JP2021018607A - System for two-dimensionally displaying output result of learning model, method, program and recording medium - Google Patents

Abstract

To provide a system for two-dimensionally displaying an output result of a learning model, a method, a program and a recording medium.SOLUTION: A two-dimensional display system displays the n-th circle corresponding to the n-th class having the n-th largest proportion in the plurality of classes to be confirmed so as not to be superimposed on the (n-1)-th circle as the circle having the same radius as the second circle with the position apart from the center of the first circle by the center-to-center distance between the first circle 14 and the second circle 15 as the center on the basis of the image data formed on the basis of the information obtained by analyzing the output result, and displays identification marks 15a, 16a, 17a... indicating such a probability that unknown data belong to the class at a position obtained by proportionally dividing the (n-1)-th straight lines such that the ratio of the length to the center of the first circle to the length to the center of the n-th circle corresponds to the ratio of the proportion of the second class to the proportion of the first class on the (n-1)-th straight lines 30, 33-37... connecting the center of the first circle to the center of the n-th circle.SELECTED DRAWING: Figure 7

Description

The present invention relates to a system, a method, a program, and a recording medium capable of displaying the output result of the learning model.

Various learning models used for machine learning such as deep running are being developed every day. The learning model can be generated by using a known machine learning algorithm such as a neural network, and the determination accuracy can be improved by inputting various learning data into the machine learning algorithm. For example, Japanese Patent No. 5858875 has conventionally developed a system capable of improving determination accuracy, in other words, visualizing the growth process of artificial intelligence.

Japanese Patent No. 5885875

However, in the conventional learning model, it is impossible for the user to confirm the validity of the output result and the criterion for deriving such an output result only by showing the output result to the user, or It was difficult.
An object of the present invention is to eliminate such drawbacks in the prior art, and when data is input, the certainty that the input data belongs to each of a plurality of classes (categories) is determined for each class. For a learning model that can be output as a ratio (also called "discrimination score") between a plurality of classes, a system, a method, a program, and a recording medium that can display the output result of the learning model in two dimensions are provided. To do.

In order to solve the above problems, the system according to one aspect of the present invention is obtained when data is input to a trained learning model in which machine learning is performed using training data prepared for each of a plurality of classes. It is a two-dimensional display system that displays the output result to be obtained in two dimensions, and the learning model determines the certainty that the data belongs to each of the plurality of the classes when the data is input. It is output as a ratio between classes, and the system is a means for analyzing an output result obtained by inputting unknown data other than the training data into the training model, and a plurality of the above items to be confirmed. The classes are ordered according to the magnitude of the proportion, and the first circle corresponding to the first class having the highest proportion among the plurality of said classes to be confirmed and the second corresponding to the second largest class. The center of the two circles is determined, and the radius is set by dividing the first straight line connecting the center of the first circle and the center of the second circle, and further, of the first class. The means for determining the ratio of the ratio to the ratio of the second class, a means for forming image data based on the information obtained by the analysis, and a plurality of items to be confirmed based on the image data. The first circle corresponding to the first class having the largest ratio in the class and the second circle corresponding to the second class having the second largest ratio are the center of the first circle and the second circle. Each is displayed as a circle having a radius obtained by dividing the first straight line connecting the centers of the circles, and the certainty that the unknown data belongs to the class is shown in the first class and the second. The identification mark shown in relation to the class of the above is on the first straight line, and the ratio of the length to the center of the first circle and the length to the center of the second circle is the second. It features a two-dimensional display system comprising means for displaying the first straight line at proportionally divided positions so as to correspond to the ratio of the class ratio to the ratio of the first class.
According to the system of this aspect, the validity of the output result of the learning model can be intuitively grasped, and the determination criteria thereof can be easily grasped.

In the system of the above aspect, the means for analysis further determines the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of classes to be confirmed, and said the first. The means for obtaining and displaying the ratio of the ratio of one class to the ratio of the nth class is further, based on the image data, the ratio is the nth largest among the plurality of classes to be confirmed. The nth circle corresponding to the nth class is the same as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the first circle and the second circle. As a circle having a radius, it is displayed so as not to overlap with the n-1th circle, and the certainty that the unknown data belongs to the class is shown in the relationship between the first class and the nth class. The identification mark shown is on the n-1 straight line connecting the center of the first circle and the center of the nth circle, and is the length to the center of the first circle and the nth circle. The ratio of the length to the center corresponds to the ratio of the second class to the ratio of the first class or the ratio of the nth class to the ratio of the first class. The straight line of the n-1th line may be displayed at a proportionally divided position (n is an integer of 3 ≦ n ≦ (the number of classes to be confirmed)).

Further, in order to solve the above-mentioned problems, the system according to another aspect of the present invention inputs data into a trained learning model in which machine learning is performed using training data prepared for each of a plurality of classes. It is a two-dimensional display system that displays the output result obtained in two dimensions, and the learning model determines the certainty that the data belongs to each of the plurality of the classes when the data is input. It is output as a ratio among the plurality of classes, and the system is a means for analyzing the output result obtained by inputting unknown data other than the training data into the training model, and confirms it. The plurality of the classes to be input are ordered according to the magnitude of the ratio, and the first circle corresponding to the first class having the largest ratio among the plurality of the classes to be confirmed and the second class having the second largest ratio. Determine the center of the second circle corresponding to and the third circle corresponding to the third largest class, and draw a straight line connecting the center of the first circle and the center of the second circle. The radius is set by dividing, and the ratio of the ratio of the first class to the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class are obtained. Means, means for forming image data based on the information obtained by the analysis, and a first class having the largest ratio among the plurality of classes to be confirmed based on the image data. The corresponding first circle and the second circle corresponding to the second largest second class and the third circle corresponding to the third largest third class are the first circle and the second. The circle is a circle having a radius obtained by dividing a straight line connecting the center of the first circle and the center of the second circle, and the third circle is the first circle. The second circle and the third circle are each other as a circle having the same radius as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the second circle. In addition to displaying each so as not to overlap, the identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the second class, and the third class is displayed. The straight line is drawn so that the ratio of the length to the center of the first circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. The radius of the length corresponding to the distance between the proportionally divided position and the center of the first circle. The second arc is concentric with the first circle and can be formed between the center of the second circle and the center of the third circle. The ratio of the length of the arc to the center of the circle and the length of the arc to the center of the third circle is apportioned so as to correspond to the ratio of the ratio of the third class to the ratio of the second class. It features a two-dimensional display system including means for displaying at a position corresponding to the changed position.
According to the system of this aspect, the validity of the output result of the learning model can be intuitively grasped, and the determination criteria thereof can be easily grasped.

