CN113592315A - Method and device for processing dispute order - Google Patents
Method and device for processing dispute order Download PDFInfo
- Publication number
- CN113592315A CN113592315A CN202110889249.2A CN202110889249A CN113592315A CN 113592315 A CN113592315 A CN 113592315A CN 202110889249 A CN202110889249 A CN 202110889249A CN 113592315 A CN113592315 A CN 113592315A
- Authority
- CN
- China
- Prior art keywords
- dispute
- order
- processing
- model
- risk
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012545 processing Methods 0.000 title claims abstract description 181
- 238000000034 method Methods 0.000 title claims abstract description 69
- 238000004891 communication Methods 0.000 claims abstract description 57
- 230000008569 process Effects 0.000 claims abstract description 22
- 238000011156 evaluation Methods 0.000 claims abstract description 15
- 238000012937 correction Methods 0.000 claims description 61
- 230000009466 transformation Effects 0.000 claims description 17
- 238000013210 evaluation model Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 9
- 238000012986 modification Methods 0.000 claims description 7
- 230000004048 modification Effects 0.000 claims description 7
- 238000012216 screening Methods 0.000 claims description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000006399 behavior Effects 0.000 description 39
- 238000010586 diagram Methods 0.000 description 14
- 239000013598 vector Substances 0.000 description 7
- 230000006870 function Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 230000004069 differentiation Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000004806 packaging method and process Methods 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000013107 unsupervised machine learning method 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
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/04—Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Physics & Mathematics (AREA)
- Finance (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Accounting & Taxation (AREA)
- Educational Administration (AREA)
- Technology Law (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
The invention discloses a method and a device for processing a dispute order, and relates to the technical field of artificial intelligence. One embodiment of the method comprises: identifying a risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order; judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value or not; if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service; if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record. The implementation method can solve the technical problems of lack of dispute risk identification and low model accuracy.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a method and a device for processing a dispute order.
Background
The existing dispute order processing methods mainly comprise the following methods:
1) processing dispute orders based on rules: the method is based on the logic of dispute business, and combines the key node state of transaction dispute and the key words of user communication record, such as 'return of goods', 'delivery' and the like, to form dispute handling rules. And when the dispute appeal of the user meets the requirements of each node required by the dispute rule, triggering a solution built in the rule system to push the user.
2) Treating dispute orders based on an unsupervised method: the method generally takes important information of dispute transaction records as the characteristics of an unsupervised machine learning method, and clusters the processing results of the whole historical dispute order according to the set number of cluster clusters. And when the user applies for a certain transaction dispute, returning a dispute processing result similar to the dispute according to an unsupervised algorithm as a dispute processing result for pushing.
In the process of implementing the invention, the inventor finds that at least the following problems exist in the prior art:
different users and different orders are in different states, dispute handling modes are completely consistent, dispute risk identification is lacked, high-risk disputes cannot be handled in time, and after-sale experience of the users is poor;
the existing method mainly utilizes the communication records of the customer and the merchant to carry out dispute processing, and a dispute processing model obtained by only using the communication records has low accuracy and poor effect.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method and an apparatus for processing a dispute order, so as to solve the technical problems of lack of dispute risk identification and low model accuracy.
To achieve the above object, according to an aspect of an embodiment of the present invention, there is provided a method for processing a dispute order, including:
identifying a risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order;
judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value or not;
if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service;
if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
Optionally, identifying a risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order includes:
inputting user attributes, user behavior records and dispute communication records corresponding to the dispute order into a trained risk identification model, and outputting a risk coefficient corresponding to the dispute order;
the risk identification model is obtained by adopting risk samples to conduct supervised training, wherein each risk sample comprises user attributes, user behavior records and dispute communication records corresponding to historical dispute orders, and whether the historical dispute orders have dispute risks.
Optionally, the risk identification model is a deep fm + TextCNN model.
Optionally, transferring the dispute single flow to a dispute handling model, and outputting a processing result of the dispute order based on the transaction link data of the dispute order, where the processing result includes:
inputting the transaction link data corresponding to the dispute order into a trained dispute handling model to obtain an output result output by the dispute handling model;
correcting the output result by adopting a correction coefficient so as to obtain a processing result of the dispute order;
the dispute handling model is obtained by adopting dispute samples for supervised training, wherein each dispute sample comprises transaction link data and a handling scheme corresponding to a historical dispute order; the output result comprises the corresponding probability of each processing scheme.
Optionally, the dispute handling model is a multi-modal model; the modality of the transaction link data includes text, pictures and structured data.
Optionally, the modifying the output result with a modification coefficient to obtain a processing result of the dispute order includes:
multiplying the correction coefficient corresponding to each processing scheme by the probability corresponding to each processing scheme to obtain the correction probability corresponding to each processing scheme;
and screening the processing scheme with the highest correction probability as the processing result of the dispute order.
Optionally, for any one of the processing schemes, the correction coefficient of the processing scheme is obtained by the following method:
resolving the scores of all services from a dispute return visit table filled by a user;
inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient;
and carrying out nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme.
Optionally, inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient, including:
and carrying out weighted summation on the scores of the services based on the score weights of the services to obtain a satisfaction coefficient.
Optionally, performing a nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme, where the method includes:
and carrying out nonlinear transformation on the satisfaction coefficient by adopting a natural logarithm formula to obtain a correction coefficient of the processing scheme.
In addition, according to another aspect of the embodiments of the present invention, there is provided an apparatus for processing a dispute order, including:
the identification module is used for identifying the risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order;
the processing module is used for judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value; if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service; if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
Optionally, the identification module is further configured to:
inputting user attributes, user behavior records and dispute communication records corresponding to the dispute order into a trained risk identification model, and outputting a risk coefficient corresponding to the dispute order;
the risk identification model is obtained by adopting risk samples to conduct supervised training, wherein each risk sample comprises user attributes, user behavior records and dispute communication records corresponding to historical dispute orders, and whether the historical dispute orders have dispute risks.
Optionally, the risk identification model is a deep fm + TextCNN model.
Optionally, the processing module is further configured to:
inputting the transaction link data corresponding to the dispute order into a trained dispute handling model to obtain an output result output by the dispute handling model;
correcting the output result by adopting a correction coefficient so as to obtain a processing result of the dispute order;
the dispute handling model is obtained by adopting dispute samples for supervised training, wherein each dispute sample comprises transaction link data and a handling scheme corresponding to a historical dispute order; the output result comprises the corresponding probability of each processing scheme.
