CN111507869A

CN111507869A - Translation quality inspection extraction method and device, computer equipment and storage medium

Info

Publication number: CN111507869A
Application number: CN202010160759.1A
Authority: CN
Inventors: 虞朋鑫
Original assignee: Wensihai Huizhike Technology Co ltd
Current assignee: Wensihai Huizhike Technology Co ltd
Priority date: 2020-03-10
Filing date: 2020-03-10
Publication date: 2020-08-07

Abstract

The application relates to a method and a device for extracting quality inspection of translated text, computer equipment and a storage medium. The method comprises the following steps: obtaining at least one translated string; obtaining a quality sampling inspection proportion according to the historical quality inspection score; the quality sampling rate is a rate for sampling the translation quality of the at least one translated word string; and extracting the at least one translated word string according to the quality sampling ratio to obtain a word string to be subjected to quality inspection. By adopting the method, the efficiency of quality inspection and extraction of the translated text can be improved.

Description

Translation quality inspection extraction method and device, computer equipment and storage medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for quality inspection and extraction of a translated text, a computer device, and a storage medium.

Background

In order to guarantee the translation quality of the translated file, quality inspection personnel is often required to perform quality inspection on translated contents submitted by the translator.

In the prior art, when quality inspection is performed on translated contents of a translator, the quality inspection personnel often need to manually extract part of documents from a plurality of documents to be quality inspected submitted by the translator or manually extract part of text contents in the documents to be quality inspected from the documents to be quality inspected for quality inspection, the flow is complex, and the time consumption of the sampling inspection process is long.

Therefore, the method for extracting the quality of the translated text in the prior art has the problem of low efficiency.

Disclosure of Invention

In view of the above, it is desirable to provide a translated text quality inspection method, apparatus, computer device and storage medium capable of improving the quality inspection efficiency of translated text, in order to solve the problem of low efficiency of the translated text quality inspection method in the prior art.

A translation quality inspection extraction method, the method comprising:

obtaining at least one translated string;

obtaining a quality sampling inspection proportion according to the historical quality inspection score; the quality sampling rate is a rate for sampling the translation quality of the at least one translated word string;

and extracting the at least one translated word string according to the quality sampling ratio to obtain a word string to be subjected to quality inspection.

In one embodiment, the extracting the at least one translated word string according to the quality sampling ratio to obtain a word string to be quality-checked includes:

determining a current string to be extracted in the at least one translated string;

generating an extraction random number aiming at the current string to be extracted, and determining an extraction numerical range according to the quality extraction ratio;

and when the extracted random number is in accordance with the extraction numerical range, determining the current string to be extracted as the string to be detected, and extracting the current string to be extracted.

In one embodiment, the obtaining the quality spot-check proportion according to the historical quality check score includes:

acquiring the historical quality inspection score; the historical quality inspection score is obtained by performing quality inspection on the historical translated text:

according to the historical quality inspection grade, determining a target quality sampling inspection proportion in a preset extraction proportion list as the quality sampling inspection proportion; wherein, the extraction ratio list records the corresponding relationship between the historical quality inspection score and the quality extraction ratio.

In one embodiment, the historical quality control score has a plurality; and the time section corresponding to the target historical quality inspection grade meets the preset condition.

and acquiring an actual spot check proportion in real time, comparing the actual spot check proportion with the quality spot check proportion, and determining whether to perform extraction operation on the current string to be extracted so as to adjust the actual spot check proportion to be in accordance with a preset target spot check proportion range.

In one embodiment, the comparing the actual sampled ratio with the quality sampled ratio to determine whether to perform a decimation operation on the current string to be decimated includes:

acquiring a reasonable deviation extraction proportion;

determining a first sampling proportion range aiming at the current string to be extracted according to the quality sampling proportion and the reasonable deviation extraction proportion;

and if the actual sampled inspection proportion is smaller than the lower limit of the first sampled inspection proportion range, determining that the current string to be extracted is the string to be detected, and performing extraction operation on the current string to be extracted.

In one embodiment, if the actual sampled proportion is greater than the upper limit of the first sampled proportion range, the extraction operation is not performed on the current string to be extracted.

In one embodiment, the obtaining the actual sampled inspection ratio in real time includes:

acquiring the current translated total word number in real time, and acquiring the current quality check word number in real time;

and determining the actual sampled inspection proportion according to the current translated total word number and the current quality inspection word number.

In one embodiment, the determining the actual sampled proportion according to the current translated total word number and the current quality check word number includes:

acquiring the word number of the word string of the current word string to be extracted;

adding the current total word number of the quality inspection and the word number of the character string to obtain a first word number, and adding the translated total word number and the word number of the character string to obtain a second word number;

and calculating the ratio of the first word number to the second word number to obtain the actual sampled ratio.

In one embodiment, the method further comprises the following steps:

acquiring a string distribution strategy aiming at the string to be checked;

and issuing the string to be checked according to the string distribution strategy.

In one embodiment, the issuing the string to be quality checked according to the string distribution policy includes:

determining a quality testing end having a preset binding relationship with the translation end as a candidate quality testing end; the translated word string is obtained by translating the word string to be translated by the translation end;

according to the number of the quality detection words of the candidate quality detection end aiming at the text to be translated, a target quality detection end is determined in the candidate quality detection end;

and issuing the string to be tested to the target quality testing end.

A translation quality inspection extraction device, the device comprising:

the text acquisition module is used for acquiring at least one translated word string;

the proportion acquisition module is used for acquiring the quality sampling inspection proportion according to the historical quality inspection grade; the quality sampling rate is a rate for sampling the translation quality of the at least one translated word string;

and the extraction module is used for extracting the at least one translated word string according to the quality sampling ratio to obtain the word string to be subjected to quality inspection.

A computer device comprising a memory and a processor, the memory storing a computer program, the processor implementing the following steps when executing the computer program:

obtaining at least one translated string;

A computer-readable storage medium, on which a computer program is stored which, when executed by a processor, carries out the steps of:

obtaining at least one translated string;

According to the method, the device, the computer equipment and the storage medium for extracting the quality inspection of the translated text, the translated text which is input by a translator when the translator translates the text to be translated is obtained; and according to the historical quality inspection grade corresponding to the translator, a quality spot inspection proportion is obtained, and at least one translated word string in the translated text is extracted according to the quality spot inspection proportion to obtain a word string to be inspected so that the quality inspection personnel can inspect the word string to be inspected, so that the translated text submitted by the translator can be subjected to quality spot inspection according to the actual translation performance of the translator, the translation content submitted by the translator does not need to be extracted manually by the quality inspection personnel, and the efficiency of quality inspection and extraction of the translated text is improved.

Drawings

FIG. 1 is a diagram illustrating an exemplary embodiment of an application environment for a quality inspection method of translated material;

FIG. 2 is a flowchart illustrating a method for quality inspection of translation according to an embodiment;

FIG. 3 is a flowchart illustrating a method for quality inspection of translated text according to another embodiment;

FIG. 4 is a block diagram of an embodiment of a translation quality inspection extraction apparatus;

FIG. 5 is a flowchart illustrating another method for quality inspection of translated text according to one embodiment;

FIG. 6 is a diagram illustrating an internal structure of a computer device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The translation quality inspection method provided by the application can be applied to the application environment shown in FIG. 1. The translation end 110 and the quality testing end 130 communicate with the translation management system 120 through a network. Wherein, the translation management system 120 obtains at least one translated string; then, the translation management system 120 obtains a quality spot check ratio according to the historical quality check score; the quality sampling inspection proportion is the proportion for sampling and inspecting the translation quality of at least one translated word string; finally, the translation management system 120 extracts at least one translated word string according to the quality sampling ratio to obtain a word string to be quality checked. The string to be quality-checked is sent to the quality checking terminal 130. In practical applications, the translation end 110 and the quality inspection end 130 may be, but are not limited to, various personal computers, notebook computers, smart phones, tablet computers, and portable wearable devices, and the translation management system 120 may be implemented by an independent server or a server cluster formed by a plurality of servers.

In one embodiment, as shown in fig. 2, a method for extracting quality inspection of a translation is provided, which includes the following steps:

in step S210, at least one translated string is obtained.

The translated word string is a translated text which is input when the translation end translates the word string to be translated. In practical application, the word string to be translated is obtained by splitting a file to be translated.

In practical applications, the word string to be translated may refer to a word string obtained by splitting a file to be translated in advance, for example, the word string to be translated is a sentence in the file to be translated.

In addition, the string to be translated can also refer to a complete file to be translated.

In a specific implementation, in the process that a translator translates a text to be translated through the translation terminal 110, the translation management system 120 receives and stores a translated text entered by the translator when the text to be translated is translated, and extracts a plurality of translated word strings in the text to be translated for quality inspection by a quality inspection worker when a preset execution condition is met. In practical applications, the translation management system 120 may extract a plurality of translated strings in the text to be translated each time the accumulation reaches a preset time period, for example, 2 hours. The translation management system 120 may also extract a plurality of translated word strings in the text to be translated after the translator task complete, for example, after the translator clicks a complete task button on the interactive interface of the translation end 110 through a mouse.

And step S220, obtaining the quality sampling inspection ratio according to the historical quality inspection score.

Wherein the quality sampling rate is a rate of sampling the translation quality of at least one translated string.

The historical quality inspection score is obtained by performing quality inspection on the historical translated text of the translator.

In a specific implementation, after the quality testing personnel completes the quality testing work of the string to be tested on the quality testing terminal 130, a quality testing report (QA) is generated, wherein each report has a quality testing score corresponding to the string to be tested.

When the translation management system 120 needs to obtain the quality spot check ratio corresponding to the translation end 110, the translation management system 120 may obtain a historical quality check score for the translation end 110; and determines the quality sampling rate for the translation end 110 according to the historical quality inspection score. Specifically, the translation management system 120 may obtain a preset matrix table, such as a QA history score/QA sampling ratio matrix table, in which the mapping relationship between the history quality inspection score and the quality sampling ratio is recorded. The translation management system 120 then looks up the quality spot check ratio in the matrix table corresponding to the translator's historical quality check score.

In step S230, at least one translated string is extracted according to the quality sampling ratio, so as to obtain a string to be quality-checked.

The string to be quality-checked may be a translated string that needs quality checking by a quality-checking person.

In a specific implementation, after the translation management system 120 obtains the quality spot check ratio corresponding to the translation end 110, the translation management system 120 extracts the strings to be subjected to quality check for the translated text from the plurality of translated strings submitted by the translation end 110 according to the quality spot check ratio.

After the translation management system 120 extracts the to-be-quality-checked strings for the translated text from the plurality of translated strings submitted by the translation terminal 110, the translation management system 120 issues the to-be-quality-checked strings to the corresponding quality check terminal, so that a quality check worker at the quality check terminal performs quality check on the translated text submitted by the translation terminal 110. Specifically, the translation management system 120 may obtain a string issuing policy for the string to be quality-checked, and issue the string to be quality-checked to the corresponding quality check end according to the string issuing policy.

According to the method for extracting the quality inspection of the translated text, at least one translated word string is obtained, the translated word string is a translated text which is input when a word string to be translated is translated at a translation end, and the word string to be translated is obtained by splitting a file to be translated; and obtaining a quality spot check proportion according to a historical quality check score corresponding to a translator, extracting at least one translated word string in the translated text according to the quality spot check proportion to obtain a word string to be checked, so that the quality check personnel can perform quality check on the word string to be checked, thus realizing the purpose of performing quality spot check on the translated text submitted by the translator according to the actual translation performance of the translator, avoiding the need of manually extracting translation contents submitted by the translator by the quality check personnel, improving the efficiency of extracting quality check of the translated text, realizing the quality control of the translated text at the word string level, and finding the translation quality problem of the translator in a timely and refined manner.

In another embodiment, the extracting at least one translated string according to the quality sampling ratio to obtain the string to be quality-checked includes: determining a current string to be extracted in at least one translated string; generating an extraction random number aiming at the current string to be extracted, and determining a target numerical range of the extraction random number according to the quality sampling ratio; and when the extracted random number is in accordance with the target numerical range, performing extraction operation on the current string to be extracted, and determining the current string to be extracted as the string to be checked.

In a specific implementation, the process that the translation management system 120 extracts the string to be quality checked for the translated text from the at least one translated string according to the quality sampling ratio specifically includes: the translation management system 120 first determines a current string to be extracted from at least one translated string; then, the translation management system 120 generates an extraction random number for the string to be extracted currently; specifically, the translation management system 120 may assign an equiprobable random number of 0-99 to the current string to be extracted through a random number generation tool written based on JAVA (a programming language). Meanwhile, the translation management system 120 determines a target numerical range for extracting random numbers according to the quality sampling ratio; for example, if the quality spot check percentage is known to be 30%, the extraction threshold is 30, i.e., the target value range is 0 to 30. When the random number generated by the random number generation tool for the extracted random number allocated to the current string to be extracted is less than 30, it indicates that the extracted random number of the current string to be extracted conforms to the target value range, and at this time, the translation management system 120 performs an extraction operation on the current string to be extracted to determine that the current string to be extracted is a string to be checked.

According to the technical scheme of the embodiment, the current string to be extracted in at least one translated string is determined; generating an extraction random number aiming at the current string to be extracted, and determining a target numerical range of the extraction random number according to the quality sampling ratio; when the extracted random number is in accordance with the target numerical range, extraction operation is performed on the current string to be extracted, the current string to be extracted is determined to be the string to be tested, extraction operation can be accurately performed on the translated string according to the preset extraction probability, the string to be tested is determined, and therefore the efficiency of quality test on the translated text is improved.

In another embodiment, obtaining a quality spot check ratio based on the historical quality check score comprises: acquiring a historical quality inspection score of a translation end; the historical quality inspection score is obtained by performing quality inspection on the historical translated text of the translation end: according to the historical quality inspection grade, determining a target quality sampling inspection proportion in a preset extraction proportion list as the quality sampling inspection proportion; wherein, the extraction ratio list records the corresponding relationship between the historical quality inspection score and the quality extraction ratio. .

The historical quality inspection score is obtained by quality inspection of the translated text of the history of the translator by a quality inspector.

In a specific implementation, after the quality testing personnel completes the quality testing work on the quality testing end 130 for the string to be tested, a quality testing report (QA) quality testing report is generated, wherein each report has a corresponding historical quality testing score.

When the translation management system 120 needs to obtain the quality spot check ratio corresponding to the translation end 110, the translation management system 120 may obtain a historical quality check score for the translation end 110; and determines the quality sampling rate for the translation end 110 according to the historical quality inspection score. Specifically, the translation management system 120 may obtain a preset extraction ratio list, such as a QA history score/QA extraction ratio matrix table, in which the mapping relationship between the history quality inspection score and the quality extraction ratio is recorded. The translation management system 120 then queries the extraction ratio list for a quality extraction ratio corresponding to the translator's historical quality inspection score.

Specifically, the historical quality control score has a plurality; each historical quality control score has a corresponding time cut, and the process of acquiring the quality sampling ratio according to the historical quality control score by the translation management system 120 specifically includes: the translation management system 120 determines a target quality sampling ratio in a preset sampling ratio list as the quality sampling ratio; finally, the translation management system 120 determines a target quality sampling ratio in a preset matrix map according to at least two target historical quality inspection scores, and the target quality sampling ratio is used as a quality sampling ratio.

The extraction ratio list records the corresponding relation between each historical quality inspection score and each quality extraction ratio.

And the time section corresponding to the target historical quality inspection score meets the preset condition. In practical application, the target historical quality inspection score can be obtained by quality inspection personnel performing quality inspection on the historical translated text of the translator for one or two times at the latest, that is, the time section corresponding to the target historical quality inspection score is the time section closest to the current time.

Therefore, the quality sampling inspection method can accurately distribute a quality sampling inspection proportion matched with the recent translation performance of the translation end to the translation end according to the target historical quality inspection grade meeting the preset conditions and the time, further carry out reasonable sampling inspection on the translated strings submitted by the translation end, and realize reasonable distribution of translation sampling inspection resources.

To facilitate understanding by those skilled in the art, as shown in table one; table one provides a list of decimation ratios in this embodiment. In this manner, the translation management system 120 can query the matrix map for the quality snapshot proportions corresponding to the translation end 110.

Watch 1

According to the technical scheme, the quality spot check proportion of the translation end is determined according to the historical quality check score of the translation end, the quality of the translated text submitted by the translator can be accurately checked according to the actual translation quality of the translator, the quality spot check proportion can be dynamically adjusted according to the actual performance of the translator, therefore, limited quality check resources are fully utilized, and the translator with quality risks is detected more, so that the translator can perform the translation quality spot check on each translator pertinently.

In another embodiment, determining a quality spot check ratio based on the translation quality score comprises: acquiring an initial sampling inspection proportion corresponding to the translation quality score; acquiring a translation quality control budget for a text to be translated; adjusting the initial sampling inspection proportion according to the translation quality inspection budget to obtain the adjusted initial sampling inspection proportion; and taking the adjusted initial sampling inspection proportion as a quality sampling inspection proportion.

The translation quality control budget may refer to a project budget invested in quality control of the text to be translated.

In a specific implementation, in the process that the translation management system 120 determines the quality spot check ratio according to the translation quality score, the initial spot check ratio corresponding to the translation quality score of the translator may be obtained according to a preset matrix table in which the mapping relationship between the translation quality score and the initial spot check ratio is recorded; then, the translation management system 120 acquires a translation quality inspection budget for the text to be translated, and adjusts the initial sampling inspection ratio according to the translation quality inspection budget to obtain an adjusted initial sampling inspection ratio; specifically, the translation management system 120 may determine a project progress for the text to be translated, and when the project progress reaches a preset progress threshold, it indicates that the translation project of the text to be translated is in a later stage; at this time, the translation management system 120 determines the total quantity of the quality-detectable strings for the text to be translated according to the translation quality detection budget, and obtains the total quantity of the quality-detected strings for the text to be translated; finally, adjusting the initial sampling inspection proportion according to the total quantity of the quality-detectable strings and the total quantity of the quality-detected strings to obtain the adjusted initial sampling inspection proportion; finally, the translation management system 120 takes the adjusted initial sampling rate as the quality sampling rate.

According to the technical scheme of the embodiment, the initial spot check proportion corresponding to the translation quality score is obtained; acquiring a translation quality control budget for a text to be translated; adjusting the initial sampling inspection proportion according to the translation quality inspection budget to obtain the adjusted initial sampling inspection proportion; and taking the adjusted initial sampling inspection proportion as a quality sampling inspection proportion, thus comprehensively considering the budget of the translation project group, ensuring the deviation between the final extraction amount of the character string to be subjected to quality inspection and the target extraction amount corresponding to the budget to be in a reasonable range, and ensuring the sustainable performance of quality inspection work.

In another embodiment, the extracting at least one translated string according to the quality sampling ratio to obtain the string to be quality-checked includes: determining a word string to be extracted currently in at least one translated word string; and acquiring the actual sampled inspection proportion in real time, and determining whether to perform extraction operation on the current word strings to be extracted according to the actual sampled inspection proportion and the quality sampled inspection proportion so as to adjust the actual sampled inspection proportion to be in accordance with a preset target sampled inspection proportion range.

In a specific implementation, the process that the translation management system 120 extracts the string to be quality checked for the translated text from the at least one translated string according to the quality sampling ratio specifically includes: the translation management system 120 determines a current string to be extracted from at least one translated string; then, the translation management system 120 obtains the actual sampled proportion in real time, compares the actual sampled proportion with the quality sampled proportion, and determines whether to perform extraction operation on the current string to be extracted according to the comparison result, so as to adjust the actual sampled proportion to meet the preset target sampled proportion range.

In the process of acquiring the actual sampled inspection ratio in real time, the translation management system 120 specifically includes: acquiring the current translated total word number in real time, and acquiring the current quality check word number in real time; and determining the actual sampled inspection ratio according to the current translated total word number and the current quality inspection word number.

Wherein, the current translated total word number may refer to the total word number of each word string in the current translated text. In practice, the translated total word number may refer to the total word number of each word string in the translated text that has been subjected to the translation quality check strategy.

Wherein, the current quality control word number may refer to the total number of words in the translated text of the translator which have been extracted to enter the quality control stage.

In a specific implementation, the process that the translation management system 120 extracts the string to be quality checked for the translated text from the at least one translated string according to the quality sampling ratio specifically includes: the translation management system 120 first obtains the current translated total word number of each word string in the translated text, and obtains the current quality check word number corresponding to the quality check word string of the translator; the translation management system 120 then determines the actual spot check ratio based on the current total number of translated words and the current quality check word number.

Specifically, the process of determining the actual sampled inspection ratio according to the current total translated words and the current quality check words by the translation management system 120 specifically includes: acquiring the word number of the word string of the current word string to be extracted; adding the current total word number of the quality inspection and the word number of the character string to obtain a first word number, and adding the translated total word number and the word number of the character string to obtain a second word number; and calculating the ratio of the first word number to the second word number to obtain the actual sampled ratio.

For example, the data table of the translation management system 120 records the total word number tw of the word string that the translator needs to perform the quality check policy at the time of performing the quality check policy at the last time, i.e. the current translated total word number, and also records the total word number qw of the word string that the translator performs the quality check by the quality check policy at the time of performing the quality check policy at the last time, i.e. the current quality check word number, and also records the word number zw of the word string to be extracted at the same time. Then, the translation management system 120 determines the sampled inspection proportion of the translated text according to the ratio of the total number of inspected words and the total number of translated words; more specifically, the translation management system 120 may also obtain the string word number zw of the string currently to be extracted. Then, the translation management system 120 adds the current total number of detected words and the word number of the string to obtain a first word number, and adds the translated total number of words and the word number of the string to obtain a second word number; the ratio of the first word number to the second word number is calculated to obtain the actual snapshot ratio rb ═ (qw + zw)/(tw + zw). Then, the translation management system 120 may compare the actual spot check ratio with the quality spot check ratio, and determine whether to perform an extraction operation on the current string to be extracted according to the comparison result, so as to adjust the actual spot check ratio to conform to a preset target spot check ratio range, thereby dynamically adjusting the quality spot check ratio and ensuring that a deviation between an extraction amount and a budget of the final string to be quality checked is within a reasonable range.

In another embodiment, determining whether to perform a extraction operation on the current string to be extracted according to the actual extracted proportion and the quality extracted proportion includes: acquiring a reasonable deviation extraction proportion; determining a first sampling proportion range aiming at the current string to be extracted according to the quality sampling proportion and the reasonable deviation extraction proportion; and when the actual sampled inspection proportion is smaller than the lower limit of the first sampled inspection proportion range, performing extraction operation on the current string to be extracted, and determining the current string to be extracted as the string to be inspected.

If the actual sampled proportion is larger than the upper limit of the first sampled proportion range, the extraction operation is not executed on the current string to be extracted.

Wherein, the reasonable deviation extraction proportion can be the allowable deviation proportion between the quality sampling proportion and the budget sampling proportion.

In a specific implementation, in the process that the translation management system 120 determines whether to perform extraction operation on the current string to be extracted according to the actual extracted proportion and the quality extracted proportion, the translation management system 120 first determines a first extracted proportion threshold value for the translated string according to the quality extracted proportion; as for the previous example, the translation management system 120 may obtain the reasonable deviation extraction ratio and note it as fd, and then the translation management system 120 determines the quality sampling ratio ra and the reasonable deviation extraction ratio, and determines the first sampling ratio range (ra-fd, ra + fd) for the current string to be extracted; when actually the sampling inspection proportion rb is at the lower limit of the first sampling inspection proportion range, namely rb < ra-fd, extracting the current character string to be extracted to determine that the current character string to be extracted is the character string to be inspected, so that the quality sampling inspection proportion can be dynamically adjusted, and the deviation between the extraction quantity and the budget of the final character string to be inspected is ensured to be in a reasonable range.

And if the actual sampled proportion is larger than the upper limit of the first sampled proportion range, namely rb > ra + fd, not performing extraction operation on the current string to be extracted.

In another embodiment, when the actual extracted proportion rb is in the first extracted proportion range (ra-fd, ra + fd), that is, ra-fd < (rb) < ═ ra + fd, the translation management system 120 first determines the current string to be extracted in at least one of the translated strings; then, the translation management system 120 generates an extraction random number for the string to be extracted currently; specifically, the translation management system 120 may assign an equiprobable random number of 0-99 to the current string to be extracted through a random number generation tool written based on JAVA (a programming language). Meanwhile, the translation management system 120 determines a target numerical range for extracting random numbers according to the quality sampling ratio; for example, if the quality spot check percentage is known to be 30%, the extraction threshold is 30, i.e., the target value range is 0 to 30. When the random number generated by the random number generation tool for the extracted random number allocated to the current string to be extracted is less than 30, it indicates that the extracted random number of the current string to be extracted conforms to the target value range, and at this time, the translation management system 120 performs an extraction operation on the current string to be extracted to determine that the current string to be extracted is a string to be checked.

It should be noted that the extracted strings enter the waiting pool for quality inspection allocation, and the un-extracted strings are bound to the next extracted string.

According to the technical scheme, the actual spot-check proportion and the quality spot-check proportion are compared, whether extraction operation is performed on the current character string to be extracted is determined according to the comparison result, so that the actual spot-check proportion is adjusted to be in accordance with the preset target spot-check proportion range, and the deviation between the extraction amount and the budget of the final character string to be quality checked is guaranteed to be within a reasonable range while the quality spot-check proportion is dynamically adjusted.

In another embodiment, the method further comprises: acquiring a string distribution strategy aiming at a string to be checked; and issuing the character string to be checked according to the character string distribution strategy.

The string distribution policy may be a policy for distributing strings to be quality-checked to each quality-check terminal.

In a specific implementation, in a process that the translation management system 120 issues a string to be quality-checked to a quality check end, the translation management system 120 may first obtain a string distribution policy for the string to be quality-checked; then, the translation management system 120 issues the string to be quality-checked to the corresponding target quality inspection terminal according to the string distribution policy.

In practical application, the string distribution strategy may be a distribution strategy for preferentially issuing the necessary test strings to the target quality inspection end, or a distribution strategy for preferentially issuing the to-be-inspected strings to the target quality inspection end having a binding relationship with the translator, or a distribution strategy for issuing the to-be-inspected strings to the target quality inspection end not having a binding relationship with the translator when the total quantity of the to-be-inspected strings at each quality inspection end is greater than a preset total quantity threshold, that is, when the QA capacity is sufficient; or, when the total amount of the quality testing strings of each quality testing end is still greater than the preset total amount threshold after the strategy of issuing the strings to be tested to the target quality testing end which does not have the binding relationship with the translator, each quality testing end can perform the allocation strategy of contending for the strings to be tested.

According to the technical scheme of the embodiment, a string distribution strategy for a string to be detected is obtained; the method and the device realize the flexible issuing of the character string to be quality checked to the target quality checking end according to the character string distribution strategy and the actual situation, and improve the efficiency of quality checking the translated text.

In another embodiment, issuing the string to be checked according to the string assignment strategy includes: determining a quality testing end having a preset binding relationship with the translation end as a candidate quality testing end; the translated string is obtained by translating the string to be translated by the translation end; determining a target quality inspection end in the candidate quality inspection ends according to the number of the quality inspection words of the candidate quality inspection ends aiming at the text to be translated; and issuing the character string to be tested to the target quality testing terminal.

In a specific implementation, in the process of issuing the word string to be quality-checked according to the word string distribution policy, the translation management system 120 firstly determines, from among the multiple quality-checking terminals, a quality-checking terminal having a preset binding relationship with the translator, as a candidate quality-checking terminal; then, the translation management system 120 determines a target quality inspection end in the candidate quality inspection ends according to the quality-inspected character string quantity of the candidate quality inspection end aiming at the text to be translated; specifically, the translation management system 120 may sort each candidate quality inspection terminal according to the amount of the detected character string corresponding to each candidate quality inspection terminal, from high to low according to the amount of the detected character string; then, the candidate quality inspection end arranged at the front n positions is used as a target quality inspection end; finally, the translation management system 120 issues the string to be quality checked to the target quality inspection terminal. In practice, N may be equal to 1.

When the string to be checked for the translated text is extracted by the translation management system 120, the translation management system 120 determines that the translated string is extracted, and the translation management system 120 may record the value of the string to be checked in the last extraction, the extraction amount (QA amount) of the string to be checked that has been extracted until the quality check policy is loaded, and the total extraction amount (QA total amount) of the string to be checked for the translated text. Then, the translation management system 120 may generate a corresponding Q TargetVersion (quality check Target version) according to the ttargettversion (translation Target version) of the string to be quality checked, generate a binding id for the Q TargetVersion of the string to be quality checked, record the binding id in a Target state and the T TargetVersion of the string to be quality checked, bind to the same T Task, and bind the binding id to a string that has not been previously extracted. If translation management system 120 determines that the translated string is not unneeded, translation management system 120 modifies the Target state of the translated string to be unneeded. The translation management system 120 extracts the highest value of the string extraction ratio that exceeds the maximum floating ratio, and the translation management system 120 also binds a corresponding binding id to the translated string.

According to the technical scheme of the embodiment, a quality inspection end with a preset binding relationship with a translator is determined to serve as a candidate quality inspection end; then according to the quality-checked character string quantity of the candidate quality-checking end aiming at the text to be translated, determining a target quality-checking end in the candidate quality-checking end; and the target quality inspection terminal is sent to the target quality inspection terminal; therefore, quality testing personnel can carry out quality testing on the allocated character strings to be tested skillfully, and the efficiency of carrying out quality testing on the translated text is improved.

In another embodiment, issuing the string to be checked according to the string assignment strategy includes: receiving a character string robbing request sent by a quality inspection end aiming at a character string to be inspected; responding the word string competition request, issuing the word string to be quality checked to the quality checking end, and returning a competition ending message to other quality checking ends to inform other quality checking ends that the word string to be quality checked is issued to the corresponding quality checking ends; the other quality inspection terminals are the quality inspection terminals which do not receive the character strings to be inspected.

The word string contention request may be a request sent from the quality inspection terminal to the translation management system to contend for the quality inspection word string.

Wherein, the other quality testing terminals are the quality testing terminals which do not receive the character strings to be tested.

In a specific implementation, when the total amount of the quality detectable word strings of each quality inspection end is greater than a preset total amount threshold, the translation management system 120 adopts an allocation strategy that each quality inspection end can contend for the word strings to be detected. Specifically, in the process of issuing the to-be-detected character string according to the character string distribution policy by the translation management system 120, when the translation management system 120 receives a character string robbing request submitted by the quality inspection terminal, the translation management system 120 responds to the character string robbing request, issues the to-be-detected character string to the quality inspection terminal, and returns a robbing end message to other quality inspection terminals to notify other quality inspection terminals that the to-be-detected character string has been issued to the corresponding quality inspection terminal; the other quality inspection terminals are the quality inspection terminals which do not receive the character strings to be inspected. In practical application, the quality inspection terminal can preferentially contend for the word string to be inspected, which is submitted by the translator having the binding relationship with the quality inspection terminal. Secondly, the quality inspection end can contend for the word string to be inspected submitted by the translator who has no binding relationship with the quality inspection end but has the overstock condition of the word string to be inspected.

It should be noted that, when the quality inspection end of the quality inspector receives the to-be-inspected character string and the accumulated time (for example, 1 hour) is up to the preset time, the quality inspection end will return the to-be-inspected character string to the to-be-inspected distribution waiting pool, and the translation management system 120 will send the to-be-inspected character string to other quality inspection ends again.

According to the technical scheme of the embodiment, when a word string contention request submitted by a quality inspection end is received; determining a quality inspection end corresponding to the word string scrambling request as a second target quality inspection end; generating a second string distribution strategy for distributing the string to be detected to the second target quality detection end according to the second target quality detection end; therefore, limited quality inspection resources can be more reasonably and fully utilized, and the efficiency of quality inspection on the translated content is further improved.

In another embodiment, as shown in fig. 3, a method for extracting quality inspection of a translation is provided, which includes the following steps: in step S310, at least one translated string is obtained. Step S320, obtaining historical quality control scores; the historical quality inspection score is obtained by performing quality inspection on the historical translated text. Step S330, determining a target quality sampling ratio in a preset sampling ratio list according to the historical quality inspection grade, wherein the target quality sampling ratio is used as the quality sampling ratio; wherein, the extraction ratio list records the corresponding relationship between the historical quality inspection score and the quality extraction ratio. In step S340, a current string to be extracted is determined in the at least one translated string. In step S350, the current translated total word number is obtained in real time, and the current quality check word number is obtained in real time. Step S360, determining the actual spot check proportion according to the current translated total word number and the current quality check word number, and determining whether to perform extraction operation on the current word string to be extracted according to the actual spot check proportion and the quality spot check proportion so as to adjust the actual spot check proportion to be in accordance with a preset target spot check proportion range. In step S370, a string distribution policy for the string to be checked is obtained. Step S380, according to the word string distribution strategy, the word string to be tested is issued. The specific limitations of the above steps can be referred to the limitations of the above method for extracting the quality inspection of the translation, and are not described herein again.

It should be understood that although the steps in the flowcharts of fig. 2 and 3 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2 and 3 may include multiple sub-steps or multiple stages that are not necessarily performed at the same time, but may be performed at different times, and the order of performing the sub-steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least some of the sub-steps or stages of other steps.

In one embodiment, as shown in fig. 4, there is provided a translation quality inspection extracting device including:

a text obtaining module 410, configured to obtain at least one translated word string;

the proportion obtaining module 420 is used for obtaining the quality sampling inspection proportion according to the historical quality inspection scores; the quality sampling rate is a rate for sampling the translation quality of the at least one translated word string;

and the extracting module 430 is configured to extract the at least one translated word string according to the quality sampling ratio to obtain a word string to be quality-checked.

In one embodiment, the extracting module 430 is specifically configured to determine a current string to be extracted in the at least one translated string; generating an extraction random number aiming at the current string to be extracted, and determining an extraction numerical range according to the quality extraction ratio; and when the extracted random number is in accordance with the extraction numerical range, determining the current string to be extracted as the string to be detected, and extracting the current string to be extracted.

In one embodiment, the ratio obtaining module 420 is specifically configured to obtain the historical quality inspection score; the historical quality inspection score is obtained by performing quality inspection on the historical translated text: according to the historical quality inspection grade, determining a target quality sampling inspection proportion in a preset extraction proportion list as the quality sampling inspection proportion; wherein, the extraction ratio list records the corresponding relationship between the historical quality inspection score and the quality extraction ratio.

In one embodiment, the extracting module 430 is specifically configured to determine, in the at least one translated word string, a word string to be currently extracted; and acquiring an actual sampled inspection proportion in real time, and determining whether to perform extraction operation on the current string to be extracted according to the actual sampled inspection proportion and the quality sampled inspection proportion so as to adjust the actual sampled inspection proportion to accord with a preset target sampled inspection proportion range.

In one embodiment, the extraction module 430 is specifically configured to obtain a reasonable deviation extraction ratio; determining a first sampling proportion range aiming at the current string to be extracted according to the quality sampling proportion and the reasonable deviation extraction proportion; and if the actual sampled inspection proportion is smaller than the lower limit of the first sampled inspection proportion range, determining that the current string to be extracted is the string to be detected, and performing extraction operation on the current string to be extracted.

In one embodiment, the extracting module 430 is specifically configured to not perform the extracting operation on the current string to be extracted if the actual extracted proportion is greater than the upper limit of the first extracted proportion range.

In one embodiment, the extracting module 430 is specifically configured to obtain the current translated total word number in real time, and obtain the current quality check word number in real time; and determining the actual sampled inspection proportion according to the current translated total word number and the current quality inspection word number.

In one embodiment, the extracting module 430 is specifically configured to obtain the word number of the word string of the current word string to be extracted; adding the current total word number of the quality inspection and the word number of the character string to obtain a first word number, and adding the translated total word number and the word number of the character string to obtain a second word number; and calculating the ratio of the first word number to the second word number to obtain the actual sampled ratio.

In one embodiment, the method further comprises the following steps: the strategy acquisition module is used for acquiring a string distribution strategy aiming at the string to be subjected to quality inspection; and the issuing module is used for issuing the character string to be checked according to the character string distribution strategy.

In one embodiment, the issuing module is specifically configured to determine a quality inspection terminal having a preset binding relationship with the translation terminal, and use the quality inspection terminal as a candidate quality inspection terminal; the translated word string is obtained by translating the word string to be translated by the translation end; according to the number of the quality detection words of the candidate quality detection end aiming at the text to be translated, a target quality detection end is determined in the candidate quality detection end; and issuing the string to be tested to the target quality testing end.

For specific limitations of the translation quality inspection extracting device, reference may be made to the above limitations of the translation quality inspection extracting method, which are not described herein again. All or part of the modules in the translation quality inspection extraction device can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

To facilitate understanding of those skilled in the art, fig. 5 is a schematic flow chart of another method for extracting quality inspection of translation; as shown in fig. 5, the translator translates the text to be translated through the translation end 110, and submits the translated text to the translation management system 120; wherein the translated text includes at least one translated word string. After receiving the translated text of the translator, the translation management system 120 executes the QA policy, which is a translation quality check extraction policy, at regular time, and determines whether the translated word string is a centered word string. If the string is a snapshot-to-center string, the quality check string is ready. The string to be quality checked is put into the waiting pool for quality check allocation, namely waiting for the QA pool. If the string is not extracted, the string is bound to the next extracted string, so that the data tracing in the later period is facilitated. Meanwhile, the translation management system 120 executes a corresponding distribution strategy of the strings to be quality-tested, and issues the strings to be quality-tested to the corresponding target quality testing end, so that quality testers can perform quality testing on the strings to be quality-tested to generate a quality testing report with historical quality testing scores, namely a QA report.

In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used for storing the extracted data of the quality inspection of the translated text. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to realize a translation quality inspection extraction method.

Those skilled in the art will appreciate that the architecture shown in fig. 6 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:

step S210, obtaining at least one translated string;

step S220, obtaining a quality sampling inspection ratio according to the historical quality inspection grade; the quality sampling rate is a rate for sampling the translation quality of the at least one translated word string;

step S230, according to the quality sampling ratio, extracting the at least one translated word string to obtain a word string to be quality checked.

In one embodiment, the processor, when executing the computer program, further performs the steps of: determining a current string to be extracted in the at least one translated string; generating an extraction random number aiming at the current string to be extracted, and determining an extraction numerical range according to the quality extraction ratio; and when the extracted random number is in accordance with the extraction numerical range, determining the current string to be extracted as the string to be detected, and extracting the current string to be extracted.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring the historical quality inspection score; the historical quality inspection score is obtained by performing quality inspection on the historical translated text: according to the historical quality inspection grade, determining a target quality sampling inspection proportion in a preset extraction proportion list as the quality sampling inspection proportion; wherein, the extraction ratio list records the corresponding relationship between the historical quality inspection score and the quality extraction ratio.

In one embodiment, the processor when executing the computer program further performs the steps of, the historical quality check score having a plurality; and the time section corresponding to the target historical quality inspection grade meets the preset condition.

In one embodiment, the processor, when executing the computer program, further performs the steps of: determining a current string to be extracted in the at least one translated string; and acquiring an actual spot check proportion in real time, comparing the actual spot check proportion with the quality spot check proportion, and determining whether to perform extraction operation on the current string to be extracted so as to adjust the actual spot check proportion to be in accordance with a preset target spot check proportion range.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring a reasonable deviation extraction proportion; determining a first sampling proportion range aiming at the current string to be extracted according to the quality sampling proportion and the reasonable deviation extraction proportion; and if the actual sampled inspection proportion is smaller than the lower limit of the first sampled inspection proportion range, determining that the current string to be extracted is the string to be detected, and performing extraction operation on the current string to be extracted.

In one embodiment, the processor further performs the following step when executing the computer program, and if the actual checked proportion is greater than the upper limit of the first checked proportion range, not performing the extraction operation on the current string to be extracted.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring the current translated total word number in real time, and acquiring the current quality check word number in real time; and determining the actual sampled inspection proportion according to the current translated total word number and the current quality inspection word number.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring the word number of the word string of the current word string to be extracted; adding the current total word number of the quality inspection and the word number of the character string to obtain a first word number, and adding the translated total word number and the word number of the character string to obtain a second word number; and calculating the ratio of the first word number to the second word number to obtain the actual sampled ratio.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring a string distribution strategy aiming at the string to be checked; and issuing the string to be checked according to the string distribution strategy.

In one embodiment, the processor, when executing the computer program, further performs the steps of: determining a quality testing end having a preset binding relationship with the translation end as a candidate quality testing end; the translated word string is obtained by translating the word string to be translated by the translation end; according to the number of the quality detection words of the candidate quality detection end aiming at the text to be translated, a target quality detection end is determined in the candidate quality detection end; and issuing the string to be tested to the target quality testing end.

In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:

step S210, obtaining at least one translated string;

In one embodiment, the computer program when executed by the processor further performs the steps of: determining a current string to be extracted in the at least one translated string; generating an extraction random number aiming at the current string to be extracted, and determining an extraction numerical range according to the quality extraction ratio; and when the extracted random number is in accordance with the extraction numerical range, determining the current string to be extracted as the string to be detected, and extracting the current string to be extracted.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring the historical quality inspection score; the historical quality inspection score is obtained by performing quality inspection on the historical translated text: according to the historical quality inspection grade, determining a target quality sampling inspection proportion in a preset extraction proportion list as the quality sampling inspection proportion; wherein, the extraction ratio list records the corresponding relationship between the historical quality inspection score and the quality extraction ratio.

In one embodiment, the computer program when executed by the processor further performs the steps of, the historical quality check score having a plurality; and the time section corresponding to the target historical quality inspection grade meets the preset condition.

In one embodiment, the computer program when executed by the processor further performs the steps of: determining a current string to be extracted in the at least one translated string; and acquiring an actual spot check proportion in real time, comparing the actual spot check proportion with the quality spot check proportion, and determining whether to perform extraction operation on the current string to be extracted so as to adjust the actual spot check proportion to be in accordance with a preset target spot check proportion range.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring a reasonable deviation extraction proportion; determining a first sampling proportion range aiming at the current string to be extracted according to the quality sampling proportion and the reasonable deviation extraction proportion; and if the actual sampled inspection proportion is smaller than the lower limit of the first sampled inspection proportion range, determining that the current string to be extracted is the string to be detected, and performing extraction operation on the current string to be extracted.

In one embodiment, the computer program further implements the following step when executed by the processor, and if the actual checked proportion is greater than the upper limit of the first checked proportion range, the extraction operation is not performed on the current string to be extracted.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring the current translated total word number in real time, and acquiring the current quality check word number in real time; and determining the actual sampled inspection proportion according to the current translated total word number and the current quality inspection word number.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring the word number of the word string of the current word string to be extracted; adding the current total word number of the quality inspection and the word number of the character string to obtain a first word number, and adding the translated total word number and the word number of the character string to obtain a second word number; and calculating the ratio of the first word number to the second word number to obtain the actual sampled ratio.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring a string distribution strategy aiming at the string to be checked; and issuing the string to be checked according to the string distribution strategy.

In one embodiment, the computer program when executed by the processor further performs the steps of: determining a quality testing end having a preset binding relationship with the translation end as a candidate quality testing end; the translated word string is obtained by translating the word string to be translated by the translation end; according to the number of the quality detection words of the candidate quality detection end aiming at the text to be translated, a target quality detection end is determined in the candidate quality detection end; and issuing the string to be tested to the target quality testing end.

It will be understood by those of ordinary skill in the art that all or a portion of the processes of the methods of the embodiments described above may be implemented by instructions associated with hardware via a computer program that may be stored on a non-volatile computer-readable storage medium, which when executed, may include the processes of the embodiments of the methods described above, wherein any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A method for extracting quality control of a translated text, the method comprising:

obtaining at least one translated string;

2. The method according to claim 1, wherein said extracting the at least one translated string according to the quality spot check ratio to obtain a string to be quality checked comprises:

3. The method of claim 1, wherein obtaining a quality spot proportion based on the historical quality inspection score comprises:

4. The method of claim 3, wherein the historical quality control score has a plurality; and the time section corresponding to the target historical quality inspection grade meets the preset condition.

5. The method according to claim 1, wherein said extracting the at least one translated string according to the quality spot check ratio to obtain a string to be quality checked comprises:

6. The method according to claim 5, wherein the comparing the actual sampled rate with the quality sampled rate to determine whether to perform a decimation operation on the current string to be decimated comprises:

acquiring a reasonable deviation extraction proportion;

7. The method according to claim 6, wherein if the actual sampled proportion is larger than the upper limit of the first sampled proportion range, then no extraction operation is performed on the current string to be extracted.

8. The method of claim 5, wherein the obtaining the actual spot proportion in real time comprises:

9. The method of claim 8, wherein determining the actual raffle ratio based on the current translated total word count and the current quality check word count comprises:

10. The method of claim 1, further comprising:

acquiring a string distribution strategy aiming at the string to be checked;

11. The method according to claim 6, wherein said issuing the string to be checked according to the string distribution policy comprises:

and issuing the string to be tested to the target quality testing end.

12. A translation quality inspection extracting device, characterized in that the device comprises:

13. A computer device comprising a memory and a processor, the memory storing a computer program, wherein the processor implements the steps of the method of any one of claims 1 to 11 when executing the computer program.

14. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 11.