In the system of the above aspect, the means for analysis further determines the center of the nth circle corresponding to the nth class having the nth largest proportion among the plurality of classes to be confirmed, and said the first. The means for obtaining and displaying the ratio of the ratio of the n class to the ratio of the n-1th class is further, based on the image data, the ratio is the nth among the plurality of classes to be confirmed. The nth circle corresponding to the large nth class is centered on a position separated from the center of the first circle by the distance between the centers of the first circle and the second circle. As a circle having the same radius as, the circle is displayed so as not to overlap with the n-1 circle, and the relationship between the first class, the n-1 class, and the nth class is described. The ratio of the length to the center of the first circle and the length to the center of the second circle is the ratio of the second class and the identification mark indicating the certainty that the unknown data belongs to the class. Concentric with the first circle having a radius of length corresponding to the distance between the position where the straight line is proportionally divided so as to correspond to the ratio to the ratio of the first class and the center of the first circle. An arc concentric with the first circle that is on the arc and can be formed between the center of the n-1 circle and the center of the nth circle to the center of the n-1 circle. The ratio of the length of the arc to the center of the nth circle is proportionally divided so as to correspond to the ratio of the ratio of the nth class to the ratio of the n-1 class. It may be displayed at the corresponding position (n is an integer of 4 ≦ n ≦ (number of classes to be confirmed)).

Further, in the means for analyzing the system of the above aspect, the radius of the first circle and the radius of the second circle are obtained by dividing the first straight line into two, and the radius of the second circle is obtained. The radius may be the same as the radius of the first circle or smaller than the radius of the first circle.

Further, in the display means of the system of the above aspect, the first circle and the nth circle may be displayed in a state of being substantially in contact with each other at one point on their circumferences (n is 2). Any of the above integers).
Further, in the display means of the system of the above aspect, the nth circle and the n + 1 circle may be displayed in a state of being substantially in contact with each other at one point on their circumferences (n is 2). Any of the above integers).

Further, in the system of the above aspect, the ratio of the n-1 class may be the same as the ratio of the n class (n is any one of 2 ≦ n ≦ (the number of classes to be confirmed). Integer).

Further, in the system of the above aspect, it is preferable to further provide a display changing means for changing the display method of the identification mark according to the size of the ratio.
Further, in the system of the above aspect, the identification mark may be displayed at different times in a plurality of the classes selected as display targets by the display changing means. ..
In the system of the above aspect, the identification mark may be displayed in a size corresponding to the size of the ratio by the display changing means.
In the system of the above aspect, the identification mark may be blinked and displayed at a speed corresponding to the magnitude of the ratio by the display changing means.

In the system of the above aspect, the means for analysis may exclude the class whose ratio is equal to or less than a predetermined threshold value from the display target.

In the system of the above aspect, the circle indicating the evaluation criterion for evaluating the identification mark may be displayed concentrically with the first circle by the display changing means.

In the system of the above aspect, the circles other than the first circle may be displayed in the vicinity of the center of each circle by a method other than the circle by the display changing means, or nothing may be displayed.

In the system of the above aspect, it is preferable that the learning data includes data for creating the learning model and data for verifying the learning model.

In order to solve the above problems, the method according to one aspect of the present invention is obtained when data is input to a trained learning model in which machine learning is performed using training data prepared for each of a plurality of classes. It is a two-dimensional display method that displays the output result to be obtained in two dimensions, and the learning model determines the certainty that the data belongs to each of the plurality of the classes when the data is input. It is output as a ratio between classes, and the method is a stage of analyzing an output result obtained by inputting unknown data other than the training data into the training model, and the plurality of said methods to be confirmed. The classes are ordered according to the magnitude of the proportion, and the first circle corresponding to the first class having the highest proportion among the plurality of said classes to be confirmed and the second corresponding to the second largest class. The center of the two circles is determined, and the radius is set by dividing the first straight line connecting the center of the first circle and the center of the second circle, and further, of the first class. The step of determining the ratio of the ratio to the proportion of the second class, the step of forming image data based on the information obtained by the analysis, and a plurality of steps to be confirmed based on the image data. The first circle corresponding to the first class having the largest ratio in the class and the second circle corresponding to the second class having the second largest ratio are the center of the first circle and the second circle. Each is displayed as a circle having a radius obtained by dividing the first straight line connecting the centers of the circles, and the certainty that the unknown data belongs to the class is shown in the first class and the second. The identification mark shown in relation to the class of the above is on the first straight line, and the ratio of the length to the center of the first circle and the length to the center of the second circle is the second. It is characterized by including a step of displaying the first straight line at a proportionally divided position so as to correspond to the ratio of the class ratio to the first class ratio.
According to the method of this aspect, the validity of the output result of the learning model can be intuitively grasped, and the determination criteria thereof can be easily grasped.

In the method of the above aspect, at the stage of the analysis, the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is further determined, and the said first. At the stage of obtaining the ratio of the ratio of the 1st class to the ratio of the nth class and displaying the ratio, the ratio is the nth largest among the plurality of the classes to be confirmed based on the image data. The nth circle corresponding to the nth class is the same as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the first circle and the second circle. As a circle having a radius, it is displayed so as not to overlap with the n-1th circle, and the certainty that the unknown data belongs to the class is shown in the relationship between the first class and the nth class. The identification mark shown is on the n-1 straight line connecting the center of the first circle and the center of the nth circle, and is the length to the center of the first circle and the nth circle. The ratio of the length to the center corresponds to the ratio of the second class to the ratio of the first class or the ratio of the nth class to the ratio of the first class. The straight line of the n-1th line may be displayed at a proportionally divided position (n is an integer of 3 ≦ n ≦ (the number of classes to be confirmed)).