Optionally, the dispute handling model is a multi-modal model; the modality of the transaction link data includes text, pictures and structured data.
Optionally, the processing module is further configured to:
multiplying the correction coefficient corresponding to each processing scheme by the probability corresponding to each processing scheme to obtain the correction probability corresponding to each processing scheme;
and screening the processing scheme with the highest correction probability as the processing result of the dispute order.
Optionally, the processing module is further configured to:
for any one processing scheme, the correction coefficient of the processing scheme is obtained by adopting the following method:
resolving the scores of all services from a dispute return visit table filled by a user;
inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient;
and carrying out nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme.
Optionally, the processing module is further configured to:
and carrying out weighted summation on the scores of the services based on the score weights of the services to obtain a satisfaction coefficient.
Optionally, the processing module is further configured to:
and carrying out nonlinear transformation on the satisfaction coefficient by adopting a natural logarithm formula to obtain a correction coefficient of the processing scheme.
According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors implement the method of any of the embodiments described above.
According to another aspect of the embodiments of the present invention, there is also provided a computer readable medium, on which a computer program is stored, which when executed by a processor implements the method of any of the above embodiments.
One embodiment of the above invention has the following advantages or benefits: the dispute order is issued to the customer service terminal if the risk coefficient of the dispute order is greater than or equal to the risk threshold, so that the dispute order is processed through the customer service terminal, and the dispute order is transferred to the dispute processing model if the risk coefficient of the dispute order is smaller than the risk threshold, and the processing result of the dispute order is output based on the transaction link data of the dispute order, so that the technical problems that dispute risk identification is lacked and the accuracy of the model is low in the prior art are solved. According to the embodiment of the invention, dispute risk identification is carried out on the dispute order, and the potential dispute order with dispute upgrading risk is identified, so that differentiation processing is carried out in advance, quick response of dispute processing is realized, potential risk is reduced, and after-sale experience of users is improved; according to the embodiment of the invention, the whole link data generated by the whole dispute transaction is concerned, the appeal intentions and the real transaction conditions of different transaction link links of the user are grasped, the user acceptance degree of the dispute processing model scheme prediction push is practically improved, and the dispute problem is more efficiently solved; and (4) scoring the satisfaction degree of the questionnaire through the return visit table and the comprehensive evaluation model, and taking a scoring result as a correction coefficient to finely adjust an output result of the model, thereby realizing self-correction and closed-loop processing of the dispute processing process.
Further effects of the above-mentioned non-conventional alternatives will be described below in connection with the embodiments.
Drawings
The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:
FIG. 1 is a schematic diagram of a main flow of a method of handling dispute orders according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a risk identification model according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a dispute handling model according to an embodiment of the present invention;
FIG. 4 is a schematic view of a main flow of a method of handling a dispute order according to a referential embodiment of the present invention;
FIG. 5 is a schematic diagram of the major modules of an apparatus for handling dispute orders according to an embodiment of the present invention;
FIG. 6 is an exemplary system architecture diagram in which embodiments of the present invention may be employed;
fig. 7 is a schematic block diagram of a computer system suitable for use in implementing a terminal device or server of an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present invention are described below with reference to the accompanying drawings, in which various details of embodiments of the invention are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
The embodiment of the invention introduces data with complete transaction links, including user attributes, user behavior records, article attributes, logistics progress, order records, article evaluation contents, article question and answer contents, dispute communication records and the like, and then takes the transaction link data as the model characteristics of the multi-mode model, so that the richness and diversity of the model characteristics are greatly enriched, and the current situations of low model accuracy and poor effect caused by insufficient model information are improved. Meanwhile, the embodiment of the invention also introduces a risk identification model, and the dispute order with the extremely large risk coefficient is preferentially processed, so that the potential dispute risk is reduced, and the after-sale experience of the user is improved. And finally, after the dispute order of the user is properly processed, adding an automatic return visit and self-correction processing process to realize the closed-loop processing capability of self-correction learning. Therefore, the embodiment of the invention can efficiently and quickly process the dispute order and improve the shopping experience of the consumer.
In the embodiment of the present invention, the transaction link data includes the whole data of the user from browsing, ordering, and dispute complaints, for example, the transaction link data may include user attributes, user behavior records, article attributes, logistics progress, order records, article evaluation contents, article question and answer contents, and dispute communication records.
Wherein, user behavior record mainly includes: user click behavior, purchase behavior, etc.; the order record mainly comprises: order number, article number, order placing time, order processing duration and other data; the article attributes mainly include: information on the type, packaging, quantity, etc. of the article; the user attributes mainly include: information such as user identity characteristics, account number grades, user figures and the like; the logistics progress mainly comprises: logistics state, logistics timeliness, logistics merchants and other information; the item evaluation content mainly comprises: the user's evaluation of the item; the article question and answer content mainly comprises: the user asks questions, the content of the answer questions and the like; the dispute communication record mainly comprises: text information of a user chatting with a merchant, picture information of customer or merchant testification and voice chatting information.
It should be noted that these transaction link data are only raw data, and cannot be calculated as features of the model, and these data need to be analyzed first. According to the embodiment of the invention, the analysis is performed in different modes according to different data modes. For the structured data, according to the data types of the characteristics, the structured data are divided into discrete characteristics (including order types, dispute types, user levels, dispute summary primary ids, dispute summary secondary ids and the like), continuous characteristics (including order amounts, article amounts, order article quantity, user purchasing power, optimized discount, logistics timeliness and the like), and sequence characteristics (including logistics states, dispute order states and the like); for unstructured data, the data is divided by modality into text, image, and voice. Where the speech is further processed for conversion to text form.
The embodiment of the invention mainly realizes the identification of the dispute risk coefficient and the calculation of the dispute processing result by two models. The risk identification model mainly identifies potential high-risk dispute orders, such as dispute orders of users with high influence such as certain large network V or potential dispute orders with dispute upgrading. The risk identification model calculates the probability (namely a risk coefficient) that the dispute order is high risk through relevant characteristics such as user attributes, user behavior records and dispute communication records, and then the risk identification model transfers the dispute order to an artificial channel to be responded by artificial customer service and timely responds to the dispute order with high risk according to a manually set risk threshold value for dispute orders larger than or equal to the risk threshold value. And for the dispute orders lower than the risk threshold, the dispute orders are transferred to a dispute handling model, the dispute handling model calculates the solution of the dispute orders according to the transaction link data, and the solution is pushed to the user by the system.
