Data analysis method and device, electronic equipment, storage medium and program product

Abstract

The invention relates to the technical field of data processing, and discloses a data analysis method and device, electronic equipment, a storage medium and a program product, the method comprises the following steps: displaying a question and answer annotation page, the question and answer annotation page displaying a target question and answer content of a question and answer model for the target question; in response to a trigger instruction of authenticity verification, obtaining a target verification result of the answer content, the target verification result being obtained by performing authenticity verification on first sub-content split from the answer content and performing statistical analysis on the first sub-verification result obtained by verification; and displaying the target verification result on the verification result page, wherein the target verification result is used for labeling the authenticity of the answer content on the question and answer labeling page. According to the method and the device, the problem of relatively long time consumption for performing authenticity verification on the answers given by the large model can be solved, so that the efficiency of performing authenticity verification and authenticity labeling on the answers given by the large model is improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a data analysis method, device, electronic device, storage medium and program product. Background Art

[0002] For business scenarios that use large models, in order to improve the authenticity of the answers given by the large models, annotators need to search each sentence in the answers given by the large models through the browser to verify the authenticity of the answers and give an authenticity score for the entire answer. Since the annotators need to search for each sentence in the answer, and different browsers will give different search results, the authenticity of each sentence needs to be confirmed after comparing multiple search results, which makes it time-consuming to verify the authenticity of the answers given by the large models. Summary of the invention

[0003] In view of this, the present disclosure provides a data analysis method, device, electronic device, storage medium and program product to solve the problem of time-consuming authenticity verification of answers given by large models.

[0004] In a first aspect, the present disclosure provides a data analysis method, the method comprising:

[0005] Displaying a question and answer annotation page, wherein the question and answer annotation page displays a target question and an answer content of the question and answer model to the target question;

[0006] In response to a trigger instruction for authenticity verification, a target verification result of the answer content is obtained, wherein the target verification result is obtained by performing authenticity verification on a first sub-content split from the answer content, and performing statistical analysis on the first sub-verification result obtained by the verification;

[0007] The target verification result is displayed on the verification result page, and the target verification result is used to mark the authenticity of the answer content on the question and answer marking page.

[0008] In a second aspect, the present disclosure provides a data analysis device, the device comprising:

[0009] A marked page display module is used to display a question and answer marked page, wherein the question and answer marked page displays a target question and an answer content of the question and answer model to the target question;

[0010] A verification result acquisition module, configured to obtain a target verification result of the answer content in response to a trigger instruction of authenticity verification, wherein the target verification result is obtained by performing authenticity verification on the first sub-content split from the answer content and performing statistical analysis on the first sub-verification result obtained by the verification;

[0011] The verification result display module is used to display the target verification result on the verification result page, and the target verification result is used to mark the authenticity of the answer content on the question and answer marking page.

[0012] In a third aspect, the present disclosure provides an electronic device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the data analysis method of the first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0013] In a fourth aspect, the present disclosure provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the data analysis method of the first aspect or any corresponding embodiment thereof.

[0014] In a fifth aspect, the present disclosure provides a computer program product, including computer instructions, which are used to enable a computer to execute the data analysis method of the first aspect or any corresponding embodiment thereof.

[0015] The data analysis method provided by the embodiment of the present disclosure has a target verification result obtained by verifying the authenticity of the first sub-content split out of the answer content and performing statistical analysis on the first sub-verification result obtained by the verification. Therefore, it is possible to divide the answer content into fine-grained parts, reduce redundant information in a single authenticity verification, and facilitate understanding and verification of the information represented by the first sub-content, thereby effectively improving the accuracy and efficiency of the authenticity verification. Then, the target verification result is displayed on the verification result page. Therefore, while reducing the time consumption of the authenticity verification, it is possible to quickly determine the authenticity of the answer content based on the target verification result, so as to mark the authenticity of the answer content and improve the efficiency of marking the authenticity of the answer content. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the specific embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 is a flowchart of a first data analysis method according to an embodiment of the present disclosure;

[0018] Figure 2 is a schematic diagram of a verification result page according to an embodiment of the present disclosure;

[0019] Figure 3is a schematic diagram of a first verification details page according to an embodiment of the present disclosure;

[0020] Figure 4 is a schematic diagram of a second first verification details page according to an embodiment of the present disclosure;

[0021] Figure 5 is a schematic diagram of a third first verification details page according to an embodiment of the present disclosure;

[0022] Figure 6 is a schematic diagram of a process for generating a target verification result according to an embodiment of the present disclosure;

[0023] Figure 7 is a flow chart of a second data analysis method according to an embodiment of the present disclosure;

[0024] Figure 8 is a schematic diagram of another verification result page according to an embodiment of the present disclosure;

[0025] Fig. 9 is a schematic diagram of a second verification details page according to an embodiment of the present disclosure;

[0026] Fig.10 is a flowchart of a third data analysis method according to an embodiment of the present disclosure;

[0027] Fig.11 is a schematic diagram of a data search page according to an embodiment of the present disclosure;

[0028] Fig.12 is a schematic diagram of a search result page according to an embodiment of the present disclosure;

[0029] Fig.13 is a schematic diagram of another search result page according to an embodiment of the present disclosure;

[0030] Fig.14 is a structural block diagram of a data analysis device according to an embodiment of the present disclosure;

[0031] Fig.15 is a structural block diagram of an electronic device according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0033] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0034] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0035] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0036] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0037] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0038] For business scenarios that use large models, in order to improve the authenticity of the answers given by the large models, annotators need to search each sentence in the answers given by the large models through the browser to verify the authenticity of the answers and give an authenticity score for the entire answer. Since the annotators need to search for each sentence in the answer, and different browsers will give different search results, the authenticity of each sentence needs to be confirmed after comparing multiple search results, which makes it time-consuming to verify the authenticity of the answers given by the large models.

[0039] In related technologies, annotators search for answers in two main ways:

[0040] The first method is for the annotator to search directly through the browser. However, this search method requires a second click to the page corresponding to each displayed item to read or understand the information in order to determine the authenticity of the answer. Moreover, the browser only displays the relevant links found in the search, making it difficult to understand the sentences in the answer and make authenticity judgments. In addition, if multiple sentences are searched in conjunction or the entire paragraph is searched, the browser's search engine will find it difficult to understand the content to be searched, and often will not be able to find the answer. Therefore, it is usually necessary to manually understand each sentence in the answer and convert it into a search question.

[0041] The second method is to search through artificial intelligence search websites. However, this search method only displays relevant links and cannot understand the sentences and make authenticity judgments. Moreover, if multiple sentences are searched in conjunction or the entire paragraph is searched, the search results will be inaccurate, especially when searching the entire paragraph, it may even be understood as a rewrite. In addition, if the marked sensitive data is searched through artificial intelligence search websites, there will also be a risk of data leakage.

[0042] In summary, the above two search methods for the answers given by the large model have low search accuracy and relatively long search time, which results in a long time spent on authenticity verification of the answers given by the large model.

[0043] In view of this, according to an embodiment of the present disclosure, a data analysis method embodiment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0044] In this embodiment, a data analysis method is provided, which can be used in a data annotation platform. Figure 1 is a flow chart of a first data analysis method according to an embodiment of the present disclosure, such as Figure 1 As shown, the process includes the following steps:

[0045] Step S101, displaying a question and answer annotation page, which displays a target question and an answer content of the question and answer model for the target question.

[0046] It can be understood that in the process of training or updating parameters of a question-answering model (such as a large model), it is usually necessary to input the target question to be asked into the question-answering model, and the question-answering model processes the target question and generates the answer content. The answer content of at least one target question can be displayed on the question-answering annotation page. Then, the user annotates the authenticity of the answer content, for example, annotates the overall answer content or annotates a single answer content. Among them, the annotation of the overall answer content includes at least one of the annotation items of whether the new rewrite is qualified, whether it will be hit, special labels, whether it is abandoned, whether it is a special sample, and the direction feature mark. The annotation for a single answer content includes the answer score, authenticity, whether the remarks are qualified, whether the speech is rejected, the text in the marked answer content, and at least one of the question types corresponding to the text. Then, the annotation results are obtained, and the question-answering model is trained or the parameters are updated according to the annotation results to improve the model performance of the question-answering model.

[0047] Step S102, in response to a trigger instruction for authenticity verification, obtain a target verification result of the answer content, wherein the target verification result is obtained by verifying the authenticity of the first sub-content split from the answer content and performing statistical analysis on the verified first sub-verification result.

[0048] Specifically, an auxiliary control is displayed on the question and answer annotation page. By interacting with the auxiliary control (e.g., clicking), a data search control and an authenticity verification control can be displayed. By interacting with the authenticity verification control, a trigger instruction for authenticity verification is generated. In response to the trigger instruction for authenticity verification, a target verification result of the answer content is obtained.

[0049] Specifically, users can select target questions for authenticity verification in advance on the data annotation platform. Then, the data annotation platform preloads the target questions and verifies the authenticity of the answers given by the question-answering model to the target questions in advance.

[0050] Specifically, the step of generating the target verification result includes: splitting the answer content to obtain at least one first sub-content. Generating a first search question based on the first sub-content. Then, performing a data search based on the first search question to obtain a corresponding first reference data source. Further, performing data analysis on the first reference data source to generate a first target answer to the first search question. Performing an authenticity verification on the first sub-content based on the first target answer to obtain a first sub-verification result. Finally, performing a statistical analysis on the result categories of the first sub-verification results of each first sub-content to obtain a target verification result.

[0051] Step S103, displaying the target verification result on the verification result page, and the target verification result is used to mark the authenticity of the answer content on the question and answer marking page.

[0052] Specifically, a drawer-like interactive form can be used to pop up a verification result page on the question-answer annotation page. Figure 2 As shown, a verification result page can be popped up in the target area of ​​the verification result page. Then, the target verification result is displayed on the verification result page. The user can determine the authenticity of the answer content through the target verification result displayed on the verification result page, and mark the authenticity of the answer content on the answer annotation page. Of course, if conditions permit, a data annotation model can also be used to mark the authenticity of the answer content based on the target verification result, and the annotation result can be displayed on the question and answer annotation page.

[0053] The data analysis method provided in this embodiment, since the target verification result is obtained by verifying the authenticity of the first sub-content split out of the answer content, and performing statistical analysis on the first sub-verification result obtained by the verification. Therefore, it is possible to divide the answer content into fine-grained parts, reduce redundant information in a single authenticity verification, and facilitate understanding and verification of the information represented by the first sub-content, thereby effectively improving the accuracy and efficiency of the authenticity verification. Then, the target verification result is displayed on the verification result page. Therefore, while reducing the time consumption of the authenticity verification, the authenticity of the answer content can be quickly determined based on the target verification result, so as to mark the authenticity of the answer content and improve the efficiency of marking the authenticity of the answer content.

[0054] In some optional implementations, the target verification result includes the result category of the authenticity verification corresponding to the first sub-content and the statistical times for different result categories. Then, displaying the target verification result on the verification result page in the above step S103 includes:

[0055] Step a1, display the answer content on the verification result page.

[0056] Specifically, Figure 2 As shown, reference Figure 2 As shown in the pre-verification content, in actual operation, the user can pre-select the answer content of multiple target questions for verification. Therefore, different display areas can be configured for different answer contents in the verification result page. For example, in the verification result page, answer content 1 "Giant pandas are pure vegetarians and mainly feed on bamboo. Red pandas are omnivorous animals. In addition to bamboo, they also eat fruits, insects and small vertebrates" and answer content 2 correspond to different display areas.

[0057] Step a2: Display the result category and the corresponding statistical times in the form of labels in the first display area corresponding to the answer content on the verification result page.

[0058] Specifically, the result categories of the authenticity verification include true and abnormal. By performing statistical analysis on the result categories corresponding to different first sub-contents of the answer content, the statistical times for different result categories can be obtained. Figure 2 As shown, in the display area corresponding to each answer content on the verification result page, different result categories and their corresponding statistical times can be displayed in the form of labels, so as to intuitively determine the authenticity verification status of different answer contents and quickly locate abnormal answer content.

