Similarity estimation device and similarity estimation method

Abstract

A similarity estimation device according to an embodiment comprises an operation log input unit, a screen-unit division unit, a GUI component position acquisition unit, a component arrangement density calculation unit, and a similarity calculation unit. The operation log input unit acquires a first operation log by a reference operator and a second operation log by an evaluation subject with respect to an application operable via a GUI screen. The screen-unit division unit analyzes the first operation log and the second operation log and divides the GUI screen into a plurality of screen units. The GUI component position acquisition unit acquires position information of an operated GUI component per each screen unit. The component arrangement density calculation unit calculates an arrangement density per each GUI component on the basis of the acquired position information, and estimates a work task section from a comparison with a predetermined threshold value. The similarity calculation unit calculates, per each screen unit, the work similarity between the reference operator and the evaluation subject on the basis of the work task section and a change in the arrangement density calculated from the first operation log, and the work task section and a change in the arrangement density calculated from the second operation log.

Description

Similarity estimation device and similarity estimation method

[0001] One aspect of the present invention relates to a similarity estimation device and a similarity estimation method.

[0002] A technology is known that acquires a log of mouse and keyboard operations by a user and calculates a frequency distribution (mean value, variance value) of key input accuracy and speed (see Patent Document 1). Applications of this technology to analyze computer operation logs and quantitatively evaluate the similarity of operation histories of different people are being considered. For example, in a task that involves operating a computer using a graphical user interface (GUI), similarity can be calculated based on indices such as a frequency distribution of key operation time and accuracy. Furthermore, for example, the similarity between an operation by a beginner and an operation by a reference expert can be estimated to evaluate the beginner's level of proficiency.

[0003] Here, proficiency can be evaluated by, for example, comparing the similarity of input speed (time required to input, time required to move between input fields, etc.). Task proficiency can be evaluated by, for example, comparing the similarity of input accuracy. Accuracy can be expressed by the number of times a specific unit is corrected by pressing backspace, the number of operations, or the number of questions asked to other staff members.

[0004] Patent No. 7184093

[0005] Existing technologies assume that operations by an expert and a novice are performed in the same order and with the same content when evaluating similarity. Therefore, unless the operations performed by an expert and a novice are almost completely identical, it is difficult to obtain a high evaluation. However, actual work involves ambiguity in tasks, such as different operational procedures leading to the same result and interruptions in operations. Existing technologies have difficulty taking such ambiguity into account, making it difficult to accurately evaluate similarity. This invention has been made in light of the above circumstances and aims to provide a technology that enables more accurate evaluation of similarity.

[0006] A similarity estimation device according to an embodiment includes an operation log input unit, a screen unit division unit, a GUI component position acquisition unit, a component placement density calculation unit, and a similarity calculation unit. The operation log input unit acquires a first operation log by a reference operator and a second operation log by an evaluation target user for an application operable via a GUI screen. The screen unit division unit analyzes the first operation log and the second operation log to divide the GUI screen into multiple screen units. The GUI component position acquisition unit acquires position information of GUI components operated for each screen unit. The component placement density calculation unit calculates placement densities for each GUI component based on the acquired position information and estimates task sections by comparing the calculated position information with a predetermined threshold. The similarity calculation unit calculates task similarities between the reference operator and the evaluation target user for each screen unit based on the transitions in placement density and task sections calculated from the first operation log and the transitions in placement density and task sections calculated from the second operation log.

[0007] According to one aspect of the present invention, the similarity can be evaluated more accurately.

[0008] FIG. 1 is a block diagram showing an example of a similarity estimation device according to an embodiment of the present invention. FIG. 2 is a diagram showing an example of a GUI screen of a business system. FIG. 3 is a diagram showing an example of an operation log. FIG. 4 is a diagram showing an example of a business task unit calculated from the operation log. FIG. 5 is a diagram showing an example of the relationship between the similarity estimation device according to the embodiment and the business system. FIG. 6 is a diagram showing an example of the flow of data in the similarity estimation device according to the embodiment.