Fig. 1 is a schematic diagram of a main flow of a method for handling a dispute order according to an embodiment of the present invention. As an embodiment of the present invention, as shown in fig. 1, the method for processing a dispute order may include:
step 101, identifying a risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order.
First, the risk coefficient of the dispute order is identified according to the user attribute, the user behavior record (user click behavior, purchase behavior, etc.), the dispute communication record (text information of the chat record between the user and the merchant, the image information of the customer or the merchant for evidence collection, and the voice chat information) and the like corresponding to the dispute order. In the embodiment of the present invention, the higher the risk coefficient of the dispute order is, the higher the probability that the dispute order is a high-risk dispute is.
Optionally, step 101 may comprise: inputting user attributes, user behavior records and dispute communication records corresponding to the dispute order into a trained risk identification model, and outputting a risk coefficient corresponding to the dispute order; the risk identification model is obtained by adopting risk samples to conduct supervised training, wherein each risk sample comprises a user attribute user behavior record and a dispute communication record corresponding to a historical dispute order, and whether the historical dispute order has a dispute risk or not. In the embodiment of the invention, the user attribute user behavior record and the dispute communication record are adopted to train the risk identification model, so that the model has the capability of considering the pre-sale attributes of the user, and whether the dispute order is the high-risk dispute can be accurately identified.
In order to input the user attributes, the user behavior records and the dispute communication records into the trained risk identification model for processing, the user attributes, the user behavior records and the dispute communication records need to be changed into a characteristic format available for the risk identification model, for example, converted into a vector representation through embedding (embedding) processing of the model. In some embodiments of the present invention, the dispute communication record is mainly a communication text, and there are two processing methods: one is to take the whole communication text of the user and the merchant as the input of the model, which is simpler and more direct; another way is to extract keywords related to risks from the communication text by extracting keywords, such as words like "public opinion exposure", "microblog", "complaint", etc., which can reflect the risk level of the customer to some extent. Compared with the first mode, the second mode can reduce the interference of irrelevant texts to a certain extent, and the recognition effect is better.
Taking the second way as an example, optionally, the risk identification model is a deep fm + TextCNN model. Because the model needs to be processed by the mixed features containing the text, in order to improve the recognition effect, a DeepFM + TextCNN model is selected for recognition, the traditional machine learning model is not suitable for processing the text content, and the embodiment of the invention mainly focuses on the local information of the text, so that the TextCNN model component is adopted to extract the key information of the text, and the key information of the text is spliced with the vectorization cross results of various structured features and input into a full connection layer, and finally the risk coefficient of the dispute order predicted by the model is obtained.
As shown in fig. 2, a user attribute, a user behavior record and a dispute communication record are input into a risk identification model through an input layer, wherein the user attribute and the user behavior record are usually structured data, the structured data are divided into discrete features, continuous features and sequence features according to the data types of the features, the discrete features, the continuous features and the sequence features are converted into feature vectors through an embedding layer (embedding lay), meanwhile, a communication text in the dispute communication record is input into the risk identification model, and text key information is extracted by using a TextCNN model component. And performing pairwise crossing combination on the four groups of vectors through the FM layer to output the characteristic vectors, and calculating the four groups of vectors through the full-connection layer to output the characteristic vectors. And then splicing the eigenvector output by the FM layer and the eigenvector output by the full connection layer together, inputting the eigenvector into the full connection layer, and finally outputting the risk coefficient of the dispute order through the output layer.
Step 102, judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value; if yes, go to step 103; if not, go to step 104.
And after the risk coefficient of the dispute order is output through the risk identification model, judging whether the risk coefficient is greater than or equal to a preset risk threshold value.
And 103, issuing the dispute order to a customer service terminal so as to process the dispute order through customer service.
If the risk coefficient is larger than or equal to the preset risk threshold, it indicates that the dispute order has a high potential risk dispute, such as dispute orders of users with high influence such as some network large V or potential dispute orders with dispute upgrade. And for the dispute orders which are larger than or equal to the risk threshold value, the dispute orders are transferred to the manual channel to be responded by the manual customer service, and the dispute orders with high risk are responded in time.
And 104, transferring the dispute single flow to a dispute handling model, and outputting a handling result of the dispute single flow based on the transaction link data of the dispute single flow.
And if the risk coefficient is smaller than the preset risk threshold value, the dispute order is less likely to have potential high-risk disputes. And for the dispute orders lower than the risk threshold, the dispute orders are transferred to a dispute handling model, the dispute handling model calculates the solution of the dispute orders according to the transaction link data, and the solution is pushed to the user by the system. Wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
Optionally, step 104 may include: inputting the transaction link data corresponding to the dispute order into a trained dispute handling model to obtain an output result output by the dispute handling model; correcting the output result by adopting a correction coefficient so as to obtain a processing result of the dispute order; the dispute handling model is obtained by adopting dispute samples for supervised training, wherein each dispute sample comprises transaction link data and a handling scheme corresponding to a historical dispute order; the output result comprises the corresponding probability of each processing scheme.
The dispute handling model realizes intelligent resolution of dispute handling and has certain self-correcting capability. Different dispute complaint solutions under various scenes can be combed in advance, and the dispute complaint solutions comprise 13 solutions such as agreement of return of goods, agreement of return of goods and no return of goods, platform responsibility compensation and general use. And training and verifying the dispute handling model by using the marked offline data of different solutions, finally completing the online of the model, and predicting the handling result of the dispute order through the online model.
Optionally, the dispute handling model is a multi-modal model. The multi-modal model is mainly designed into different model components according to different modes, and the model structure is continuously optimized, and the multi-modal model of the embodiment of the invention is shown in fig. 3. As shown in fig. 3, transaction link data whose modalities include text (voice data converted into text), picture, and structured data is input into the dispute processing model through the input layer. For text features, phonetic text features and structured data, the three modal data are firstly converted into low-dimensional dense vectors through an embedding layer (embedding layer), then position codes are added through a position coding layer (position layer) so as to be convenient for understanding the position information of the text, then high-dimensional features and fine features are converted and extracted through a representation layer (transformer block), the dimension of the dimension reduction layer is reduced, and finally the three modal data are output to a full connection layer of the model. For picture features, key features are extracted through 128 convolution kernels, so that the model can well obtain the key features of data in different modes, the key features are subjected to dimensionality reduction through a pooling layer and are finally output to a full-connection layer of the model, and finally, the probability corresponding to various solutions is output through an output layer.
In order to improve the calculation accuracy, the embodiment of the invention also introduces a correction coefficient on the basis of the model output result. Optionally, the correction coefficients corresponding to different processing schemes are not consistent, and the correction coefficient is mainly a means for increasing the output result of the service interventionable model.
