Research and development cycle measuring and calculating method, device and equipment and storage medium
By extracting the time factor parameters of each stage of the R&D process and using machine learning models to predict the R&D cycle, the problem of inaccurate prediction of the R&D cycle is solved, and the accurate and reasonable schedule from demand to online is achieved, ensuring the quality of software products.
Patent Information
- Application Number
- CN202311368433.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-20
- Publication Date
- 2025-01-14
AI Technical Summary
In the existing technology, the prediction of R&D cycles is inaccurate, which affects the quality of development products, making it difficult to reasonably arrange R&D resources and cycles, affecting the quality of software products online.
By extracting the time factor parameters of multiple stages of the R&D process involved in the requirements description, and using them as input data of the model, using a preset machine learning model to predict the time spent in different R&D stages, and finally summarizing and calculating the total R&D cycle. Establish a calculation model to arrange schedules accurately and reasonably.
By clarifying the factors that affect time at each stage and establishing a calculation model, the scheduling of requirements from proposal to launch is more accurate and reasonable, thereby ensuring the quality of demand launch.
Smart Images
Figure CN119313270A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of artificial intelligence technology, and specifically to a research and development cycle calculation method, device, equipment and storage medium. Background Art
[0002] In order to quickly meet new customer needs, special attention should be paid to when the business needs can be put online. To this end, it is necessary to more carefully and accurately calculate the time or key nodes of each link in the R&D life cycle, and scientifically plan the version based on the urgency of business needs to achieve steady online launch.
[0003] In the existing technology, in order to achieve accurate calculation of the time taken for each link or key nodes, the link cycles are mostly calculated one by one through manual experience. However, this model often leads to problems such as uneven busyness and idleness in each version cycle and unreasonable scheduling due to inaccurate estimation, which affects the quality of R&D and online launch. It is urgent to explore a new method that can more accurately and intelligently calculate the time taken for links in different life cycles according to the characteristics of business needs, so as to achieve a more reasonable arrangement of R&D resources and cycles, thereby effectively ensuring the quality of software products. Summary of the invention
[0004] In view of the above problems, the present disclosure provides a method, device, equipment, medium and program product for improving the accuracy of R&D cycle calculation.
[0005] According to the first aspect of the present disclosure, a method for calculating a research and development cycle is provided, wherein the research and development cycle includes N research and development stages, where N is a positive integer, and the method includes: receiving requirement description information; executing preprocessing logic on the requirement description information to obtain time factor parameters corresponding to the N research and development stages; based on the time factor parameters corresponding to the N research and development stages, using a preset machine learning model, calculating the time of the research and development stages to obtain predicted values of the time of the N stages; and summarizing the predicted values of the time of the N stages to obtain a predicted value of the research and development cycle.
[0006] According to an embodiment of the present disclosure, the preprocessing logic is executed on the requirement description information to obtain the time factor parameters corresponding to the N R&D stages, including: for any R&D stage, intelligent word segmentation is performed on the requirement description information to obtain a word segmentation result set, the word segmentation result set includes M word segmentation results, M is a positive integer; obtain a keyword library corresponding to the R&D stage, the keyword library includes K keywords, K is a positive integer; based on the M word segmentation results and K keywords, pairwise matching of the word segmentation results and the keywords is performed to obtain matching results, the matching results include L successfully matched keywords, L is a positive integer less than or equal to K; and based on the L successfully matched keywords, a time factor parameter is formed.
[0007] According to an embodiment of the present disclosure, each of the K keywords corresponds to an unassigned time factor, and forming a time factor parameter based on the L successfully matched keywords includes: assigning values to the time factors corresponding to the L successfully matched keywords to form the time factor parameter.
[0008] According to an embodiment of the present disclosure, the training method of the preset machine learning model includes: obtaining training samples, the training samples including samples of requirement description information and characterization parameters of stage timing, the samples of requirement description information and the characterization parameters of stage timing are in one-to-one correspondence according to the research and development stages; executing the preprocessing logic on the samples of the requirement description information to obtain samples of timing factor parameters corresponding to the N research and development stages; based on the characterization parameters of the stage timing, forming training samples of the research and development cycle, the training samples of the research and development cycle include the training stage timing corresponding to the N research and development stages; and training the machine learning model based on the samples of N timing factor parameters and the training stage timing corresponding to the N research and development stages.
[0009] According to an embodiment of the present disclosure, the training samples include historical training samples and training samples in use.
[0010] According to an embodiment of the present disclosure, after aggregating the predicted values of the time taken for the N stages to obtain the predicted value of the R&D cycle, it also includes: obtaining the actual value of the R&D cycle; calculating an error value based on the actual value of the R&D cycle and the predicted value of the R&D cycle; and starting the dynamic configuration logic when the error value is greater than a preset error threshold.
[0011] According to an embodiment of the present disclosure, the dynamic configuration logic includes: receiving adjustment instructions from personnel; and executing the adjustment instructions, wherein the adjustment instructions include performing modification operations on the keyword library and / or the factor parameters, and the modification operations include adding, deleting and modifying.
[0012] According to an embodiment of the present disclosure, the characterization parameters of the stage time include J sub-characterization parameters, J is a positive integer, and the training samples of the R&D cycle are formed based on the characterization parameters of the stage time, including: obtaining a preset mapping relationship, the preset mapping relationship includes a one-to-one mapping relationship between the sub-characterization parameters and the training R&D sub-stages; based on the J sub-characterization parameters and the preset mapping relationship, deriving J training R&D sub-stages; and summarizing the J training R&D sub-stages to obtain the training samples of the R&D stage.
