Software development target analysis device, method, and storage medium

Abstract

The application provides an analysis device, method and storage medium for software development target. The analysis device for software development target comprises a data acquisition module and an analysis module. The data acquisition module is used for acquiring multi-source data about software development target of a user. The analysis module comprises a value point extraction unit, a fact point extraction unit, a correlation analysis unit and a purpose induction unit. The value point extraction unit is configured to extract at least one value point pursued by the user from the multi-source data. The fact point extraction unit is configured to extract a fact point of at least one fact from the multi-source data. The correlation analysis unit is configured to perform semantic analysis on each value point and each fact point to determine the correlation of each fact point to each value point. The purpose induction unit is configured to determine a key value point with a total weight greater than a preset first threshold according to the correlation of each fact point to each value point; and determine the software development target according to the key value point and the corresponding related fact point.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer technology, and in particular to an analysis device for software development target, an analysis method for software development target, and a computer readable storage medium. BACKGROUND

[0002] In the modern software development life cycle, customer demand analysis is the core cornerstone of determining the success of the project. Accurate understanding and transformation of the real and deep purpose of the customer is the key to ensuring the final value of the software product. However, the current technical means used in the industry have significant limitations in dealing with software customer demand.

[0003] The traditional manual analysis method relies on product managers or demand analysts to collect information through conference interviews, questionnaire surveys and other methods. Its effectiveness is severely restricted by the personal experience, communication skills and subjective judgment of the analysis personnel. Not only is it inefficient and costly, but it is also extremely easy to cause demand distortion due to information omission and understanding bias, which poses a huge risk to the project. Project and task management tools such as JIRA, Trello can only manage explicit demand items and cannot automatically analyze, mine and refine deep purposes from raw and ambiguous customer communication information. In addition, when applying primary natural language processing technology to analyze customer demand, since keyword extraction and word frequency statistics cannot understand the semantic relationship and context between words, surface phenomena are often misjudged as core demands. Moreover, the sentiment analysis of primary natural language processing technology can only determine the sentiment polarity of the text, and it is difficult to reveal the specific reasons and customer's potential goals behind the sentiment. General chat robots or question and answer systems are limited by preset rule libraries and are difficult to effectively induce and reason complex deep business purposes. In addition, although existing CRM systems are good at recording customer information and interaction history, their analysis capabilities are limited to macro statistics and lack a deep mining mechanism for the "why" behind the customer's specific product function demand.

[0004] In order to overcome the above-mentioned defects of the prior art, there is an urgent need in the field for an improved analysis device for software development target, which can be used to deeply mine the development target in the software development process according to the user demand, thereby improving the efficiency and automation level in the process of analyzing the development target, enhancing the objectivity and accuracy of the analysis result, and further reducing the risk of the development project. SUMMARY

[0005] The following presents a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements of all aspects nor delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that is presented later.

[0006] In order to overcome the above-mentioned defects existing in the prior art, the application provides an analysis device of a software development target, an analysis method of a software development target and a computer readable storage medium, which can extract value points and fact points in multi-source data of a user about a software development target, establish an association relationship between the value points and the fact points, and mine development targets in a software development process according to user requirements, so as to improve the efficiency and automation level in the process of analyzing development targets, enhance the objectivity and accuracy of analysis results, and further reduce the risk of a development project.

[0007] Specifically, the analysis device of a software development target according to the first aspect of the application comprises a data acquisition module and an analysis module. The data acquisition module is configured to acquire multi-source data of a user about a software development target. The analysis module comprises a value point extraction unit, a fact point extraction unit, an association analysis unit and a purpose induction unit. The value point extraction unit is configured to extract at least one value point pursued by the user from the multi-source data. The fact point extraction unit is configured to extract at least one fact point of at least one fact from the multi-source data. The association analysis unit is configured to perform semantic analysis on each of the extracted value points and each of the extracted fact points to determine the association of each of the fact points to each of the value points. The purpose induction unit is configured to determine a key value point whose total weight is greater than a preset first threshold according to the association of each of the fact points to each of the value points, and determine the software development target according to the key value point and the corresponding related fact points.

[0008] Further, in some embodiments of the application, the multi-source data is selected from at least one of project management data, online documents, emails and conference recordings. A business data acquisition unit is configured in the data acquisition module and connected to a project management system, an online document sharing system, a mailbox server and a conference recording transcription system of the user's company to acquire the multi-source data of the user about the software development target.

[0009] Further, in some embodiments of the application, a preprocessing unit is further configured in the data acquisition module and configured to perform a deduplication process on the acquired multi-source data, remove HTML tags from the data after the deduplication process, perform uniform encoding processing on the data after the removal of the HTML tags, and divide the data after the uniform encoding processing into a plurality of semantically complete paragraphs.

[0010] Further, in some embodiments of the present application, the value point extraction unit is configured with a pre-trained fine-tuned large language model, which is configured to: according to a preset prompt engineering, ask each of the multi-source data one by one to extract at least one value point pursued by the user.

[0011] Further, in some embodiments of the present application, the data acquisition module is configured with a user data acquisition unit for acquiring feature data of a plurality of users for user classification. The value point extraction unit is further configured to: according to a preset prompt engineering, ask each of the multi-source data of each type of user one by one to extract at least one value point pursued by the corresponding type of user.

[0012] Further, in some embodiments of the present application, the fact point extraction unit is configured with a pre-trained named entity recognition model, which is configured to: according to a plurality of pre-set dimensions of pre-named entities, analyze each of the multi-source data one by one, and form at least one fact point of a fact according to the identified named entities. The plurality of pre-set dimensions are selected from at least one of a person name, an organization name, a post name, a product name, an action name, a degree name, and a time length.

[0013] Further, in some embodiments of the present application, the fact point extraction unit is further configured to: determine a reliability weight according to the data source and / or explicitness of the fact point; determine an importance weight according to the number of occurrences of the fact point in the multi-source data; determine a fact point weight of the corresponding fact point according to the reliability weight and the importance weight; and retain valid fact points whose fact point weights reach a preset second threshold, and filter out invalid fact points whose fact point weights are less than the second threshold.

[0014] Further, in some embodiments of the present application, the fact point extraction unit is further configured with a pre-trained explicitness classification model, which is configured to: in response to forming the fact point of at least one fact, perform semantic analysis on the fact point via the explicitness classification model to determine the corresponding explicitness classification.

[0015] Further, in some embodiments of the present application, the association analysis unit is configured with a pre-trained association classification model. The step of performing semantic analysis on each of the extracted value points and each of the extracted fact points to determine the association of each of the fact points to each of the value points comprises: combining each of the value points with each of the fact points into a high-dimensional vector to calculate the correlation degree of the two; inputting the high-dimensional vector with a correlation degree reaching a preset third threshold into the association classification model to determine the association label of the fact vector to the corresponding value vector. A positive association label indicates that the fact vector supports the corresponding value vector. A negative association label indicates that the fact vector opposes the corresponding value vector.

[0016] Further, in some embodiments of the present application, the step of determining the key value point with a total weight greater than a preset first threshold according to the association of each of the fact points to each of the value points comprises: for each of the value points, marking the fact point weight of the fact point corresponding to each positive association label of the value point as positive and calculating the cumulative value as the total weight of the value point; and determining the value point with a total weight greater than the first threshold as the key value point.

[0017] Further, in some embodiments of the present application, the step of determining the software development goal according to the key value point and the corresponding relevant fact point comprises: filling the value vector V1 of the key value point and the fact vector F1 of the corresponding relevant fact point into a preset software development goal template respectively to output the software development goal.

[0018] Further, in some embodiments of the present application, the purpose induction unit is further configured to: for each of the value points, mark the fact point weight of the fact point corresponding to each negative association label of the value point as negative and calculate the cumulative value as the total weight of the value point; and determine the value point with a total weight less than the fourth threshold as the effective value point.

[0019] Further, in some embodiments of the present application, the step of determining the software development goal according to the key value point and the corresponding relevant fact point comprises: filling the value vector V2 of the value point which is both the key value point and the effective value point and the fact vector F2 of the corresponding relevant fact point into a preset software development goal template respectively to output the software development goal.

[0020] Further, in some embodiments of the present application, the analysis device further comprises a display module configured to provide a display interface with at least one tab page to display the software development goal and its corresponding value points, fact points, supporting relationships and / or opposing relationships. The supporting relationships and / or the opposing relationships are displayed in the form of a relationship graph, and the weight between the value point and the first fact point supporting it and / or the second fact point opposing it is shown.

[0021] Further, in some embodiments of the present application, the display module further supports an editing function and is configured to add or delete value points and / or fact points in the display interface and / or adjust the relationship between at least one value point and the corresponding fact point according to the user's editing operation on the relationship graph.

[0022] Further, in some embodiments of the present application, the analysis device further comprises a modification assistance module connected to the display module and the purpose induction unit of the analysis module, and configured to obtain a modification instruction provided by the user based on natural language and / or images; perform natural language analysis and / or image analysis on the modification instruction to determine at least one editing operation instruction on the relationship graph; transmit the at least one editing operation instruction to the display module to implement the corresponding editing operation; and transmit the at least one editing operation instruction to the purpose induction unit to re-determine the key value point.

[0023] In addition, the analysis method of the software development goal according to the second aspect of the present application comprises the following steps: obtaining multi-source data of the user about the software development goal; extracting at least one value point pursued by the user and at least one fact fact point from the multi-source data; performing semantic analysis on each of the extracted value points and fact points to determine the relevance of each fact point to each value point; determining a key value point with a total weight greater than a preset first threshold according to the relevance of each fact point to each value point; and determining the software development goal according to the key value point and its corresponding relevant fact points.

[0024] In addition, the computer readable storage medium according to the third aspect of the present application has computer instructions stored thereon. When the computer instructions are executed by a processor, the analysis method of the software development goal according to the second aspect of the present application is implemented. BRIEF DESCRIPTION OF DRAWINGS

[0025] The above features and advantages of the present application can be better understood by reading the following detailed description of embodiments of the present application in conjunction with the accompanying drawings, in which: in the drawings, components are not necessarily drawn to scale, and components having similar related properties or features can have the same or similar reference labels.

[0026] Figure 1 A schematic diagram of a software development objective analysis apparatus provided according to some embodiments of the present invention is shown.

[0027] Figure 2 A flowchart illustrating a method for analyzing software development goals according to some embodiments of the present invention is shown.

[0028] Figure 3 A schematic diagram illustrating the principle of a software development goal analysis method provided according to some embodiments of the present invention is shown.

[0029] Figure label:

[0030] 10 Data Acquisition Modules

[0031] 11 Business Data Acquisition Unit

[0032] 12 preprocessing units

[0033] 13 User Data Acquisition Unit

[0034] 20 Analysis Modules

[0035] 21 Value Point Extraction Unit

[0036] 22 Fact Point Extraction Unit

[0037] 23 Correlation Analysis Unit

[0038] 24 Objective Summary Unit

[0039] 30 display modules

[0040] 40 Modify auxiliary modules Detailed Implementation

[0041] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Although the description of the present invention is presented in conjunction with preferred embodiments, this does not mean that the features of the invention are limited to these embodiments. On the contrary, the purpose of describing the invention in conjunction with embodiments is to cover other options or modifications that may be derived based on the claims of the present invention. To provide a thorough understanding of the invention, many specific details will be included in the following description. The invention may also be implemented without using these details. Furthermore, to avoid confusion or obscuring the focus of the invention, some specific details will be omitted in the description.

[0042] In the description of the present application, it should be noted that unless specifically stated and limited otherwise, the terms "mounting", "connected", "connection" should be understood broadly, for example, can be fixed connection, can also be detachable connection, or integrally connected; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0043] In addition, "upper", "lower", "left", "right", "top", "bottom", "horizontal", "vertical" used in the following description should be understood as the orientation shown in the section and the related drawings. The relative terms are only for the convenience of description, and do not mean that the device described should be manufactured or operated in a particular orientation, so it should not be understood as a limitation on the present application.

[0044] It can be understood that although the terms "first", "second", "third" and the like are used herein to describe various components, regions, layers and / or parts, these components, regions, layers and / or parts should not be limited by these terms, and these terms are only used to distinguish different components, regions, layers and / or parts. Therefore, the first component, region, layer and / or part discussed below can be referred to as the second component, region, layer and / or part without departing from some embodiments of the present application.

[0045] As described above, the traditional manual analysis method relies on product managers or requirement analysts to collect information through conference interviews, questionnaires and other methods, which is severely limited by the personal experience, communication skills and subjective judgment of the analysis personnel. Not only is it inefficient and costly, but it is also extremely easy to cause demand distortion due to information omission and understanding bias, which poses a huge risk to the project. Project and task management tools such as JIRA, Trello can only manage explicit demand items and cannot automatically analyze, mine and refine deep purposes from original and ambiguous customer communication information. In addition, when applying primary natural language processing technology to analyze customer demand, since keyword extraction and word frequency statistics cannot understand the semantic relationship and context between words, surface phenomena are often misjudged as core demands. Moreover, the sentiment analysis of primary natural language processing technology can only judge the sentiment polarity of the text, and it is difficult to reveal the specific reasons and customer's potential goals behind the sentiment. General chat robots or question and answer systems are limited by preset rule libraries and are difficult to effectively induce and reason complex deep business purposes. In addition, although existing CRM systems are good at recording customer information and interaction history, their analysis capabilities are limited to macro statistics and lack a deep mining mechanism for the "why" behind customer specific product function demand.

[0046] In order to overcome the above-mentioned defects in the prior art, the application provides a software development target analysis device, a software development target analysis method and a computer readable storage medium, which can extract value points and fact points in multi-source data of a user about a software development target, establish an association relationship between the value points and the fact points, and deeply mine development targets in a software development process according to user requirements, so as to improve the efficiency and automation level in the analysis of the development targets, enhance the objectivity and accuracy of the analysis results, and further reduce the risk of a development project.

[0047] In some non-limiting embodiments, the above-mentioned software development target analysis device provided by the first aspect of the application can be implemented based on the above-mentioned software development target analysis method provided by the second aspect of the application.

[0048] Please refer to Figure 1 . Figure 1 A structural schematic diagram of the software development target analysis device provided by some embodiments of the application is shown.

[0049] In Figure 1 the embodiments shown, the above-mentioned software development target analysis device provided by the first aspect of the application includes a data acquisition module 10 and an analysis module 20. Here, the data acquisition module 10 is configured to acquire multi-source data of a user about a software development target. The analysis module 20 includes a value point extraction unit 21, a fact point extraction unit 22, an association analysis unit 23 and a purpose induction unit 24.

[0050] In addition, in some non-limiting embodiments, the above-mentioned software development target analysis device provided by the first aspect of the application includes a memory and a processor. Here, the memory includes but is not limited to the computer readable storage medium provided by the third aspect described above, and computer instructions are stored on the memory. The processor is connected to the memory and is configured to execute the computer instructions stored on the memory to implement the software development target analysis method provided by the second aspect of the application.

[0051] The working principle of the above-mentioned software development target analysis device will be described below in combination with some embodiments of a software development target analysis method. Those skilled in the art can understand that these embodiments of the analysis method are only some non-limiting embodiments provided by the application, which are intended to clearly demonstrate the main idea of the application and provide some specific schemes for facilitating the public to implement, rather than limit all functions or all working modes of the analysis device. Similarly, the software development target analysis device is also only a non-limiting embodiment provided by the application, which does not limit the execution subject and execution order of each step in these software development target analysis methods.

[0052] Please refer toFigure 2 and Figure 3 . Figure 2 A flowchart of a method for analyzing software development goals is shown according to some embodiments of the present application. Figure 3 A schematic diagram of a method for analyzing software development goals is shown according to some embodiments of the present application.

[0053] As shown in Figure 2 and Figure 3 , the first aspect of the present application provides an analysis device for the above-mentioned software development goals, which can first acquire multi-source data about software development goals of a user via a data acquisition module 10.

[0054] Further, in the embodiment shown in Figure 1 , the data acquisition module 10 is configured with a business data acquisition unit 11 connected to the project management system (e.g. JIRA system) of the user's company, online document sharing system (e.g. Confluence knowledge base), email server, conference recording transcription system, to acquire multi-source data about software development goals of the user through API interface, web crawler or file upload, etc. In this case, the multi-source data is selected from at least one of project management data, online documents, emails, and conference recordings.

[0055] For example, the business data acquisition unit 11 can acquire tasks defined by the customer in the current work project in the JIRA system, and the data nature is the customer's thinking on how to decompose and define the project development work. In this case, the JIRA system is a project management system, and its basic functions are to set up, assign and track tasks.

[0056] For another example, the business data acquisition unit 11 can acquire online documents of project descriptions of the customer for the project development work in the Confluence knowledge base, which are not limited to business scenarios, requirement definitions, interface definitions, and business logic. In this case, the Confluence knowledge base is an online document sharing system, which is usually used for sharing summary information within or between software teams.

[0057] In addition, in the embodiment shown in Figure 1In the illustrated embodiment, the data collection module 10 is further configured with a preprocessing unit 12 for preprocessing the acquired multi-source data. Specifically, in the preprocessing process, the preprocessing unit 12 can first perform deduplication processing on the acquired multi-source data. Then, the preprocessing unit 12 can remove HTML tags from the deduplicated data. After that, the preprocessing unit 12 can perform uniform encoding processing (e.g., uniform encoding to UTF-8) on the data from which the HTML tags are removed. After that, the preprocessing unit 12 can split the uniformly encoded data into multiple semantically complete paragraphs.

[0058] Further, in some embodiments, the preprocessing unit 12 can perform hierarchical segmentation on the uniformly encoded data. For example, the preprocessing unit 12 can split a semi-structured long document according to the hierarchy of its original hierarchical structure, i.e., large sections, small sections, paragraphs, and sentences.

[0059] Alternatively, in some embodiments, the preprocessing unit 12 is configured with a pre-trained fine-tuned large language model (e.g., GPT-4). Specifically, in the process of splitting the data, the preprocessing unit 12 can utilize the understanding ability of the large language model to perform the splitting. For example, the preprocessing unit 12 can issue a splitting instruction to the large language model, i.e., “Please split the following long text into multiple semantically coherent and complete parts. Each part should revolve around a core theme. Please use the special marker [END_OF_CHUNK] at the end of each part to identify the splitting point”.

[0060] Those skilled in the art can understand that the above-mentioned embodiments of preprocessing the acquired multi-source data are only some non-limiting embodiments provided by the present application, which are intended to clearly demonstrate the main idea of the present application and provide some specific schemes for the public to implement, but not to limit the protection scope of the present application.

[0061] Alternatively, in some embodiments, those skilled in the art can also not perform deduplication, HTML tag removal, uniform encoding, or splitting preprocessing operations on the acquired multi-source data, but directly utilize it for subsequent operations.

[0062] Alternatively, in some embodiments, those skilled in the art can also select one or more of the above preprocessing operations or adopt other preprocessing operations according to the actual data situation of the acquired multi-source data.

[0063] After that, as shown in FIG. 2, the data collection module 10 is further configured with a data analysis unit 14 for analyzing the preprocessed data. Specifically, in the process of analyzing the data, the data analysis unit 14 can first perform data cleaning on the preprocessed data. For example, the data analysis unit 14 can remove the noise in the preprocessed data, such as redundant information, irrelevant information, and incorrect information. Figure 2 and Figure 3As shown, the analysis device for the above-mentioned software development target provided by the first aspect of the present application can extract at least one value point pursued by the user from the above-mentioned multi-source data via a value point extraction unit 21 of the analysis module 20.

[0064] Further, in some preferred embodiments, the above-mentioned value point extraction unit 21 is configured with a pre-trained fine-tuned large language model (for example: GPT, BERT, etc.). Specifically, the value point extraction unit 21 can ask each of the multi-source data one by one according to the preset prompt engineering to extract at least one value point pursued by the user.

[0065] For example, in some embodiments, the above-mentioned value point extraction unit 21 can ask the large language model "what core expectations or success criteria does this text express?". Then, the value point extraction unit 21 can extract the value point V1: "improve the collaboration efficiency of the sales team".

[0066] In addition, Figure 1 In the embodiments shown, the above-mentioned data collection module 10 is configured with a user data acquisition unit 13 for acquiring feature data of multiple users for user classification. For example, the above-mentioned feature data includes fields (for example: semiconductors, construction, finance, etc.), positions (for example: engineers, testers, sales, etc.), gender, age, education, etc. Specifically, the above-mentioned value point extraction unit 21 can ask each of the multi-source data of each type of user one by one according to the preset prompt engineering to extract at least one value point pursued by the corresponding category of user.

[0067] In this way, by classifying users based on objective data, the present application can represent the needs and values of corresponding categories of users according to the characteristics of user needs and values of different fields, positions, genders, ages, and education, thereby accurately analyzing their software development targets.

[0068] In addition, as Figure 2 and Figure 3 As shown, the analysis device for the above-mentioned software development target provided by the first aspect of the present application can extract at least one fact point via a fact point extraction unit 22.

[0069] Further, in some preferred embodiments, the above-mentioned fact point extraction unit 22 is configured with a pre-trained named entity recognition model (Named Entity Recognition, NER). Specifically, the fact point extraction unit 22 can parse each of the multi-source data one by one according to a plurality of pre-named entities of multiple preset dimensions, and form at least one fact point according to the identified named entities.

[0070] Further, in some embodiments, the plurality of preset dimensions are selected from at least one of a person name, an organization name, a position name, a product name, an action name, a degree name, and a time length.

[0071] In addition, in some embodiments, the fact point extraction unit 22 can further determine a reliability weight according to a data source (e.g., a project management system, an online document sharing system, a mailbox server, a conference recording transcription system) and / or explicitness (e.g., a verb name “complain” “praise”; a degree name “very” “possibly”; a time length “1 minute” “1 hour”) of the fact point.

[0072] Then, the fact point extraction unit 22 can determine an importance weight according to a number of times the fact point appears in the multi-source data.

[0073] Further, the fact point extraction unit 22 can determine a fact point weight of the corresponding fact point according to the reliability weight and the importance weight.

[0074] Further, the fact point extraction unit 22 can retain the valid fact points whose fact point weights reach a preset second threshold (e.g., 0.3) and filter out the invalid fact points whose fact point weights are less than the second threshold.

[0075] For example, the fact point extraction unit 22 extracts a fact point F1: “Sales A complains that he spends 1 hour every day manually entering customer visit records into a customer relationship management system (CRM)”. Then, the fact point extraction unit 22 can determine a reliability weight of 0.9 according to a source (e.g., an interview recording transcript) and explicitness of the fact point F1.

[0076] Further, the fact point extraction unit 22 can determine an importance weight of 0.8 according to that the fact point F1 is mentioned in multiple places.

[0077] Further, the fact point extraction unit 22 can determine a fact point weight of the corresponding fact point F1 according to the reliability weight and the importance weight:

[0078]

[0079] wherein, , is a hyperparameter, .

[0080] Further, in response to the fact point weight of the fact point F1 being greater than the second threshold, the fact point extraction unit 22 can retain the fact point F1.

[0081] On the contrary, for another fact point F2: "I heard that B company uses deepseek for sales forecasting". Via the fact point extraction unit 22 determines that the fact point weight of the fact point is less than the second threshold value mentioned above, the fact point extraction unit 22 can filter out the fact point F2.

[0082] In addition, in some embodiments, the fact point extraction unit 22 mentioned above is also configured with a pre-trained explicitness classification model. Specifically, in response to forming at least one fact, the fact point extraction unit 22 can perform semantic analysis on the fact point via the explicitness classification model to determine the corresponding explicitness classification (for example: "explicit" (1), "more explicit" (0.5), "not explicit" (0.3)).

[0083] Further, as shown in Figure 2 and Figure 3 The first aspect of the present application provides an analysis device for the above-mentioned software development target, which can perform semantic analysis on each value point and each fact point via the correlation analysis unit 23 to determine the correlation of each fact point to each value point.

[0084] Further, in some preferred embodiments, the correlation analysis unit 23 mentioned above is configured with a pre-trained correlation classification model. Specifically, the correlation analysis unit 23 mentioned above can first combine each value point with each fact point into a high-dimensional vector (V, F) to calculate the correlation between the two (for example: cosine similarity, Euclidean distance, etc.). Here, the fact point is preferably an effective fact point.

[0085] Those skilled in the art can understand that the above-mentioned embodiment of calculating the correlation between the value point and the fact point based on the cosine similarity is only some non-limiting embodiments provided by the present application, which aims to clearly demonstrate the main idea of the present application and provide some specific schemes for the public to implement, but not for limiting the protection scope of the present application.

[0086] Alternatively, in other embodiments, those skilled in the art can also determine the correlation between the value point and the fact point based on the Euclidean distance, Pearson correlation coefficient, Jaccard similarity or adjusted cosine similarity.

[0087] Then, the correlation analysis unit 23 inputs the high-dimensional vector with a correlation reaching a preset third threshold value (for example: 0.7) into the correlation classification model to determine the correlation label of the fact vector F to the corresponding value vector V, wherein the positive correlation label indicates that the fact vector F supports the corresponding value vector V, and the negative correlation label indicates that the fact vector F opposes the corresponding value vector V.

[0088] In addition, in some embodiments, the above-mentioned correlation analysis model can also determine whether the fact point F has positive correlation with the value point V or negative correlation with the value point V. For example, in the customer demand data of "hope that the system can automatically synchronize data", the fact point F supports the value point V, i.e. has positive correlation with it. For example, in "data entry is too slow and leads to low efficiency", the fact point F is a counterexample of the value point V, i.e. has negative correlation with it.

[0089] Further, as shown in Figure 2 and Figure 3 , the first aspect of the present application provides an analysis device for the above-mentioned software development target, which can determine the key value point whose total weight is greater than a preset first threshold (for example: 2.5) via the purpose induction unit 24 according to the correlation of each fact point with each value point.

[0090] Specifically, the purpose induction unit 24 can mark the fact point weight of the fact point corresponding to each positive correlation label of each value point as positive, and calculate the cumulative value as the total weight of the value point. Here, the above-mentioned fact point is preferably an effective fact point. Then, the purpose induction unit 24 can determine the value point whose total weight is greater than the first threshold as the key value point KV.

[0091] Further, as shown in Figure 2 and Figure 3 , the purpose induction unit 24 can determine the software development target according to the key value point KV and the relevant fact point corresponding thereto.

[0092] Specifically, in the process of determining the software development target, the purpose induction unit 24 can fill the value vector V1 of the key value point and the fact vector F1 of the relevant fact point corresponding thereto into the preset software development target template, respectively, to output the software development target. For example: "In order to achieve the purpose of [V1], the problem of [F1] needs to be solved".

[0093] In addition, in some preferred embodiments, the above-mentioned purpose induction unit 24 can mark the fact point weight of the fact point corresponding to each negative correlation label of each value point as negative, and calculate the cumulative value as the total weight of the value point. Here, the fact point is preferably an effective fact point. Then, the purpose induction unit 24 can determine the value point whose total weight is less than the fourth threshold (for example: 0.1) as the effective value point AV.

[0094] Specifically, in the process of determining the software development goal, the purpose induction unit 24 can fill the value vector V2 of the value point which is both a key value point KV and an effective value point AV and the fact vector F2 of the corresponding related fact point into the preset software development goal template, respectively, to output the software development goal. For example, "in order to achieve the purpose of [V2], the problem of [F2] needs to be solved".

[0095] In addition, in Figure 1 In the embodiment shown, the first aspect of the present application provides the above-mentioned software development goal analysis device further comprising a display module 30 for providing a display interface with at least one tab page to display the software development goal and its corresponding value points, fact points, supporting relationships and / or opposing relationships.

[0096] Further, in some embodiments, the supporting relationships and / or opposing relationships are displayed in the form of a relationship graph, and the weights between the value points and the first fact points supporting them and / or the second fact points opposing them are shown.

[0097] Still further, the display module 30 also supports editing functions. Specifically, the above-mentioned display module 30 can add or delete value points and / or fact points and / or adjust the relationships between at least one value point and the corresponding fact points in the display interface according to the user's editing operations on the relationship graph.

[0098] In addition, in Figure 1 In the embodiment shown, the first aspect of the present application provides the above-mentioned software development goal analysis device further comprising a modification assistance module 40. The modification assistance module 40 is connected to the display module 30 and the purpose induction unit 24 of the analysis module.

[0099] Specifically, the above-mentioned modification assistance module 40 can first obtain the modification instructions provided by the user based on natural language and / or images.

[0100] Then, the modification assistance module 40 can perform natural language analysis and / or image analysis on the modification instructions to determine at least one editing operation instruction on the relationship graph. Here, the editing operation instruction can be a single editing operation instruction, or a plurality of editing operation instructions for batch modification under certain conditions. For example, the modification assistance module 40 can analyze the instruction to delete nodes that meet certain conditions, and use the pre-trained large language model configured therein to generate description text as required.

[0101] After that, the modification assistance module 40 can transmit the at least one editing operation instruction to the display module to implement the corresponding editing operation.

[0102] After that, the modification auxiliary module 40 can transmit at least one editing operation instruction to the target summary unit 24 to re-determine the key value point KV and feed back the judgment result to the display module 30 in real time.

[0103] To sum up, the above-mentioned software development target analysis device, software development target analysis method and computer readable storage medium provided by the application can extract value points and fact points in multi-source data about software development targets of a user, establish the correlation between the value points and the fact points, and deeply mine the development targets in the software development process according to the user demand, so as to improve the efficiency and automation level in the analysis of the development targets, enhance the objectivity and accuracy of the analysis results, and further reduce the risk of the development project.

[0104] Although the methods are illustrated and described as a series of acts for simplicity, it is understood and appreciated that the methods are not limited by the order of acts, as some acts may, in accordance with one or more embodiments, occur simultaneously or in different order than shown and described herein, or omitted altogether, depending on the circumstances.

[0105] Those skilled in the art will understand that information, signals, and data can be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that can be referenced throughout the above description can be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0106] Those skilled in the art will further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans can implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

[0107] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In the alternative, the processor and the storage medium can reside as discrete components in a user terminal.

[0108] In one or more exemplary embodiments, the functions described can be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functions can be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media can be any available media that can be accessed by a computer. By way of example, and not limitation, such computer-readable media can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. Also, any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. Disk and disc, as used herein, include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray® disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

[0109] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be readily apparent to those skilled in the art, and the generic principles defined herein can be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples described herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A software development goal analysis device, characterized in that, include: The data acquisition module is used to acquire multi-source data from users regarding software development goals; as well as The analysis module includes a value point extraction unit, a fact point extraction unit, a correlation analysis unit, and a purpose summarization unit. The value point extraction unit is configured to extract multiple value points sought by the user from the multi-source data. The fact point extraction unit is configured to extract fact points of multiple facts from the multi-source data. The association analysis unit is configured with a pre-trained association classification model and is configured to: combine each value point with each fact point to form a high-dimensional vector to calculate the correlation between the two; and input the high-dimensional vector with a correlation reaching a preset third threshold into the association classification model to determine the association label of the fact vector to the corresponding value vector, wherein a positive association label indicates that the fact vector supports the corresponding value vector, and a negative association label indicates that the fact vector opposes the corresponding value vector. The purpose induction unit is configured to: for each value point, mark the fact point weights of the fact points corresponding to its positive correlation labels as positive, and calculate their cumulative value as the total weight of the value point; determine the value points whose total weight is greater than a preset first threshold as key value points; and determine the software development goal based on the key value points and their corresponding related fact points.

2. The analytical apparatus as described in claim 1, characterized in that, The multi-source data is selected from at least one of the following: project management data, online documents, emails, and meeting recordings. The data acquisition module is equipped with a business data acquisition unit, which connects to the user company's project management system, online document sharing system, email server, and meeting recording transcription system to obtain multi-source data from the user regarding the software development goals.

3. The analytical apparatus as described in claim 2, characterized in that, The data acquisition module is also equipped with a preprocessing unit, which is configured as follows: The acquired multi-source data is deduplicated; Remove HTML tags from the data that has undergone the deduplication process; The data with the HTML tags removed is then subjected to uniform encoding processing; as well as The data processed by the unified encoding is divided into multiple semantically complete paragraphs.

4. The analytical apparatus as described in claim 1, characterized in that, The value point extraction unit is equipped with a pre-trained and fine-tuned large language model, which is configured as follows: Based on the preset prompts, questions are asked one by one about each of the multiple sources of data in order to extract the multiple value points pursued by the user.

5. The analytical apparatus as described in claim 4, characterized in that, The data acquisition module is equipped with a user data acquisition unit, which is used to acquire feature data of multiple users for user classification. The value point extraction unit is further configured to: ask questions to each of the users in the multi-source data of each type of user according to the preset prompts, so as to extract multiple value points pursued by the corresponding type of user.

6. The analytical apparatus as claimed in claim 1, characterized in that, The fact point extraction unit is equipped with a pre-trained named entity recognition model, which is configured as follows: Based on pre-named entities of multiple preset dimensions, each entity in the multi-source data is parsed one by one, and fact points of multiple facts are formed based on the identified named entities. The multiple preset dimensions are selected from at least one of person name, organization name, job name, product name, action name, degree name, and time length.

7. The analytical apparatus as described in claim 6, characterized in that, The fact point extraction unit is further configured to: A reliability weight is determined based on the data source and / or clarity of the fact points; An importance weight is determined based on the frequency of occurrence of the fact points in the multi-source data; Based on the reliability weight and the importance weight, the fact point weight of the corresponding fact point is determined; as well as Valid fact points whose fact point weights reach a preset second threshold are retained, while invalid fact points whose fact point weights are less than the second threshold are filtered out.

8. The analytical apparatus as described in claim 7, characterized in that, The fact point extraction unit is also equipped with a pre-trained explicitness classification model, which is configured as follows: In response to fact points that constitute multiple facts, semantic analysis is performed on the fact points via the explicitness classification model to determine the corresponding explicitness classification.

9. The analytical apparatus as claimed in claim 1, characterized in that, The step of determining the software development goal based on the key value points and their corresponding relevant fact points includes: The value vector of the key value points V 1 and its corresponding fact vector F 1. Fill in the preset software development goal templates to output the software development goals.

10. The analytical apparatus as claimed in claim 1, characterized in that, The objective induction unit is further configured to: For each value point, the fact point weights of the fact points corresponding to its reverse correlation labels are marked as negative, and their cumulative values ​​are calculated as the total weight of the value point. as well as Value points whose total weight is less than the fourth threshold are identified as valid value points.

11. The analytical apparatus as claimed in claim 10, characterized in that, The step of determining the software development goal based on the key value points and their corresponding relevant fact points includes: The value vector of the value point that is both the key value point and the effective value point. V 2 and the fact vector of its corresponding related fact points F 2. Fill in the preset software development goal templates to output the software development goals.

12. The analytical apparatus as claimed in claim 1, characterized in that, Also includes: The display module is used to provide a display interface with at least one tab to display the software development goals, as well as their corresponding value points, fact points, supporting relationships and / or opposing relationships, wherein the supporting relationships and / or the opposing relationships are displayed in the form of a relationship diagram, and the weights between the value points and the first fact point supporting them and / or the second fact point opposing them are shown.

13. The analytical apparatus as claimed in claim 12, characterized in that, The display module also supports editing functions and is configured as follows: Based on the user's editing operations on the relationship diagram, value points and / or fact points can be added or removed on the display interface, and / or the relationship between at least one value point and its corresponding fact point can be adjusted.

14. The analytical apparatus as claimed in claim 13, characterized in that, The analysis device further includes a modification assistance module, which is connected to the display module and the purpose summarization unit of the analysis module, and is configured as follows: Obtain user-provided modification instructions based on natural language and / or images; The modification instructions are parsed using natural language and / or image parsing to determine at least one edit operation instruction for the relationship graph; The at least one editing operation instruction is transmitted to the display module to realize the corresponding editing operation; as well as The at least one editing operation instruction is transmitted to the target summarization unit to redefine the key value point.

15. A method for analyzing software development goals, characterized in that, Includes the following steps: Obtain multi-source data from users regarding software development goals; Extract multiple value points pursued by the user from the multi-source data, and extract fact points of multiple facts; Each value point is combined with each fact point to form a high-dimensional vector to calculate the correlation between the two. A high-dimensional vector with a relevance reaching a preset third threshold is input into an association classification model to determine the association label between the fact vector and the corresponding value vector. A positive association label indicates that the fact vector supports the corresponding value vector, and a negative association label indicates that the fact vector opposes the corresponding value vector. For each value point, the fact point weights of the fact points corresponding to its positive correlation labels are marked as positive, and their cumulative values ​​are calculated as the total weight of the value point. Value points with a total weight greater than a preset first threshold are identified as key value points; and The software development goals are determined based on the key value points and their corresponding relevant fact points.

16. A computer-readable storage medium storing computer instructions thereon, characterized in that, When the computer instructions are executed by the processor, the analysis method for software development objectives as described in claim 15 is implemented.

Images