Further, in order to solve the above-mentioned problems, in the method according to another aspect of the present invention, data is input to a trained learning model in which machine learning is performed using training data prepared for each of a plurality of classes. It is a two-dimensional display method that displays the output result obtained at the time of the operation in two dimensions, and the learning model determines the certainty that the data belongs to each of the plurality of the classes when the data is input. It is output as a ratio among the plurality of classes, and the method is a stage of analyzing an output result obtained by inputting unknown data other than the training data into the training model, and confirms the result. The plurality of the classes to be input are ordered according to the magnitude of the ratio, and the first circle corresponding to the first class having the largest ratio among the plurality of the classes to be confirmed and the second class having the second largest ratio. Determine the center of the second circle corresponding to and the third circle corresponding to the third largest class, and draw a straight line connecting the center of the first circle and the center of the second circle. The radius is set by dividing, and the ratio of the ratio of the first class to the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class are obtained. The step, the step of forming image data based on the information obtained by the analysis, and the first class having the largest ratio among the plurality of classes to be confirmed based on the image data. The corresponding first circle and the second circle corresponding to the second largest second class and the third circle corresponding to the third largest third class are the first circle and the second. The circle is a circle having a radius obtained by dividing a straight line connecting the center of the first circle and the center of the second circle, and the third circle is the first circle. The second circle and the third circle are each other as a circle having the same radius as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the second circle. In addition to displaying each so as not to overlap, the identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the second class, and the third class is displayed. The straight line is drawn so that the ratio of the length to the center of the first circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. The first having a radius of length corresponding to the distance between the apportioned position and the center of the first circle. An arc concentric with the first circle, which is on an arc concentric with the circle 1 and can be formed between the center of the second circle and the center of the third circle, is the arc of the second circle. The ratio of the length of the arc to the center to the length of the arc to the center of the third circle is proportionally divided so as to correspond to the ratio of the ratio of the third class to the ratio of the second class. It is characterized by having a stage of displaying at a corresponding position.
According to the method of this aspect, the validity of the output result of the learning model can be intuitively grasped, and the determination criteria thereof can be easily grasped.

In the method of the above aspect, at the stage of the analysis, the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is further determined, and the first At the stage of obtaining the ratio of the ratio of the n class to the ratio of the n-1th class and displaying the ratio, the ratio is the nth among the plurality of classes to be confirmed based on the image data. The nth circle corresponding to the large nth class is centered on a position separated from the center of the first circle by the distance between the centers of the first circle and the second circle. As a circle having the same radius as, the circle is displayed so as not to overlap with the n-1 circle, and the relationship between the first class, the n-1 class, and the nth class is described. The ratio of the length to the center of the first circle and the length to the center of the second circle is the ratio of the second class and the identification mark indicating the certainty that the unknown data belongs to the class. Concentric with the first circle having a radius of length corresponding to the distance between the position where the straight line is proportionally divided so as to correspond to the ratio to the ratio of the first class and the center of the first circle. An arc concentric with the first circle that is on the arc and can be formed between the center of the n-1 circle and the center of the nth circle to the center of the n-1 circle. The ratio of the length of the arc to the center of the nth circle is proportionally divided so as to correspond to the ratio of the ratio of the nth class to the ratio of the n-1 class. It may be displayed at the corresponding position (n is an integer of 4 ≦ n ≦ (number of classes to be confirmed)).

In order to solve the above problems, the program according to one aspect of the present invention is obtained when data is input to a trained learning model in which machine learning is performed using training data prepared for each of a plurality of classes. It is a two-dimensional display program that displays the output result to be output in two dimensions, and the learning model determines the certainty that the data belongs to each of the plurality of the classes when the data is input. It is output as a ratio between classes, and the program is a function to analyze the output result obtained by inputting unknown data other than the training data into the training model in the computer, and should be confirmed. The plurality of the classes are ordered according to the magnitude of the proportion, and the first circle corresponding to the first class having the largest proportion among the plurality of the classes to be confirmed and the second class having the second largest proportion. The center of the corresponding second circle is determined, and the radius is set by dividing the first straight line connecting the center of the first circle and the center of the second circle, and further, the first The function of obtaining the ratio of the ratio of the class to the ratio of the second class, the function of forming image data based on the information obtained by the analysis, and the confirmation based on the image data. The first circle corresponding to the first class having the largest proportion among the plurality of said classes to be power and the second circle corresponding to the second largest class are defined as the center of the first circle. The circles having a radius obtained by dividing the first straight line connecting the centers of the second circles are displayed as circles, and the certainty that the unknown data belongs to the class is referred to as the first class. The identification mark shown in relation to the second class is the ratio of the length to the center of the first circle and the length to the center of the second circle on the first straight line. It is characterized by a two-dimensional display program for realizing a function of displaying the first straight line at a proportionally divided position so as to correspond to the ratio of the ratio of the second class to the ratio of the first class.
According to the program of this aspect, the validity of the output result of the learning model can be intuitively grasped, and the determination criteria thereof can be easily grasped.

In the program of the above aspect, the computer, in the function of the analysis, further determines the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed. In the function of obtaining the ratio of the ratio of the first class to the ratio of the nth class, and in the function of displaying, the ratio is further among the plurality of classes to be confirmed based on the image data. The second circle, which corresponds to the nth largest class n, is centered on a position separated from the center of the first circle by the distance between the centers of the first circle and the second circle. As a circle having the same radius as the circle of, it is displayed so as not to overlap with the n-1th circle, and the certainty that the unknown data belongs to the class is shown in the first class and the nth class. The identification mark shown in relation to is on the n-1 straight line connecting the center of the first circle and the center of the nth circle, and the length to the center of the first circle and the first. The ratio of the length to the center of the circle of n is the ratio of the ratio of the second class to the ratio of the first class or the ratio of the ratio of the nth class to the ratio of the first class. A function of displaying the n-1th straight line at a proportionally divided position (n is an integer of 3 ≦ n ≦ (the number of classes to be confirmed)) may be realized.

In addition, in order to solve the above problem, the output result obtained when data is input to the trained learning model in which machine learning is performed using the training data prepared for each of a plurality of classes is displayed in two dimensions. A two-dimensional display program for displaying, when data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class. The program is a function of analyzing an output result obtained by inputting unknown data other than the training data into the training model in the computer, and the ratio of a plurality of the classes to be confirmed. A first circle corresponding to the first class having the highest ratio among the plurality of said classes to be confirmed and a second circle corresponding to the second largest class. The center of the third circle corresponding to the third largest third class is determined, and the radius is set by dividing the straight line connecting the center of the first circle and the center of the second circle. Further, the function of obtaining the ratio of the ratio of the first class to the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class is obtained by the analysis. The function of forming image data based on the obtained information, and the first circle and 2 corresponding to the first class having the largest ratio among the plurality of classes to be confirmed based on the image data. The second circle corresponding to the second largest class and the third circle corresponding to the third largest class, and the first circle corresponding to the first circle and the second circle. As a circle having a radius obtained by dividing a straight line connecting the center of the second circle and the center of the second circle, and for the third circle, between the center of the first circle and the center of the second circle. As a circle centered on a position away from the center of the first circle by a distance and having the same radius as the second circle, the second circle and the third circle are displayed so as not to overlap each other. At the same time, an identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the second class, and the third class is placed up to the center of the first circle. The position where the straight line is proportionally divided so that the ratio of the length to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class and the first. The first having a radius of length corresponding to the distance from the center of the circle. An arc concentric with the first circle that is concentric with the circle and can be formed between the center of the second circle and the center of the third circle to the center of the second circle. The ratio of the length of the arc to the center of the third circle corresponds to the position prorated so as to correspond to the ratio of the ratio of the third class to the ratio of the second class. It features a two-dimensional display program for realizing a function of displaying at a position.
According to the program of this aspect, the validity of the output result of the learning model can be intuitively grasped, and the determination criteria thereof can be easily grasped.

In the program of the above aspect, the computer, in the function of the analysis, further determines the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed. In the function of obtaining the ratio of the ratio of the nth class to the ratio of the n-1 class, the function of displaying, and further, among the plurality of classes to be confirmed based on the image data, the said The nth circle corresponding to the nth class having the nth largest ratio is centered on a position separated from the center of the first circle by the distance between the centers of the first circle and the second circle. As a circle having the same radius as the second circle, it is displayed so as not to overlap with the n-1 circle, and the first class, the n-1 class, and the nth class are displayed. In this relationship, the identification mark indicating the certainty that the unknown data belongs to the class is the ratio of the length to the center of the first circle to the length to the center of the second circle in the second class. The first having a radius of length corresponding to the distance between the position where the straight line is proportionally divided so as to correspond to the ratio of the ratio to the ratio of the first class and the center of the first circle. An arc concentric with the first circle, which is on an arc concentric with the circle and can be formed between the center of the n-1 circle and the center of the nth circle, is the n-1. The ratio of the length of the arc to the center of the circle to the length of the arc to the center of the nth circle corresponds to the ratio of the ratio of the nth class to the ratio of the n-1 class. The function of displaying at the position corresponding to the proportionally divided position (n is an integer of 4 ≦ n ≦ (the number of classes to be confirmed)) may be realized.

In order to solve the above problems, according to one aspect of the present invention, a computer-readable recording medium on which a program is recorded is provided.

According to the present invention, it is possible to provide a system, a method, a program, and a recording medium capable of displaying the output result of a learning model in two dimensions, which eliminates the above-mentioned drawbacks in the prior art.

It is a figure which showed roughly the manufacturing method of the learning model which can be used for the system by one Embodiment of this invention. It is a figure which shows typically an example of the hardware composition of the system by one Embodiment of this invention together with a learning model and a display device which are peripheral devices. It is a figure which shows an example of the display which can be displayed on a display device using a system. It is a flowchart of the operation executed by the analysis part. It is a flowchart of the operation executed by the analysis unit to perform the linear display. It is a figure which shows an example of the linear display when it is determined that there are two display classes. It is a figure which shows an example of the linear display when it is determined that the display class is 3 or more. It is a figure which shows an example of the linear display when it is determined that the display class is 3 or more. It is a figure which shows an example of the linear display when it is determined that the display class is 3 or more. It is a figure which shows an example of the linear display when it is determined that the display class is 3 or more. It is a flowchart of the operation executed by the analysis part to perform the arc display. It is a figure which shows an example of an arc display. It is a figure which shows an example of an arc display.

Hereinafter, one preferred embodiment of the present invention will be described with reference to the accompanying drawings. Only suitable embodiments are shown for convenience of explanation, but of course, this is not intended to limit the invention.

FIG. 1 is a diagram schematically showing a method for producing a learning model 1 that can be used in a system according to an embodiment of the present invention. In the learning model 1, for example, which fruit image data such as apple, banana, grape, etc. is the input image data, in other words, which "class (category)" the image data belongs to. Can be output, and the certainty of belonging to each "class" can be output as a ratio among a plurality of classes for each class. Of course, the learning model 1 is not limited to the determination of fruits, and various types of learning models such as those for diagnosing a disease name from a roentgen photograph of a viewer and those for identifying a person from a fingerprint or a face. Is included.

These learning models 1 can be produced by using, for example, a known machine learning algorithm such as a neural network. The learning data (teacher data) to be input to the machine learning algorithm is prepared in advance by the creator of the learning model 1. For example, in the learning model 1 used for determining the fruit shown in FIG. 1, the training data is an image related to the fruit, such as image data of various fruits photographed and image data of paintings of various fruits by hand. It will be data. FIG. 1 shows, as an example, an apple image data 11A, a banana image data 11B, and a grape image data 11C. At least 1000 image data shall be prepared for each class. A correct answer score is associated with these image data. The associated correct answer score has a numerical parameter from 0 to 1. The correct answer score is basically evaluated and set by human hands for each of the image data 11A, 11B, and 11C. As the training data, in addition to the production data for producing the learning model, the verification data for verifying the learning model produced by using the production data can also be used. By using the verification data, the accuracy of the learning model 1 can be further improved. When the learning data is input to the learned learning model learned using the learning data, an output result having a correct answer rate of almost 100% is basically obtained.

FIG. 2 schematically shows an example of the hardware configuration of the system 2 according to the embodiment of the present invention together with the learning model 1 as a peripheral device.

When "unknown data", that is, "data other than the training data used to create the trained learning model 1 and to verify it" is input, the learning model 1 is "unknown". It outputs which fruit, for example, apple, banana, grape, etc., that is, belongs to the "class", and further, the certainty of belonging to each "class" is determined between multiple classes for each class. Output as "ratio" in. In the example shown in FIG. 2, the learning model 1 determines that a certain "unknown data" has a high proportion of belonging to any of the classes "apple", "banana", and "grape", and these An example is shown in which the certainty of belonging to each of the three classes is output as 0.6, 0.3, and 0.1 as the ratios among the three classes for each class. In addition, "No." in the table shown in FIG. 2 is merely a display for convenience (hereinafter, the same applies).

By using system 2, the certainty that such "unknown data" belongs to each of a plurality of classes (in this example, apples, bananas, grapes, etc.) can be determined for each class and the ratio among the plurality of classes. In particular, according to the system 2, these output results can be displayed on the display device 5 in two dimensions, and thus in a manner that allows the user to intuitively grasp the output results. it can.

The display device 5 is sufficient as long as it can display based on image data and can display the output result of the learning model 1 in two dimensions, and is a commonly used display or tablet. , Other visuals Various devices can be used.

FIG. 3 shows an example of a display that can be displayed on the display device 5 by using the system 2. FIG. 3A is an overall image of the display, and FIG. 3B is a partially enlarged view thereof. This display example is a display example of "unknown data" including a plurality of image data considered to be the closest to the "apple", and each class is displayed as one circle 14 to 20, respectively. There is. The circle 14 corresponding to the class with the highest certainty, in other words, the class with the highest ratio among multiple classes, the first circle 14 corresponding to "apple" in this example is the other class "Bana". It is placed in the center of the second circle 15 and the third circle 16 corresponding to the grapes and "grape", and such a display intuitively indicates that it is most likely to be an "apple". It has become a thing. Further, a plurality of image data included in the "unknown data" make the "identification marks" 14a, 15a, 16a corresponding to each image data a picture of an "apple" having the highest certainty. The pictures are displayed at positions away from the center of the first circle 14 according to the ratio among the plurality of classes. In the example of FIG. 3, the closer to the center of the first circle 14, the greater the certainty that the identification data is an “apple”. Furthermore, although the certainty of being an "apple" is the highest, it corresponds to "banana" when the certainty of being a "bana" or "grape" other than "apple" is also relatively high. The identification marks 15a, 16a, and 17a are displayed at positions relatively close to the center of the second circle 15 and the third circle 16 corresponding to the "grape", respectively. In this way, by displaying the output result of the learning model 1 in two dimensions using the system 2, the user can intuitively grasp the validity of the output result of the learning model 1. Further, as is well shown in FIG. 3 (b), additional circles or areas indicating evaluation criteria for evaluating the identification mark, such as "pass area" and "fail area", and further, their By displaying the "semi-passing area" provided between them concentrically with the first circle 14, it is possible to easily grasp the degree to which each "identification mark" belongs to a certain class based on a predetermined criterion. Can be done.

See FIG. 2 again. The system 2 mainly includes an input unit 3 and a main body 4. The input unit 3 is sufficient as long as it can input instructions from the user to the main body 4, and includes various input interfaces such as a keyboard, a display, and a scanner. The user can instruct at least the following items through the input unit 3.
1) "Class to be confirmed", that is, among a plurality of classes, the class to be displayed on the display device 5, in other words, the validity of the output result of the learning model 1 is confirmed by displaying on the display device 5. A class determined by the user to be what to do,
2) Among the "classes to be confirmed", the "threshold value", which is the value of the ratio for selecting the class for which the user has determined that the display on the display device 5 may be omitted because the ratio value is low,
3) Display method in the display device 5, for example, the color and pattern of the circle representing each class, the size of the radius of the circle, the necessity of displaying the circle of the class other than the class having the largest ratio, the color and shape of the identification mark. , And size, how to display the identification mark, etc.
4) Other useful indications, such as the criteria shown in FIG.

The main body 4 generates display data necessary for display on the display device 5 based on the instruction content input from the input unit 3 and the output result obtained from the learning model 1. Based on this display data, a two-dimensional display is performed on the display device 5.

The main body 4 has a general computer configuration, for example, a memory, a controller, a bus, a communication interface, a CPU 42 that executes a program that is software that realizes each function, and the computer (or CPU) can read the program and various data. A ROM or storage device recorded in the computer, a RAM for developing the program, etc. (referred to as “recording medium”) 41, and further, a receiving unit 43, an analysis unit 200, an image forming unit 300, and a display changing unit 400. , And the output unit 44.

The receiving unit 43 receives various data from the input unit 3 and the learning model 1.

FIG. 4 shows a flowchart of the operation executed by the analysis unit 200. The analysis unit 200 analyzes the information (S202) acquired through the reception unit 43, for example, the instruction content input through the input unit 3 and the output result obtained from the learning model 1, using the CPU 42 or the like, and displays an image. Get the information you need to do. At the time of analysis, first, the "class to be confirmed", that is, the class to be displayed on the display device 5 among the plurality of classes is confirmed (S203). Classes other than "classes to be confirmed", that is, classes that do not need to be displayed, are basically not processed after that. Next, when the "threshold value", that is, the value of the ratio is low through the input unit 3 or the like, and the value of the ratio for selecting the class determined by the user that the display on the display device 5 may be omitted is set. To compare each percentage of "classes to be confirmed" with such "thresholds" (S204). Classes below the "threshold value" are excluded from the display target (S205), and no further processing is performed. For the classes to be displayed even after the above processing (hereinafter referred to as "display class"), their "ratio" are compared with each other and ordered according to the magnitude of the "ratio" (S206). .. At the time of ordering, the sizes of adjacent "ratio" may be the same, and in this case, one of the classes may be treated as having a large or small "ratio". Next, it is determined whether or not there are two or more "display classes" (S207). If there are not two or more, that is, if there is one display class, the process ends. On the other hand, when there are two or more, it is further determined whether the display class is three or more (S208). If there are not three or more, that is, if there are two display classes, the process proceeds to "A". On the other hand, when there are three or more, it is determined whether or not "straight line display" is instructed through the input unit 3 or the like (S209). If "straight line display" is instructed, the process proceeds to "A". On the other hand, if "straight line display" is not instructed, in other words, if "arc display" is selected, the process proceeds to "B".

1. 1. Linear display In FIG. 5, it is determined in S208 of FIG. 4 that there are two "display classes", or in S208 of FIG. 4, it is determined that there are three or more "display classes", and in S209, "linear display" is determined. Is determined to be instructed, and when the process proceeds to the process “A”, the flowchart of the operation executed by the analysis unit 200 is shown. Here, the "straight line display" means a display method in which the "straight line" is divided and shown according to the ratio between the classes. FIG. 6 shows an example of linear display when it is determined that there are two “display classes” in S208 of FIG. 4, and FIGS. 7 to 10 show that the “display class” is 3 in S208 of FIG. An example of the linear display when it is determined that the number is one or more and the "straight line display" is instructed in S209 is shown. FIG. 6 shows an example of the linear display together with the output result of the learning model on which the linear display is based, and FIG. 7 shows an example of the linear display with the output of the learning model on which the linear display is based. It is shown together with the result and the calculation method, (a) shows the whole image of the linear display, and (b) shows the partially enlarged view.

1-1. Linear display with two display classes With reference to FIGS. 5 and 6, a process when it is determined that there are two “display classes” in S208 of FIG. 4 will be described. First, it is confirmed whether or not the display method is instructed by the user or the like through the input unit 3 (S210). How the user displays the identification mark 15a, that is, the mark indicating the certainty that unknown data belongs to the display class in the relationship between the first class and the second class, for example, an apple related to the class. It is possible to instruct through the input unit 3 whether to display with or as a simple point, the display speed, the size of the display, and the like. When the display method is instructed, the information is transmitted to the display change unit 400.

Next, the centers of the first circle 14 corresponding to the first class having the largest proportion among the display classes and the second circle 15 corresponding to the second largest second class are determined, and the centers are determined. The radii of each of the first circle 14 and the second circle 15 are set, and the position where the identification mark 15a should be displayed is calculated (S211).

The radii of the first circle 14 and the second circle 15 are obtained, for example, by dividing the first circle 14 and the second circle 15 by dividing the first straight line 30 connecting the centers 14A and 15B thereof. be able to. The first straight line 30 may be divided into two, three or the like. The division method follows, for example, an instruction input by the user through the input unit 3. FIG. 6 shows an example of two divisions, particularly an equal division. In this case, the first circle 14 and the second circle 15 are displayed in a state where they are substantially in contact with each other at one point on their circumferences. Here, "substantially in contact with one point" means, as shown in the figure, a state in which they overlap and are in contact with each other to the extent that they cannot be seen as double lines when visually observed by the user. The linear distances of the divided pieces 31 and 32 correspond to the radii of the first circle 14 and the second circle 15, respectively. In this case, the size of the radius 31 of the first circle 14 and the size of the radius 32 of the second circle 15 obtained by the division are both smaller than the length of the first straight line 30. The division method is not particularly limited, but from the viewpoint of easy viewing of the identification mark 15a and the like, the radius 32 of the second circle 15 is the same as the radius 31 of the first circle 14 or the radius 31 of the first circle 14. It is preferably smaller. The centers 14A and 15B of the first circle 14 and the second circle 15 thus, for example, the first circle 14 and the second circle 15 touched substantially at one point on their circumferences. It is determined so that the state is formed, and at least the circles of the first circle 14 and the circles of the second circle 15 can be displayed so as not to overlap each other.

The identification mark 15a shall be displayed on the first straight line 30 connecting the centers 14A and 15A of the first circle 14 and the second circle 15, and the position on the first straight line 30 shall be the first circle 14. The ratio of the length to the center 14A of the second circle 15 to the center 15A of the second circle 15 is the ratio of the second class, "0.30" in the illustrated example, and the ratio of the first class, shown in the figure. In the example, the first straight line 30 is proportionally divided so as to correspond to the ratio of "0.70". The analysis unit 200 obtains the ratio between the ratio of the first class and the ratio of the second class, and calculates the proportional division position.

After that, the number of display classes is confirmed (S212). Since there are two display classes here, the process ends.

The information obtained by the above processing of the analysis unit 200 is then transmitted to the image forming unit 300.

1-2. Linear display with three or more display classes With reference to FIGS. 5 and 7 to 10, it is determined in S208 of FIG. 4 that the "display class" is three or more, and "linear display" is instructed in S209. The processing when it is determined that the case is satisfied will be described. In the example of FIG. 7, in addition to the first circle 14 and the second circle 15, four additional circles 16 to 20, and in the example of FIG. 8, two additional circles 16 and 17. Further, in the example of FIG. 9, eight additional circles 16 to 24 are included. In FIGS. 8 and 9, the drawings are simplified for convenience.
In the flowchart of FIG. 5, the processing up to S212 is the same as <straight line display with two display classes> described with reference to FIGS. 5 and 6. Therefore, the description thereof is omitted here. In S212, since the number of display classes is 3 or more here, the process proceeds to S213, and 3 is set in n. Next, in S214, the center of the third circle 16 corresponding to the third class having the third largest ratio in the display class is determined, the radius is set, and the identification mark 16a, that is, the display class. The position where the mark indicating the certainty to which the unknown data belongs in the relationship between the first class and the third class should be displayed is calculated (S214).

The radius of the third circle 16 is the same as that of the n-1th circle, that is, the second circle 15.
Further, the center 16A of the third circle 16 is separated from the center 14A of the first circle 14 by the distance between the centers of the first circle 14 and the second circle 15, that is, the length of the first straight line 30. At this time, the position is such that the third circle 16 can be displayed so as not to overlap with the second circle 15. The circumference of the displayed third circle 16 and the circumference of the second circle 15 may be substantially in contact with each other at one point as shown in FIGS. 7, 9, and 10, and FIG. As shown in, it does not have to be in contact.

The identification mark 16a is displayed on the n-1 straight line connecting the center of the first circle and the center of the third circle, that is, on the second straight line 33, and the position on the second straight line 33 is For example, according to the instruction input through the input unit 3, the ratio of the length of the first circle 14 to the center 14A and the length of the third circle 16 to the center 16A is A) the ratio of the second class and the second. The position where the second straight line 33 is proportionally divided so as to correspond to the ratio of the ratio of one class, or B) to correspond to the ratio of the ratio of the third class to the ratio of the first class. To do. For example, if there is no particular instruction through the input unit 3, B) may be used. 7 to 9 show an example of the case where an instruction is given through the input unit 3, that is, A). In the case of A), the identification mark 15a displayed on the first straight line 30 and the identification mark 16a displayed on the second straight line 33 are both located on the concentric circles of the center 14A of the first circle. Unless the display method is changed, the relationship between the ratio of the second class and the ratio of the third class cannot be recognized. The method of changing the display may be obtained through, for example, the input unit 3.

Next, in S215, only 1 is added to n, and the processes of S214 and S215 are repeated until the value of n reaches the display class (S216). The process ends when the value of n reaches the display class.
The information obtained by the above processing of the analysis unit 200 is then transmitted to the image forming unit 300.

2. 2. Arc display In FIG. 11, when it is determined that "straight line display" is not instructed in S209 of FIG. 4, in other words, when it is determined that "arc display" is selected and the process proceeds to "B". , The flowchart of the operation executed by the analysis unit 200 is shown. Here, the "arc display" means a display method in which the "arc" is divided and shown according to the ratio between the classes. 12 and 13 are both examples of arc display obtained through the operation of the flowchart shown in FIG. FIG. 12 shows an example of the arc display together with the output result of the learning model on which the arc display is based. FIG. 12A shows the whole image of the arc display, and FIG. 12B shows the part thereof. An enlarged view is shown. FIG. 13 is a diagram showing an arc display in the same manner as in FIG. 12, but here, an example in which the fourth class to the seventh class below the threshold value are deleted through S204 and S205 of FIG. Shown.

2-1. Arc display with three display classes With reference to FIGS. 11 and 12, the analysis unit 200 when it is determined that "arc display" is selected in S209 of FIG. 4 and there are three display classes. The processing in is described. First, it is confirmed whether or not the display method is instructed by the user or the like through the input unit 3 (S220). The user can instruct through the input unit 3 how to display the identification mark 156a, for example, whether to display it with an apple related to the class or just a point, its display speed, the size of the display, and the like. it can. When the display method is instructed, the information is transmitted to the display change unit 400.

Next, the first circle 14 corresponding to the first class having the largest proportion among the display classes, the second circle 15 corresponding to the second largest second class, and the third largest third The center of each of the third circles 16 corresponding to the class is determined, the radius is set, and the identification mark 156a, that is, the certainty that unknown data belongs to the display class is determined by the first class and the second. The position where the mark to be displayed in relation to the class and the third class should be displayed is calculated (S221).

The radius of the first circle 14 and the second circle 15 can be obtained by dividing the first straight line 30 connecting the centers 14A and 15B thereof. The first straight line 30 may be divided into two, three or the like. The example of FIG. 12 shows an example of two divisions, particularly an equal division. The linear distances of the divided pieces 31 and 32 correspond to the radii of the first circle 14 and the second circle 15, respectively. Through the input unit 3, the user can instruct the division method, more specifically, the size of the radius of the first circle 14 and the second circle 15. The size of the radius 31 of the first circle 14 and the size of the radius 32 of the second circle 15 obtained by the division are both smaller than the length of the first straight line 30. The dividing method is not particularly limited, but from the viewpoint of easy viewing of the identification mark 156a and the like, the radius 32 of the second circle 15 is the same as the radius 31 of the first circle 14 or the radius 31 of the first circle 14. It is preferably smaller. The centers 14A and 15B of the first circle 14 and the second circle 15 thus, for example, the first circle 14 and the second circle 15 touched substantially at one point on their circumferences. It is determined so that the state is formed, and at least the circles of the first circle 14 and the circles of the second circle 15 can be displayed so as not to overlap each other.

The radius of the third circle 16 is the same as that of the second circle 15. In the example of FIG. 12, the radius of the first circle 14 is the same.
Further, the center 16A of the third circle is separated from the center 14A of the first circle 14 by the distance between the centers of the first circle 14 and the second circle 15, that is, the length of the first straight line 30. The position is set, and at this time, the position where the third circle 16 and the second circle 15 can be displayed so as not to overlap each other. As shown in FIG. 12, the circumference of the displayed third circle 16 and the circumference of the second circle 15 may or may not touch substantially at one point.

The identification mark 156a is displayed on an arc 36 concentric with the first circle, and the radius of the circle forming the arc 36 is the length to the center 14A of the first circle 14 and the second circle. The ratio of the lengths to the center 15A of 15 is the proportion of the second class, i.e. "0.20" in the illustrated example, and the proportion of the first class, i.e. "0.70" in the illustrated example. The length corresponds to the distance 35 between the position 38 obtained by proportionally dividing the straight line 30 and the center 14A of the first circle 14 so as to correspond to the ratio with. In this arc 36, the identification mark 156a forms a second arc 37 concentric with the first circle 14, which may be formed between the center 15A of the second circle 15 and the center 16A of the third circle 16. The ratio of the length of the arc 37a to the center 15A of the circle 15 and the length of the arc 37b to the center 16A of the third circle 16 is the ratio of the third class, that is, "0.10" in the illustrated example. , The ratio of the second class, that is, the position corresponding to the position proportionally divided so as to correspond to the ratio with "0.20" in the illustrated example. In order to determine the position of the identification mark 156a, the analysis unit 200 obtains the ratio of the ratio of the first class to the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class. , Calculate this apportioned position.

After that, the number of display classes is confirmed (S222). Since there are three display classes here, the process ends.
The information obtained by the above processing of the analysis unit 200 is then transmitted to the image forming unit 300.

2-2. Arc display with four or more display classes While further referring to FIG. 11, the analysis unit when it is determined that "arc display" is selected in S209 of FIG. 4 and the display class is four or more. The processing in 200 will be described. An example of the arc display obtained through this process is not particularly shown. When there are four or more display classes, 4 is set in n in S223. Next, the center of the fourth circle corresponding to the fourth class having the fourth largest ratio in the display class is determined, the radius is set, and the identification mark, that is, the unknown data is displayed in the display class. The position where the mark indicating the certainty of belonging in relation to the first class, the third class, and the fourth class should be displayed is calculated (S224).

