operation situation I1 to be detected and tag system operation conditions I2
Similarity；It is determined when the similarity is greater than predetermined similarity threshold values and reforms user caused by making mistakes and reform business；
According to it is described reform user and reform business and carry out ticket reform to obtain correct ticket, the correct ticket eliminates described
Mistake in vicious daily ticket.
2. the method according to claim 1, wherein before starting ticket reforms mechanism further include:
It is transmitted to predetermined server by the daily ticket generated in expense mechanism is calculated in ticket, to daily ticket in predetermined server
Regular in real time and sequence is carried out, the merger ticket with after sequence regular in real time can be mentioned after ticket reforms mechanism starting
It takes and is reformed in ticket.
3. the method according to claim 1, wherein
The scale invariant feature conversion SIFT feature for obtaining error message includes: to extract the scale invariant feature conversion of error message
SIFT feature set includes that scale invariant feature converts SIFT feature in the SIFT feature set, wraps in the error message
Containing user information and business information；
It includes: based on SIFT feature that the structure feature between every two SIFT feature is constructed based on the relationship between SIFT feature
Between relationship construct each structure feature, structure feature includes relative distance between two SIFT features, relative angle
And relative scalar；
Determine that the distance between two described structure features are the weighted sums of relative distance, relative angle and relative scalar；
System operation situation I1 and tag system operation to be detected is determined according to the structure feature set that the structure feature is formed
Similarity between situation I2 includes: to determine two kinds for system operation situation I1 to be detected and tag system operation conditions I2
The distance between structure feature set of system operation situation, so that it is determined that system operation situation I1 to be detected and tag system fortune
Similarity between row situation I2 includes the structure feature in the structure feature set；
The similarity be greater than predetermined similarity threshold values when determine make mistakes caused by reform user and reform business include: by
The similarity is compared with predetermined similarity threshold, is determined when the similarity is greater than predetermined similarity threshold values and is needed weight
It does, so that it is determined that reforming user caused by making mistakes and reforming business.
4. according to the method described in claim 3, it is characterized in that, the scale invariant feature conversion SIFT for extracting error message is special
Collection is closed and is specifically included:
Extract the SIFT feature set of error messagedik,θik,sikIndicate the i-th germline
K-th of SIFT feature of operation conditions of uniting is in the position of corresponding error message, angle and scale.
5. according to the method described in claim 3, it is characterized in that, constructing each structure based on the relationship between SIFT feature
Feature, structure feature include that relative distance, relative angle and the relative scalar between two SIFT features specifically include:
Structure feature is constructed based on the relationship between SIFT featureWithWherein,
dij, θijAnd sijRespectively indicate relative distance in operating condition between i-th of SIFT feature and j-th of SIFT feature, relative angle
Degree and relative scalar, dkl, θklAnd sklRespectively indicate the phase in operating condition between k-th of SIFT feature and first of SIFT feature
It adjusts the distance, relative angle and relative scalar.
6. according to the method described in claim 5, it is characterized in that, the scale invariant feature conversion SIFT for extracting error message is special
Before collection is closed further include:
Whole SIFT features is clustered, the difference of classification is clustered according to belonging to each SIFT feature, uses vkIndicate i-th kind
K-th of SIFT feature of operation conditionsCorresponding error message.
7. according to the method described in claim 6, it is characterized in that, determine the distance between two structure features be it is opposite away from
Weighted sum from, relative angle and relative scalar includes:
Determine dl, dθ, dsRelative distance, relative angle and the relative scalar between two structure features are respectively indicated, then dl=|
dij-dkl|, dθ=| θij-θkl| and ds=| sij-skl|；
The distance between two structure featuresIt is then the weighting of the relative distance, relative angle and relative scalar
WithAnd error message vi=vk, vj=vl；Alternatively, when error message v is not presenti
=vk, vj=vlWhen,Wherein, wl, wθ, wsIt is weight, the weight can be adjusted to be suitable for difference
Applicable cases.
8. according to the method described in claim 3, it is characterized in that, being transported for system operation situation I1 to be detected and tag system
Row situation I2 determines the distance between the structure feature set of two kinds of system operation situations, so that it is determined that examining system to be checked is run
Similarity between situation I1 and tag system operation conditions I2 includes:
Structure feature formation first structure characteristic set GOF1 is extracted from system operation situation I1 to be detected to indicate operation shape
Condition extracts the second structure feature formation structure feature set GOF2 from tag system operation conditions I2 to indicate operation conditions；
The distance between the structure feature set for determining two kinds of system operation situations is first structure characteristic set GOF1 minimum value
Sum and the second structure feature set GOF2 minimum value and average value
So that it is determined that the similarity between system operation situation I1 to be detected and tag system operation conditions I2 is
9. the method according to claim 1, wherein further include:
User's cumulant for reforming ticket corresponding with business of reforming and account negative increment are reformed described in calculating, in billing and accounting system
In check and write off it is described reform expense caused by ticket and increase the amount of money, complete the reduction of billing and accounting system.
10. a kind of device for reforming ticket characterized by comprising
Ticket reforms start unit, reforms mechanism for starting ticket；
Ticket reforms selection unit, needs to reform reforming for ticket for filtering out from error message and user and reforms business,
The error message is to calculate information corresponding to the vicious daily ticket that generates in expense mechanism in ticket, comprising: it is poor to obtain
The scale invariant feature of wrong information converts SIFT feature；Constructed based on the relationship between SIFT feature every two SIFT feature it
Between structure feature；System operation situation I1 and label to be detected are determined according to the structure feature set that the structure feature is formed
Similarity between system operation situation I2；Caused weight of making mistakes is determined when the similarity is greater than predetermined similarity threshold values
It is user and reforms business；
Ticket reforms unit, reforms to obtain correct ticket for reforming user according to and reforming business progress ticket, described
Correct ticket eliminates the mistake in the vicious daily ticket.