Optionally, the modifying the output result with a modification coefficient to obtain a processing result of the dispute order includes: multiplying the correction coefficient corresponding to each processing scheme by the probability corresponding to each processing scheme to obtain the correction probability corresponding to each processing scheme; and screening the processing scheme with the highest correction probability as the processing result of the dispute order. It should be noted that, if the correction probability of each processing scheme is lower than the preset probability threshold, the general processing scheme is used as the processing result of the dispute order.
After the processing result of the dispute order is obtained, finding out a specific processing scheme corresponding to the processing result according to the mapping table, pushing the specific processing scheme to the customer and the merchant, and after the scheme is pushed to the customer, requiring the customer to feed back whether to accept the processing scheme, if the processing scheme is accepted, resolving disputes according to the scheme, wherein the dispute problem is reasonably resolved, and at the moment, the system closes the dispute order; if the merchant or the customer does not accept the scheme, the difficulty of resolving the dispute order is high, the dispute order flows back to the manual customer service channel, and the manual customer service performs order following resolution.
Optionally, for any one of the processing schemes, the correction coefficient of the processing scheme is obtained by the following method: resolving the scores of all services from a dispute return visit table filled by a user; inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient; and carrying out nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme. Regardless of the manner in which the dispute is handled and resolved, the system will trigger the customer to be pushed a return visit form with reward properties or interest after the dispute order is finally closed. The return visit form is that a unified return visit table is given according to different dispute handling schemes, after the user finishes the return visit, the values of all services are analyzed from the dispute return visit table, the analyzed value data are input into the comprehensive evaluation model to obtain the satisfaction degree score, and the satisfaction degree coefficient is stored in a database of the system and used as the characteristic of the dispute handling model to correct the model.
Optionally, inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient, including: and carrying out weighted summation on the scores of the services based on the score weights of the services to obtain a satisfaction coefficient.
For example, the satisfaction factor may be calculated using the following comprehensive evaluation model:
satisfaction coefficient 0.35+ dispute service score 0.2+ dispute resolution aging 0.2+ purchase experience 0.1+ number of items to be improved 0.15
Optionally, performing a nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme, where the method includes: and carrying out nonlinear transformation on the satisfaction coefficient by adopting a natural logarithm formula to obtain a correction coefficient of the processing scheme. And the satisfaction coefficient is subjected to nonlinear transformation and then is used as a correction coefficient of the dispute handling model to finely adjust the output result of the model, so that the output result is more consistent with the handling scene of the dispute order. Therefore, the whole dispute processing process realizes a closed-loop dispute intelligent processing scheme with self-correcting capability through dispute full-link data, an effective processing scheme can be provided more objectively and fairly, and the resolution efficiency and the customer experience of disputes are greatly improved.
Optionally, performing nonlinear transformation on the satisfaction coefficient by using a natural logarithm formula to obtain a correction coefficient of the processing scheme, where the method includes:
calculating a correction factor for the treatment plan using the following formula:
y=ln(x+1)
wherein y is a correction coefficient and x is a satisfaction coefficient.
Alternatively, the correction factor of the processing scheme may also be calculated using the following formula:
y=ln(x+2)
wherein y is a correction coefficient and x is a satisfaction coefficient.
Alternatively, the correction factor of the processing scheme may also be calculated using the following formula:
y=ln(x+2.5)
wherein y is a correction coefficient and x is a satisfaction coefficient.
It should be noted that, in the embodiment of the present invention, the formula of the nonlinear transformation is not limited, and an appropriate formula may be selected as needed to perform the nonlinear transformation.
According to the various embodiments described above, it can be seen that the technical means of the embodiments of the present invention, which are lacked in dispute risk identification and low in model accuracy in the prior art, are solved by issuing the dispute ticket to the customer service terminal to process the dispute ticket through the customer service if the risk coefficient of the dispute ticket is greater than or equal to the risk threshold, transferring the dispute ticket to the dispute processing model if the risk coefficient of the dispute ticket is smaller than the risk threshold, and outputting the processing result of the dispute ticket based on the transaction link data of the dispute ticket. According to the embodiment of the invention, dispute risk identification is carried out on the dispute order, and the potential dispute order with dispute upgrading risk is identified, so that differentiation processing is carried out in advance, quick response of dispute processing is realized, potential risk is reduced, and after-sale experience of users is improved; according to the embodiment of the invention, the whole link data generated by the whole dispute transaction is concerned, the appeal intentions and the real transaction conditions of different transaction link links of the user are grasped, the user acceptance degree of the dispute processing model scheme prediction push is practically improved, and the dispute problem is more efficiently solved; and (4) scoring the satisfaction degree of the questionnaire through the return visit table and the comprehensive evaluation model, and taking a scoring result as a correction coefficient to finely adjust an output result of the model, thereby realizing self-correction and closed-loop processing of the dispute processing process.
Fig. 4 is a schematic diagram of a main flow of a method of handling a dispute order according to a referential embodiment of the present invention. As another embodiment of the present invention, as shown in fig. 4, the method for processing a dispute order may include:
the transaction link data corresponding to each historical dispute order is obtained from the database, and the transaction link data includes the whole data from browsing, ordering and dispute complaints of the user, for example, the transaction link data may include user attributes, user behavior records, article attributes, logistics progress, order records, article evaluation contents, article question and answer contents, dispute communication records, and the like. Wherein, user behavior record mainly includes: user click behavior, purchase behavior, etc.; the order record mainly comprises: order number, article number, order placing time, order processing duration and other data; the article attributes mainly include: information on the type, packaging, quantity, etc. of the article; the user attributes mainly include: information such as user identity characteristics, account number grades, user figures and the like; the logistics progress mainly comprises: logistics state, logistics timeliness, logistics merchants and other information; the item evaluation content mainly comprises: the user's evaluation of the item; the article question and answer content mainly comprises: the user asks questions, the content of the answer questions and the like; the dispute communication record mainly comprises: text information of a user chatting with a merchant, picture information of customer or merchant testification and voice chatting information.
And training a risk identification model and a dispute handling model by adopting transaction link data corresponding to the historical dispute order, wherein the risk identification model is a DeepFM + TextCNN model, and the dispute handling model is a multi-modal model. The risk identification model is obtained by adopting risk samples to conduct supervised training, wherein each risk sample comprises a user attribute user behavior record and a dispute communication record corresponding to a historical dispute order, and whether the historical dispute order has dispute risks. The dispute handling model is obtained by adopting dispute samples to carry out supervised training, and each dispute sample comprises transaction link data and a handling scheme corresponding to the historical dispute order.
And inputting the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order into the trained risk identification model, thereby outputting a risk coefficient corresponding to the dispute order.
And judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value.
And if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service.
And if not, inputting the transaction link data corresponding to the dispute order into the trained dispute handling model to obtain an output result output by the dispute handling model.
And correcting the output result by adopting a correction coefficient so as to obtain the processing result of the dispute order, wherein the output result comprises the probability corresponding to each processing scheme.
And screening the processing scheme with the highest correction probability as the processing result of the dispute order.
Finding out a specific processing scheme corresponding to a processing result according to a mapping table, pushing the specific processing scheme to a customer and a merchant, after the scheme is pushed to the customer, feeding back whether to accept the processing scheme by the customer, if the processing scheme is accepted, solving disputes according to the scheme, representing that dispute problems are reasonably solved, and closing a dispute order by the system; if the merchant or the customer does not accept the scheme, the difficulty of resolving the dispute order is high, the dispute order flows back to the manual customer service channel, and the manual customer service performs order following resolution.
Regardless of the manner in which the dispute is handled and resolved, the system will trigger the customer to be pushed a return visit form with reward properties or interest after the dispute order is finally closed. The return visit form is that a unified return visit table is given according to different dispute handling schemes, after the user finishes the return visit, the values of all services are analyzed from the dispute return visit table, the analyzed value data are input into the comprehensive evaluation model to obtain the satisfaction degree score, and the satisfaction degree coefficient is stored in a database of the system and used as the characteristic of the dispute handling model to correct the model.
In addition, in one embodiment of the present invention, the detailed implementation of the method for handling dispute orders is described in detail in the above-mentioned method for handling dispute orders, and therefore, the repeated content is not described again.
Fig. 5 is a schematic diagram of main modules of an apparatus for processing a dispute order according to an embodiment of the present invention, and as shown in fig. 5, the apparatus 500 for processing a dispute order includes an identification module 501 and a processing module 502; the identification module 501 is configured to identify a risk coefficient of a dispute order according to a user attribute, a user behavior record and a dispute communication record corresponding to the dispute order; the processing module 502 is configured to determine whether a risk coefficient of the dispute order is greater than or equal to a risk threshold; if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service; if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
Optionally, the identifying module 501 is further configured to:
inputting user attributes, user behavior records and dispute communication records corresponding to the dispute order into a trained risk identification model, and outputting a risk coefficient corresponding to the dispute order;
the risk identification model is obtained by adopting risk samples to conduct supervised training, wherein each risk sample comprises user attributes, user behavior records and dispute communication records corresponding to historical dispute orders, and whether the historical dispute orders have dispute risks.
Optionally, the risk identification model is a deep fm + TextCNN model.
Optionally, the processing module 502 is further configured to:
inputting the transaction link data corresponding to the dispute order into a trained dispute handling model to obtain an output result output by the dispute handling model;
correcting the output result by adopting a correction coefficient so as to obtain a processing result of the dispute order;
the dispute handling model is obtained by adopting dispute samples for supervised training, wherein each dispute sample comprises transaction link data and a handling scheme corresponding to a historical dispute order; the output result comprises the corresponding probability of each processing scheme.
Optionally, the dispute handling model is a multi-modal model; the modality of the transaction link data includes text, pictures and structured data.
Optionally, the processing module 502 is further configured to:
multiplying the correction coefficient corresponding to each processing scheme by the probability corresponding to each processing scheme to obtain the correction probability corresponding to each processing scheme;
and screening the processing scheme with the highest correction probability as the processing result of the dispute order.
Optionally, the processing module 502 is further configured to:
for any one processing scheme, the correction coefficient of the processing scheme is obtained by adopting the following method:
resolving the scores of all services from a dispute return visit table filled by a user;
inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient;
and carrying out nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme.
Optionally, the processing module 502 is further configured to:
and carrying out weighted summation on the scores of the services based on the score weights of the services to obtain a satisfaction coefficient.
Optionally, the processing module 502 is further configured to:
and carrying out nonlinear transformation on the satisfaction coefficient by adopting a natural logarithm formula to obtain a correction coefficient of the processing scheme.
According to the various embodiments described above, it can be seen that the technical means of the embodiments of the present invention, which are lacked in dispute risk identification and low in model accuracy in the prior art, are solved by issuing the dispute ticket to the customer service terminal to process the dispute ticket through the customer service if the risk coefficient of the dispute ticket is greater than or equal to the risk threshold, transferring the dispute ticket to the dispute processing model if the risk coefficient of the dispute ticket is smaller than the risk threshold, and outputting the processing result of the dispute ticket based on the transaction link data of the dispute ticket. According to the embodiment of the invention, dispute risk identification is carried out on the dispute order, and the potential dispute order with dispute upgrading risk is identified, so that differentiation processing is carried out in advance, quick response of dispute processing is realized, potential risk is reduced, and after-sale experience of users is improved; according to the embodiment of the invention, the whole link data generated by the whole dispute transaction is concerned, the appeal intentions and the real transaction conditions of different transaction link links of the user are grasped, the user acceptance degree of the dispute processing model scheme prediction push is practically improved, and the dispute problem is more efficiently solved; and (4) scoring the satisfaction degree of the questionnaire through the return visit table and the comprehensive evaluation model, and taking a scoring result as a correction coefficient to finely adjust an output result of the model, thereby realizing self-correction and closed-loop processing of the dispute processing process.
It should be noted that, in the implementation of the device for processing dispute orders according to the present invention, the details of the method for processing dispute orders are described in detail above, and therefore, the repeated details are not described herein.
Fig. 6 illustrates an exemplary system architecture 600 of a method for handling dispute orders or a device for handling dispute orders, to which embodiments of the present invention may be applied.
As shown in fig. 6, the system architecture 600 may include terminal devices 601, 602, 603, a network 604, and a server 605. The network 604 serves to provide a medium for communication links between the terminal devices 601, 602, 603 and the server 605. Network 604 may include various types of connections, such as wire, wireless communication links, or fiber optic cables, to name a few.
A user may use the terminal devices 601, 602, 603 to interact with the server 605 via the network 604 to receive or send messages or the like. The terminal devices 601, 602, 603 may have installed thereon various communication client applications, such as shopping applications, web browser applications, search applications, instant messaging tools, mailbox clients, social platform software, etc. (by way of example only).