[0013] The second aspect of the present disclosure provides a device for calculating a research and development cycle, wherein the research and development cycle includes N research and development stages, where N is a positive integer, and the device includes: a demand receiving module, used to receive demand description information; a preprocessing module, used to execute preprocessing logic on the demand description information to obtain time factor parameters corresponding to the N research and development stages; a predicted value calculation module for stage time, used to calculate the time of the research and development stage based on the time factor parameters corresponding to the N research and development stages through a preset machine learning model to obtain predicted values of the time of the N stages; and a summary module, used to summarize the predicted values of the time of the N stages to obtain a predicted value of the research and development cycle.
[0014] According to an embodiment of the present disclosure, the preprocessing module includes a word segmentation unit, a keyword library acquisition unit, a matching unit and a time factor parameter formation unit. The word segmentation unit is used to perform intelligent word segmentation on the requirement description information for any R&D stage to obtain a word segmentation result set, and the word segmentation result set includes M word segmentation results, where M is a positive integer; the keyword library acquisition unit is used to obtain a keyword library corresponding to the R&D stage, and the keyword library includes K keywords, where K is a positive integer; the matching unit is used to perform pairwise matching of the word segmentation results and the keywords based on the M word segmentation results and the K keywords to obtain matching results, and the matching results include L successfully matched keywords, where L is a positive integer less than or equal to K; and the time factor parameter formation unit is used to form a time factor parameter based on the L successfully matched keywords.
[0015] According to an embodiment of the present disclosure, each of the K keywords corresponds to an unassigned time factor, and the time factor parameter forming unit includes a time factor parameter assignment subunit, which is used to assign time factors corresponding to the L successfully matched keywords to form the time factor parameter.
[0016] According to an embodiment of the present disclosure, the device further includes a sample acquisition module, a R&D cycle sample formation module and a training module, wherein the acquisition module is used to acquire training samples, wherein the training samples include samples of requirement description information and characterization parameters of stage duration, and the samples of requirement description information and the characterization parameters of stage duration are in one-to-one correspondence according to the R&D stages; the preprocessing module is used to execute the preprocessing logic on the samples of the requirement description information to obtain samples of the time factor parameters corresponding to the N R&D stages; the R&D cycle sample formation module is used to form training samples of the R&D cycle based on the characterization parameters of the stage duration, and the training samples of the R&D cycle include the training stage duration corresponding to the N R&D stages; and the training module is used to train the machine learning model based on samples of the N time factor parameters and the training stage duration corresponding to the N R&D stages.
[0017] According to an embodiment of the present disclosure, the training samples include historical training samples and training samples in use.
[0018] According to an embodiment of the present disclosure, the device further includes an actual value acquisition module, an error calculation module and a dynamic configuration module, wherein the actual value acquisition module is used to acquire the actual value of the R&D cycle; the error calculation module is used to calculate the error value based on the actual value of the R&D cycle and the predicted value of the R&D cycle; and the dynamic configuration module is used to start the dynamic configuration logic when the error value is greater than a preset error threshold.
[0019] According to an embodiment of the present disclosure, the dynamic configuration module includes an adjustment instruction receiving unit and an adjustment instruction execution unit, wherein the adjustment instruction receiving unit is used to receive adjustment instructions from personnel; and the adjustment instruction execution unit is used to execute the adjustment instructions, wherein the adjustment instructions include modifying the keyword library and / or the factor parameters, and the modification operations include adding, deleting and modifying.
[0020] According to an embodiment of the present disclosure, the characterization parameter of the stage time includes J sub-characterization parameters, J is a positive integer, and the R&D cycle sample formation module includes a mapping relationship acquisition unit, a sub-cycle derivation unit and a training sample summary unit of the R&D cycle, the mapping relationship acquisition unit is used to obtain a preset mapping relationship, and the preset mapping relationship includes a one-to-one mapping relationship between the sub-characterization parameters and the training R&D sub-stages; the sub-cycle derivation unit is used to derive J training R&D sub-stages based on the J sub-characterization parameters and the preset mapping relationship; and the training sample summary unit of the R&D cycle is used to summarize the J training R&D sub-stages to obtain the training samples of the R&D stage.
[0021] The third aspect of the present disclosure provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above-mentioned R&D cycle estimation method.
[0022] The fourth aspect of the present disclosure further provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above-mentioned R&D cycle estimation method.
[0023] In the embodiments of the present disclosure, in order to solve the technical problem that the quality of developed products is affected by inaccurate prediction of the R&D cycle. By extracting the factor parameters of the time required for multiple stages in the R&D process in the demand description and using the factor parameters as the input data of the model, the predicted time required for different R&D stages is predicted through a machine learning model. Finally, these predicted time requirements are summarized to calculate the total R&D cycle. The embodiments of the present disclosure can at least clarify the factors that affect the time required at each stage and establish a measurement model to make the scheduling of requirements from proposal to launch more accurate and reasonable, thereby ensuring the quality of the online demand. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:
[0025] Figure 1 The following schematically shows an application scenario diagram of the R&D cycle calculation method according to an embodiment of the present disclosure;
[0026] Figure 2 A flowchart of a method for calculating a development cycle according to an embodiment of the present disclosure is schematically shown;
[0027] Figure 3 A flowchart of a method for calculating a time factor parameter according to an embodiment of the present disclosure is schematically shown;
[0028] Figure 4 A flowchart of a model training method according to an embodiment of the present disclosure is schematically shown;
[0029] Figure 5 A flowchart schematically illustrates an iterative method of a model according to an embodiment of the present disclosure;
[0030] Figure 6 A flowchart of a dynamic configuration method according to an embodiment of the present disclosure is schematically shown;
[0031] Figure 7 A flowchart schematically shows a method for generating training samples in a research and development cycle according to an embodiment of the present disclosure;
[0032] Figure 8 A schematic diagram of a structure of a device for calculating a development cycle according to an embodiment of the present disclosure is shown; and
[0033] Fig. 9 A block diagram of an electronic device suitable for implementing a method for calculating a development cycle according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION
[0034] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.
[0035] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.
[0036] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.
[0037] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).
[0038] Before the embodiments of the present disclosure are disclosed in detail, the key technical terms involved in the embodiments of the present disclosure are disclosed one by one, as shown below:
[0039] Natural Language Processing (NLP) is an important field in the fields of computer science and artificial intelligence. Natural Language Processing studies the theories and methods of effective communication between humans and computers using natural language.
[0040] Jieba word segmentation technology: a Python Chinese word segmentation component that can implement word segmentation in different modes.
[0041] Machine learning is a multi-disciplinary subject that studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning models include at least unsupervised learning models and supervised learning models.
[0042] Backpropagation Neural Network (BPNN) is a neural network model trained using the backpropagation algorithm. In BPNN, information is forward propagated through the connections between layers, and the error is reversely propagated and the weights are updated through the backpropagation algorithm to minimize the error between the predicted output and the actual output.
[0043] In a typical scenario, the R&D life cycle of a product has multiple different stages (or links), and among the multiple different stages, there are some important stages (or important links) that play a decisive role in the launch cycle. In the existing R&D process, after the business department proposes requirements and establishes a project, the development department usually estimates the time required for the above-mentioned multiple important stages or important links based on past R&D experience and arranges the version launch plan.
[0044] However, the department's calculations based solely on past experience make it difficult to accurately assess the implementation time of each link, and further, it is difficult to reasonably arrange the corresponding R&D resources and cycles, making it difficult to ensure the quality of software products.
[0045] In order to solve the technical problems existing in the prior art, an embodiment of the present disclosure provides a method for calculating the R&D cycle, wherein the R&D cycle includes N R&D stages, where N is a positive integer, and the method includes: receiving requirement description information; executing preprocessing logic on the requirement description information to obtain time factor parameters corresponding to the N R&D stages; based on the time factor parameters corresponding to the N R&D stages, using a preset machine learning model, calculating the time of the R&D stage to obtain predicted values of the time of the N stages; and summarizing the predicted values of the time of the N stages to obtain a predicted value of the R&D cycle.
[0046] In the embodiments of the present disclosure, in order to solve the technical problem that the quality of developed products is affected by inaccurate prediction of the R&D cycle. By extracting the factor parameters of the time required for multiple stages in the R&D process in the demand description and using the factor parameters as the input data of the model, the predicted time required for different R&D stages is predicted through a machine learning model. Finally, these predicted time requirements are summarized to calculate the total R&D cycle. The embodiments of the present disclosure can at least clarify the factors that affect the time required at each stage and establish a measurement model to make the scheduling of requirements from proposal to launch more accurate and reasonable, thereby ensuring the quality of the online demand.
[0047] Figure 1 The application scenario diagram of the R&D cycle estimation method according to an embodiment of the present disclosure is schematically shown.
[0048] like Figure 1As shown, the application scenario 100 according to this embodiment may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.
[0049] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, social platform software, etc. (only examples).
[0050] The terminal devices 101, 102, 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.
[0051] The server 105 may be a server that provides various services, such as a background management server (only an example) that provides support for websites browsed by users using the terminal devices 101, 102, and 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.
[0052] It should be noted that the R&D cycle calculation method provided in the embodiment of the present disclosure can generally be executed by the server 105. Accordingly, the R&D cycle calculation device provided in the embodiment of the present disclosure can generally be set in the server 105. The R&D cycle calculation method provided in the embodiment of the present disclosure can also be executed by a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the R&D cycle calculation device provided in the embodiment of the present disclosure can also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101, 102, 103 and / or the server 105.
[0053] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.
[0054] The following will be based on Figure 1 The scene described by Figure 2 to Figure 7 The R&D cycle calculation method of the disclosed embodiment is described in detail.
[0055] Figure 2 The flowchart of the method for calculating the R&D cycle according to an embodiment of the present disclosure is schematically shown.
[0056] like Figure 2 As shown, the R&D cycle estimation method of this embodiment includes operations S210 to S240 , and the method can be executed by the server 105 .
[0057] According to an embodiment of the present disclosure, the R&D cycle includes N R&D stages, where N is a positive integer.
[0058] In operation S210 , requirement specification information is received.
[0059] In operation S220, preprocessing logic is executed on the requirement description information to obtain time factor parameters corresponding to the N research and development stages.
[0060] In a typical scenario, taking the financial field as an example, the R&D life cycle of financial products at least includes R&D stages such as demand proposal, demand confirmation, demand development, product acceptance, online operation, offline exit, etc. In other words, in the embodiments of the present disclosure, the above-mentioned N R&D stages can be full R&D stages such as demand proposal, demand confirmation, demand development, product acceptance, online operation, offline exit, etc.; the above-mentioned N R&D stages can also be some important (decisive) stages in the full R&D stages such as demand proposal, demand confirmation, demand development, product acceptance, online operation, offline exit, etc.
[0061] Throughout the entire life cycle, the three R&D stages of demand confirmation, demand development, and product acceptance play a decisive role in the launch cycle. Therefore, after the business department proposes a demand and establishes a project, it is necessary to focus on estimating the time spent in the above three R&D stages, summarize and calculate the total time spent, and arrange a version launch plan. Therefore, it is necessary to obtain the characteristics of the corresponding stages for the R&D stages of demand confirmation, demand development, and product acceptance, so that they can be used as input data for the machine learning model later. Among them, the business requirements specification includes at least project data, business requirements use case specifications, and software requirements specifications. Project data includes function points, stage, and emergency requirements.
[0062] The demand department submits the business demand specification (i.e. the demand specification information mentioned above) so that the relevant R&D department can implement it according to the content specified in the business demand specification. Among them, for the business demand specification, at least the following aspects are included: 1. The functional points involved in the business demand specification; 2. The organizational structure involved in the requirements in the business demand specification, for example, multiple departments involved; 3. The different types of business requirements involved in the business demand specification, for example, the description of new business requirements, the description of customer and account-related business requirements, and the description of emergency business requirements. Among them, the function point number is directly assigned by the time factor parameter through the data after searching the keyword "function point". The function point number is comprehensively evaluated by the upstream department according to a specific calculation method, aiming to evaluate the complexity of the function development in the new requirements. It will not be repeated here. The organizational structure is the assignment of the number of split calculation departments. The business type depends on the field and will not be explained in detail here.
[0063] By extracting the contents related to the above three aspects in the business requirements specification, the time factor parameters for each R&D stage can be formed. Among them, for business requirements, different business requirements are included in different fields, and accordingly, the impact of business requirements on the overall R&D time consumption is also different. It should be understood that the constraints on the impact between business requirements and time consumption by relevant personnel in different fields are also different. This is due to the characteristics of the business itself. Therefore, the types of business requirements will not be repeated here.
[0064] It should be noted that the above-mentioned functional points, organizational structure and business requirements need to be formed in accordance with the characteristics of different R&D stages, through different combinations, to form time factor parameters for different stages. This process (i.e., the above-mentioned operation S220) can be achieved through manual participation, and this process (i.e., the above-mentioned operation S220) can also be achieved through intelligent means.
[0065] The intelligent means of operation S220 will be disclosed in detail below, as follows:
[0066] Figure 3 The flowchart of the method for calculating the time factor parameter according to an embodiment of the present disclosure is schematically shown.
[0067] like Figure 3 As shown, the method for calculating the time factor parameter of this embodiment includes operations S310 to S340, and the operations S310 to S340 can at least partially execute the above-mentioned operation S230.
[0068] In operation S310, for any R&D stage, intelligent word segmentation is performed on the requirement description information to obtain a word segmentation result set, where the word segmentation result set includes M word segmentation results, where M is a positive integer.
[0069] Specifically, the above intelligent word segmentation can be implemented by using Jieba word segmentation technology, which is a Chinese word segmentation component that divides a Chinese character sequence into individual words. Word segmentation is the process of recombining a continuous character sequence into a word sequence according to certain specifications.
[0070] In operation S320, a keyword library corresponding to the research and development stage is obtained, where the keyword library includes K keywords, where K is a positive integer.
[0071] The keyword library is pre-set and adjustable, and the keywords in the keyword library are different for different research and development stages.
[0072] In operation S330, based on the M word segmentation results and the K keywords, pairwise matching of the word segmentation results and the keywords is performed to obtain matching results, where the matching results include L successfully matched keywords, where L is a positive integer less than or equal to K.
[0073] Specifically, the keyword is used to match the word segmentation results one by one, and the matching result is recorded, for example, the result is a complete match success or a complete match failure.
[0074] In operation S340, a time factor parameter is formed based on the L successfully matched keywords.
[0075] Specifically, for example, the time factor may be assigned a value of 0 or 1 according to whether the match is successful or failed.
[0076] Of course, the above matching process can be a complete correspondence between keywords and word segmentation results, for example, the keyword "function points" is completely matched with the word segmentation result "function points"; the above matching process can also be based on the matching method of word vectors. For example, the keyword "involving accounts" can be matched through word vectors to produce more fuzzy matches such as the word segmentation result "involving accounts", that is, the same semantics can be matched through word vector technology.
[0077] Since the R&D cycle is obtained by summarizing the time consumed in each R&D stage, or the R&D cycle is obtained by summarizing the time consumed in each important R&D stage, and since there are certain differences between the time consumption logics corresponding to each R&D cycle, this difference will affect the relationship between the time factor parameters and the time consumption, so the acquisition of the time factor parameters is based on the R&D stage, rather than the overall R&D cycle. Therefore, in the process of obtaining the time factor parameters, the keywords involved in the keyword library for each R&D stage are also different. Taking the software demand development in the financial scenario as an example, the specific details are as follows:
[0078] Keywords for the demand confirmation stage include: stage information (demand confirmation), business estimated function point information, whether it is an urgent demand, organizational information (including whether it crosses business departments, crosses R&D departments, and whether it involves a third party), business type information (including whether it is a new demand);
[0079] Keywords for the demand development stage include: preset stage information (demand development), software product function points, organizational information (whether it crosses R&D departments, whether it involves third parties), business type information (whether it is customer-oriented, whether it involves accounts);
[0080] The keyword library for the product acceptance stage includes: preset stage information (product development), software product function points, business type information (whether it is customer-oriented, whether it involves accounts, whether it is a new function), and calculation complexity information (whether it involves upstream and downstream joint testing, whether it involves batch testing, whether it involves batch testing, whether it involves performance testing).
[0081] According to an embodiment of the present disclosure, each of the K keywords corresponds to an unassigned time factor, and forming a time factor parameter based on the L successfully matched keywords includes: assigning values to the time factors corresponding to the L successfully matched keywords to form the time factor parameter.
[0082] It is understandable that, from the perspective of the R&D stage, an uninitialized time factor parameter corresponding to the R&D stage is pre-set, and the time factor parameter includes multiple time factors (factors that affect the time) and the assignment bits corresponding to the time factors (the assignment bits can be, for example, 0 or 1). The above operations S310 to S340 are equivalent to initializing the time factor parameter, and through keyword matching, a value of 1 is assigned if the match is successful, and a value of 0 is assigned if the match is unsuccessful, and finally the time factor parameter is formed.
[0083] In operation S230, based on the time factor parameters corresponding to the N research and development stages, the time of the research and development stages is calculated through a preset machine learning model to obtain predicted values of the time of the N stages.
[0084] Specifically, by using the time factor parameters as input data for machine learning, the time required for each R&D stage can be output.
[0085] Just like the one-to-one correspondence between the time factor parameters and the R&D stages mentioned above, the machine learning models also correspond one-to-one with the R&D stages for the same reason, that is, each of the N R&D stages has its own machine learning model, and the parameters of the machine learning model of each R&D stage are set in advance.
[0086] The machine learning model may be a supervised learning model, such as a BP neural network model, or an unsupervised learning model. For a supervised learning model, the training or iteration logic is as follows:
[0087] Figure 4 A flowchart of a method for training a model according to an embodiment of the present disclosure is schematically shown.
[0088] like Figure 4 As shown, the model training method of this embodiment includes operations S410 to S440.
[0089] In operation S410, a training sample is obtained, wherein the training sample includes a sample of requirement specification information and a characterization parameter of a stage time consumption, and the sample of requirement specification information and the characterization parameter of the stage time consumption are in one-to-one correspondence according to the research and development stage.
[0090] Specifically, the data of the training samples can be either historical data or data obtained during use; the training samples include samples of requirement description information and characterization parameters of stage time consumption, wherein the samples of requirement description information can be either historical requirement description information or requirement description information obtained during the actual production process.
[0091] Correspondingly, the other data used as the "result" in supervised learning is the characterization parameter of the stage time consumption, rather than the actual stage time consumption. The advantage of doing so is that relevant personnel have found that the process of demand development has its own characteristics, including a small amount of actual data for training and a certain degree of customization of the R&D needs (that is, the correlation between different demand development processes is weak, resulting in a weak correlation between each demand development cycle). This causes the estimated period of each development cycle to be more affected by irrelevant items. Furthermore, in the embodiments of the present disclosure, a method is designed to calculate the hypothetical stage time consumption through the characterization parameter of the stage time consumption, which is a more ideal value, and then use the hypothetical stage time consumption as training data. Similarly, the characterization parameter of the stage time consumption is also calculated based on the R&D stage.
[0092] In operation S420, the preprocessing logic is executed on the sample of the requirement specification information to obtain samples of the time factor parameters corresponding to the N research and development stages.
[0093] The preprocessing logic is consistent with the preprocessing logic in the above operation S220, and is intended to calculate the training value of the time factor parameter. The method of matching and then copying is also adopted, which will not be described in detail here.
[0094] In operation S430, a training sample of a development cycle is formed based on the characterization parameter of the stage time consumption, and the training sample of the development cycle includes the training stage time consumption corresponding to the N development stages.
[0095] Specifically, the time consumption corresponding to the parameters in each R&D stage according to certain calculation rules is as follows:
[0096] Figure 7 The flowchart of the method for generating training samples in the R&D cycle according to the embodiment of the present disclosure is schematically shown.
[0097] like Figure 7 As shown, the method for generating training samples in the development cycle of this embodiment includes operations S710 to S730, and operations S710 to S730 can at least partially execute the above operation S430.
[0098] According to an embodiment of the present disclosure, the characterization parameter of the stage time consumption includes J sub-characterization parameters, where J is a positive integer.
[0099] In operation S710, a preset mapping relationship is obtained, where the preset mapping relationship includes a one-to-one mapping relationship between the sub-characterization parameters and the training development sub-stages.
[0100] In operation S720, J training development sub-stages are obtained based on the J sub-characterization parameters and the preset mapping relationship.
[0101] In operation S730, the J training and development sub-stages are summarized to obtain training samples of the development stage.
[0102] Specifically, the corresponding R&D sub-cycle is obtained through the preset mapping relationship between each sub-characterization parameter and the time consumption in the characterization parameter, and then each R&D sub-cycle is accumulated to obtain the training sample of the R&D stage.
[0103] In other words, operations S710 to S730 can be understood as, when the parameters of a certain action in the characterization parameters are implemented, they are integrated, and according to the interval in which the integration is located, the corresponding time consumption is obtained (the time consumption includes positive time consumption and negative time consumption), and finally all the time consumption is accumulated to obtain the research and development stage.
[0104] Among them, the sub-characterization parameters in the characterization parameters can be divided into different categories according to the characteristics of the corresponding R&D stages, and the mapping relationship under the category is executed, as shown below:
[0105] For the time taken in the demand confirmation phase: Benchmark time calculation: if the function points are less than 200, then 1 week; 200-400 function points 1.5 weeks, 400 function points and above 2 weeks; organizational structure impact calculation: if the number of departments is <= 3, then 0.5 weeks; if 3<the number of departments is <= 5, then 1 week; if the number of departments is > 5, then 1.5 weeks; for new business requirements, the time taken is +0.5, and for urgent requirements, -0.5 weeks;
[0106] For the time taken in the demand development phase: Benchmark time calculation: if the function points are less than 200, then 3 weeks; 200-400 function points, 4 weeks, 400 function points and above 6 weeks; organizational structure impact calculation: if the number of departments is <= 3, then 1 week; if 3 < the number of departments <= 5, then 1.5 weeks; if the number of departments is > 5, then 2 weeks; for customer and account-related business needs, the time taken is +0.5 respectively;
[0107] For the time taken in the product acceptance phase: Benchmark time calculation: if the function points are less than 200, it will be 2 weeks; 200-400 function points will be 2.5 weeks, and 3 weeks if the function points are more than 400; calculation of the impact of requirement type: for new business requirements, the time will be increased by +0.5 weeks; calculation of the impact of calculation complexity: if it involves upstream and downstream joint testing, the time will be increased by +0.5 weeks; if it involves a third party, the time will be increased by 0.7 weeks; if it involves batch testing, the time will be increased by 1 week; if it involves performance calculation, the time will be increased by 0.8 weeks.
[0108] In operation S440, the machine learning model is trained based on samples of N timing factor parameters and training phase times corresponding to N development phases.
[0109] According to an embodiment of the present disclosure, the training samples include historical training samples and training samples in use.
[0110] Specifically, the machine learning model of the corresponding R&D stage is trained through the samples of the time factor parameters of each R&D stage and the training stage time.
[0111] In operation S240, the predicted values of the time taken for the N stages are summarized to obtain a predicted value of the R&D cycle.
[0112] Specifically, the predicted values of multiple different stages may be accumulated and summed, and the obtained total value may be used as the predicted value of the R&D cycle.
[0113] In the embodiments of the present disclosure, in order to solve the technical problem that the quality of developed products is affected by inaccurate prediction of the R&D cycle. By extracting the factor parameters of the time required for multiple stages in the R&D process in the demand description and using the factor parameters as the input data of the model, the predicted time required for different R&D stages is predicted through a machine learning model. Finally, these predicted time requirements are summarized to calculate the total R&D cycle. The embodiments of the present disclosure can at least clarify the factors that affect the time required at each stage and establish a measurement model to make the scheduling of requirements from proposal to launch more accurate and reasonable, thereby ensuring the quality of the online demand.
[0114] It is understandable that after using a relatively ideal training time, the model needs to be corrected based on the actual time consumed in each development stage, so that the model can be continuously iterated. The specific steps are as follows:
[0115] Figure 5 The flowchart of the iterative method of the model according to the embodiment of the present disclosure is schematically shown.
[0116] like Figure 5 As shown, the iterative method of the model in this embodiment includes operations S510 to S530, and operations S510 to S530 are performed after the above-mentioned operation S240.
[0117] In operation S510 , an actual value of the development cycle is acquired.
[0118] In operation S520 , an error value is calculated based on the actual value of the R&D period and the predicted value of the R&D period.
[0119] In operation S530 , when the error value is greater than a preset error threshold, the dynamic configuration logic is enabled.
[0120] Specifically, in order to make the subsequent test results more credible, the actual online time information of the demand is included in the subsequent measurement process. For example, the actual time is compared with the measurement result. If the error is less than 0.5 weeks, the relevant data is included in the training set of the next round of measurement; if the error is greater than 0.5 weeks, a large error instruction is sent to guide human intervention, thereby starting the dynamic adjustment logic.
[0121] Figure 6 The flowchart of the dynamic configuration method according to the embodiment of the present disclosure is schematically shown.
[0122] like Figure 6 As shown, the dynamic configuration method of this embodiment includes operations S610 to S620, and operations S610 to S620 can at least partially execute the above-mentioned operation S530.
[0123] In operation S610 , an adjustment instruction from a person is received.
[0124] In operation S620, the adjustment instruction is executed, and the adjustment instruction includes performing a modification operation on the keyword library and / or the factor parameters, and the modification operation includes adding, deleting, and modifying.
[0125] Specifically, manual analysis can be performed to determine whether the factors affecting the time consumption are reasonable, and then the factors / keywords can be dynamically adjusted. For example, a configuration interface is reserved to support dynamic adjustment of the time value of existing influencing factors and dynamic update of the measurement model (including adding / deleting the time value of influencing factors and mapping relationships).
[0126] In the embodiments of the present disclosure, in the process of software product evolution, dynamic adjustment of factors affecting time is supported to meet the demand for faster and more accurate R&D launch as the R&D maturity model continues to improve.
[0127] Based on the above R&D cycle calculation method, the present disclosure also provides a R&D cycle calculation device. Figure 8 The device is described in detail.
[0128] Figure 8 The structural block diagram of the R&D cycle estimation device according to an embodiment of the present disclosure is schematically shown.
[0129] like Figure 8 As shown, the R&D cycle estimation device 800 of this embodiment includes a demand receiving module 810 , a preprocessing module 820 , a stage time consumption prediction value calculation module 830 and a summary module 840 .
[0130] The receiving module 810 is used to receive the requirement description information. In one embodiment, the receiving module 810 can be used to perform the operation S210 described above, which will not be described in detail here.
[0131] The preprocessing module 820 is used to perform preprocessing logic on the requirement description information to obtain the time factor parameters corresponding to the N R&D stages. In one embodiment, the preprocessing module 820 can be used to perform the operation S220 described above, which will not be repeated here.
[0132] The predicted value calculation module 830 for stage time consumption is used to calculate the time consumption of the R&D stage based on the time factor parameters corresponding to the N R&D stages through a preset machine learning model to obtain the predicted value of the time consumption of the N stages. In one embodiment, the predicted value calculation module 830 for stage time consumption can be used to perform the operation S230 described above, which will not be repeated here.
[0133] The summarizing module 840 is used to summarize the predicted values of the time consumption of the N stages to obtain a predicted value of the R&D cycle. In one embodiment, the summarizing module 840 can be used to perform the operation S240 described above, which will not be described in detail here.
[0134] In the embodiments of the present disclosure, in order to solve the technical problem that the quality of developed products is affected by inaccurate prediction of the R&D cycle. By extracting the factor parameters of the time required for multiple stages in the R&D process in the demand description and using the factor parameters as the input data of the model, the predicted time required for different R&D stages is predicted through a machine learning model. Finally, these predicted time requirements are summarized to calculate the total R&D cycle. The embodiments of the present disclosure can at least clarify the factors that affect the time required at each stage and establish a measurement model to make the scheduling of requirements from proposal to launch more accurate and reasonable, thereby ensuring the quality of the online demand.
[0135] According to an embodiment of the present disclosure, the preprocessing module includes a word segmentation unit, a keyword library acquisition unit, a matching unit and a time factor parameter formation unit. The word segmentation unit is used to perform intelligent word segmentation on the requirement description information for any R&D stage to obtain a word segmentation result set, and the word segmentation result set includes M word segmentation results, where M is a positive integer; the keyword library acquisition unit is used to obtain a keyword library corresponding to the R&D stage, and the keyword library includes K keywords, where K is a positive integer; the matching unit is used to perform pairwise matching of the word segmentation results and the keywords based on the M word segmentation results and the K keywords to obtain matching results, and the matching results include L successfully matched keywords, where L is a positive integer less than or equal to K; and the time factor parameter formation unit is used to form a time factor parameter based on the L successfully matched keywords.
[0136] According to an embodiment of the present disclosure, each of the K keywords corresponds to an unassigned time factor, and the time factor parameter forming unit includes a time factor parameter assignment subunit, which is used to assign time factors corresponding to the L successfully matched keywords to form the time factor parameter.
[0137] According to an embodiment of the present disclosure, the device further includes a sample acquisition module, a R&D cycle sample formation module and a training module, wherein the acquisition module is used to acquire training samples, wherein the training samples include samples of requirement description information and characterization parameters of stage duration, and the samples of requirement description information and the characterization parameters of stage duration are in one-to-one correspondence according to the R&D stages; the preprocessing module is used to execute the preprocessing logic on the samples of the requirement description information to obtain samples of the time factor parameters corresponding to the N R&D stages; the R&D cycle sample formation module is used to form training samples of the R&D cycle based on the characterization parameters of the stage duration, and the training samples of the R&D cycle include the training stage duration corresponding to the N R&D stages; and the training module is used to train the machine learning model based on samples of the N time factor parameters and the training stage duration corresponding to the N R&D stages.
[0138] According to an embodiment of the present disclosure, the training samples include historical training samples and training samples in use.
[0139] According to an embodiment of the present disclosure, the device further includes an actual value acquisition module, an error calculation module and a dynamic configuration module, wherein the actual value acquisition module is used to acquire the actual value of the R&D cycle; the error calculation module is used to calculate the error value based on the actual value of the R&D cycle and the predicted value of the R&D cycle; and the dynamic configuration module is used to start the dynamic configuration logic when the error value is greater than a preset error threshold.
[0140] According to an embodiment of the present disclosure, the dynamic configuration module includes an adjustment instruction receiving unit and an adjustment instruction execution unit, wherein the adjustment instruction receiving unit is used to receive adjustment instructions from personnel; and the adjustment instruction execution unit is used to execute the adjustment instructions, wherein the adjustment instructions include modifying the keyword library and / or the factor parameters, and the modification operations include adding, deleting and modifying.
[0141] According to an embodiment of the present disclosure, the characterization parameter of the stage time includes J sub-characterization parameters, J is a positive integer, and the R&D cycle sample formation module includes a mapping relationship acquisition unit, a sub-cycle derivation unit and a training sample summary unit of the R&D cycle, the mapping relationship acquisition unit is used to obtain a preset mapping relationship, and the preset mapping relationship includes a one-to-one mapping relationship between the sub-characterization parameters and the training R&D sub-stages; the sub-cycle derivation unit is used to derive J training R&D sub-stages based on the J sub-characterization parameters and the preset mapping relationship; and the training sample summary unit of the R&D cycle is used to summarize the J training R&D sub-stages to obtain the training samples of the R&D stage.
[0142] According to an embodiment of the present disclosure, any multiple modules of the demand receiving module 810, the preprocessing module 820, the predicted value calculation module 830 of the stage time consumption, and the summary module 840 can be combined in one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functions of one or more of these modules can be combined with at least part of the functions of other modules and implemented in one module. According to an embodiment of the present disclosure, at least one of the demand receiving module 810, the preprocessing module 820, the predicted value calculation module 830 of the stage time consumption, and the summary module 840 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, at least one of the demand receiving module 810, the preprocessing module 820, the stage time prediction value calculation module 830 and the summary module 840 can be at least partially implemented as a computer program module, and when the computer program module is executed, the corresponding function can be executed.
[0143] Fig. 9 A block diagram of an electronic device suitable for implementing a method for calculating a development cycle according to an embodiment of the present disclosure is schematically shown.
[0144] like Fig. 9 As shown, the electronic device 900 according to an embodiment of the present disclosure includes a processor 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage part 908 into a random access memory (RAM) 903. The processor 901 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 901 may also include an onboard memory for caching purposes. The processor 901 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.
[0145] In RAM 903, various programs and data required for the operation of electronic device 900 are stored. Processor 901, ROM 902 and RAM 903 are connected to each other via bus 904. Processor 901 performs various operations of the method flow according to the embodiment of the present disclosure by executing the program in ROM 902 and / or RAM 903. It should be noted that the program can also be stored in one or more memories other than ROM 902 and RAM 903. Processor 901 can also perform various operations of the method flow according to the embodiment of the present disclosure by executing the program stored in the one or more memories.
[0146] According to an embodiment of the present disclosure, the electronic device 900 may further include an input / output (I / O) interface 905, which is also connected to the bus 904. The electronic device 900 may further include one or more of the following components connected to the I / O interface 905: an input portion 906 including a keyboard, a mouse, etc.; an output portion 907 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 908 including a hard disk, etc.; and a communication portion 909 including a network interface card such as a LAN card, a modem, etc. The communication portion 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 910 as needed, so that a computer program read therefrom is installed into the storage portion 908 as needed.
[0147] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present disclosure is implemented.
[0148] According to an embodiment of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus, or a device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 902 and / or RAM 903 described above and / or one or more memories other than ROM 902 and RAM 903.
[0149] The embodiment of the present disclosure also includes a computer program product, which includes a computer program, and the computer program contains program code for executing the method shown in the flowchart. When the computer program product is run in a computer system, the program code is used to enable the computer system to implement the item recommendation method provided by the embodiment of the present disclosure.
[0150] The above functions defined in the system / device of the embodiment of the present disclosure are performed when the computer program is executed by the processor 901. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.
[0151] In one embodiment, the computer program may be based on a tangible storage medium such as an optical storage device, a magnetic storage device, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of a signal on a network medium, and downloaded and installed through the communication part 909, and / or installed from a removable medium 911. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.
[0152] In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 909, and / or installed from the removable medium 911. When the computer program is executed by the processor 901, the above functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the system, device, means, module, unit, etc. described above can be implemented by a computer program module.
[0153] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect through the Internet).
[0154] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0155] It will be appreciated by those skilled in the art that the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways, even if such combinations and / or combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments and / or claims of the present disclosure may be combined and / or combined in a variety of ways without departing from the spirit and teachings of the present disclosure. All of these combinations and / or combinations fall within the scope of the present disclosure.
[0156] The embodiments of the present disclosure are described above. However, these embodiments are only for illustrative purposes and are not intended to limit the scope of the present disclosure. Although the embodiments are described above separately, this does not mean that the measures in the various embodiments cannot be used in combination to advantage. The scope of the present disclosure is defined by the attached claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make a variety of substitutions and modifications, which should all fall within the scope of the present disclosure.
Claims
1. A method for calculating a research and development cycle, wherein the research and development cycle includes N research and development stages, where N is a positive integer. The method comprises: Receive requirement description information; Execute preprocessing logic on the requirement description information to obtain time factor parameters corresponding to the N R&D stages; Based on the time factor parameters corresponding to the N R&D stages, the time of the R&D stages is calculated through a preset machine learning model to obtain predicted values of the time of the N stages; as well as The predicted values of the time taken for the N stages are summarized to obtain a predicted value of the R&D cycle.
2. The method according to claim 1, wherein: The performing of preprocessing logic on the requirement description information to obtain the time factor parameters corresponding to the N R&D stages includes: For any R&D stage, intelligent word segmentation is performed on the requirement description information to obtain a word segmentation result set, wherein the word segmentation result set includes M word segmentation results, where M is a positive integer; Obtain a keyword library corresponding to the research and development stage, wherein the keyword library includes K keywords, where K is a positive integer; Based on the M word segmentation results and the K keywords, performing pairwise matching of the word segmentation results and the keywords to obtain matching results, wherein the matching results include L successfully matched keywords, where L is a positive integer less than or equal to K; and A time factor parameter is formed based on the L successfully matched keywords.
3. The method according to claim 2, wherein: Each of the K keywords corresponds to an unassigned time factor, The forming of the time factor parameter based on the L successfully matched keywords includes: The time factors corresponding to the L successfully matched keywords are assigned values to form the time factor parameters.
4. The method according to claim 2 or 3, wherein: The training method of the preset machine learning model includes: Acquire training samples, wherein the training samples include samples of requirement description information and characterization parameters of stage time, and the samples of requirement description information and the characterization parameters of stage time correspond to each other according to the R&D stages; Executing the preprocessing logic on the sample of the requirement specification information to obtain samples of the time factor parameters corresponding to the N research and development stages; Based on the characterization parameters of the stage durations, forming a training sample of a research and development cycle, wherein the training sample of the research and development cycle includes the training stage durations corresponding to the N research and development stages; and The machine learning model is trained based on samples of N timing factor parameters and training phase times corresponding to N research and development phases.
5. The method according to claim 4, wherein: The training samples include historical training samples and training samples in use.
6. The method according to claim 5, wherein: After the predicted values of the time taken for the N stages are summarized to obtain the predicted value of the R&D cycle, the method further includes: Get the actual value of the R&D cycle; calculating an error value based on an actual value of the development period and a predicted value of the development period; and When the error value is greater than a preset error threshold, the dynamic configuration logic is enabled.
7. The method according to claim 6, wherein: The dynamic configuration logic includes: Receive adjustment instructions from personnel; and The adjustment instruction is executed, and the adjustment instruction includes performing a modification operation on the keyword library and / or the factor parameters, and the modification operation includes adding, deleting and modifying.
8. The method according to claim 4, wherein: The characterization parameters of the stage time include J sub-characterization parameters, where J is a positive integer. The forming of training samples of the R&D cycle based on the characterization parameters of the stage time includes: Obtaining a preset mapping relationship, wherein the preset mapping relationship includes a one-to-one mapping relationship between the sub-characterization parameters and the training and development sub-stages; Based on the J sub-characterization parameters and the preset mapping relationship, deriving J training and development sub-stages; and The J training and development sub-stages are summarized to obtain training samples of the development stage.
9. A device for calculating a development cycle, wherein the development cycle includes N development stages, where N is a positive integer. The device comprises: A demand receiving module, used for receiving demand description information; A preprocessing module, used to execute preprocessing logic on the requirement description information to obtain time factor parameters corresponding to the N R&D stages; A module for calculating the predicted value of the stage time, which is used to calculate the time of the research and development stage based on the time factor parameters corresponding to the N research and development stages through a preset machine learning model to obtain the predicted value of the time of the N stages; as well as The summarizing module is used to summarize the predicted values of the time taken by the N stages to obtain a predicted value of the R&D cycle.
10. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to execute the method according to any one of claims 1 to 8.
11. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 8.