JPWO2020054073A1 - Analyzer, analysis method and program - Google Patents

Abstract

Provided is an analyzer 100 that analyzes a personalization level indicating the degree of personalization in a software development process. The analyzer 100 applies the personalization level prediction model 22 that receives the metric information 7 regarding the development process and outputs the personalization level based on the metric information to the metric information 7 to determine the personalization level. A control unit 1 for determining is provided.

Description

The present invention relates to an analyzer, an analysis method and a program for analyzing the personalization level indicating the degree of personalization in the software development process.

An analysis technique is known that analyzes source code and source code metrics and presents to users the parts that need to be modified, maintainability, and reusability of source code created by others. As a result, even a person other than a specific person in charge can modify and manage the source code.

For example, Patent Document 1 discloses an analyzer that presents to the user the priority of source code that needs to be modified by performing multivariate analysis of source code metrics and program development project information.

Patent Document 2 presents a standard and concrete technique for standardization of source code to a developer by using source code metric information, and evaluates the result of standardization carried out by the developer. We disclose a source code control system that can notify developers of the best implementation results.

Patent Document 3 discloses an object-oriented program quality evaluation device that extracts classes that lead to a decrease in maintainability and reusability by analyzing module-based metrics.

JP-A-2017-204090 Japanese Unexamined Patent Publication No. 2009-086814 Japanese Unexamined Patent Publication No. 11-338691

In software development, a personalized state may occur in which technology, progress management, construction period management, etc. depend on a specific person. When a personalized state occurs, if that particular person is suddenly absent, for example, another person may not understand the specifications, technology, construction period, etc., and development may not be able to proceed. Therefore, the occurrence of personalization can lead to lost opportunities.

The prior art can solve the personalized state of the coding process by presenting to the user the parts that need to be modified, maintainability, reusability, etc. of the source code created by another person. However, as described above, personalization in software development occurs not only in the coding process but also in various processes such as a specification design process, an external design process, and an internal design process.

The present invention has been made to solve the above problems, and an object of the present invention is to provide an analyzer, an analysis method, and a program for analyzing the personalization level indicating the degree of personalization in the software development process. To do.

One aspect of the present invention provides an analyzer that analyzes the level of personalization, which indicates the degree of personalization in the software development process. This analyzer is a control unit that receives metric information about the development process and applies a personalization level prediction model that outputs the personalization level based on the metric information to the metric information to determine the personalization level. And.

Another aspect of the present invention provides an analytical system that analyzes the level of personalization, which indicates the degree of personalization in the software development process. This analysis system is connected to a learning device that generates a personalization level prediction model based on sample information about metric information and the personalization level given to the sample information, and a learning device via a network. It includes a server that stores the personalization level prediction model received from the learning device, and a control device that is connected to the server via a network and receives the personalization level prediction model from the server. The control device receives the metric information about the development process and applies the personalization level prediction model to the metric information about the development process to determine the personalization level.

Another aspect of the present invention provides an analytical method for analyzing the level of personalization, which indicates the degree of personalization in the software development process. In this analysis method, the step of receiving the metric information about the development process by the control unit and the personalization level prediction model in which the control unit outputs the personalization level based on the metric information are applied to the metric information. Includes steps to determine the personalization level.

Yet another aspect of the present invention provides a program for causing the control unit to execute the above analysis method.

According to the present invention, it is possible to determine the personalization level of the development process based on the metric information regarding the software development process.

It is a block diagram which shows the structure of the analyzer which concerns on Embodiment 1 of this invention. It is a figure which shows the flow of data in the analyzer which concerns on Embodiment 1 of this invention. It is a flowchart which shows the flow of the analysis of the personalization level by the analyzer which concerns on Embodiment 1 of this invention. It is a figure which shows an example of a metric information database. It is a figure which shows an example of the personalization level prediction model. It is a block diagram which shows the structure of the analyzer which concerns on Embodiment 2 of this invention. It is a figure which shows the flow of data in the analyzer which concerns on Embodiment 2 of this invention. It is a flowchart which shows the flow of the analysis of the personalization level by the analyzer which concerns on Embodiment 2 of this invention. It is a flowchart which shows the flow of the analysis of the personalization level by the analyzer which concerns on Embodiment 2 of this invention. It is a figure which shows an example of a sample database. It is a figure which shows the flow of data in the analyzer which concerns on Embodiment 3 of this invention. It is a flowchart which shows the flow of the analysis of the personalization level by the analyzer which concerns on Embodiment 3 of this invention. It is a figure which shows the example of the display for confirming to the user whether or not the redefinition of the personalization level is necessary. It is a figure which shows the example of the dialog box for a user to enter a redefined personalization level.

Hereinafter, the analyzer according to the embodiment of the present invention will be described with reference to the drawings. In each embodiment, the same configuration is designated by the same reference numerals, and duplicate description will be omitted.

Embodiment 1.
[Constitution]
FIG. 1 is a block diagram showing a configuration of an analyzer 100 according to a first embodiment of the present invention. The analyzer 100 includes a control unit 1, an auxiliary storage unit 2, a main storage unit 3, an input unit 4, a display unit 5, and a communication interface (I / F) 6.

The control unit 1 is a processing device that performs various calculations and controls the entire operation of the analyzer 100. The control unit 1 includes a general-purpose processor such as a CPU that realizes a predetermined function by executing a program. The control unit 1 includes a metric analysis unit 11 that performs an operation for analyzing the metric information described later.

The auxiliary storage unit 2 is a medium for recording various information. Specifically, the auxiliary storage unit 2 is realized by a semiconductor memory device such as a ROM, a flash memory, an SSD, a disk device such as a hard disk, or other storage device alone or in combination thereof. The auxiliary storage unit 2 stores a metric information database (DB) 21, a personalization level prediction model 22, an operating system (OS), a control program executed by the control unit 1, and the like, which will be described later.

The main storage unit 3 is a medium for temporarily recording various information. The main storage unit 3 loads the control program stored in the auxiliary storage unit 2, and the control unit 1 sequentially reads and executes the loaded program. Further, the main storage unit 3 loads at least a part of the OS stored in the auxiliary storage unit 2, and the control unit 1 executes the loaded OS to realize the control function. The main storage unit 3 is composed of, for example, a volatile memory such as SRAM or DRAM.

The input unit 4 is an input interface that converts the contents of an input instruction or operation received from the user into an electric signal and transmits it to the control unit 1. The input unit 4 includes a mouse, a keyboard, a touch panel, buttons, a microphone for voice input, and the like.

The display unit 5 is a display device that displays various information. The display unit 5 is composed of, for example, a liquid crystal display, an organic EL display, a projector, or the like. Further, the analyzer 100 may include an output device such as a speaker or a printer in addition to the display unit 5.

The communication I / F6 is an interface circuit or module for enabling communication connection between the analyzer 100 and an external device via a network. Communication I / F6 communicates according to standards such as IEEE802.3, IEEE802.11 or Wi-Fi, LTE, 3G, 4G, 5G.

[motion]
FIG. 2 is a diagram showing a data flow in the analyzer 100. FIG. 3 is a flowchart showing the flow of analysis of the personalization level by the analyzer 100. The operation of the analyzer 100 will be described with reference to FIGS. 2 and 3.

Metric information 7 is input to the analyzer 100 via the input unit 4 (S101). However, this is an example, and the means for making the metric information 7 available to the control unit 1 is not limited to inputting the metric information 7 via the input unit 4. For example, the metric information 7 may be temporarily stored in an external storage device such as a server, and the analyzer 100 may download the metric information 7 via the communication I / F6.

Here, the metric is the quality and productivity of at least one of the requirements analysis, specification design, external design, internal design, coding, unit test, combination test, comprehensive test, system test, etc. in software development. , Refers to the criteria for evaluating reliability, etc.

The metric information 7 includes, for example, the number of pages of the document for explaining the development process, the document density, the ratio of figures and tables in the document, the review man-hours, the review density, the number of reviewers, the number of defect indications, the complexity of the source code, and the number of defects. , Information indicating the difficulty level of the process. Alternatively, for example, in the test design process, the metric information 7 is information indicating the number of testers, the test density, the defect detection rate, and the like.

Further, the metric information 7 may include a review density for each person in charge, a review speaking time for each person in charge, a defect detection rate for each person in charge, or a maximum value and a minimum value of these values.

The control unit 1 classifies the metric information 7 received via the input unit 4 into a database, and stores the metric information 7 in the metric information database 21 (S102). FIG. 4 is a diagram showing an example of the metric information database 21. If the development process can be decomposed into M stages, the metric information 7 can be organized in the form of metric groups (x ₁ , x ₂ , x ₃ , ..., x _k , ..., X _M ). x _k is a vector representing a group of metrics in the k-th stage development process (development process k). When the development process includes L metrics, it can be expressed as x _k = (x _k1 , x _k2 , x _k3 , ..., X _kL ). Therefore, the metric information database 21 is represented as a table as shown in FIG. 4, and can be represented in a matrix format.

Next, as shown in FIGS. 2 and 3, the metric analysis unit 11 reads the metric information database 21 and the personalization level prediction model 22, and displays the metric information 7 stored in the metric information database 21 as a person. The personalization level for each development process is calculated by applying it to the explanatory variable x of the conversion level prediction model 22 (S103).

FIG. 5 is a diagram showing an example of the personalization level prediction model 22. The personalization level prediction model 22 is a statistical model for obtaining the personalization level from the metric information 7. In FIG. 4, the personalization level prediction model 22 is composed of M multiple regression equations corresponding to each of M development processes. In the multiple regression equation, the objective variable y = (y ₁ , y ₂ , y ₃ , ..., y _k , ..., y _M ) representing the personalization level is the explanatory variable x = representing the metric information 7. It is represented by (x ₁ , x ₂ , x ₃ , ..., x _k , ..., x _M ) and the partial regression coefficients α and β. For example, the objective variable y _k of the development process k included in the personalization level prediction model 22 is represented by the following multiple regression equation (1) as shown in FIG.

y _k = α _k + β _k1 x _k1 + β _k2 x _k2 + β _k3 x _k3 + ... + β _kL x _kL (1)

The partial regression coefficients α and β are determined by performing multiple regression analysis, for example, based on sample data of metric information and metric information in the past development process. The type and number of metrics in each development process do not have to be the same. When the number of metrics acquired in the development process k is G, which is less than L, for example, in the equation (1), (LG) of the partial regression coefficients β _{k1 to} β _kL. The coefficient can be considered as 0.

The personalization level prediction model 22 may be stored in the auxiliary storage unit 2 from the outside via the input unit 4 of the analyzer 100 shown in FIG. Further, for example, the personalization level prediction model 22 may be downloaded from an external storage device such as a server via the communication I / F6 and stored in the auxiliary storage unit 2.

The control unit 1 displays the calculated personalization level for each development process on the display unit 5 (S104). As a result, the user can recognize the personalization level for each development process. The personalization level displayed on the display unit 5 represents the degree of personalization for each development process, for example, by a numerical value of 1 to 10.

The display unit 5 may display the personalization level of all development processes. Further, the display unit 5 may display the personalization levels of all the development processes in descending order of personalization level. As a result, the user can recognize the advanced development process of personalization just by looking at the display unit 5.

Alternatively, the calculated personalization level for each development process may be uploaded to an external storage device such as a server. Then, for example, the personalization level of the development process is downloaded to the PC used by the manager of a specific development process, so that the manager can confirm the personalization level of the development process managed by himself / herself. Can be done.

The display unit 5 may display the values of the explanatory variables x and the partial regression coefficients α and β. This allows the user to recognize the extent to which a particular metric affects a particular development process. For example, when β _k3 is the largest among the partial regression coefficients (β _k1 , β _k2 , β _k3 , ..., β _kL ) of the development process k represented by the equation (1), it is represented by x _k3. It can be seen that the metrics have a great influence on personalization.

[Effects, etc.]
As described above, the analyzer 100 analyzes the personalization level indicating the degree of personalization in the software development process. The control unit 1 of the analyzer 100 receives the metric information 7 regarding the software development process via the input unit 4. The control unit 1 applies the personalization level prediction model 22 that outputs the personalization level based on the metric information 7 to the metric information 7 to determine the personalization level in the development process.

With the analyzer 100, it is possible to obtain the personalization level of the software development process. This makes it possible to quantify the degree of personalization. Further, the analysis result by the analyzer 100 can be used for data processing for grasping the personalized state of the software development process, such as listing the development processes having a high personalization level.

The analyzer 100 may further include a display unit 5 that displays the personalization level determined by the control unit 1. This makes it possible to visualize the personalization level of the software development process. Visualization of the personalization level allows development process managers to appropriately undertake mitigation or elimination of personalization status. Therefore, for example, if a specific person involved in software development is suddenly absent, the risk that development cannot be carried out can be minimized.

Embodiment 2.
FIG. 6 is a block diagram showing the configuration of the analyzer 200 according to the second embodiment of the present invention. Compared to the first embodiment, the control unit 1 of the analyzer 200 further includes a learning unit 12 that generates a personalization level prediction model 22.

FIG. 7 is a diagram showing a data flow in the analyzer 200. 8 and 9 are flowcharts showing the flow of analysis of the personalization level by the analyzer 200. The operation of the analyzer 200 will be described with reference to FIGS. 7 to 9.

Metric information 7 is input to the analyzer 200 via the input unit 4 (S201). The control unit 1 classifies the metric information 7 received via the input unit 4 into a database, and stores the metric information 7 in the metric information database 21 (S202). Steps S201 and S202 are the same as steps S101 and S102 of the first embodiment.

In the second embodiment, the learning unit 12 of the control unit 1 generates the personalization level prediction model 22 (S203). Note that step S203 does not have to be executed after step S202, and may be executed before, for example, step S201.

FIG. 9 is a flowchart showing the details of step S203 for generating the personalization level prediction model 22. Hereinafter, step S203 for generating the personalization level prediction model 22 will be described with reference to FIG.

First, the learning unit 12 acquires a sample of metric information and a sample of personalization level from the sample database 8 via the input unit 4 (S211). FIG. 10 is a diagram showing an example of the sample database 8. In the sample database 8, the metrics group in the past development such as the development projects A, B, and C are stored in association with the personalization level given by the user. For example, the metric group x _Ak = (x _Ak1 , x _Ak2 , x _Ak3 , ..., X _AkL ) in the development process k of the development project A and the personalization level y _Ak given by the user to the development process k. Is stored in the sample database 8 as a pair.

The metric group stored in the sample database 8 is not limited to those acquired in the past development. For example, the metric group stored in the sample database 8 may be created by the user based on a theory, a rule of thumb, or the like.

Next, the learning unit 12 calculates the partial regression coefficient by the least squares method with the acquired personalization level as the objective variable and the metrics group as the explanatory variable for each development process (S212). The calculated partial regression coefficient is stored in the auxiliary storage unit 2 as the personalization level prediction model 22 (S213).

As described above, in the second embodiment, the personalization level prediction model 22 is generated by the learning unit 12 of the control unit 1 instead of being stored in the auxiliary storage unit 2 in advance as in the first embodiment. ..

Returning to FIG. 8, the metric analysis unit 11 of the control unit 1 reads the metric information database 21 and the personalization level prediction model 22 after step S203, and assigns the metric information 7 stored in the metric information database 21 to the genus. The personalization level for each development process is calculated by applying it to the explanatory variables of the humanization level prediction model 22 (S204). Next, the control unit 1 displays the calculated personalization level for each development process on the display unit 5 (S205). As a result, the user can recognize the personalization level for each development process.

As described above, the analyzer 200 includes the learning unit 12 in the control unit 1 in addition to the configuration of the analyzer 100 of the first embodiment. The learning unit 12 generates the personalization level prediction model 22 based on the sample of the metric information stored in the sample database 8 and the personalization level assigned to the sample. As described above, unlike the analyzer 100 of the first embodiment, the analyzer 200 can generate the personalization level prediction model 22 by itself.

Embodiment 3.
FIG. 11 is a diagram showing a data flow in the analyzer 300 according to the third embodiment of the present invention. Since the configuration of the analyzer 300 is the same as the configuration of the analyzer 200 shown in FIG. 6, the illustration is omitted.

FIG. 12 is a flowchart showing the flow of analysis of the personalization level by the analyzer 300. Hereinafter, the operation of the analyzer 300 will be described with reference to FIGS. 11 to 14.

Since steps S301 to S305 are the same as steps S201 to S205 of the second embodiment shown in FIG. 8, the description thereof will be omitted.

After step S305 for displaying the personalization level on the display unit 5, the control unit 1 displays on the display unit 5 a display for confirming to the user whether or not the personalization level needs to be redefined (S306). .. The display displayed on the display unit 5 in step S306 is, for example, a dialog box as shown in FIG.

The user confirms whether or not the personalization level displayed in step S305 is correct, and determines whether or not it is necessary to redefine the personalization level (S307). For example, if the user determines that the personalization level needs to be redefined, he or she clicks the "Yes" button in the dialog box as shown in FIG.

When the user determines that the personalization level needs to be redefined, the control unit 1 displays a dialog box as shown in FIG. 14, for example, on the display unit 5. The user inputs the redefined personalization level (hereinafter, referred to as “correction level”) 310 by using the input unit 4 such as a keyboard (S308).

Next, the control unit 1 associates the modification level 310 input by the user with the metric information and stores it in the sample database 8 (S309). After that, the process proceeds to step S303 again, and the learning unit 12 of the control unit 1 reads the updated sample database 8 and regenerates and updates the personalization level prediction model 22. After that, steps S304 to S307 are executed again.

As described above, in the third embodiment, the control unit 1 of the analyzer 300 acquires the correction level 310 given to the metric information 7 by the user. The learning unit 12 updates the personalization level prediction model 22 based on the metric information 7 and the modification level 310.

The modification level 310 is a personalization level redefined by the user when the personalization level calculated by the metric analysis unit 11 differs from the theory or the user's empirical rule. By providing such feedback, the analyzer 300 can improve the accuracy of the personalization level prediction model 22.

Modification example.
Although the present invention has been described above using a plurality of embodiments, the features described in the respective embodiments may be freely combined. In addition, various improvements, design changes and deletions may be added to the above embodiments, and there are various modifications of the present invention.

For example, in the above embodiment, an example in which the personalization level is calculated by performing multiple regression analysis based on sample data of metric information and metric information in the past development process has been described. However, the personalization level may be determined by multivariate analysis other than multiple regression analysis such as logistic regression analysis.

Looking at the second embodiment and the third embodiment, the learning unit 12 performs a multivariate analysis using the sample information related to the metric information and the personalization level given to the sample information, thereby performing the genus. The humanization level prediction model 22 may be generated.

Alternatively, the personalization level prediction model 22 uses a neural network, decision tree learning, a support vector machine, or the like to obtain sample data of metric information, metric information in the past development process, and the personalization level given by the user. It may be a learning model generated by learning.

Further, in each of the above embodiments, the analyzer including the control unit 1, the auxiliary storage unit 2, the main storage unit 3, the input unit 4, the display unit 5, and the communication I / F 6 has been described. However, these components do not have to be contained within one analyzer and may be, for example, separate instruments connected via a network. Therefore, in one aspect, the present invention includes a metric analysis device that performs metric analysis, a learning device that generates a personalization level prediction model 22, and a server that stores the generated personalization level prediction model 22. It may be a provided analysis system.

1 Control unit, 2 Auxiliary storage unit, 3 Main storage unit, 4 Input unit, 5 Display unit, 6 Communication interface, 7 Metrics information, 8 Sample database, 11 Metrics analysis unit, 12 Learning unit, 21 Metrics information database, 22 genus Humanized level prediction model, 100 analyzers.

Claims (10)

An analyzer that analyzes the level of personalization, which indicates the degree of personalization in the software development process.
A control unit that determines the personalization level by applying a personalization level prediction model that receives metric information related to the development process and outputs the personalization level based on the metric information to the metric information. When,
An analyzer equipped with. The control unit further includes a learning unit that generates the personalization level prediction model based on the sample information regarding the metric information and the personalization level given to the sample information, according to claim 1. Analysis equipment. The learning unit generates the personalization level prediction model by performing multivariate analysis using the sample information regarding the metric information and the personalization level given to the sample information, according to claim 2. The analyzer described. The analyzer according to claim 2, wherein the learning unit generates the personalization level prediction model by machine learning using the sample information regarding the metric information and the personalization level given to the sample information. .. The control unit further acquires the personalization level given to the metric information by the user.
The analyzer according to any one of claims 2 to 4, wherein the learning unit updates the personalization level prediction model based on the metric information and the personalization level given to the metric information. .. The analyzer according to any one of claims 1 to 5, further comprising a display unit for displaying the personalization level determined by the control unit. The analyzer according to any one of claims 1 to 6, further comprising a storage unit for storing the personalization level prediction model. An analysis system that analyzes the level of personalization, which indicates the degree of personalization in the software development process.
A learning device that generates a personalization level prediction model based on the sample information regarding the metric information and the personalization level given to the sample information.
A server connected to the learning device via a network and storing the personalization level prediction model received from the learning device.
It is provided with a control device that is connected to the server via a network and receives the personalization level prediction model from the server.
The control device is an analysis system that receives metric information about the development process, applies the personalization level prediction model to the metric information about the development process, and determines the personalization level. An analysis method that analyzes the level of personalization, which indicates the degree of personalization in the software development process.
A step in which the control unit receives metric information related to the development process,
A step in which the control unit applies a personalization level prediction model that outputs the personalization level based on the metric information to the metric information to determine the personalization level.
Analytical methods including. A program for causing the control unit to execute the analysis method according to claim 9.

Images