The terminal devices 601, 602, 603 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, and the like.
The server 605 may be a server providing various services, such as a background management server (for example only) providing support for shopping websites browsed by users using the terminal devices 601, 602, 603. The background management server may analyze and otherwise process the received data such as the item information query request, and feed back a processing result (for example, target push information, item information — just an example) to the terminal device.
It should be noted that the method for processing the dispute order provided by the embodiment of the present invention is generally executed by the server 605, and accordingly, the apparatus for processing the dispute order is generally disposed in the server 605.
It should be understood that the number of terminal devices, networks, and servers in fig. 6 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
Referring now to FIG. 7, shown is a block diagram of a computer system 700 suitable for use with a terminal device implementing an embodiment of the present invention. The terminal device shown in fig. 7 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.
As shown in fig. 7, the computer system 700 includes a Central Processing Unit (CPU)701, which can perform various appropriate actions and processes in accordance with a program stored in a Read Only Memory (ROM)702 or a program loaded from a storage section 708 into a Random Access Memory (RAM) 703. In the RAM703, various programs and data necessary for the operation of the system 700 are also stored. The CPU 701, the ROM 702, and the RAM703 are connected to each other via a bus 704. An input/output (I/O) interface 705 is also connected to bus 704.
The following components are connected to the I/O interface 705: an input portion 706 including a keyboard, a mouse, and the like; an output section 707 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage section 708 including a hard disk and the like; and a communication section 709 including a network interface card such as a LAN card, a modem, or the like. The communication section 709 performs communication processing via a network such as the internet. A drive 710 is also connected to the I/O interface 705 as needed. A removable medium 711 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 710 as necessary, so that a computer program read out therefrom is mounted into the storage section 708 as necessary.
In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication section 709, and/or installed from the removable medium 711. The computer program performs the above-described functions defined in the system of the present invention when executed by the Central Processing Unit (CPU) 701.
It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer programs according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules described in the embodiments of the present invention may be implemented by software or hardware. The described modules may also be provided in a processor, which may be described as: a processor includes an identification module and a processing module, where the names of the modules do not in some way constitute a limitation on the modules themselves.
As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be separate and not incorporated into the device. The computer readable medium carries one or more programs which, when executed by a device, implement the method of: identifying a risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order; judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value or not; if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service; if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
According to the technical scheme of the embodiment of the invention, as the technical means that if the risk coefficient of the dispute order is greater than or equal to the risk threshold, the dispute order is sent to the customer service terminal so as to be processed by the customer service, and if the risk coefficient of the dispute order is smaller than the risk threshold, the dispute order is transferred to the dispute processing model, and the processing result of the dispute order is output based on the transaction link data of the dispute order is adopted, the technical problems that dispute risk identification is lacked and the model accuracy rate is low in the prior art are solved. According to the embodiment of the invention, dispute risk identification is carried out on the dispute order, and the potential dispute order with dispute upgrading risk is identified, so that differentiation processing is carried out in advance, quick response of dispute processing is realized, potential risk is reduced, and after-sale experience of users is improved; according to the embodiment of the invention, the whole link data generated by the whole dispute transaction is concerned, the appeal intentions and the real transaction conditions of different transaction link links of the user are grasped, the user acceptance degree of the dispute processing model scheme prediction push is practically improved, and the dispute problem is more efficiently solved; and (4) scoring the satisfaction degree of the questionnaire through the return visit table and the comprehensive evaluation model, and taking a scoring result as a correction coefficient to finely adjust an output result of the model, thereby realizing self-correction and closed-loop processing of the dispute processing process.
The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (12)
1. A method for processing dispute orders, comprising:
identifying a risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order;
judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value or not;
if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service;
if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
2. The method as claimed in claim 1, wherein identifying the risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order comprises:
inputting user attributes, user behavior records and dispute communication records corresponding to the dispute order into a trained risk identification model, and outputting a risk coefficient corresponding to the dispute order;
the risk identification model is obtained by adopting risk samples to conduct supervised training, wherein each risk sample comprises user attributes, user behavior records and dispute communication records corresponding to historical dispute orders, and whether the historical dispute orders have dispute risks.
3. The method of claim 2, wherein the risk identification model is a deep fm + TextCNN model.
4. The method as claimed in claim 1, wherein transferring the dispute single flow to a dispute handling model, and outputting a result of the dispute order handling based on transaction link data of the dispute order comprises:
inputting the transaction link data corresponding to the dispute order into a trained dispute handling model to obtain an output result output by the dispute handling model;
correcting the output result by adopting a correction coefficient so as to obtain a processing result of the dispute order;
the dispute handling model is obtained by adopting dispute samples for supervised training, wherein each dispute sample comprises transaction link data and a handling scheme corresponding to a historical dispute order; the output result comprises the corresponding probability of each processing scheme.
5. The method of claim 4, wherein the dispute handling model is a multi-modal model; the modality of the transaction link data includes text, pictures and structured data.
6. The method as claimed in claim 4, wherein the modifying the output result by using a modification factor to obtain the result of processing the dispute order comprises:
multiplying the correction coefficient corresponding to each processing scheme by the probability corresponding to each processing scheme to obtain the correction probability corresponding to each processing scheme;
and screening the processing scheme with the highest correction probability as the processing result of the dispute order.
7. The method of claim 4, wherein for any processing scheme, the correction coefficients of the processing scheme are obtained by:
resolving the scores of all services from a dispute return visit table filled by a user;
inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient;
and carrying out nonlinear transformation on the satisfaction coefficient to obtain a correction coefficient of the processing scheme.
8. The method of claim 7, wherein inputting the scores of the services into a comprehensive evaluation model to output a satisfaction coefficient comprises:
and carrying out weighted summation on the scores of the services based on the score weights of the services to obtain a satisfaction coefficient.
9. The method of claim 7, wherein performing a non-linear transformation on the satisfaction index to obtain a modification index for the processing scheme comprises:
and carrying out nonlinear transformation on the satisfaction coefficient by adopting a natural logarithm formula to obtain a correction coefficient of the processing scheme.
10. An apparatus for processing dispute orders, comprising:
the identification module is used for identifying the risk coefficient of the dispute order according to the user attribute, the user behavior record and the dispute communication record corresponding to the dispute order;
the processing module is used for judging whether the risk coefficient of the dispute order is greater than or equal to a risk threshold value; if so, issuing the dispute order to a customer service terminal so as to process the dispute order through the customer service; if not, transferring the dispute single stream to a dispute processing model, and outputting a processing result of the dispute single based on the transaction link data of the dispute single; wherein the transaction link data includes at least two of: user attribute, user behavior record, article attribute, logistics progress, order record, article evaluation content, article question and answer content and dispute communication record.
11. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
the one or more programs, when executed by the one or more processors, implement the method of any of claims 1-9.
12. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110889249.2A CN113592315B (en) | 2021-08-04 | 2021-08-04 | Method and device for processing dispute sheets |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110889249.2A CN113592315B (en) | 2021-08-04 | 2021-08-04 | Method and device for processing dispute sheets |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113592315A true CN113592315A (en) | 2021-11-02 |
CN113592315B CN113592315B (en) | 2024-08-16 |
Family
ID=78254769
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110889249.2A Active CN113592315B (en) | 2021-08-04 | 2021-08-04 | Method and device for processing dispute sheets |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113592315B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114358984A (en) * | 2021-12-31 | 2022-04-15 | 城云科技(中国)有限公司 | Dispute management method and device, readable storage medium and electronic device |
CN116469576A (en) * | 2023-03-27 | 2023-07-21 | 中关村科学城城市大脑股份有限公司 | Dispute event identification method, device, electronic equipment and computer readable medium |
Citations (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102385732A (en) * | 2011-09-23 | 2012-03-21 | 陈晓亮 | Network trading dispute processing system and method thereof |
CN104376471A (en) * | 2013-08-14 | 2015-02-25 | 世纪禾光科技发展(北京)有限公司 | Electronic commerce after-sale dispute handling method and system |
KR20160049442A (en) * | 2015-07-29 | 2016-05-09 | 한경정보기술주식회사 | Dispute Management System For Consumer Of Small Or Medium Sized Business |
CN106776651A (en) * | 2015-11-24 | 2017-05-31 | 阿里巴巴集团控股有限公司 | The processing method and processing device of business dispute |
WO2017107794A1 (en) * | 2015-12-22 | 2017-06-29 | 阿里巴巴集团控股有限公司 | Method and device for risk identification |
CN107220353A (en) * | 2017-06-01 | 2017-09-29 | 深圳追科技有限公司 | A kind of intelligent customer service robot satisfaction automatic evaluation method and system |
CN108694650A (en) * | 2018-05-24 | 2018-10-23 | 阿里巴巴集团控股有限公司 | Transaction Disputes processing method and processing device and electronic equipment |
CN109858702A (en) * | 2019-02-14 | 2019-06-07 | 中国联合网络通信集团有限公司 | Client upgrades prediction technique, device, equipment and the readable storage medium storing program for executing complained |
CN110349039A (en) * | 2019-06-13 | 2019-10-18 | 中国平安人寿保险股份有限公司 | Complaint risk appraisal procedure, system, computer equipment and readable storage medium storing program for executing |
CN110442712A (en) * | 2019-07-05 | 2019-11-12 | 阿里巴巴集团控股有限公司 | Determination method, apparatus, server and the text of risk try system |
CN110533368A (en) * | 2019-08-20 | 2019-12-03 | 合肥维天运通信息科技股份有限公司 | It is a kind of that logistics platform order dispute processing system is carried without vehicle based on AR |
CN110795587A (en) * | 2018-08-01 | 2020-02-14 | 深圳云天励飞技术有限公司 | Medical alarm early warning method, service platform and computer readable storage medium |
CN110858348A (en) * | 2018-08-23 | 2020-03-03 | 阿里巴巴集团控股有限公司 | Logistics information processing method and device |
CN110991837A (en) * | 2019-11-21 | 2020-04-10 | 支付宝(杭州)信息技术有限公司 | Online taxi appointment transaction data processing method, device and system |
WO2020086334A1 (en) * | 2018-10-26 | 2020-04-30 | Microsoft Technology Licensing, Llc | Machine learning system for taking control actions |
CN111092999A (en) * | 2019-11-13 | 2020-05-01 | 泰康保险集团股份有限公司 | Data request processing method and device |
CN111311205A (en) * | 2020-05-09 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Transaction auditing processing method and device |
CN111489095A (en) * | 2020-04-15 | 2020-08-04 | 腾讯科技(深圳)有限公司 | Risk user management method and device, computer equipment and storage medium |
CN111833070A (en) * | 2019-04-10 | 2020-10-27 | 阿里巴巴集团控股有限公司 | After-sale processing method and device and electronic equipment |
CN112529584A (en) * | 2020-12-08 | 2021-03-19 | 支付宝(杭州)信息技术有限公司 | Transaction dispute data processing method, device, equipment and storage medium |
CN112561535A (en) * | 2020-12-08 | 2021-03-26 | 支付宝(杭州)信息技术有限公司 | Transaction dispute data processing method, device, equipment and storage medium |
CN112651635A (en) * | 2020-12-28 | 2021-04-13 | 长沙市到家悠享网络科技有限公司 | Risk identification method and device, electronic equipment and storage medium |
CN112668859A (en) * | 2020-12-23 | 2021-04-16 | 平安普惠企业管理有限公司 | Big data based customer risk rating method, device, equipment and storage medium |
CN112712270A (en) * | 2020-12-31 | 2021-04-27 | 深圳前海微众银行股份有限公司 | Information processing method, device, equipment and storage medium |
CN112825174A (en) * | 2019-11-15 | 2021-05-21 | 阿里云计算有限公司 | Litigation prediction method, device, system and computer storage medium |
CN113077196A (en) * | 2021-06-07 | 2021-07-06 | 泰豪信息技术有限公司 | Dispute event processing method and device, readable storage medium and computer equipment |
-
2021
- 2021-08-04 CN CN202110889249.2A patent/CN113592315B/en active Active
Patent Citations (26)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102385732A (en) * | 2011-09-23 | 2012-03-21 | 陈晓亮 | Network trading dispute processing system and method thereof |
CN104376471A (en) * | 2013-08-14 | 2015-02-25 | 世纪禾光科技发展(北京)有限公司 | Electronic commerce after-sale dispute handling method and system |
KR20160049442A (en) * | 2015-07-29 | 2016-05-09 | 한경정보기술주식회사 | Dispute Management System For Consumer Of Small Or Medium Sized Business |
CN106776651A (en) * | 2015-11-24 | 2017-05-31 | 阿里巴巴集团控股有限公司 | The processing method and processing device of business dispute |
WO2017107794A1 (en) * | 2015-12-22 | 2017-06-29 | 阿里巴巴集团控股有限公司 | Method and device for risk identification |
CN107220353A (en) * | 2017-06-01 | 2017-09-29 | 深圳追科技有限公司 | A kind of intelligent customer service robot satisfaction automatic evaluation method and system |
CN108694650A (en) * | 2018-05-24 | 2018-10-23 | 阿里巴巴集团控股有限公司 | Transaction Disputes processing method and processing device and electronic equipment |