[0009] In an embodiment, a similarity estimation device is disclosed that estimates the similarity between operations performed by different operators on a business system, for example. The business system is a computer on which an application operable via a GUI screen is installed, and the similarity estimation device estimates the similarity between the operators and the business system based on an operation log of the application.

[0010] 1 is a block diagram showing an example of a similarity estimation device according to an embodiment of the present invention. The similarity estimation device 1 is a computer including a processor 10, a memory 20, a storage 30, and an input / output unit 60. The input / output unit 60 accesses a database 200 constructed in the cloud, for example, to exchange various types of data.

[0011] The database 200 stores operation logs (operation logs 200a and 200b) of a plurality of operators for a business system. In the embodiment, the operation log 200a is an operation log (first operation log) of a reference operator (reference operator), and the operation log 200b is an operation log (second operation log) of an evaluation subject.

[0012] The processor 10 is an arithmetic device such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit), and realizes processing functions according to the embodiment in accordance with a program loaded from the storage 30 to the memory 20. The memory 20 is a semiconductor memory such as a ROM (Read Only Memory) or a RAM (Random Access Memory).

[0013] The storage 30 is a non-volatile memory such as a hard disk drive (HDD) or a solid state drive (SSD), and stores a program 30a for implementing the processing according to the embodiment in addition to basic software such as an operating system (OS). That is, the program 30a can be installed in the similarity estimation device 1. The storage 30 also stores a plurality of operation logs 30b and 30c.

[0014] The processor 10 includes, as processing functions according to an embodiment of the present invention, an operation log input unit 10a, a screen-unit division unit 10b, a GUI component position acquisition unit 10c, a component placement density calculation unit 10d, and a similarity calculation unit 10e. The processor 10 includes the operation log input unit 10a, the screen-unit division unit 10b, the GUI component position acquisition unit 10c, the component placement density calculation unit 10d, and the similarity calculation unit 10e. These functional blocks are realized by the processor 10 executing a program loaded into the memory 20.

[0015] The operation log input unit 10a acquires operation logs 200a and 200b from the database 200 via the input / output unit 60 and stores them (operation logs 30b and 30c) in the storage 30. The operation log 30b is the comparison standard, and the operation log 30c is the evaluation target.

[0016] The screen unit dividing unit 10b analyzes the acquired operation logs 30b and 30c and divides the GUI screen of the business system into a plurality of screen units.

[0017] The GUI component position acquisition unit 10c acquires position information of the operated GUI component for each screen unit set by the screen unit division unit 10b. Here, the position information of the GUI component is recorded in each of the operation logs 30b and 30c.

[0018] The component placement density calculation unit 10d calculates the placement density of each GUI component based on the position information acquired by the GUI component position acquisition unit 10c, and estimates the task section by comparing the calculated placement density with a predetermined threshold (density threshold) that is provided externally, for example.

[0019] The similarity calculation unit 10e calculates the transition in placement density and task sections individually from each of the operation logs 30b and 30c.The similarity calculation unit 10e then has a function of calculating and outputting task similarity for each task section from the transition in placement density of GUI components for the comparison standard and the evaluation target, and the task section.In other words, the similarity calculation unit 10e calculates the task similarity between the reference operator and the evaluation target for each screen.The calculated task similarity 300 is output via the input / output unit 60 and presented to an operator, etc.

[0020] FIG. 2 is a diagram showing an example of a GUI screen for a business system. For example, consider a GUI screen with two operation areas: "Search" and "Details." In FIG. 2, the "Search" area has input fields for "Member Code," "Member Name," and "Contact Information." After entering the necessary information and clicking the "Search" button, input fields for "Account," "Address," "Contact Information," and "Transaction Amount" are displayed in the "Details" area. After entering further necessary information and clicking the "Register" button, the necessary information is entered into the business system.

[0021] By evaluating the similarity between this series of operations by an expert and the operations by the subject, it is possible to evaluate, for example, the proficiency of the subject. In order to accurately evaluate the proficiency of the subject, it is important to be able to specify the similarity with high precision. Below, we will explain in detail the technology that can meet such needs.

[0022] 3 is a diagram showing an example of an operation log. The operation log is operation history information of terminal operations performed by an operator, such as mouse and keyboard operations, operations on an active window, and operations on GUI components within an application screen.

[0023] Figure 3(a) shows an example of an operation log by a beginner (evaluation subject), and Figure 3(b) shows an example of an operation log by an expert (reference operator). For each record in the operation log, the GUI component, its operation content, and its position in the vertical direction (Y axis) of the screen are recorded. For example, in record #4, the beginner and the expert performed different operations.

[0024] The placement density of GUI components is analyzed from each operation log. Placement density is a value that indicates the density of objects, and takes a high value when objects are in the same place or appear in close proximity.

[0025] Generally, operation components related to the same type of task are placed close to each other on the screen. Based on this, it is possible to evaluate the similarity of task units by comparing operations in a set space of GUI components.

[0026] 4A and 4B are diagrams showing an example of a task unit calculated from an operation log. Fig. 4A shows a graph corresponding to the operation log of a beginner, and Fig. 4B shows a graph corresponding to the operation log of an expert. In Fig. 4, the vertical axis scale of the graph for the beginner is enlarged to twice that of the expert. The horizontal axis of the graph is a number indicating the position on the screen (Y axis), corresponding to Fig. 3.

[0027] The line graph represents the change in the placement density of GUI components, and the bar graph represents the number of GUI components. The placement density of GUI components is calculated from the records of each of the operation logs 30b and 30c, and the section where the density continues at a certain threshold is defined as a business task unit. In Figure 4, two business task units AR1 and AR2 are obtained. Furthermore, by comparing the density changes in each business task unit AR1 and AR2, the similarity between the business tasks can be calculated.

[0028] Here, the comparison method can be any method that can numerically evaluate the difference in density transition, such as using the difference in placement density for each placement position or the area enclosed by the density transition and the line segment corresponding to the threshold value.

[0029] 5 is a diagram illustrating an example of the relationship between a similarity estimation device and a business system according to an embodiment. As shown in FIG. 5, the business system 100 is accessible from multiple operation terminals 400a and 400b and is provided as a system that can be used in common by multiple users. For example, an expert operates an application of the business system 100 using the operation terminal 400a, and an operation log 200a is obtained. Similarly, an evaluation target, who is the comparison target, operates an application of the business system 100 using the operation terminal 400b, and an operation log 200b is obtained.

[0030] The operation logs 200a and 200b are read by the similarity estimation device 1. The similarity estimation device 1 calculates the similarity for each task from the operation logs 200a and 200b, and estimates the similarity between the tasks performed by the expert and the evaluation target from the combination of task tasks and the similarity. The calculated task similarity 300 is output to the outside.

[0031] 6 is a diagram showing an example of a data flow in the similarity estimation device according to the embodiment. The operation logs 200a and 200b are read into the processor 10 by the operation log input unit 10a of the similarity estimation device 1 and passed to the screen-unit division unit 10b.

[0032] The screen-based division unit 10b analyzes the acquired operation logs 200a, 200b and classifies them into system screen units. The GUI component position acquisition unit 10c acquires position information of operated GUI components from the input screen-based operation logs.

[0033] The component placement density calculation unit 10d aggregates the position information of the GUI components and calculates the transition of the component placement density. Furthermore, the component placement density calculation unit 10d estimates the task section based on an externally specified "density threshold."

[0034] The similarity calculation unit 10e calculates task similarity for each task task section from the transition of the placement density of GUI components between the comparison standard and the comparison target, and the task task section.

[0035] The similarity can be calculated by density difference as shown in formula (1). When the value on the left side is small, the similarity is evaluated as high. Similarity = Σ Ｙ {Comparison reference placement density (Y) - Comparison target placement density (Y)} ２ ... (1) Alternatively, the similarity can be calculated by comparing the areas as shown in equation (2). As in (1), the smaller the value on the left side, the higher the similarity is evaluated to be. Similarity = Σ Ｙ {Comparison standard placement density (Y) - threshold} - Σ Ｙ {Comparison target placement density (Y) - threshold} ... (2)

[0036] As described above, in the embodiment, the similarity of terminal tasks is estimated by using the placement positions of the operated GUI components and evaluating the similarity of operation situations for each spatial unit in which each GUI component is placed. That is, in the embodiment, the placement density of GUI components recorded in the operation log is calculated based on the principle that "operation components related to similar tasks are placed in close positions on the screen." Then, a section where the placement density is continuous at a certain threshold is defined as a task unit, and the density transition within that section is compared to determine the similarity of task tasks. This makes it possible to evaluate the similarity of tasks even if the procedures and operation contents are not identical. In other words, even if the operation procedures and operation contents are different, the similarity of operations between an expert and a novice can be accurately evaluated. Ultimately, it becomes possible to evaluate the proficiency of a novice with high accuracy.

[0037] Furthermore, it is possible to evaluate the similarity in the task execution between operators, even for tasks that require complex operations that allow discretion. Furthermore, because it is possible to evaluate the similarity in task execution between both experts and novices, it is possible to grasp the proficiency level by determining whether the task of a novices approaches that of an expert. Furthermore, because it is possible to evaluate the similarity in the task execution between operators performing the same task, it is possible to extract different operation groups as wasteful tasks. For these reasons, according to the embodiment, it is possible to more accurately evaluate proficiency.

[0038] It should be noted that the present invention is not limited to the above-described embodiments. In other words, the present invention is not limited to the above-described embodiments as they are, and the components can be modified and embodied in the implementation stage without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriately combining multiple components disclosed in the above-described embodiments. For example, some components may be deleted from all the components shown in the embodiments. Furthermore, components from different embodiments may be appropriately combined.

[0039] 1...Similarity estimation device 10...Processor 10a...Operation log input unit 10b...Screen unit division unit 10c...GUI component position acquisition unit 10d...Component placement density calculation unit 10e...Similarity calculation unit 20...Memory 30...Storage 30a...Program 30b...Operation log 30c...Operation log 60...Input / output unit 100...Business system 200...Database 200a...Operation log 200b...Operation log 300...Business similarity 400a...Operation terminal 400b...Operation terminal AR1, AR2...Business task unit.

Claims

1. A similarity estimation device comprising: an operation log input unit that acquires a first operation log by a reference operator and a second operation log by a subject to evaluation for an application that can be operated via a GUI (Graphical User Interface) screen; a screen unit division unit that analyzes the first operation log and the second operation log and divides the GUI screen into a plurality of screen units; a GUI component position acquisition unit that acquires position information of GUI components operated for each screen unit; a component placement density calculation unit that calculates a placement density for each GUI component based on the acquired position information and estimates a task section by comparing it with a predetermined threshold; and a similarity calculation unit that calculates a task similarity between the reference operator and the subject to evaluation for each screen unit based on the transition of the placement density and the task section calculated from the first operation log and the transition of the placement density and the task section calculated from the second operation log.

2. The similarity estimation device according to claim 1, wherein the similarity calculation unit calculates the task similarity based on a difference in placement density for each placement position of the GUI components.

3. The similarity estimation device according to claim 1, wherein the similarity calculation unit calculates the task similarity using an area enclosed by a graph showing the placement density of the GUI components and a line segment corresponding to the threshold value.

4. A similarity estimation method comprising the steps of: a computer acquiring a first operation log by a reference operator and a second operation log by a subject to evaluation for an application operable via a GUI (Graphical User Interface) screen; a computer analyzing the first operation log and the second operation log and dividing the GUI screen into a plurality of screen units; a computer acquiring position information of GUI components operated for each screen unit; a computer calculating the placement density of each GUI component based on the acquired position information and estimating task sections by comparing with a predetermined threshold; and a computer calculating the task similarity between the reference operator and the subject to evaluation for each screen unit based on the transition in placement density and task sections calculated from the first operation log and the transition in placement density and task sections calculated from the second operation log.

Images