[0059] For example, Figure 2 As shown, there is 1 true and 1 abnormal in answer content 1, that is, there is 1 true first sub-content and 1 abnormal first sub-content in answer content 1. There are 5 true and 0 abnormal in answer content 2, that is, there are 5 true first sub-contents and 0 abnormal first sub-contents in answer content 2.

[0060] Step a3: displaying a first viewing control in the first display area, where the first viewing control is used to view authenticity verification details corresponding to the target verification result.

[0061] It can be understood that the first viewing control is displayed in the display area corresponding to each answer content. Therefore, the authenticity verification details corresponding to the target verification result can be deeply understood through the first viewing control.

[0062] The data analysis method provided in this embodiment displays the result category and the corresponding statistical times in the form of a label in the first display area corresponding to the answer content on the verification result page. Therefore, it is possible to quickly locate the answer content with abnormalities according to the statistical times of different result categories, thereby improving the efficiency of authenticity verification. At the same time, displaying the first viewing control for viewing the authenticity verification details corresponding to the target verification result in the first display area can also facilitate in-depth understanding of the authenticity verification details of each answer content, thereby improving the credibility of the target verification result.

[0063] In some optional embodiments, the data analysis method of the present disclosure further includes:

[0064] Step b1, in response to a viewing instruction for a target verification result, obtaining a first sub-content and a first sub-verification result corresponding to the first sub-content.

[0065] Specifically, a viewing instruction for the target verification result may be generated in response to an interactive operation on the first viewing control. A viewing instruction for the target verification result may also be generated in response to an interactive operation on the first display area. Thus, in response to the viewing instruction for the target verification result, the first sub-content of the answer content and the first sub-verification result of the first sub-content are acquired.

[0066] It should be noted that the first sub-verification result is the authenticity verification result of the first sub-content.

[0067] Step b2: displaying the first sub-content and the corresponding first sub-verification result on the first verification details page.

[0068] Specifically, the first sub-content is obtained by splitting the answer content according to a preset identifier. Optionally, the preset identifier is a period. Alternatively, the first sub-content is obtained by splitting the answer content according to semantic information of the answer content.

[0069] It should be noted that, in actual operation, a single answer content can be split into at least one first sub-content, and in the first verification details page, different first sub-contents and corresponding first sub-verification results can be displayed in different display areas.

[0070] For example, Figure 3 As shown, assuming that the first sub-content obtained after splitting the answer content "Giant pandas are pure vegetarians, mainly feeding on bamboo. Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates" includes "Giant pandas are pure vegetarians, mainly feeding on bamboo" and "Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates". Then, "Giant pandas are pure vegetarians, mainly feeding on bamboo" and "Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates" can be displayed separately on the first verification details page. The corresponding first sub-verification result is displayed in the display area corresponding to the first sub-content on the first verification details page. For example, the first sub-verification result of "Giant pandas are pure vegetarians, mainly feeding on bamboo" is abnormal, and the first sub-verification result of "Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates" is true.

[0071] The data analysis method provided in this embodiment displays the first sub-content and the corresponding first sub-verification result on the first verification details page, so that it is easy to understand the authenticity of each first sub-content in the answer content, so as to improve the accuracy of authenticity verification.

[0072] In some optional embodiments, the first sub-content and the corresponding first sub-verification result are displayed on the first verification details page in the above step b2, including: if the first sub-verification result is abnormal, obtaining the first reason for determining that the first sub-verification result is abnormal, and displaying the first sub-content, as well as the first sub-verification result and the first reason corresponding to the first sub-content on the first verification details page.

[0073] Specifically, if the first sub-check result is abnormal, the first sub-check result and its corresponding first reason are displayed in the display area corresponding to the first sub-content on the first check details page. Figure 3 As shown, assuming that the first sub-verification result of "Giant pandas are pure vegetarians, mainly feeding on bamboo" is abnormal, then the display area corresponding to the first sub-content will display "Giant pandas are pure vegetarians, mainly feeding on bamboo" as the first reason for the abnormality. For example, the first reason why "Giant pandas are pure vegetarians, mainly feeding on bamboo" is abnormal is "Giant pandas are not vegetarians, but omnivorous animals."

[0074] In the data analysis method provided in this embodiment, if the first sub-verification result is abnormal, the first reason why the first sub-content is abnormal is displayed on the first verification details page. Therefore, it is easy to find out the reason why the first sub-content is abnormal.

[0075] In some optional implementations, if the first sub-verification result is true, the first sub-content and the first sub-verification result corresponding to the first sub-content are displayed on the first verification details page.

[0076] For example, Figure 3 As shown, assuming that the first sub-verification result of "Red pandas are omnivorous animals. In addition to bamboo, they also eat fruits, insects and small vertebrates" is true, then the display area corresponding to the first sub-content will display "Red pandas are omnivorous animals. In addition to bamboo, they also eat fruits, insects and small vertebrates" and its first sub-verification result "True".

[0077] In some optional implementations, the second display area corresponding to the first reason in the first verification details page further displays an index control, where the index control is generated based on the first reference data source of the first sub-verification result.

[0078] For example, Figure 4 As shown, an index control "①" is displayed behind the first reason "Giant pandas are not vegetarians, but omnivores." The user can click the index control to determine the first reference data source of the first sub-check result.

[0079] The data analysis method provided in this embodiment displays an index control generated based on the first reference data source of the first sub-check result in the first check details page. Therefore, it is easy to quickly locate the data source referenced by the abnormal first sub-check result, thereby improving data search efficiency.

[0080] In some optional embodiments, the data analysis method of the present disclosure further includes:

[0081] Step c1, in response to a viewing instruction for the first sub-verification result, obtain the first reference data source of the first sub-verification result, the first search question for searching the first reference data source, and the first target answer for the first search question, wherein the first search question is generated based on the first sub-content, and the first target answer is obtained by performing data analysis on the first reference data source.

[0082] It should be noted that in the actual operation steps, after obtaining the answer content of the question-answer model, the answer content is split to obtain the first sub-content. Then, a first search question is generated based on the first sub-content. And data search is performed according to the first search question to obtain the first reference data source. Then, data analysis is performed on the first reference data source to obtain the first target answer. The authenticity of the first sub-content is verified according to the first target answer to obtain the first sub-verification result. Finally, all the first sub-verification results are statistically analyzed to obtain the target verification result of the answer content.

[0083] It should be noted that when searching for data according to the first search question, at least one reference data source will be searched. Based on the correlation between the first search question and the at least one reference data source, a first preset number of reference data sources can be screened out as the first reference data source, starting from the reference data source with the highest correlation. Optionally, the first preset number is 5. In addition, the first preset number can also be dynamically adjusted to 3, 4, etc. according to actual conditions. Alternatively, based on the correlation between the first search question and the at least one reference data source, a preset proportion of reference data sources are screened out from the reference data sources to serve as the first reference data source.

[0084] Step c2: displaying the first search question, the first target answer and the first reference data source corresponding to the first sub-verification result on the first verification details page.

[0085] The data analysis method provided in this embodiment displays the first search question, the first target answer, and the first reference data source corresponding to the first sub-verification result on the first verification details page. Therefore, the authenticity of the first sub-verification result can be improved. At the same time, since the answer content is split into the first sub-content, the first search question is generated based on the first sub-content, and the first reference data source is obtained by searching the data based on the first search question. Therefore, compared with the full-text comprehension search in the related art, the data analysis method of the present disclosure can provide fine-grained understanding and search by splitting the answer content, and can more accurately search for the first reference data source used to verify the authenticity of the first sub-content, thereby effectively improving the accuracy of the first sub-verification result and the target verification result. In addition, the first target answer is obtained by performing data analysis on the first reference data source, and the first target answer is displayed on the first verification details page, which can facilitate a quick understanding of the data related to the first sub-content in the first reference data source and improve the readability of the search data.

[0086] It should be noted that at least one first search question may be generated for the first sub-content. Figure 4 As shown, assuming that the first search question for the first sub-content "Giant pandas are pure vegetarians and mainly feed on bamboo" is "What are the characteristics of the giant panda's diet" and "What is the main food of the giant panda", the first target answer for "What are the characteristics of the giant panda's diet" is "Giant pandas are omnivorous animals", and the first target answer for "What is the main food of the giant panda" is "The main food of the giant panda is bamboo". For the first search question belonging to the same first sub-content, it can be displayed in the display area corresponding to the first sub-content on the first verification details page. And the first target answer and the first reference data source corresponding to the first search question are displayed in the display area corresponding to the first search question on the first verification details page.

[0087] In some optional implementations, the first search question is generated based on a semantic rewriting result of the first sub-content.

[0088] Specifically, the step of generating the first search question includes: semantically rewriting the first sub-content to obtain a semantically rewritten result of the first sub-content, and then generating at least one first search question based on the semantically rewritten result of the first sub-content.

[0089] The data analysis method provided in this embodiment generates a first search question based on the semantic rewriting result of the first sub-content. Therefore, it is possible to filter out redundant information in the first sub-content and convert the first sub-content into semantic information that can be understood by the data annotation platform, thereby effectively improving the accuracy of data search and further improving the accuracy of authenticity verification.

[0090] In some optional implementations, the data analysis method of the present disclosure further includes: in response to a rewrite view instruction for the first sub-content, obtaining a semantic rewriting result of the first sub-content; and displaying the semantic rewriting result on the first verification details page.

[0091] Specifically, Figure 4 As shown, in the display area corresponding to the first sub-content on the first verification details page, a rewriting control is displayed. In response to the interactive operation of the rewriting control, a rewriting viewing instruction of the first sub-content is generated to display the semantic rewriting result corresponding to the first sub-content.

[0092] For example, Figure 5 As shown in the figure, the semantic rewriting result of the first sub-content "Giant pandas are pure vegetarians, mainly eating bamboo" is "Giant pandas mainly eat bamboo and are vegetarians."

[0093] The data analysis method provided in this embodiment improves the accuracy of data search through semantic rewriting, and at the same time displays the semantic rewriting result corresponding to the first sub-content on the first verification details page, which can facilitate increasing the credibility of the first search question and the corresponding first reference data source.

[0094] It should be noted that after the semantic rewriting result is displayed on the first verification details page, a first retracting control is displayed in the display area corresponding to the semantic rewriting result on the first verification details page. In response to the interactive operation of the first retracting control, the semantic rewriting result is not displayed on the first verification details page.

[0095] As an example, in order to facilitate understanding of the target verification result generation process, the target verification result generation process is described below with a specific example. Figure 6 As shown, the process includes:

[0096] Process 1: Get the answer content of the question-answering model.

[0097] Process 2: split the answer content to obtain at least one first sub-content, and generate a first search question based on the semantic rewriting result of the at least one first sub-content.

[0098] For example Figure 6 As shown, the answer content is split, and based on the semantic rewriting result of the split first sub-content, the first search question 1, the first search question 2, ..., the first search question N are obtained, where N is the number of the first search questions.

[0099] Process 3: Perform data search based on the first search question to obtain at least one reference data source for the first search question.

[0100] For example Figure 6As shown, for the first search question 1, reference data sources 11, reference data sources 12, ..., and reference data source M are searched. Where M is the number of reference data sources corresponding to the first search question 1. For the first search question 2, reference data sources 21, reference data sources 22, ..., and reference data source P are searched. Where P is the number of reference data sources corresponding to the first search question 2.

[0101] Process 4, for each first search question, the searched reference data sources are ranked by relevance, and a first preset number of reference data sources with the highest credibility are screened out as the first reference data sources.

[0102] For example Figure 6 As shown, for the reference data sources of the first search question 1, the first reference data source 11, the first reference data source 12, the first reference data source 13, the first reference data source 14, and the first reference data source 15 are screened out. For the reference data sources of the first search question 2, the first reference data source 21, the first reference data source 22, the first reference data source 23, the first reference data source 24, and the first reference data source 25 are screened out.

[0103] Process 5, for the first search question, performing data analysis based on the corresponding first reference data sources to obtain a first target answer to the first search question.

[0104] For example Figure 6 As shown, for the first reference data source of the first search question 1, the first target answer 1 is generated. For the first reference data source of the first search question 2, the first target answer 2 is generated. Similarly, for the first reference data source of the first search question N, the first target answer N is generated.

[0105] Process 6, statistical analysis is performed on all first target answers to obtain target verification results.

[0106] This embodiment provides a second data analysis method that can be used in a data annotation platform. Figure 7 is a flow chart of a second data analysis method according to an embodiment of the present disclosure, such as Figure 7 As shown, the process includes the following steps:

[0107] Step S701, displaying a question and answer annotation page, the question and answer annotation page displays the target question and the answer content of the question and answer model for the target question. See the above step S101 for details, which will not be repeated here.

[0108] Step S702, in response to the authenticity verification trigger instruction, obtain the target verification result of the answer content, the target verification result is obtained by verifying the authenticity of the first sub-content split from the answer content, and performing statistical analysis on the first sub-verification result obtained by the verification. See the above step S102 for details, which will not be repeated here.

[0109] Step S703: Display the target verification result on the verification result page. The target verification result is used to mark the authenticity of the answer content on the question and answer marking page. See the above step S103 for details, which will not be repeated here.

[0110] Step S704, in response to the input operation of the target verification content in the search control of the verification result page, generates a detection instruction, the detection instruction is used to obtain the detection result, the detection result includes the second sub-content split out of the target verification content, the second reference data source corresponding to the second sub-content and the second sub-verification result, the second reference data source is obtained based on the data search for the second sub-content, and the second sub-verification result is obtained based on the authenticity verification of the second sub-content based on the second reference data source.

[0111] Specifically, Figure 8 As shown, a search control for real-time search is displayed on the verification result page, and a detection instruction is generated in response to an input operation of the search control. In response to the detection instruction, the target verification result is split to obtain at least one second sub-content. The second sub-content is semantically rewritten to obtain the semantic rewriting result of the second sub-content. A second search question is generated based on the semantic rewriting result of the second sub-content. Data search is performed based on the second search question to obtain a second reference data source. Data analysis is performed on the second reference data source to obtain a third target answer. The second sub-content is verified for authenticity based on the third target answer to obtain a second sub-verification result.

[0112] Step S705 : Displaying the second sub-content, the second sub-verification result and the second reference data source corresponding to the target verification content on the second verification details page.

[0113] It should be noted that the target detection content is also displayed on the second verification details page.

[0114] The data analysis method provided in this embodiment displays a search control on the verification result page to provide real-time verification services. The second sub-content obtained by splitting the target verification content, the second sub-verification result corresponding to the second sub-content, and the second reference data source are displayed on the second verification details page. Therefore, different authenticity verification requirements and data search requirements in the authenticity verification process can be met.

[0115] For example, Figure 8As shown in the figure, enter "Giant pandas are pure vegetarians and mainly feed on bamboo. Red pandas are omnivorous animals. In addition to bamboo, they also eat fruits, insects and small vertebrates." in the search control on the verification result page. Then a detection instruction will be generated and the detection result will be obtained. Fig. 9 The second verification details page shown shows that "Giant pandas are pure vegetarians, mainly feeding on bamboo. Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates." and the corresponding second sub-content "Giant pandas are pure vegetarians, mainly feeding on bamboo" and "Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates." At the same time, in the display area corresponding to the second sub-content on the second verification details page, the second sub-verification result and the second reference source corresponding to the second sub-verification result are displayed. For example, the second sub-verification result of "Giant pandas are pure vegetarians, mainly feeding on bamboo" is abnormal. The second sub-verification result of "Red pandas are omnivores, and in addition to bamboo, they also eat fruits, insects and small vertebrates" is "true".

[0116] It should be noted that the data analysis method and display method for the target verification content, the second sub-content, the second sub-verification result and the second reference data source are respectively similar to the data analysis method and display method for the above-mentioned answer content, the first sub-content, the first sub-verification result and the first reference data source. You may refer to the data analysis method and display method for the above-mentioned answer content, the first sub-content, the first sub-verification result and the first reference data source, and they will not be repeated here.

[0117] This embodiment provides a third data analysis method that can be used in a data annotation platform. Fig.10 is a flow chart of a third data analysis method according to an embodiment of the present disclosure, such as Fig.10 As shown, the process includes the following steps:

[0118] Step S1001, displaying a question and answer annotation page, which displays the target question and the answer content of the question and answer model for the target question. See the above step S101 for details, which will not be repeated here.

[0119] Step S1002, in response to the trigger instruction of authenticity verification, obtain the target verification result of the answer content, the target verification result is obtained by verifying the authenticity of the first sub-content split from the answer content, and performing statistical analysis on the first sub-verification result obtained by verification. See the above step S102 for details, which will not be repeated here.

[0120] Step S1003, displaying the target verification result on the verification result page, the target verification result is used to mark the authenticity of the answer content on the question and answer marking page. The verification result page also displays a data search control.

[0121] Step S1004: in response to a trigger instruction for a data search control, display a data search page.

[0122] Specifically, Fig.11 As shown, in response to a trigger instruction for a data search control, a data search page may be displayed. The data search page displays at least one of a search control and a search question recommended based on the answer content. In addition, a history record of a historical search, such as history record 1, is also displayed to facilitate tracing back the content of the historical search.

[0123] Step S1005, in response to the target search content determined on the data search page, generates a search instruction, the search instruction is used to obtain search results, the search results include a third reference data source obtained based on the target search content, a content summary of the third reference data source, and a second target answer for the target search content, the second target answer is obtained based on data analysis of the third reference data source.

[0124] Specifically, in response to an input operation of the target search content in the search control of the data search page, the target search content is determined. Alternatively, in response to a selection operation of a second search question on the data search page, the target search content is determined. Further, a search instruction is generated according to the target search content to obtain search results.

[0125] It should be noted that the third reference data source is obtained by performing a data search based on the target search content, and the content summary is obtained by performing a data analysis on the third reference data source.

[0126] It should be noted that, in actual operation, a full network search may be performed based on the target search content to obtain reference data sources related to the target search content. Then, according to the relevance of the searched reference data sources to the target search content, a second preset number of reference data sources may be screened out from the searched reference data sources to serve as the third reference data source. Optionally, the second preset number is 10. For example, from the searched reference data sources, the 10 reference data sources with the highest credibility are screened out as the third reference data source.

[0127] Specifically, a language model may be used to perform data analysis on the third reference data source to obtain a content summary of the third reference data and a second target answer to the target search content.

[0128] Step S1006, displaying the target search content, the third reference data source, the content summary and the second target answer on the search result page.

[0129] For example, Fig.12 As shown, from top to bottom, the third reference data source, the content summary of the third reference data source, and the second target answer for the target search content (e.g., "How to use Java to write a chess game") may be displayed in sequence. At the same time, in order to facilitate users to distinguish the third reference data source, the content summary, and the second target answer, a dividing line or prompt words may be used to assist users in determining the third reference data source, the content summary, and the second target answer. For example, the display area corresponding to the prompt word "full text abstract" displays the content summary, and the display area corresponding to the prompt word "answer" displays the second target answer.

[0130] The data analysis method provided in this embodiment displays a data search control on the verification result page, and displays a data search page in response to a trigger instruction for the data search control. Therefore, different data search needs can be met. In addition, in response to the target search content determined on the data search page, a search instruction is generated to obtain a third reference data source, a content summary of the third reference data source, and a second target answer for the target search content. Compared with the search based on a search engine in the related art, the data analysis method disclosed in the present invention can more efficiently understand the general content of each third reference data source, as well as the answer to the target search content through the content summary and the second target answer. Therefore, it can effectively reduce the time for clicking to view and understand the content of each third reference data source, and improve the efficiency of data search and data understanding.

[0131] In some optional implementations, the data search page displays a second search question recommended based on the answer content. Then, in response to the target search content determined on the data search page, generating a search instruction in step S1005 includes: in response to a selection instruction for the second search question, determining the second search question as the target search content, and generating a search instruction.

[0132] Specifically, data analysis is performed on the answer content based on the language model to obtain the second search question.

[0133] For example, Fig.11 As shown, the data search page displays the second search questions "How to write a chess game using Java?" and "Is there any sample code for a chess game written in Java?". The second search question as the target search content can be selected from the displayed second search questions to generate a search instruction.

[0134] The data analysis method provided in this embodiment displays a second search question recommended based on the answer content on the data search page. Therefore, it is possible to provide multiple search questions based on the answer content. Then, in response to the selection instruction for the second search question, the second search question is determined as the target search content, and a search instruction is generated. Therefore, the data search efficiency can be further improved through the recommended second search question.

[0135] In some optional implementations, the third display area of ​​the search result page corresponding to the third reference data source also displays a second viewing control, and the second viewing control is used to view fragment content related to the target search content in the third reference data source.

[0136] For example, reference Fig.12 The view control corresponding to the prompt word "original text related fragment" (i.e., the second view control) is provided. The user can view the fragment content related to the target search content in the third reference data source by clicking on the view control.

[0137] In the data analysis method provided in this embodiment, the search result page also displays a section for viewing the fragment content related to the target search content in the third reference data source. Therefore, the fragment content related to the target search content in the third reference data source can be displayed through the second viewing control, which is convenient for quickly understanding the key information in the third reference data source, so as to quickly compare the content of different third reference data sources.

[0138] In some optional embodiments, the data analysis method of the present disclosure further includes:

[0139] Step d1: In response to an interactive instruction of the second viewing control, a segment label is displayed in the third display area.

[0140] Specifically, based on the relevance between the fragment contents in the third reference data source and the target search content, a third preset number of fragment contents with the highest relevance can be screened out. Fragment labels can be generated based on the screened fragment contents. The fragment labels can also be sorted based on the relevance of the screened fragment contents. For example Fig.13 As shown, there are six fragment labels, namely, fragment 1, fragment 2, fragment 3, fragment 4, fragment 5, and fragment 6. The higher the fragment label is ranked, the higher the relevance of the fragment content corresponding to the fragment label to the target search content is.

[0141] Step d2: in response to a selection instruction for a segment tag, display the segment content corresponding to the segment tag.

[0142] For example, Fig.13As shown, assuming that the user needs to view the content of segment 1, the user can click the segment label of segment 1, so that the content of segment 1 is displayed in the area below the segment label.

[0143] The data analysis method provided in this embodiment displays a segment label in the third display area in response to an interactive instruction of the second viewing control. In response to a selection instruction of the segment label, the segment content corresponding to the segment label is displayed. Therefore, by selecting the segment label, different segment contents in the third reference data source can be viewed to quickly understand the relevant content of the third reference data source, thereby improving data search efficiency and data understanding efficiency.

[0144] It should be noted that when the search result page displays the fragment label or fragment content, a collapse control is displayed in the third display area. In response to the interactive operation of the collapse control in the third display area, the fragment label and fragment content are not displayed in the third display area.

[0145] As a specific application example, a target application is installed on the data annotation platform, and the target application is configured with a language model for data analysis. In the target application, the answer content of the question-and-answer model can be verified for authenticity, and the authenticity of the answer content can be annotated. The annotator can select the annotation data through the target application, and the annotation data includes the target question and the answer content of the question-and-answer model for the target question. Then, the target application uses the data analysis method of the present disclosure to verify the authenticity of the answer content, and displays the target verification result on the verification result page. The standard staff annotates the authenticity of the answer content according to the content displayed on the verification result page.

[0146] It is worth noting that, compared to the full-text comprehension-style retrieval in the related art, the data analysis method disclosed in the present invention can achieve a more fine-grained understanding and search through a splitting method, and can more accurately give the authenticity results of the search content (such as answer content, target verification content, target search content, etc.). Not only that, compared to the search interaction that only gives relevant links, the data can be analyzed through a language model, and a search question understood by the language model can be given. Answers are generated based on the content of the reference data source searched for the search question, which can effectively improve the user's understanding of the search content and answers, and help users more quickly clarify the authenticity of the search content. Therefore, after adopting the data analysis method disclosed in the present invention, the search engine and the understanding ability of the language model can be used to verify the authenticity of the content of the input text. Users only need to search with one click to verify the authenticity of a sentence or an entire paragraph of text.

[0147] In this embodiment, a data analysis device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can implement a combination of software and / or hardware of a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0148] This embodiment provides a data analysis device, such as Fig.14 As shown, including:

[0149] The annotation page display module 1401 is used to display a question and answer annotation page, where the question and answer annotation page displays the target question and the answer content of the question and answer model to the target question;

[0150] The verification result acquisition module 1402 is used to obtain a target verification result of the answer content in response to the authenticity verification trigger instruction, where the target verification result is obtained by performing authenticity verification on the first sub-content split from the answer content and performing statistical analysis on the first sub-verification result obtained by the verification;

[0151] The verification result display module 1403 is used to display the target verification result on the verification result page. The target verification result is used to mark the authenticity of the answer content on the question and answer marking page.

[0152] In some optional implementations, the target verification result includes the result category of the authenticity verification corresponding to the first sub-content and the statistical times for different result categories. Then, the verification result display module 1403 includes:

[0153] A first display unit, used to display the answer content on the verification result page;

[0154] A second display unit, used to display the result category and the corresponding statistical number in the form of a label in the first display area corresponding to the answer content on the verification result page;

[0155] The third display unit is used to display a first viewing control in the first display area, where the first viewing control is used to view authenticity verification details corresponding to the target verification result.

[0156] In some optional embodiments, the data analysis device of the present disclosure further includes:

[0157] A first detail viewing module, configured to obtain a first sub-content and a first sub-verification result corresponding to the first sub-content in response to a viewing instruction for a target verification result;

[0158] The first detail display module is used to display the first sub-content and the corresponding first sub-verification result on the first verification detail page.

[0159] In some optional implementations, the first details display module includes:

[0160] The fourth display unit is used to obtain the first reason for determining that the first sub-verification result is abnormal if the first sub-verification result is abnormal, and display the first sub-content, as well as the first sub-verification result and the first reason corresponding to the first sub-content on the first verification details page.

[0161] In some optional implementations, the second display area corresponding to the first reason in the first verification details page further displays an index control, where the index control is generated based on the first reference data source of the first sub-verification result.

[0162] In some optional embodiments, the data analysis device of the present disclosure further includes:

[0163] A second detail viewing module is used to obtain, in response to a viewing instruction for the first sub-verification result, a first reference data source of the first sub-verification result, a first search question for searching the first reference data source, and a first target answer for the first search question, wherein the first search question is generated based on the first sub-content, and the first target answer is obtained by performing data analysis on the first reference data source;

[0164] The second detail display module is used to display the first search question, the first target answer and the first reference data source corresponding to the first sub-verification result on the first verification detail page.

[0165] In some optional implementations, the first search question is generated based on a semantic rewriting result of the first sub-content.

[0166] In some optional embodiments, the data analysis device of the present disclosure further includes:

[0167] A rewriting result viewing module, configured to obtain a semantic rewriting result of the first sub-content in response to a rewriting viewing instruction for the first sub-content;

[0168] The rewriting result display module is used to display the semantic rewriting result on the first verification details page.

[0169] In some optional embodiments, the data analysis device of the present disclosure further includes:

[0170] A data verification module, for generating a detection instruction in response to an input operation of a target verification content in a search control on a verification result page, the detection instruction being used to obtain a detection result, the detection result including a second sub-content split out of the target verification content, a second reference data source corresponding to the second sub-content, and a second sub-verification result, the second reference data source being obtained by performing a data search based on the second sub-content, and the second sub-verification result being obtained by performing an authenticity verification on the second sub-content based on the second reference data source;

[0171] The verification information display module is used to display the second sub-content, the second sub-verification result and the second reference data source corresponding to the target verification content on the second verification details page.

[0172] In some optional embodiments, the data analysis device of the present disclosure further includes:

[0173] A search response module, used to display a data search page in response to a trigger instruction for a data search control;

[0174] A data search module, for generating a search instruction in response to a target search content determined on a data search page, the search instruction being used to obtain search results, the search results including a third reference data source obtained by searching based on the target search content, a content summary of the third reference data source, and a second target answer for the target search content, the second target answer being obtained by performing data analysis based on the third reference data source;

[0175] The search information display module is used to display the target search content, the third reference data source, the content summary and the second target answer on the search result page.

[0176] In some optional implementations, the third display area of ​​the search result page corresponding to the third reference data source also displays a second viewing control, and the second viewing control is used to view fragment content related to the target search content in the third reference data source.

[0177] In some optional embodiments, the data analysis device of the present disclosure further includes:

[0178] a segment viewing module, configured to display a segment label in a third display area in response to an interactive instruction of the second viewing control;

[0179] The segment display module is used to display the segment content corresponding to the segment tag in response to a selection instruction of the segment tag.

[0180] In some optional implementations, the data search page displays a second search question based on the answer content recommendation. The data search module includes:

[0181] The question selection unit is used to determine the second search question as the target search content and generate a search instruction in response to a selection instruction for the second search question.

[0182] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0183] The data analysis device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0184] The present disclosure also provides an electronic device having the above Fig.14 The data analysis device shown.

[0185] See also Fig.15 , Fig.15 is a structural block diagram of an electronic device provided by an optional embodiment of the present disclosure, such as Fig.15 As shown, the electronic device includes: one or more processors 1501, memory 1502, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process instructions executed in the electronic device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple electronic devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Fig.15 A processor 1501 is taken as an example.

[0186] The processor 1501 may be a central processing unit, a network processor or a combination thereof. The processor 1501 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable logic gate array, a general purpose array logic or any combination thereof.

[0187] The memory 1502 stores instructions executable by at least one processor 1501 so as to enable at least one processor 1501 to implement the method shown in the above embodiment.

[0188] The memory 1502 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 1502 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 1502 may optionally include a memory remotely arranged relative to the processor 1501, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0189] The memory 1502 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 1502 may also include a combination of the above types of memory.

[0190] The electronic device also includes an input device 1503 and an output device 1504. The processor 1501, the memory 1502, the input device 1503 and the output device 1504 may be connected via a bus or other means. Fig.15 The example of connecting through bus is taken in the following.

[0191] The input device 1503 can receive input digital or character information, and generate key signal input related to the user settings and function control of the electronic device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator rod, one or more mouse buttons, a trackball, a joystick, etc. The output device 1504 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0192] The embodiments of the present disclosure also provide a computer-readable storage medium. The above-mentioned method according to the embodiments of the present disclosure can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium and downloaded through a network, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0193] A part of the present disclosure may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present disclosure through the operation of the computer. Those skilled in the art should understand that the existence of computer program instructions in computer-readable media includes, but is not limited to, source files, executable files, installation package files, etc., and accordingly, the way in which computer program instructions are executed by a computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to the computer.

[0194] Although the embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A data analysis method, characterized in that: The method comprises: Displaying a question-and-answer annotation page, wherein the question-and-answer annotation page displays a target question and an answer content of the question-and-answer model to the target question; In response to a trigger instruction for authenticity verification, obtaining a target verification result of the answer content, the target verification result being obtained by performing authenticity verification on a first sub-content split from the answer content and performing statistical analysis on the verified first sub-verification result; The target verification result is displayed on the verification result page, and the target verification result is used to mark the authenticity of the answer content on the question and answer marking page.

2. The data analysis method according to claim 1, characterized in that The target verification result includes the result category of the authenticity verification corresponding to the first sub-content and the statistical number of times for different result categories; The displaying of the target verification result on the verification result page includes: Displaying the answer content on the verification result page; In the first display area of ​​the verification result page corresponding to the answer content, the result category and the corresponding statistical number are displayed in the form of a label; A first viewing control is displayed in the first display area, where the first viewing control is used to view authenticity verification details corresponding to the target verification result.

3. The data analysis method according to claim 1, wherein: The method further comprises: In response to a viewing instruction for the target verification result, acquiring the first sub-content and a first sub-verification result corresponding to the first sub-content; The first sub-content and the corresponding first sub-verification result are displayed on a first verification details page.

4. The data analysis method according to claim 3, characterized in that: The displaying of the first sub-content and the corresponding first sub-verification result on the first verification details page includes: If the first sub-verification result is abnormal, a first reason for determining that the first sub-verification result is abnormal is obtained, and the first sub-content, as well as the first sub-verification result and the first reason corresponding to the first sub-content are displayed on the first verification details page.

5. The data analysis method according to claim 4, characterized in that: An index control is further displayed in a second display area corresponding to the first reason in the first verification details page. The index control is generated based on a first reference data source of the first sub-verification result.

6. The data analysis method according to claim 3, characterized in that: The method further comprises: In response to a viewing instruction for the first sub-verification result, obtaining a first reference data source for the first sub-verification result, a first search question for searching the first reference data source, and a first target answer for the first search question, wherein the first search question is generated based on the first sub-content, and the first target answer is obtained by analyzing data of the first reference data source; The first search question, the first target answer, and the first reference data source corresponding to the first sub-verification result are displayed on the first verification details page.

7. The data analysis method according to claim 6, characterized in that: The first search question is generated based on a semantic rewriting result of the first sub-content.

8. The data analysis method according to claim 7, characterized in that: The method further comprises: In response to a rewriting viewing instruction for the first sub-content, obtaining a semantic rewriting result of the first sub-content; The semantic rewriting result is displayed on the first verification details page.

9. The data analysis method according to claim 1, wherein: The method further comprises: In response to an input operation of target verification content in a search control on the verification result page, generating a detection instruction, the detection instruction being used to obtain a detection result, the detection result including a second sub-content split from the target verification content, a second reference data source corresponding to the second sub-content, and a second sub-verification result, the second reference data source being obtained based on a data search for the second sub-content, and the second sub-verification result being obtained based on an authenticity verification of the second sub-content performed by the second reference data source; The second sub-content, the second sub-verification result, and the second reference data source corresponding to the target verification content are displayed on the second verification details page.

10. The data analysis method according to claim 1, wherein: The verification result page also displays a data search control, and the method further includes: In response to a trigger instruction for the data search control, displaying a data search page; generating a search instruction in response to target search content determined on the data search page, the search instruction being used to obtain search results, the search results including a third reference data source obtained by searching based on the target search content, a content summary of the third reference data source, and a second target answer to the target search content, the second target answer being obtained based on data analysis of the third reference data source; The target search content, the third reference data source, the content summary, and the second target answer are displayed on a search result page.

11. The data analysis method according to claim 10, characterized in that: The search result page further displays a second viewing control in a third display area corresponding to the third reference data source. The second viewing control is used to view fragment content related to the target search content in the third reference data source.

12. The data analysis method according to claim 11, characterized in that: The method further comprises: In response to an interaction instruction of the second viewing control, displaying a segment label in the third display area; In response to a selection instruction for the segment tag, the segment content corresponding to the segment tag is displayed.

13. The data analysis method according to claim 10, characterized in that: The data search page displays a second search question recommended based on the answer content; and generating a search instruction in response to the target search content determined on the data search page includes: In response to a selection instruction for the second search question, the second search question is determined as the target search content, and the search instruction is generated.

14. A data analysis device, characterized in that: The device comprises: A marking page display module is used to display a question and answer marking page, wherein the question and answer marking page displays a target question and an answer content of the question and answer model to the target question; a verification result acquisition module, configured to obtain, in response to a trigger instruction for authenticity verification, a target verification result of the answer content, the target verification result being obtained by performing authenticity verification on a first sub-content split from the answer content and performing statistical analysis on the verified first sub-verification result; The verification result display module is used to display the target verification result on the verification result page, and the target verification result is used to mark the authenticity of the answer content on the question and answer marking page.

15. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the data analysis method according to any one of claims 1 to 13 by executing the computer instructions.

16. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the data analysis method according to any one of claims 1 to 13.

17. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the data analysis method according to any one of claims 1 to 13.

Images