CN110795587A (en) * | 2018-08-01 | 2020-02-14 | 深圳云天励飞技术有限公司 | Medical alarm early warning method, service platform and computer readable storage medium |
CN110858348A (en) * | 2018-08-23 | 2020-03-03 | 阿里巴巴集团控股有限公司 | Logistics information processing method and device |
WO2020086334A1 (en) * | 2018-10-26 | 2020-04-30 | Microsoft Technology Licensing, Llc | Machine learning system for taking control actions |
CN109858702A (en) * | 2019-02-14 | 2019-06-07 | 中国联合网络通信集团有限公司 | Client upgrades prediction technique, device, equipment and the readable storage medium storing program for executing complained |
CN111833070A (en) * | 2019-04-10 | 2020-10-27 | 阿里巴巴集团控股有限公司 | After-sale processing method and device and electronic equipment |
CN110349039A (en) * | 2019-06-13 | 2019-10-18 | 中国平安人寿保险股份有限公司 | Complaint risk appraisal procedure, system, computer equipment and readable storage medium storing program for executing |
CN110442712A (en) * | 2019-07-05 | 2019-11-12 | 阿里巴巴集团控股有限公司 | Determination method, apparatus, server and the text of risk try system |
CN110533368A (en) * | 2019-08-20 | 2019-12-03 | 合肥维天运通信息科技股份有限公司 | It is a kind of that logistics platform order dispute processing system is carried without vehicle based on AR |
CN111092999A (en) * | 2019-11-13 | 2020-05-01 | 泰康保险集团股份有限公司 | Data request processing method and device |
CN112825174A (en) * | 2019-11-15 | 2021-05-21 | 阿里云计算有限公司 | Litigation prediction method, device, system and computer storage medium |
CN110991837A (en) * | 2019-11-21 | 2020-04-10 | 支付宝(杭州)信息技术有限公司 | Online taxi appointment transaction data processing method, device and system |
CN111489095A (en) * | 2020-04-15 | 2020-08-04 | 腾讯科技(深圳)有限公司 | Risk user management method and device, computer equipment and storage medium |
CN111311205A (en) * | 2020-05-09 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Transaction auditing processing method and device |
CN112529584A (en) * | 2020-12-08 | 2021-03-19 | 支付宝(杭州)信息技术有限公司 | Transaction dispute data processing method, device, equipment and storage medium |
CN112561535A (en) * | 2020-12-08 | 2021-03-26 | 支付宝(杭州)信息技术有限公司 | Transaction dispute data processing method, device, equipment and storage medium |
CN112668859A (en) * | 2020-12-23 | 2021-04-16 | 平安普惠企业管理有限公司 | Big data based customer risk rating method, device, equipment and storage medium |
CN112651635A (en) * | 2020-12-28 | 2021-04-13 | 长沙市到家悠享网络科技有限公司 | Risk identification method and device, electronic equipment and storage medium |
CN112712270A (en) * | 2020-12-31 | 2021-04-27 | 深圳前海微众银行股份有限公司 | Information processing method, device, equipment and storage medium |
CN113077196A (en) * | 2021-06-07 | 2021-07-06 | 泰豪信息技术有限公司 | Dispute event processing method and device, readable storage medium and computer equipment |
Non-Patent Citations (3)
Title |
---|
WING YIU等: "Logistic Regression Modeling of Construction Negotiation Outcomes", IEEE TRANSACTIONS ON ENGINEERING MANAGEMENT * |
孙凡: "平台企业财务欺诈成因、机理与防范对策", 会计之友 * |
邵兵家;罗旋;杨海峰;: "C2C电子商务纠纷对交易评价结果的影响", 统计与决策, no. 19 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114358984A (en) * | 2021-12-31 | 2022-04-15 | 城云科技(中国)有限公司 | Dispute management method and device, readable storage medium and electronic device |
CN116469576A (en) * | 2023-03-27 | 2023-07-21 | 中关村科学城城市大脑股份有限公司 | Dispute event identification method, device, electronic equipment and computer readable medium |
CN116469576B (en) * | 2023-03-27 | 2023-10-27 | 中关村科学城城市大脑股份有限公司 | Dispute event identification method, device, electronic equipment and computer readable medium |
Also Published As
Publication number | Publication date |
---|---|
CN113592315B (en) | 2024-08-16 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109492772B (en) | Method and device for generating information | |
CN110119445B (en) | Method and device for generating feature vector and classifying text based on feature vector | |
EP4242955A1 (en) | User profile-based object recommendation method and device | |
CN113592315B (en) | Method and device for processing dispute sheets | |
CN110929136A (en) | Personalized recommendation method and device | |
CN112925911B (en) | Complaint classification method based on multi-modal data and related equipment thereof | |
CN109711733A (en) | For generating method, electronic equipment and the computer-readable medium of Clustering Model | |
CN112036954A (en) | Item recommendation method and device, computer-readable storage medium and electronic device | |
CN109190123A (en) | Method and apparatus for output information | |
CN110866625A (en) | Promotion index information generation method and device | |
CN113887214B (en) | Willingness presumption method based on artificial intelligence and related equipment thereof | |
JP2023554210A (en) | Sort model training method and apparatus for intelligent recommendation, intelligent recommendation method and apparatus, electronic equipment, storage medium, and computer program | |
CN108062423B (en) | Information-pushing method and device | |
CN110555747A (en) | method and device for determining target user | |
CN117235264A (en) | Text processing method, apparatus, device and computer readable storage medium | |
CN116933800A (en) | Template-based generation type intention recognition method and device | |
CN113742564A (en) | Target resource pushing method and device | |
US20220383125A1 (en) | Machine learning aided automatic taxonomy for marketing automation and customer relationship management systems | |
CN113780610B (en) | Customer service portrait construction method and device | |
CN110334283A (en) | Information recommendation method, device, server and storage medium | |
CN114529309A (en) | Information auditing method and device, electronic equipment and computer readable medium | |
CN107483595A (en) | Information-pushing method and device | |
CN111784377B (en) | Method and device for generating information | |
CN113609397A (en) | Method and device for pushing information | |
CN115329183A (en) | Data processing method, device, storage medium and equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |