CN116245258B - Quality accident risk prediction method and device - Google Patents

Quality accident risk prediction method and device Download PDF

Info

Publication number
CN116245258B
CN116245258B CN202310521326.8A CN202310521326A CN116245258B CN 116245258 B CN116245258 B CN 116245258B CN 202310521326 A CN202310521326 A CN 202310521326A CN 116245258 B CN116245258 B CN 116245258B
Authority
CN
China
Prior art keywords
information
product
event
decoding
available
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
Application number
CN202310521326.8A
Other languages
Chinese (zh)
Other versions
CN116245258A (en
Inventor
冯蕾
禄雨薇
支云杰
杨景娜
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China National Institute of Standardization
Original Assignee
China National Institute of Standardization
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China National Institute of Standardization filed Critical China National Institute of Standardization
Priority to CN202310521326.8A priority Critical patent/CN116245258B/en
Publication of CN116245258A publication Critical patent/CN116245258A/en
Application granted granted Critical
Publication of CN116245258B publication Critical patent/CN116245258B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • 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
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06395—Quality analysis or management
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Development Economics (AREA)
  • Marketing (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Educational Administration (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application discloses a quality accident risk prediction method and device, which aim at the research and development stage of a product, and predict the risk possibly faced by the product based on available information which is formed in the research and development stage and can characterize the product. However, the positioning and details of the product may not be clear in the development stage, and may negatively affect the predicted results. In view of this, the method in the present specification introduces usability, so that the recognition model can pay attention to the process of processing the available information, and output the first decoding information from the decoding end, which is not yet embodied in the design concept in the target product, of the user. So that the first decoded information can represent information that the "future" target product should have.

Description

Quality accident risk prediction method and device
Technical Field
The application relates to the technical field of risk prediction based on a specific computer model, in particular to a quality accident risk prediction method and device.
Background
With the development of information technology, the demand of users for products has been converted from the demand of pure use value to the demand with composivity aiming at multiple aspects such as use value, emotion value, identity value and the like. That is, the product quality is not only reflected in the content of the use value of the product itself and the exertion degree of the use value, but also reflected in the emotion support of the user on the product and the feeling of the identity experience brought by the product to the user. It can be seen that the quality of a product depends to some extent on objective factors such as the use value of the product, and also on subjective factors such as the emotional value that the product can bring to the user.
For example, if a product is priced with a cost as a reference, the audience of the product is positioned in a speaker-like manner, and if a deviation occurs in some aspect of the pricing or the audience positioning, or if there is a mismatch between the pricing and the audience, a quality accident caused by the foregoing "subjective factor" will occur. If these risks could be predicted during the development phase of the product, it would be advantageous to reduce the costs consumed for risk avoidance.
In view of this, how to predict the risk of product quality accidents by combining with 'subjective factors' becomes a problem to be solved urgently.
Disclosure of Invention
The embodiment of the application provides a quality accident risk prediction method and device, so as to at least partially solve the technical problems.
The embodiment of the application adopts the following technical scheme:
in a first aspect, an embodiment of the present application provides a quality accident risk prediction method, including:
acquiring product information of a target product as available information;
determining availability of the available information, wherein the availability is positively correlated with the attribute representation capability of the available information on the attribute of the target product and is positively correlated with a preset weight corresponding to an information dimension contained in the available information;
Inputting the available information into a coding end of a pre-trained identification model to obtain first coding information output by the coding end; the coding end and the decoding end of the identification model are obtained in a co-training mode;
inputting the first coding information and the availability into the decoding end to obtain first decoding information output by the decoding end;
searching event information of an event matched with the first decoding information from a preset historical event library; the event information comprises product information of a reference product for which the event belongs and feedback information of a user for the reference product;
and predicting the quality accident risk of the target product based on the event information.
In an alternative embodiment of the present specification, the recognition model is trained by:
determining product information of a product causing a quality accident from the historical product information;
disassembling the product information of the product causing the quality accident to obtain undetermined information, so that each undetermined information records effective data in only one of the associated dimensions and the target dimension; the target dimension is the information dimension which is most concerned by the quality accident corresponding to the undetermined information; the associated dimension is an information dimension related to, but not identical to, the target dimension;
Carrying out blurring processing on the data recorded by the undetermined information in the target dimension and the associated dimension to obtain sample information;
determining the availability of the sample information so that the availability of the sample information is inversely related to the degree of blurring;
inputting the sample information into a coding end of an identification model to be trained to obtain first to-be-determined coding information output by the coding end;
inputting the first to-be-determined coding information into a decoding end of the recognition model to be trained to obtain first to-be-determined decoding information output by the decoding end;
inputting the availability of the first to-be-determined coding information and the sample information into a decoding end of the recognition model to be trained to obtain second to-be-determined decoding information output by the decoding end;
and training by taking the maximum similarity between the second to-be-determined decoding information and the sample information and taking the maximum similarity between the first to-be-determined decoding information and the to-be-determined information as a training target until convergence.
In an alternative embodiment of the present specification, determining the availability of the available information includes:
extracting a field matched with a preset standard information dimension from the available information to serve as an available field;
Determining the highest definition degree in the available fields as a target field;
searching a weight value corresponding to the target field from a preset weight list to serve as target weight;
determining the comprehensive refinement degree according to the refinement degree of each available field contained in the available information;
and weighting the comprehensive refinement degree by the target weight to obtain the availability.
In an alternative embodiment of the present specification, determining the comprehensive refinement according to the refinement of each available field included in the available information includes:
summing the refinement degrees of the fields to obtain the comprehensive refinement degree.
In an optional embodiment of the present disclosure, searching event information of an event matching the first decoding information from a preset historical event library includes:
for each information dimension, searching an event with highest information dimension matching degree corresponding to the first decoding information from the historical event library as an available event;
and taking the event information of the available event as event information of an event matched with the first decoding information.
In an alternative embodiment of the present specification, predicting a quality accident risk of the target product based on the event information includes:
determining the degree of risk of a quality incident that each available incident has historically caused;
taking the availability of the product information of the reference product corresponding to the available event to which the maximum risk degree belongs as a weight, and weighting the maximum risk degree to obtain a first risk degree;
taking the risk degree of the reference product with the maximum similarity with the target product in the reference products corresponding to each available event as a second risk degree;
and obtaining the quality accident risk of the target product according to the first risk degree and the second risk degree.
In a second aspect, embodiments of the present application further provide a quality accident risk prediction apparatus, the apparatus including:
an information acquisition module configured to: acquiring product information of a target product as available information;
the availability determination module is configured to: determining availability of the available information, wherein the availability is positively correlated with the attribute representation capability of the available information on the attribute of the target product and is positively correlated with a preset weight corresponding to an information dimension contained in the available information;
A first encoded information generation module configured to: inputting the available information into a coding end of a pre-trained identification model to obtain first coding information output by the coding end; the coding end and the decoding end of the identification model are obtained in a co-training mode;
a first decoding information generation module configured to: inputting the first coding information and the availability into the decoding end to obtain first decoding information output by the decoding end;
a search module configured to: searching event information of an event matched with the first decoding information from a preset historical event library; the event information comprises product information of a reference product for which the event belongs and feedback information of a user for the reference product;
a prediction module configured to: and predicting the quality accident risk of the target product based on the event information.
In a third aspect, embodiments of the present application further provide an electronic device, including:
a processor; and
a memory arranged to store computer executable instructions which, when executed, cause the processor to perform the method steps of the first aspect.
In a fourth aspect, embodiments of the present application also provide a computer-readable storage medium storing one or more programs, which when executed by an electronic device comprising a plurality of application programs, cause the electronic device to perform the method steps of the first aspect.
The above-mentioned at least one technical scheme that this application embodiment adopted can reach following beneficial effect:
products typically go through a phase from development to market. In the development period of the product, if the risk of quality accidents possibly facing the product can be predicted and targeted adjustment is performed, the risk can be possibly avoided. With the development of commercial economy, quality accidents of products are not only limited to risks caused by the use value of the products, but also extend to other additional value aspects provided by users on the products, for example, if the products cannot meet the emotional demands of the users, negative effects are brought to the release of the products, and further product quality accidents are caused. The method in this specification focuses on the development stage of the product, and predicts the risk that the product is likely to face based on the available information that has been formed in the development stage and that characterizes the product. However, the positioning and details of the product may not be clear in the development stage, and may negatively affect the predicted results. In view of this, the method in the present specification introduces usability, so that the recognition model can pay attention to the process of processing the available information, and output the first decoding information from the decoding end, which is not yet embodied in the design concept in the target product, of the user. So that the first decoded information can represent information that the "future" target product should have. Further, the method in the present specification predicts the risk that the target product may face based on the historical event, rather than predicting the target product only, and can focus more on what will happen in the future rather than the product information (after all, the current target product is not formed) of the product, and the accident that may happen to the product in different fields may possibly affect the target product, and this situation can also perform risk prediction by the method in the present specification.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute an undue limitation to the application. In the drawings:
fig. 1 is a process schematic diagram of a quality accident risk prediction method according to an embodiment of the present disclosure;
fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.
Detailed Description
The invention will be described in further detail below with reference to the drawings by means of specific embodiments. Wherein like elements in different embodiments are numbered alike in association. In the following embodiments, numerous specific details are set forth in order to provide a better understanding of the present application. However, one skilled in the art will readily recognize that some of the features may be omitted, or replaced by other elements, materials, or methods in different situations. In some instances, some operations associated with the present application have not been shown or described in the specification to avoid obscuring the core portions of the present application, and may not be necessary for a person skilled in the art to describe in detail the relevant operations based on the description herein and the general knowledge of one skilled in the art.
Furthermore, the described features, operations, or characteristics of the description may be combined in any suitable manner in various embodiments. Also, various steps or acts in the method descriptions may be interchanged or modified in a manner apparent to those of ordinary skill in the art. Thus, the various orders in the description and drawings are for clarity of description of only certain embodiments, and are not meant to be required orders unless otherwise indicated.
The numbering of the components itself, e.g. "first", "second", etc., is used herein merely to distinguish between the described objects and does not have any sequential or technical meaning. The terms "coupled" and "connected," as used herein, are intended to encompass both direct and indirect coupling (coupling), unless otherwise indicated.
The following describes in detail the technical solutions provided by the embodiments of the present application with reference to the accompanying drawings.
As shown in fig. 1, the quality accident risk prediction method in the present specification includes the following steps:
s100: product information of the target product is obtained as available information.
The products (including, but not limited to, target products and reference products) in this specification may be physical products having a specific physical form, or may be virtual (e.g., a service) products. The information dimension is the amount used to describe a certain attribute of information. Some information may contain multiple information dimensions. Illustratively, a product is an athletic shoe, the information dimension of which may include, but is not limited to: the selling price of the product, the material of the product, the target crowd of the product, the selling mode of the product, the propaganda term of the product and the like.
The information in this specification may be made up of several fields.
S102: and determining the availability of the available information, wherein the availability is positively correlated with the attribute representation capability of the available information on the attribute of the target product and is positively correlated with a preset weight corresponding to the information dimension contained in the available information.
Availability in this specification is an abstract quantity that can be used to predictably measure the true extent of risk predicted by subsequent steps for a target product. The true degree of the risk level is positively correlated with the availability of the available information. That is, the higher the availability, the more realistic the predicted risk level.
The characterizability in this specification is used to characterize how well the available information describes the target product. Taking athletic shoes as an example, if a product is "athletic shoe" in the information dimension of product type, then it will be less illustrative of the product in at least this dimension; while another product is "basketball shoes" in the information dimension of the product type, it is more descriptive of the product in at least this dimension. Because of the multitude of information dimensions, the characterization capability may be a combination of the degree of interpretation in each dimension (e.g., may be combined by summing). The preset weight corresponding to the information dimension can be set by the user according to the service requirement, and is usually determined by the capability of the user to bear the risk. For example, if the risk bearing capacity of a user is higher in one of the information dimensions of product pricing, then the weight corresponding to that information dimension may be lower.
In an alternative embodiment of the present disclosure, the technical means for determining the availability of the available information is:
and extracting a field matched with a preset standard information dimension (the standard information dimension comprises a target dimension and an associated dimension which will be described later) from the available information, and taking the field as an available field. According to a preset refinement identification rule, determining the highest refinement degree in the available fields as a target field; searching a weight value corresponding to the target field from a preset weight list (formulated according to the business rule) to serve as a target weight; and determining the comprehensive refinement degree according to the refinement degree of each available field contained in the available information (for example, the refinement degree of each available field can be determined respectively, the refinement degree of each available field is normalized based on the number of the available fields, and then the normalized results are summed to obtain the comprehensive refinement degree). And weighting the comprehensive refinement degree by the target weight to obtain the availability.
S104: and inputting the available information into a coding end of a pre-trained identification model to obtain first coding information output by the coding end.
The recognition model in the present specification includes an encoding end and a decoding end. The coding end and the decoding end are obtained in a co-training mode. The function of the encoding end can be simply summarized as extracting the characteristics of the information, and the function of the decoding end can be simply summarized as restoring the characteristics output by the encoding end so as to calculate the information input by the encoding end. The first encoded information obtained by this step can be simply understood as a feature that enables generalization of the available information.
In the related art, a model including both the encoding end and the decoding end is applicable to the present specification where conditions allow. In an alternative embodiment of the present description, the recognition model may be a self-encoder.
S106: and inputting the first coding information and the availability into a decoding end of the identification model to obtain first decoding information output by the decoding end.
Typically, the artificial intelligence model is trained in the same manner as on-line, and the recognition model in this specification also introduces usability in the training process. On one hand, the obtained decoding information can amplify the significant degree difference between the features through the interference about the characterization capability in the availability; on the other hand, the user-focused feature in the obtained first decoding information can be made more convex by the weight in the availability. In addition, based on the subsequent model training method in the specification, the difference between the available information of the first decoding information is not too large, so that the authenticity of the content transmitted by the first decoding information compared with the available information can be ensured.
S108: and searching event information of the event matched with the first decoding information from a preset historical event library.
The event information in the present specification includes product information of a reference product to which an event to which the event belongs, and feedback information of a user to the reference product.
The reference product is a product that has been historically formed. The feedback of the user to the reference product may be a comment of the product by the user or may be feedback expressed by a market behavior (for example, sales of the reference product, etc.).
This step is directed to events that occur historically, rather than to some reference product. The historical events in this specification are only for negative events, such as a star speaking a reference product, which is totally put down for replacement packaging due to the occurrence of a disadvantaged behaviour of the star. The factors that cause the event to have a depolarization nature are many, and besides the problem of the speaker, the factors can be such as that the merchant kills the words and publicizes the words and the words related to sensitive words and the like. With respect to the negative definition, adjustments may be made by the user based on actual business rules.
Technical means that can be used to determine the degree of matching in the related art are applicable to the present specification, as conditions allow. Illustratively, if greater than the match threshold, then match; if not, the two are not matched.
This step measures the degree of matching between event information of a historical event and available information, and aims to predict whether the historical event, or an event similar to the historical event, is likely to occur on a target product.
In an optional embodiment of the present disclosure, for each information dimension, an event with the highest matching degree with the corresponding information dimension of the first decoding information is found from the historical event library, and the event is used as an available event; and taking the event information of the available event as event information of an event matched with the first decoding information.
The available events thus obtained may not be unique, and in an alternative embodiment of the present description, subsequent steps may be performed for each available event.
S110: and predicting the quality accident risk of the target product based on the event information.
This step aims at assuming how much risk the event described by the event information would be on the target product.
Products typically go through a phase from development to market. In the development period of the product, if the risk of quality accidents possibly facing the product can be predicted and targeted adjustment is performed, the risk can be possibly avoided. With the development of commercial economy, quality accidents of products are not only limited to risks caused by the use value of the products, but also extend to other additional value aspects provided by users on the products, for example, if the products cannot meet the emotional demands of the users, negative effects are brought to the release of the products, and further product quality accidents are caused. The method in this specification focuses on the development stage of the product, and predicts the risk that the product is likely to face based on the available information that has been formed in the development stage and that characterizes the product. However, the positioning and details of the product may not be clear in the development stage, and may negatively affect the predicted results. In view of this, the method in the present specification introduces usability, so that the recognition model can pay attention to the process of processing the available information, and output the first decoding information from the decoding end, which is not yet embodied in the design concept in the target product, of the user. So that the first decoded information can represent information that the "future" target product should have. Further, the method in the present specification predicts the risk that the target product may face based on the historical event, rather than predicting the target product only, and can focus more on what will happen in the future rather than the product information (after all, the current target product is not formed) of the product, and the accident that may happen to the product in different fields may possibly affect the target product, and this situation can also perform risk prediction by the method in the present specification.
How the recognition model in the specification can be obtained by training will now be described. The model training process is as follows:
s200: product information of a product that caused a quality accident is determined from the historical product information.
In an alternative embodiment, this step is based on a product information repository in which product information for products that have historically had a history of quality incidents has been maintained. The product information base may maintain the product information of the historical product by creating a table in the product information base, the table having the information dimension as a column and the unique identifier of the product information as a row.
S202: and disassembling the product information of the product causing the quality accident to obtain the undetermined information, so that each undetermined information records effective data in only one of the associated dimensions and the target dimension.
The target dimension in the specification is the information dimension which is most concerned by the quality accident corresponding to the undetermined information; the associated dimension is an information dimension that is related to, but not identical to, the target dimension.
For example, certain product information may be expressed as [ unique identification, target group of products, sales of products, propaganda term of products, selling price of products ]. The pending information obtained by splitting may be: pending information 1: unique identifier 1, null, sales of the product ], pending information 2: unique identifier 2, target group of products, sales mode of products, null ], and pending information 3: [ unique identifier 3, null, promotional term for product, null ]. In the subsequent step, model training is performed with samples obtained based on the undetermined information, so that the model learns to distinguish and predict the input information for each dimension. In the present specification, valid data is data in which usable contents are recorded, and a portion of corresponding invalid data may be denoted by "null".
Under the condition that the product sales mode is the target dimension, the selling price of the product and the target crowd of the product are all associated dimensions of the target dimension. The undetermined information 1 only comprises the selling mode of the product and the selling price of the product, and the undetermined information 2 only comprises the target crowd of the product and the selling mode of the product. The propaganda term of the product is not related to other dimensions, so that the pending information 3 only contains the propaganda term of the product. In the preset business rule, the selling price of the product cannot be used as a target dimension, and the information to be determined aiming at the selling price of the product is not available.
In an optional embodiment of the present disclosure, the selection of the target dimension and the correspondence between the target dimension and the associated dimension are preset based on a business rule. In another alternative embodiment of the present disclosure, the target dimension is an information dimension within a user controllable factor range, and for a target uncontrollable factor range, the target dimension cannot be used as the target dimension; the associated dimension is an information dimension which is high in frequency of occurrence with the target dimension and is in the range of the user controllable factor when quality accidents occur historically.
Although the product information is split in the step, the association relationship between the target dimension and the association dimension is constructed, so that the action relationship between the dimensions is more prominent.
S204: and carrying out blurring processing on the data recorded by the undetermined information in the target dimension and the associated dimension to obtain sample information.
In an optional embodiment of the present disclosure, blurring is performed on both the target dimension and the associated dimension; in another alternative embodiment of the present description, one of the target dimension and the associated dimension may be obfuscated. The goal of the blurring process is to make the content of the information communication less subtle and accurate. Alternatively, the way the blurring process may be generic to the concept, e.g., one information dimension is "football shoes", and the result after the blurring process is "sports shoes". The manner of blurring may also be a bias to the concept, e.g., a "football shoe" after blurring results in a "men's shoe".
In the related art, all the technical means capable of implementing blurring processing of information are applicable to the present specification (for example, blurring cog method). In addition, blurring processing means based on graph technology can be adopted.
S206: the availability of the sample information is determined such that the availability is inversely related to the degree of blurring processing.
The technical means for determining the availability in this step is the same as the manner for determining the availability in the foregoing step, and will not be described in detail herein.
S208: and inputting the sample information into a coding end of the identification model to be trained to obtain first to-be-coded information output by the coding end.
S210: and inputting the first to-be-determined coding information into a decoding end of the recognition model to be trained to obtain the first to-be-determined decoding information output by the decoding end.
S212: and inputting the availability of the first to-be-determined coding information and the sample information into a decoding end of the recognition model to be trained to obtain second to-be-determined decoding information output by the decoding end.
S214: and training by taking the maximum similarity between the second to-be-determined decoding information and the sample information and taking the maximum similarity between the first to-be-determined decoding information and the to-be-determined information as a training target until convergence.
The similarity between the first to-be-determined decoding information and the to-be-determined information is maximized as a training target, so that the processing effect of the coding end on the input information can be ensured, further, the coding end can obtain excellent feature extraction capability, and the accuracy of the identification effect can be ensured. The similarity between the second undetermined decoding information and the sample information is maximized as a training target, so that on one hand, the second undetermined decoding information output by the decoding end and the input sample information can be ensured to keep higher consistency; on the other hand, the availability is introduced, so that the decoding end can identify the blurring degree of the input first to-be-encoded information, and the effective identification can be performed even if the input is blurring. In the research and development scene, the condition that the product information is ambiguous in a certain aspect is easy to happen, but the quality accidents which can be referred to in history are limited, and the model training method of the specification can effectively adapt to the research and development scene of the product and can effectively utilize the quality accidents which occur in history.
Further, in an alternative embodiment of the present specification, in predicting a quality accident risk of a target product, the following steps may be performed:
s300: the degree of risk of a historically induced quality incident for each available event is determined.
The step is a process of quantifying risks of quality accidents, and technical means for quantifying in the related art are applicable to the specification under the condition of permission.
In addition, the risk level can be determined according to the self bearing capacity of the user, for example, if a quality accident causes the taking of a reference product, the risk level of the quality accident is higher; if a quality incident causes a sales of the reference product to slip down, the risk of the quality incident is low.
S302: and weighting the maximum risk degree by taking the availability degree of the reference product corresponding to the available event to which the maximum risk degree belongs as a weight to obtain a first risk degree.
The technical means for determining the availability in this step may be the same as the method of the previous step.
S304: and taking the risk degree of the reference product with the maximum similarity with the target product in the reference products corresponding to the available events as a second risk degree.
The technical means for determining the similarity in the related art are applicable to the present specification as the conditions allow. Optionally, the similarity is determined from product information.
S306: and obtaining the quality accident risk of the target product according to the first risk degree and the second risk degree.
In an alternative embodiment of the present specification, the sum of the first risk level and the second risk level may be taken as the risk level of the target product in case of a quality accident. In another alternative embodiment of the present specification, the greater of the first risk level and the second risk level may be taken as the risk level that the target product is exposed to in case of a quality accident.
Optionally, the accident reason of each available event can be determined by integrating and analyzing each available event, and the accident reason is used as the accident reason of the target product under the condition of quality accidents.
Further, the present specification also provides a quality accident risk prediction apparatus, the apparatus comprising:
an information acquisition module configured to: acquiring product information of a target product as available information;
the availability determination module is configured to: determining availability of the available information, wherein the availability is positively correlated with the attribute representation capability of the available information on the attribute of the target product and is positively correlated with a preset weight corresponding to an information dimension contained in the available information;
A first encoded information generation module configured to: inputting the available information into a coding end of a pre-trained identification model to obtain first coding information output by the coding end; the coding end and the decoding end of the identification model are obtained in a co-training mode;
a first decoding information generation module configured to: inputting the first coding information and the availability into the decoding end to obtain first decoding information output by the decoding end;
a search module configured to: searching event information of an event matched with the first decoding information from a preset historical event library; the event information comprises product information of a reference product for which the event belongs and feedback information of a user for the reference product;
a prediction module configured to: and predicting the quality accident risk of the target product based on the event information.
The apparatus can perform the method in any of the foregoing embodiments, and can obtain the same or similar technical effects, which are not described herein.
Fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present application. Referring to fig. 2, at the hardware level, the electronic device includes a processor, and optionally an internal bus, a network interface, and a memory. The Memory may include a Memory, such as a Random-Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least 1 disk Memory. Of course, the electronic device may also include hardware required for other services.
The processor, network interface, and memory may be interconnected by an internal bus, which may be an ISA (Industry Standard Architecture ) bus, a PCI (Peripheral Component Interconnect, peripheral component interconnect standard) bus, or EISA (Extended Industry Standard Architecture ) bus, among others. The buses may be classified as address buses, data buses, control buses, etc. For ease of illustration, only one bi-directional arrow is shown in FIG. 2, but not only one bus or type of bus.
And the memory is used for storing programs. In particular, the program may include program code including computer-operating instructions. The memory may include memory and non-volatile storage and provide instructions and data to the processor.
The processor reads the corresponding computer program from the nonvolatile memory into the memory and then runs the computer program to form a quality accident risk prediction device on a logic level. And the processor is used for executing the program stored in the memory and particularly executing any one of the quality accident risk prediction methods.
The quality accident risk prediction method disclosed in the embodiment shown in fig. 1 of the present application can be applied to a processor or implemented by the processor. The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or by instructions in the form of software. The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; but also digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The disclosed methods, steps, and logic blocks in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be embodied directly in hardware, in a decoded processor, or in a combination of hardware and software modules in a decoded processor. The software modules may be located in a random access memory, flash memory, read only memory, programmable read only memory, or electrically erasable programmable memory, registers, etc. as well known in the art. The storage medium is located in a memory, and the processor reads the information in the memory and, in combination with its hardware, performs the steps of the above method.
The electronic device may also execute a quality accident risk prediction method in fig. 1, and implement the functions of the embodiment shown in fig. 1, which are not described herein.
The embodiments also provide a computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by an electronic device comprising a plurality of application programs, perform any of the quality incident risk prediction methods described above.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.
The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.
Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.
It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.

Claims (8)

1. A method of predicting a risk of a quality accident, the method comprising:
acquiring product information of a target product as available information;
determining availability of the available information, wherein the availability is positively correlated with the attribute representation capability of the available information on the attribute of the target product and is positively correlated with a preset weight corresponding to an information dimension contained in the available information;
Inputting the available information into a coding end of a pre-trained identification model to obtain first coding information output by the coding end; the coding end and the decoding end of the identification model are obtained in a co-training mode;
inputting the first coding information and the availability into the decoding end to obtain first decoding information output by the decoding end;
searching event information of an event matched with the first decoding information from a preset historical event library; the event information comprises product information of a reference product for which the event belongs and feedback information of a user for the reference product;
predicting the quality accident risk of the target product based on the event information;
wherein, the recognition model is a self-encoder, and the recognition model is obtained by training the following steps:
determining product information of a product causing a quality accident from the historical product information;
disassembling the product information of the product causing the quality accident to obtain undetermined information, so that each undetermined information records effective data in only one of the associated dimensions and the target dimension; the target dimension is the information dimension which is most concerned by the quality accident corresponding to the undetermined information; the associated dimension is an information dimension related to, but not identical to, the target dimension;
Carrying out blurring processing on the data recorded by the undetermined information in the target dimension and the associated dimension to obtain sample information;
determining the availability of the sample information so that the availability of the sample information is inversely related to the degree of blurring;
inputting the sample information into a coding end of an identification model to be trained to obtain first to-be-determined coding information output by the coding end;
inputting the first to-be-determined coding information into a decoding end of the recognition model to be trained to obtain first to-be-determined decoding information output by the decoding end;
inputting the availability of the first to-be-determined coding information and the sample information into a decoding end of the recognition model to be trained to obtain second to-be-determined decoding information output by the decoding end;
and training by taking the maximum similarity between the second to-be-determined decoding information and the sample information and taking the maximum similarity between the first to-be-determined decoding information and the to-be-determined information as a training target until convergence.
2. The method of claim 1, wherein determining the availability of the available information comprises:
extracting a field matched with a preset standard information dimension from the available information to serve as an available field;
Determining the highest definition degree in the available fields as a target field;
searching a weight value corresponding to the target field from a preset weight list to serve as target weight;
determining the comprehensive refinement degree according to the refinement degree of each available field contained in the available information;
and weighting the comprehensive refinement degree by the target weight to obtain the availability.
3. The method of claim 2, wherein determining the aggregate level of refinement from the levels of refinement for each available field contained in the available information comprises:
summing the refinement degrees of the fields to obtain the comprehensive refinement degree.
4. The method of claim 1, wherein finding event information of an event matching the first decoding information from a preset historical event library comprises:
for each information dimension, searching an event with highest information dimension matching degree corresponding to the first decoding information from the historical event library as an available event;
and taking the event information of the available event as event information of an event matched with the first decoding information.
5. The method of claim 4, wherein predicting a quality accident risk for the target product based on the event information comprises:
determining the degree of risk of a quality incident that each available incident has historically caused;
taking the availability of the product information of the reference product corresponding to the available event to which the maximum risk degree belongs as a weight, and weighting the maximum risk degree to obtain a first risk degree;
taking the risk degree of the reference product with the maximum similarity with the target product in the reference products corresponding to each available event as a second risk degree;
and obtaining the quality accident risk of the target product according to the first risk degree and the second risk degree.
6. A quality accident risk prediction apparatus, the apparatus comprising:
an information acquisition module configured to: acquiring product information of a target product as available information;
the availability determination module is configured to: determining availability of the available information, wherein the availability is positively correlated with the attribute representation capability of the available information on the attribute of the target product and is positively correlated with a preset weight corresponding to an information dimension contained in the available information;
A first encoded information generation module configured to: inputting the available information into a coding end of a pre-trained identification model to obtain first coding information output by the coding end; the coding end and the decoding end of the identification model are obtained in a co-training mode;
a first decoding information generation module configured to: inputting the first coding information and the availability into the decoding end to obtain first decoding information output by the decoding end;
a search module configured to: searching event information of an event matched with the first decoding information from a preset historical event library; the event information comprises product information of a reference product for which the event belongs and feedback information of a user for the reference product;
a prediction module configured to: predicting the quality accident risk of the target product based on the event information;
wherein, the recognition model is a self-encoder, and the recognition model is obtained by training the following steps:
determining product information of a product causing a quality accident from the historical product information;
disassembling the product information of the product causing the quality accident to obtain undetermined information, so that each undetermined information records effective data in only one of the associated dimensions and the target dimension; the target dimension is the information dimension which is most concerned by the quality accident corresponding to the undetermined information; the associated dimension is an information dimension related to, but not identical to, the target dimension;
Carrying out blurring processing on the data recorded by the undetermined information in the target dimension and the associated dimension to obtain sample information;
determining the availability of the sample information so that the availability of the sample information is inversely related to the degree of blurring;
inputting the sample information into a coding end of an identification model to be trained to obtain first to-be-determined coding information output by the coding end;
inputting the first to-be-determined coding information into a decoding end of the recognition model to be trained to obtain first to-be-determined decoding information output by the decoding end;
inputting the availability of the first to-be-determined coding information and the sample information into a decoding end of the recognition model to be trained to obtain second to-be-determined decoding information output by the decoding end;
and training by taking the maximum similarity between the second to-be-determined decoding information and the sample information and taking the maximum similarity between the first to-be-determined decoding information and the to-be-determined information as a training target until convergence.
7. An electronic device, comprising:
a processor; and
a memory arranged to store computer executable instructions which, when executed, cause the processor to perform the method of any of claims 1 to 5.
8. A computer readable storage medium storing one or more programs, which when executed by an electronic device comprising a plurality of application programs, cause the electronic device to perform the method of any of claims 1-5.
CN202310521326.8A 2023-05-10 2023-05-10 Quality accident risk prediction method and device Active CN116245258B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310521326.8A CN116245258B (en) 2023-05-10 2023-05-10 Quality accident risk prediction method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310521326.8A CN116245258B (en) 2023-05-10 2023-05-10 Quality accident risk prediction method and device

Publications (2)

Publication Number Publication Date
CN116245258A CN116245258A (en) 2023-06-09
CN116245258B true CN116245258B (en) 2023-08-04

Family

ID=86631696

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310521326.8A Active CN116245258B (en) 2023-05-10 2023-05-10 Quality accident risk prediction method and device

Country Status (1)

Country Link
CN (1) CN116245258B (en)

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110020882A (en) * 2018-12-11 2019-07-16 阿里巴巴集团控股有限公司 A kind of event prediction method and apparatus
US11119895B2 (en) * 2019-08-19 2021-09-14 International Business Machines Corporation Risk-focused testing
EP3975080A1 (en) * 2020-09-29 2022-03-30 Siemens Aktiengesellschaft Automated risk driven patch management
CN113722461B (en) * 2021-09-02 2022-05-06 中科雨辰科技有限公司 Target event extraction data processing system
CN115935265B (en) * 2023-03-03 2023-05-26 支付宝(杭州)信息技术有限公司 Method for training risk identification model, risk identification method and corresponding device

Also Published As

Publication number Publication date
CN116245258A (en) 2023-06-09

Similar Documents

Publication Publication Date Title
CN110245935B (en) Payment channel recommendation method, device and equipment
US10242415B2 (en) Method and system for determining content treatment
US10430727B1 (en) Systems and methods for privacy-preserving generation of models for estimating consumer behavior
CN110688974B (en) Identity recognition method and device
CN106919611B (en) Product information pushing method and device
CN114298417A (en) Anti-fraud risk assessment method, anti-fraud risk training method, anti-fraud risk assessment device, anti-fraud risk training device and readable storage medium
CN110874491A (en) Privacy data processing method and device based on machine learning and electronic equipment
Galvão et al. Forecasting movie box office profitability
CN109697454B (en) Cross-device individual identification method and device based on privacy protection
US20190347443A1 (en) Detection and prevention of privacy violation due to database release
WO2023016147A1 (en) Multi-target prediction method and apparatus, device, storage medium, and program product
CN110751515A (en) Decision-making method and device based on user consumption behaviors, electronic equipment and storage medium
KR100955523B1 (en) Method and system for case based music recommendation using context-awareness
CN106844402B (en) Data processing method and device
CN116245258B (en) Quality accident risk prediction method and device
CN116109338B (en) Electric business analysis method and system based on artificial intelligence
CN111275071B (en) Prediction model training method, prediction device and electronic equipment
CN116542747A (en) Product recommendation method and device, storage medium and electronic equipment
CN113297462A (en) Data processing method, device, equipment and storage medium
CN114840762A (en) Recommended content determining method and device and electronic equipment
CN111311372A (en) User identification method and device
CN113159834A (en) Commodity information sorting method, device and equipment
CN116596638B (en) Information recommendation method based on numerical processing model
CN113240422B (en) Financial product transfer method based on block chain and related device
CN110705439A (en) Information processing method, device 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