11. device according to claim 10, which is characterized in that ticket reforms selection unit and includes:
Characteristic extracting module, the scale invariant feature for extracting error message convert SIFT feature set, the SIFT feature
Include SIFT feature in set, includes the information of user and business in the error message；
Structure feature module, for constructing each structure feature based on the relationship between SIFT feature, structure feature includes two
Relative distance, relative angle and relative scalar between a SIFT feature；
Structure feature spacing module, for determining that the distance between two structure features are relative distance, relative angle and opposite
The weighted sum of scale；
Similarity module, for determining two kinds of systems for system operation situation I1 to be detected and tag system operation conditions I2
The distance between structure feature set of operation conditions, so that it is determined that system operation situation I1 to be detected and tag system run shape
Similarity between condition I2 includes structure feature in the structure feature set；
Module is chosen, for the similarity to be compared with predetermined similarity threshold, is greater than predetermined phase in the similarity
It is determined when like bottom valve value and needs to reform, so that it is determined that reforming user caused by making mistakes and reforming business.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510040759.7A CN105991298B (en) | 2015-01-27 | 2015-01-27 | A kind of method and apparatus for reforming ticket |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510040759.7A CN105991298B (en) | 2015-01-27 | 2015-01-27 | A kind of method and apparatus for reforming ticket |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105991298A CN105991298A (en) | 2016-10-05 |
| CN105991298B true CN105991298B (en) | 2019-03-12 |
Family
ID=57034048
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510040759.7A Active CN105991298B (en) | 2015-01-27 | 2015-01-27 | A kind of method and apparatus for reforming ticket |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105991298B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108123810A (en) * | 2016-11-28 | 2018-06-05 | 中国移动通信集团公司 | A kind of error bills processing method and processing device |
| CN108521527B (en) * | 2018-03-14 | 2020-12-11 | 北京思特奇信息技术股份有限公司 | Ticket difference detection method, system, computer storage medium and computer equipment |
| CN113095517B (en) * | 2019-12-23 | 2023-09-19 | 中国移动通信集团湖南有限公司 | Method and device for monitoring and early warning charging error codes and electronic equipment |
| CN111182485A (en) * | 2020-01-02 | 2020-05-19 | 兰州嘉文电子科技有限公司 | Single ticket total traffic control system and method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101415168A (en) * | 2007-10-17 | 2009-04-22 | 华为技术有限公司 | Method and apparatus for checking charging result and configuration data |
| CN103167202A (en) * | 2011-12-19 | 2013-06-19 | 中国移动通信集团甘肃有限公司 | Method and device for auditing phone bills with different sources |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090177612A1 (en) * | 2008-01-08 | 2009-07-09 | Value Creation Institute | Method and Apparatus for Analyzing Data to Provide Decision Making Information |
-
2015
- 2015-01-27 CN CN201510040759.7A patent/CN105991298B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101415168A (en) * | 2007-10-17 | 2009-04-22 | 华为技术有限公司 | Method and apparatus for checking charging result and configuration data |
| CN103167202A (en) * | 2011-12-19 | 2013-06-19 | 中国移动通信集团甘肃有限公司 | Method and device for auditing phone bills with different sources |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105991298A (en) | 2016-10-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105991298B (en) | A kind of method and apparatus for reforming ticket | |
| CN106408410A (en) | Automatic account checking method and device | |
| CN106384280B (en) | Artificial intelligence accounting system and calculation method thereof | |
| CN110264039A (en) | A kind of generation method and equipment of working report | |
| CN107657500A (en) | Stock recommends method and server | |
| CN109039710B (en) | Routing data auditing method, device, server and storage medium | |
| CN106682137A (en) | Intelligent stock investment adviser questioning-answering interaction method and intelligent stock investment adviser questioning-answering interaction system | |
| CN110119877A (en) | A kind of choosing method and equipment of target employee | |
| CN110401780A (en) | A kind of method and device identifying fraudulent call | |
| CN110418173A (en) | Determine method, apparatus, server and the storage medium of exception account | |
| CN107689009A (en) | Annuity data processing method, device, server and storage medium | |
| CN108345670A (en) | A kind of service hot spot discovery method being used for 95598 electric power work orders | |
| CN110119948A (en) | Based on when variable weight dynamic combined power consumer credit assessment method and system | |
| CN113379530A (en) | User risk determination method and device and server | |
| CN109474923A (en) | Object identifying method and device, storage medium | |
| CN110222971A (en) | It is a kind of to be classified the method and device for improving trading efficiency based on trading card | |
| CN109918645A (en) | Method, apparatus, computer equipment and the storage medium of depth analysis text | |
| US10956914B2 (en) | System and method for mapping a customer journey to a category | |
| CN110569237A (en) | System and method for realizing real-time data cleaning processing | |
| CN110309737A (en) | A kind of information processing method applied to cigarette sales counter, apparatus and system | |
| CN110245879A (en) | A kind of risk rating method and device | |
| CN112637793B (en) | Scene charging method, system, electronic equipment and storage medium based on 5G | |
| CN110750302B (en) | Assembly line account making method for accountant | |
| CN112037049A (en) | Bank card charging method and device | |
| CN109308565A (en) | The recognition methods of crowd's performance ratings, device, storage medium and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |