CN113158050A - Suspected problem review, review and screening method and system and storage medium - Google Patents
Suspected problem review, review and screening method and system and storage medium Download PDFInfo
- Publication number
- CN113158050A CN113158050A CN202110442881.2A CN202110442881A CN113158050A CN 113158050 A CN113158050 A CN 113158050A CN 202110442881 A CN202110442881 A CN 202110442881A CN 113158050 A CN113158050 A CN 113158050A
- Authority
- CN
- China
- Prior art keywords
- suspected
- comment
- screened
- audited
- audit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012216 screening Methods 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 49
- 238000012552 review Methods 0.000 title claims description 79
- 238000012550 audit Methods 0.000 claims abstract description 128
- 238000004364 calculation method Methods 0.000 claims description 28
- 238000004422 calculation algorithm Methods 0.000 claims description 11
- 230000008859 change Effects 0.000 claims description 11
- 238000004088 simulation Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 230000008901 benefit Effects 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 238000009795 derivation Methods 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/01—Social networking
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Economics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method, a system and a storage medium for reviewing and screening suspected questions, wherein the method comprises the following steps: placing the suspected problem comments to be checked and screened currently into a checking queue; according to the data characteristics of the topics to which each suspected question comment to be audited and screened belongs in the audit queue, firstly ordering the topics, and then distributing corresponding weight to each suspected question comment to be audited and screened under each topic; and performing priority ranking on all the suspected problem comments to be audited and screened according to the audit weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited and screened with high weight in front of the audit queue, and performing audit screening preferentially. Compared with the prior art, the method and the device can enable the auditing work to be more optimized and reasonable, improve the user experience and enable the auditing efficiency to be higher and more convenient.
Description
Technical Field
The invention relates to the technical field of computer internet, in particular to a method, a system and a storage medium for reviewing and screening suspected problem comments.
Background
With the development of the era, various topics or news are not beyond the ear every day, community APPs composed of a plurality of main topic areas begin to be popularized and popular and become main media for people to publish opinions, when watching various topics or news, users can publish their own opinions because of a certain opinion, a certain event, a certain mechanism or a certain person, at the moment, some bad network comments are easy to be ignited by unfavorable party flaring, network violence is caused slightly, certain harm is caused to people, bad influence is caused even to the country seriously, and a great pressure is also caused to APP operation teams of a community platform, related bottom lines are likely to be involved with the APPs and even companies, so that the auditing of the comments is of great importance, but some companies can not kill one before by mistake, the method has the advantages that some topics are hardly commented under, the topics are not hot enough, but are too strict in auditing, even the idea of making people go out of business is generated, and no matter how to express the view of the topics, the 'footprints' are difficult to leave under the topics, and the user is very disliked.
In addition, a large amount of time is consumed for auditing, the basis can be screened out through an official sensitive vocabulary library, continuous intellectualization is realized along with the current auditing, meanwhile, the auditing is very strict, a lot of companies are normal, only the expression mode which is witnessed can be audited, the auditing is enabled to lose humanization, and the user loses experience, so that some companies also need to perform secondary auditing on suspected problematic comments by a perfect auditing team, and the hope that a bad person is not put in and a good person is not caused is achieved.
In the sensitive vocabulary screening technique of today, the screening and the discernment of sensitive vocabulary can only be carried out through the forbidden word storehouse of official party to intelligent general, but because coverage is very wide, and the screening is comparatively strict, in order not to influence user's experience, still generally have special audit team to carry out the secondary audit to the comment that is intercepted, like this, if this comment is not bad comment in the discovery of secondary audit, can also release its comment and disclose, be unlikely to make user experience sense reduce.
However, since review teams are made of manpower, threads are single, and reviews suspected of being intercepted are published in a time sequence, so the review efficiency is not very efficient in such a manner, because each review under a hot topic has a certain value, not only the flow is driven, but also users can pay more attention to the review state (especially some large-V users) of the users, and review the review even if the users pay more attention to the review state, other negative counter responses (such as attacking APP with rhythm of large V) may be caused, and troubles are also caused to the APP operation team, for example, since review reviews are reviewed in time sequence, but the current review reviews are not hot topics, and the suspected problem reviews of the hot topics are still behind, and when the review turns to review the hot topic, the benefit from this audit cannot be guaranteed to be maximized. Not only the user loses experience (the comment has no problem), but also the hot topic reduces a certain flow unconsciously.
Disclosure of Invention
The invention mainly aims to provide a method, a system and a storage medium for auditing and screening suspected problem comments, and aims to improve the efficiency of auditing and screening the suspected problem comments and improve the user experience.
In order to achieve the purpose, the invention provides a suspected problem review screening method, which comprises the following steps:
placing the suspected problem comments to be checked and screened currently into a checking queue;
distributing corresponding audit priority weights to each suspected question comment to be audited and screened according to the data characteristics of the topic to which each suspected question comment to be audited and screened belongs in the audit queue;
and performing priority ranking on all the suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with high audit priority weight in front of the audit queue, and performing audit screening preferentially.
According to a further technical scheme of the invention, the step of assigning a corresponding audit priority weight to each suspected question comment to be audited and screened according to the data characteristics of the topic to which each suspected question comment to be audited and screened belongs in the audit queue comprises the following steps:
constructing a suspected question comment weight calculation model according to a large amount of simulation data and data characteristics of topics to which all suspected question comments to be audited and screened belong;
the step of assigning a corresponding review priority weight to each suspected question comment to be reviewed and screened according to the data characteristics of the topic to which each suspected question comment to be reviewed and screened belongs in the review queue includes:
and putting the data characteristics of the topic to which each suspected question comment to be audited and screened belongs into the suspected question comment weight calculation model, calculating the weight according to the given value in the suspected question comment weight calculation model, and distributing corresponding audit priority weight to each suspected question comment to be audited and screened.
According to a further technical scheme of the present invention, the method for prioritizing all suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened includes the steps of:
detecting whether the data characteristics of the topics to which the suspected question comments to be audited belong in the audit queue change or not within preset time;
if the data characteristics of the topic change, judging whether the review priority weight distance of the suspected problem comments with the changed data characteristics of the topic is larger than the last updating time threshold value or not;
if the data feature of the topic is larger than the preset updating time threshold, updating the auditing priority weight of the suspected question comments with changed data features of the topic, and updating the auditing queue;
if the data characteristics of the topic are less than or equal to the preset updating time threshold, the suspected problem comments of the topic with changed data characteristics are put into a cache queue to wait for auditing and screening.
A further technical solution of the present invention is that the step of updating the audit queue includes:
in the updating process, updating the auditing priority weight of the suspected problem comment to be audited on the basis of the original auditing queue; alternatively, the first and second electrodes may be,
and the newly-built audit queue stores the updated audit priority weight of the suspected problem comment to be audited, and the original audit queue is deleted after the update is finished.
The invention further adopts the technical scheme that the suspected problem comments to be audited with high audit priority weight are placed in front of the audit queue, and the step of preferentially performing audit screening comprises the following steps:
and placing the suspected problem comments to be audited with high audit priority weight in front of the audit queue, and adopting a pair-wise algorithm to audit and screen.
The invention further adopts the technical scheme that the method for auditing and screening by adopting the pair-wise algorithm comprises the following steps of:
commenting pair on the suspected problem to be audited with the high audit priority weight;
and judging which suspected problem comment in the suspected problem comment pair should be preferentially checked and screened and which suspected problem comment is discarded according to a preset rule.
The technical scheme of the invention is that the step of judging which suspected comment of the suspected problem comments pair should be preferentially checked and screened according to a preset rule and which suspected comment of the suspected problem comments is discarded comprises the following steps:
when the suspected question comment pair is suspected bad comment data which is intelligently screened and filtered, discarding the suspected question comment pair;
and when the suspected question comment pair is not subjected to screening and filtering, setting the comment data without any data characteristics as the highest-level audit priority weight.
The invention further adopts the technical scheme that the data characteristics of the topics comprise the heat degree of the topics, the time for publishing the topics, the attention degree after the topics are published and the user characteristic parameters of topic publishers.
In order to achieve the above object, the present invention further provides a suspected problem review screening system, where the system includes a memory, a processor, and a suspected problem review screening program stored in the memory, and the suspected problem review screening program is executed by the processor to perform the steps of the method described above.
To achieve the above object, the present invention further provides a computer-readable storage medium, where a suspected problem review screening program is stored, and when executed by a processor, the suspected problem review screening program performs the steps of the method described above.
The method for reviewing and screening the suspected problems has the advantages that: according to the technical scheme, the suspected problem comments to be checked and screened at present are put into the checking queue; according to the data characteristics of the topics to which each suspected question comment to be audited and screened belongs in the audit queue, assigning corresponding audit priority weights to each suspected question comment to be audited and screened; and performing priority ranking on all suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with the high audit priority weight in front of the audit queue, and performing audit screening preferentially, so that the audit work is optimized and reasonable, the user experience is improved, and the audit efficiency is higher and more convenient.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the structures shown in the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a review screening method for suspected problems according to a first embodiment of the present invention;
fig. 2 is a schematic flow chart of a suspected problem review screening method according to a second embodiment of the present invention.
The objects, features and advantages of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In order to solve the problem of performing priority ranking on suspected problems during secondary audit of the suspected problem comments, the audit work of an audit team becomes more optimized, more scientific and reasonable, the user experience degree is improved, and the audit efficiency is higher and more convenient, the invention provides an audit screening method of the suspected problem comments, and particularly relates to a method for performing secondary priority screening on the suspected problem comments screened based on an intelligent sensitive lexicon.
Referring to fig. 1, a review screening method for suspected problems according to a first embodiment of the present invention includes the following steps:
and step S10, placing the suspected question comments to be checked and screened currently into a checking queue.
In this embodiment, the suspected problem comments to be currently screened and reviewed may be suspected problem comments screened based on the intelligent sensitive thesaurus, and the suspected problem comment screening proposed in this embodiment may refer to secondary screening and review of the suspected problem comments screened by the intelligent sensitive thesaurus.
And step S20, according to the data characteristics of the topic to which each suspected question comment to be audited and screened belongs in the audit queue, assigning a corresponding audit priority weight to each suspected question comment to be audited and screened.
Wherein the suspected question comment may include: comment data published for articles, news, or some network information, or comment data published based on some published comment data, i.e. nested comment data, or reply comment data for the published comment data, etc., which may be text data, graphic data, and/or voice data, etc.
In this embodiment, a process of reviewing suspected problem comments mainly based on text-type data is used for explanation, however, the review method using each type of comment provided by the present invention is not limited to the contents in this embodiment.
In this embodiment, the data characteristics of the topic to which the suspected question comment to be reviewed and screened belongs include: the time of topic publication, the heat of the topic, the time of topic publication, the attention after topic publication, the user characteristic parameters of the topic publisher, and the like. The popularity of the topic refers to whether the aging of the topic includes the current popular policy, hotwords and the like, and the attention after the topic is published refers to the attention rate, the praise amount, the comment amplification and the like after the topic is published.
And step S30, performing priority ranking on all suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with high audit priority weight in front of the audit queue, and performing audit screening preferentially.
In this embodiment, after assigning a corresponding review priority weight to each suspected problem comment to be reviewed and screened, all suspected problem comments to be reviewed and screened are prioritized according to the review priority weight of each suspected problem comment to be reviewed and screened, and the suspected problem comment to be reviewed and screened with a high review priority weight is arranged in front of the review queue for being reviewed and screened preferentially, for example, the suspected problem comment with a high topic popularity or the suspected problem comment with a high topic attention is arranged in front of the review queue for being reviewed and screened preferentially, so that review work can be optimized and reasonable, user experience is improved, and review efficiency is higher and more convenient.
According to the technical scheme, the suspected problem comments to be checked and screened at present are put into the checking queue; distributing corresponding audit priority weights to each suspected question comment to be audited and screened according to the data characteristics of the topic to which each suspected question comment to be audited and screened belongs in the audit queue; and performing priority ranking on all suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with the high audit priority weight in front of the audit queue, and performing audit screening preferentially, so that the audit work is optimized and reasonable, the user experience is improved, and the audit efficiency is higher and more convenient.
Further, referring to fig. 2, fig. 2 is a schematic flow chart of a second embodiment of the method for auditing and screening suspected problem comments according to the present invention, and the difference between this embodiment and the first embodiment shown in fig. 1 is that, in this embodiment, the step S20 of allocating, according to the data characteristic of the topic to which each suspected problem comment to be audited and screened belongs in the audit queue, a corresponding audit priority weight to each suspected problem comment to be audited and screened includes:
and S200, constructing a suspected problem comment weight calculation model according to the data characteristics of a large amount of simulation data and the questions to which all suspected problem comments to be checked and screened belong.
Specifically, the suspected problem comment weight calculation model can be constructed through data features of topic data to which suspected problem comments belong, a feature vector can be obtained through extracting data features of suspected problem comment data to be audited in an audit queue, the feature vector is sent to the priority weight calculation model, and the priority scoring model can query and export corresponding audit priority weight scores according to the feature vector, so that the priority weight calculation scores of the suspected problem comment topics to be audited are obtained.
In the step S20, the step of assigning a corresponding review priority weight to each suspected question comment to be reviewed and screened according to the data feature of the topic to which each suspected question comment to be reviewed and screened belongs in the review queue includes:
step S201, putting the data characteristics of the topic to which each suspected question comment to be audited and screened belongs into the suspected question comment weight calculation model, performing weight calculation according to the corresponding score in the suspected question comment weight calculation model, and distributing corresponding audit priority weight to each suspected question comment to be audited and screened.
According to the technical scheme, the suspected problem comments to be checked and screened at present are put into the checking queue; distributing corresponding audit priority weights to each suspected question comment to be audited and screened according to the data characteristics of the topic to which each suspected question comment to be audited and screened belongs in the audit queue; constructing a suspected question comment weight calculation model according to a large amount of simulation data and data characteristics of topics to which all suspected question comments to be audited and screened belong; putting the data characteristics of the topic to which each suspected problem comment to be audited and screened belongs into the suspected problem comment weight calculation model, performing weight calculation according to the given score in the suspected problem comment weight calculation model, and distributing corresponding audit priority weight to each suspected problem comment to be audited and screened; and performing priority ranking on all the suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with the high audit priority weight in front of the audit queue, and performing audit screening preferentially, so that the audit work is optimized and reasonable, the user experience is improved, and the audit efficiency is higher and more convenient.
Further, based on the first embodiment shown in fig. 1, a third embodiment of the suspected problem comment review screening method of the present invention is proposed, where the difference between this embodiment and the first embodiment shown in fig. 1 is that, in the step S30, all suspected problem comments to be reviewed and screened are prioritized according to the review priority of each suspected problem comment to be reviewed and screened, and the suspected problem comment to be reviewed and screened with a high review priority is placed in front of the review queue, and after the step of preferentially performing review screening, the method further includes:
step S40, in a preset time, detecting whether the data characteristics of the topic to which the suspected question comment to be audited in the audit queue belongs change.
In the embodiment, when suspected problem comments are audited and screened, for example, a large amount of data of suspected problem comments can be generated in a short time for a certain hot topic, the comment data volume of the audit queue can be changed continuously, when a certain text data is reviewed, a large amount of comment data can be generated in a short time, the comment data volume of the audit queue can be changed continuously, in addition, the comment data with given weight priority calculation also has the condition that the data characteristics of the comment data are changed, therefore, in order to improve the audit efficiency, the audit priority score of the comment data in the audit queue can be changed along with the change of the data characteristics, after the suspected problem comments to be audited with high priority are placed in front of the audit queue for preferential audit screening, and detecting whether the data characteristics of the topics to which the suspected question comments to be audited belong in the audit queue are changed or not within a preset time.
If the data characteristics of the topic to which the suspected question comments to be reviewed belong in the review queue change, step S50 is executed to determine whether the review priority weight distance of the suspected question comments with the changed data characteristics of the topic is greater than the preset update time threshold.
The preset update time threshold may be set according to an average value of the variation of the audit data under a normal condition, or may be set according to a requirement of an actual audit worker.
If the distance between the audit priority weight of the suspected problem comment with the changed data feature of the topic and the last update time is greater than the preset update time threshold, step S60 is executed, the audit priority weight of the suspected problem comment with the changed data feature of the topic is updated, and the audit queue is updated.
In this embodiment, the following two schemes may be adopted in the step of updating the audit queue:
in the updating process, the auditing priority weight of the suspected problem comment to be audited is updated on the basis of the original auditing queue.
Or, the newly-built audit queue stores the updated audit priority weight of the suspected problem comment to be audited, and the original audit queue is deleted after the update is finished.
In other embodiments, it can also be understood that a large amount of review data may be generated in a short time for a certain topic, where there are suspected problem review data, and the change of data characteristics also changes greatly, and the continuous updating of the priority weight score brings unnecessary trouble to the review work of the reviewer (such as the repeated review of the same suspected problem review may occur).
If the distance between the review priority weight of the suspected problem comments with the changed data characteristics of the topic and the last update time is less than or equal to the preset update time threshold, step S70 is executed, the suspected problem comments with the changed data characteristics of the topic are placed in a cache queue, and the review and screening are waited.
Further, a fourth embodiment of the suspected problem comment auditing and screening method according to the present invention is provided based on the first embodiment shown in fig. 1, and the difference between this embodiment and the first embodiment shown in the drawing is that, in step S30, the suspected problem comment to be audited with a high auditing priority is placed in front of the auditing queue, and the step of preferentially performing auditing and screening includes:
and placing the suspected problem comments to be audited with high audit priority weight in front of the audit queue, and adopting a pair-wise algorithm to audit and screen.
The pair-wise algorithm is a pairing method, and the basic idea of the method is to compare samples pairwise to construct a partial order document pair and learn sorting from the comparison, because for a query keyword, the most important fact is not to accurately estimate the relevance of a certain document, but to correctly estimate the relative relation between a group of documents, one pair is a combination, too many combinations are provided, and the input cost is too high. The pair-wise algorithm optimizes the traditional full orthogonal design method on the basis of mathematical statistical analysis, and properly improves the efficiency.
For example: there is a capability array [7,9,11,13,15], calculated with an optimal combination value of 20, the two combinations 7 plus 13 and 11 plus 9 conform to the pairwise, and the return value of ([7,9,11,13,15],20) should be the sum of 0+3+1+2, i.e. 6.
The core flow of the pair-wise algorithm is as follows:
1. obtaining dimensions and factors for full arrangement;
2. obtaining a specific every two factor combination (band position) of case;
3. judging whether the combination of every two factors in the case appears on the upper surface, and if the combination of every two factors in the case appears, deleting the combination; if all cases do not appear, the case is reserved;
4. filtering again by using a pairwise algorithm according to different sequences;
5. obtaining 2 groups of data and finding out the same cases;
6. cases are added in order of dimension.
In this embodiment, the step of placing the suspected problem review to be reviewed with a high review priority weight in front of the review queue and performing review screening by using a pair-wise algorithm includes:
and commenting pair on the suspected problems to be audited with high audit priority.
And judging which suspected problem comment in the suspected problem comment pair should be preferentially checked and screened and which suspected problem comment is discarded according to a preset rule.
The step of judging which suspected problem comment in the suspected problem comment pair should be preferentially checked and screened according to a preset rule and which suspected problem comment is discarded includes:
when the suspected question comment pair is suspected bad comment data which is intelligently screened and filtered, discarding the suspected question comment pair;
and when the suspected question comment pair is not subjected to screening and filtering, setting the comment data without any data characteristics as the highest-level audit priority weight.
Specifically, the suspected problem comments pair are then marked as to which one of the comments pair should be preferentially reviewed, and the poorly judged pair is discarded by being lost. It should be noted that there may be two cases in the determination here, one of which is: when the to-be-audited data in the audit queue are suspected bad comment data which are intelligently screened and filtered, bad judgment pair is lost; the other is as follows: and when the data to be checked in the checking queue is not subjected to screening and filtering, the comment data without any characteristics can be set as the highest priority.
After a certain number of labeled pairs are obtained through the rules, any one sequencing model F (x) capable of correctly sequencing all the pairs is learned by using a plurality of existing pair-wise learning algorithms, namely: xl, x2, if the auditor considers xl to be higher priority than x2, F (xl) should be greater than F (x 2).
In this embodiment, the suspected problem comment weight calculation model is finally used to assign review priority scores to suspected problem comments to be reviewed and screened, that is, a comment X is given, and the priority of the comment X is f (X), that is: by reviewing the comment data X in the queue, the review priority score f (X) of the comment X can be obtained through the suspected problem comment weight calculation model.
Hardware functional modules related to the suspected problem review screening method are introduced below.
The hardware function module related to the suspected question review screening method comprises the following steps:
a receiving module: and the auditing queue is used for receiving the suspected problem comments to be audited and screened currently.
An assignment module: and weight calculation for giving the suspected question comments to be audited and screened in the audit queue to audit priority weight.
An auditing module: and the system is used for sorting the audit queue according to the audit priority weight and then carrying out secondary audit on the sorted suspected problem comments.
The assignment module is divided into a construction module and a derivation module.
Constructing a module: a large amount of simulation data and various data characteristics are prepared in advance to construct an intelligent suspected question comment weight calculation model.
A derivation module: and the method is used for deriving the auditing priority weight of the suspected problem comment to be audited according to the suspected problem comment weight calculation model, giving the weight calculation of the auditing priority weight to each suspected problem comment and deriving an auditing queue.
The auditing module is divided into a detecting module and an updating module.
A detection module: and the system is used for detecting whether the data characteristics of the topics to which the suspected question comments to be audited and screened belong in the audit queue change within preset time, and entering an updating module if the data characteristics change.
An update module: and updating the weight calculation of the audit priority weight of the suspected question comments audited and screened. The updating module is divided into:
a preset time judgment module: judging whether the auditing priority weight calculation of the auditing queue is larger than a set updating time threshold value or not from the last updating time; if yes, updating the auditing queue, and performing weight calculation of the auditing priority weight again; if not, the audit data is put into a cache queue to wait for audit.
The method for reviewing and screening the suspected problems has the advantages that: according to the technical scheme, the suspected problem comments to be checked and screened at present are put into the checking queue; according to the data characteristics of the topics to which each suspected question comment to be audited and screened belongs in the audit queue, assigning corresponding audit priority weights to each suspected question comment to be audited and screened; and performing priority ranking on all suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with the high audit priority weight in front of the audit queue, and performing audit screening preferentially, so that the audit work is optimized and reasonable, the user experience is improved, and the audit efficiency is higher and more convenient.
In order to achieve the above object, the present invention further provides a suspected problem review screening system, where the system includes a memory, a processor, and a suspected problem review screening program stored in the memory, and when the suspected problem review screening program is executed by the processor, the steps of the method according to the above embodiment are executed, which are not described herein again.
In order to achieve the above object, the present invention further provides a computer-readable storage medium, where a suspected question review screening program is stored on the computer-readable storage medium, and when the suspected question review screening program is executed by a processor, the steps of the method according to the above embodiment are performed, which is not described herein again.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, and all modifications and equivalents of the present invention, which are made by the contents of the present specification and the accompanying drawings, or directly/indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (10)
1. A suspected question review screening method is characterized by comprising the following steps:
placing the suspected problem comments to be checked and screened currently into a checking queue;
according to the data characteristics of the topics to which each suspected question comment to be audited and screened belongs in the audit queue, firstly, sorting the questions, and then, distributing corresponding weights to each suspected question comment to be audited and screened under each topic so as to perform priority sorting;
and performing priority ranking on all the suspected problem comments to be audited and screened according to the audit weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited and screened with high audit weight in front of the audit queue, and performing audit screening preferentially.
2. The method for auditing and screening suspected problem comments according to claim 1, wherein the step of assigning a corresponding audit priority weight to each suspected problem comment to be audited and screened according to the data characteristics of the topic to which each suspected problem comment to be audited and screened belongs in the audit queue comprises:
constructing a suspected question comment weight calculation model according to a large amount of simulation data and data characteristics of topics to which all suspected question comments to be audited and screened belong;
the step of assigning a corresponding review priority weight to each suspected question comment to be reviewed and screened according to the data characteristics of the topic to which each suspected question comment to be reviewed and screened belongs in the review queue includes:
and putting the data characteristics of the topic to which each suspected question comment to be audited and screened belongs into the suspected question comment weight calculation model, performing weight calculation according to the given value in the suspected question comment weight calculation model, performing priority ranking on the topics, and then allocating corresponding weight to each suspected question comment to be audited and screened under the topic for performing priority ranking.
3. The method for auditing and screening suspected problem comments according to claim 1, wherein the steps of prioritizing and ranking all suspected problem comments to be audited and screened according to the audit priority weight of each suspected problem comment to be audited and screened, placing the suspected problem comment to be audited with a high audit priority weight in front of the audit queue, and preferentially auditing and screening further comprise:
detecting whether the data characteristics of the topics to which the suspected question comments to be audited belong in the audit queue change or not within preset time;
if the data characteristics of the topic change, judging whether the auditing priority weight of the suspected problem comment of which the data characteristics of the topic change is larger than a preset updating time threshold value from the last updating time;
if the data feature of the topic is larger than the preset updating time threshold, updating the auditing priority weight of the suspected question comments with changed data features of the topic, and updating the auditing queue;
if the data characteristics of the topics are less than or equal to the preset updating time threshold, the suspected problem comments with the changed data characteristics of the topics are placed in a cache queue, and the examination and screening are waited.
4. The method for reviewing and screening comments on suspected problems as recited in claim 3, wherein the step of updating the review queue comprises:
in the updating process, updating the auditing priority weight of the suspected problem comment to be audited on the basis of the original auditing queue; or newly building an empty queue to store the updated suspected problem comment list to be checked, and deleting the original checking queue after updating.
5. The suspected question comment auditing and screening method according to claim 1, wherein the step of placing the suspected question comment to be audited with a high auditing priority weight in front of the auditing queue and preferentially performing auditing and screening comprises:
and placing the suspected problem comments to be audited with high audit priority weight in front of the audit queue, and adopting a pair-wise pairing algorithm to audit and screen.
6. The suspected question comment auditing and screening method according to claim 5, wherein the step of placing the suspected question comment to be audited with a high auditing priority weight in front of the auditing queue and adopting a pair-wise pairing algorithm for auditing and screening comprises:
firstly, combining the data characteristics of the current topic and the suspected question comments of the current topic to be used as pair;
pairing and comparing the pair according to the rules of a pairing method, constructing a partial order document pair, and learning and sorting from the comparison, because for a query keyword, the most important is that whether the relevance of a certain document is accurately estimated or not, but the relative relation between a group of documents can be correctly estimated to judge which suspected problem comment in the suspected problem comment pair should be preferentially checked and screened and which suspected problem comment is discarded.
7. The method for auditing and screening suspected problem comments according to claim 6, wherein the step of determining which suspected problem comment in the suspected problem comment pair should be audited and screened with priority and which suspected problem comment is discarded according to a preset rule comprises:
when the suspected question comment pair is suspected bad comment data which is intelligently screened and filtered, discarding the suspected question comment pair;
when the suspected question comment pair is not screened and filtered, setting comment data without any data characteristics as the highest-level audit priority weight;
when there is no data feature in the suspected question comment pair, comments that are liked more per unit time may be set as a priority review.
8. The suspected question review screening method of any one of claims 1 to 7, wherein the data features of the topics include the heat degree of the topic, the time duration of topic publication, the attention degree of the topic after topic publication, and the user feature parameters of the topic publisher.
9. A suspected problem review screening system comprising a memory, a processor, and a suspected problem review screening program stored on the memory, the suspected problem review screening program when executed by the processor performing the steps of the method of any of claims 1 to 8.
10. A computer-readable storage medium having stored thereon a suspected problem review screening program which, when executed by a processor, performs the steps of the method of any of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110442881.2A CN113158050A (en) | 2021-04-23 | 2021-04-23 | Suspected problem review, review and screening method and system and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110442881.2A CN113158050A (en) | 2021-04-23 | 2021-04-23 | Suspected problem review, review and screening method and system and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113158050A true CN113158050A (en) | 2021-07-23 |
Family
ID=76869977
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110442881.2A Pending CN113158050A (en) | 2021-04-23 | 2021-04-23 | Suspected problem review, review and screening method and system and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113158050A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113590791A (en) * | 2021-07-30 | 2021-11-02 | 北京壹心壹翼科技有限公司 | Method, device, equipment and storage medium for optimizing underwriting inquiry strategy |
| CN114372700A (en) * | 2022-01-07 | 2022-04-19 | 京东科技信息技术有限公司 | Data sampling detection method and device |
| CN114710692A (en) * | 2022-03-22 | 2022-07-05 | 上海哔哩哔哩科技有限公司 | Multimedia file processing method and device |
| CN114996165A (en) * | 2022-08-01 | 2022-09-02 | 飞狐信息技术（天津）有限公司 | Business data auditing method and device, storage medium and electronic equipment |
-
2021
- 2021-04-23 CN CN202110442881.2A patent/CN113158050A/en active Pending
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113590791A (en) * | 2021-07-30 | 2021-11-02 | 北京壹心壹翼科技有限公司 | Method, device, equipment and storage medium for optimizing underwriting inquiry strategy |
| CN113590791B (en) * | 2021-07-30 | 2023-11-24 | 北京壹心壹翼科技有限公司 | Nuclear insurance query strategy optimization method, device, equipment and storage medium |
| CN114372700A (en) * | 2022-01-07 | 2022-04-19 | 京东科技信息技术有限公司 | Data sampling detection method and device |
| CN114710692A (en) * | 2022-03-22 | 2022-07-05 | 上海哔哩哔哩科技有限公司 | Multimedia file processing method and device |
| CN114710692B (en) * | 2022-03-22 | 2024-03-01 | 上海哔哩哔哩科技有限公司 | Multimedia file processing method and device |
| CN114996165A (en) * | 2022-08-01 | 2022-09-02 | 飞狐信息技术（天津）有限公司 | Business data auditing method and device, storage medium and electronic equipment |
| CN114996165B (en) * | 2022-08-01 | 2022-12-13 | 飞狐信息技术（天津）有限公司 | Business data auditing method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113158050A (en) | Suspected problem review, review and screening method and system and storage medium | |
| US9268766B2 (en) | Phrase-based data classification system | |
| US9460458B1 (en) | Methods and system of associating reviewable attributes with items | |
| CN103336766B (en) | Short text garbage identification and modeling method and device | |
| CN110377804A (en) | Method for pushing, device, system and the storage medium of training course data | |
| Das et al. | Manipulation among the arbiters of collective intelligence: How Wikipedia administrators mold public opinion | |
| US20110087656A1 (en) | Apparatus for question answering based on answer trustworthiness and method thereof | |
| CN107153847A (en) | Predict method and computing device of the user with the presence or absence of malicious act | |
| CN102884527A (en) | Automatic frequently asked question compilation from community-based question answering archive | |
| CN103150374A (en) | Method and system for identifying abnormal microblog users | |
| CN109670937A (en) | Risk subscribers recognition methods, user equipment, storage medium and device | |
| CN112104642B (en) | Abnormal account number determination method and related device | |
| CN109063000A (en) | Question sentence recommended method, customer service system and computer readable storage medium | |
| CN103218356A (en) | Question quality judging method and system facing open platform | |
| CN106708817A (en) | Information searching method and device | |
| US20230050197A1 (en) | Systems and methods quantifying trust perceptions of entities within social media documents | |
| CN104750760A (en) | Application software recommending method and device | |
| CN109255000A (en) | A kind of the dimension management method and device of label data | |
| CN109960719A (en) | A kind of document handling method and relevant apparatus | |
| CN108920909B (en) | Counterfeit mobile application program discrimination method and system | |
| CN103810204A (en) | Information search method and information search device | |
| CN106997340A (en) | The generation of dictionary and the Document Classification Method and device using dictionary | |
| CN105574112A (en) | Comment information processing method and system of communication process | |
| CN111882224A (en) | Method and device for classifying consumption scenes | |
| CN110209804B (en) | Target corpus determining method and device, storage medium and electronic device |
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 |