The radius of the fourth circle is the same as that of the second circle 15.
Further, the center of the fourth circle is the distance between the centers of the first circle 14 and the second circle 15, that is, a position separated from the center 14A of the first circle 14 by the length of the first straight line 30. At this time, the position is set so that the fourth circle can be displayed so as not to overlap with the third circle 16. The circumference of the displayed fourth circle and the circumference of the third circle may or may not be substantially in contact with each other at one point.

The identification mark shall be displayed on an arc concentric with the first circle, and the radius of the circle forming this arc shall be the length up to the center 14A of the first circle 14 and the center of the second circle 15. The position 38, which is the proportional division of the straight line 30, and the center 14A of the first circle 14, so that the ratio of the lengths up to 15A corresponds to the ratio of the second class to the ratio of the first class, The length corresponds to the distance between 35. In this arc, the identification mark forms an arc concentric with the first circle 14, which can be formed between the center 16A of the third circle 16 and the center of the fourth circle, and the center 16A of the third circle 16. The position corresponding to the position where the ratio of the length of the arc to the center of the fourth circle is proportionally divided so as to correspond to the ratio of the ratio of the fourth class to the ratio of the third class. And. In order to determine the position of the identification mark, the analysis unit 200 obtains the ratio of the ratio of the third class to the ratio of the fourth class, and calculates the proportional division position.

Next, in S225, only 1 is added to n, and the processes of S224 and S225 are repeated until the value of n reaches the display class (S226). The process ends when the value of n reaches the display class.

The information obtained by the above processing of the analysis unit 200 is then transmitted to the image forming unit 300.

The image forming unit 300 is image data necessary for performing two-dimensional display based on the information obtained by the analysis by the analysis unit 200, the instruction content input from the input unit 3 obtained through the analysis unit 200, and the like. To form. For example, the center coordinates of the circle for displaying each class when displayed by the display device 5 are determined, and the coordinates for displaying the identification mark are determined. Further, data necessary for displaying an image recorded on a recording medium 41 or the like is collected. The image data formed by the image forming unit 300 is transmitted to the display device 5 through the output unit 44, and the display device 5 performs two-dimensional display based on the image data.

The display changing unit 24 operates the image data formed by the image forming unit 300 based on the instruction content and the like input from the input unit 3 obtained through the analysis unit 200, and changes the display method. If there is no need to change the display, no special processing is required. For example, in S214, the ratio of the length of the first circle 14 to the center 14A and the length of the third circle 16 to the center 16A is set as the ratio of the second class according to the instruction input through the input unit 3. When the position where the second straight line 33 is apportioned so as to correspond to the ratio to the ratio of the first class (case A above), the ratio of the second class and the ratio of the third class The display method of the identification marks 15a and 16a can be changed according to the size of the ratio according to the instruction content and the like input from the input unit 3 so that the two can be compared. Further, for example, as shown in the examples of FIGS. 7 and 8, the identification mark may be displayed in a size corresponding to the size of the ratio, or as shown in FIG. 10, the time may be displayed between the display classes. They may be displayed differently (in the figure, the arrow "39" means the order in which the identification marks are displayed, and further, the identification marks 15a to 20a may be displayed in the clockwise order). .. Further, the identification mark may be blinked and displayed at a speed corresponding to the size of the ratio. Further, of course, these display methods may be appropriately combined. Furthermore, the determination criteria shown in FIG. 3 may be displayed.

The present invention is not limited to the above-described embodiment, and various other modifications can be made.
For example, in the above-described embodiment, the circles are displayed in descending order of the ratio, but it is not always necessary to display them in such an order. They may be displayed in any order as long as they do not overlap each other.
Further, the classes other than the first class do not necessarily have to be displayed in a circle as in the first class. Furthermore, the circles corresponding to the classes other than the first class may be displayed by a method other than the circle near the center of each circle, or nothing may be displayed. As the display other than the circle, for example, the display by the characters representing the class shown in FIG. 6 and the like can be used.
Further, the system 2 may instruct the content of the instruction from the user instructed through the input unit 3.
Further, in the above-described embodiment, the learning model 1 is described as being separate from the system 2, while the display device 5 is described as being included in the system 2, but the learning model 1 may be included in the system 2. However, conversely, the display device 5 does not have to be included in the system 2. In the latter case, the output unit 44 in the main body 4 substantially functions as a display means.

Yet another aspect, feature and effect of the present invention is facilitated from the following detailed description, merely showing a number of specific embodiments and examples, including the best aspects intended to carry out the present invention. Will be clear. The present invention can also be configured in other and different embodiments, and many details thereof can be modified in various obvious respects without departing from the spirit and scope of the present invention. Therefore, the drawings and description are merely examples and are not limited thereto.

1 Learning model 2 System 3 Input unit 4 Main unit 5 Display device (display means)
44 Output unit (display means)
200 Analysis unit (analysis means)
300 Image forming unit (image data forming means)
400 Display change part

Claims (25)

It is a two-dimensional display system that displays the output result obtained when data is input to a trained training model that has undergone machine learning using the training data prepared for each of a plurality of classes in two dimensions. When data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class.
A means for analyzing an output result obtained by inputting unknown data other than the training data into the training model, in which a plurality of the classes to be confirmed are ordered according to the magnitude of the ratio and a plurality of the classes to be confirmed are to be confirmed. The center of the first circle corresponding to the first class having the largest proportion among the classes and the second circle corresponding to the second largest class is determined, and the first circle is defined. The radius is set by dividing the first straight line connecting the center of the circle and the center of the second circle, and the ratio of the ratio of the first class to the ratio of the second class is obtained. means,
A means for forming image data based on the information obtained by the analysis, and a first class having the largest ratio among the plurality of classes to be confirmed based on the image data. The first circle and the second circle corresponding to the second largest second class were obtained by dividing the first straight line connecting the center of the first circle and the center of the second circle. Each is displayed as a circle having a radius, and an identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class and the second class is displayed on the first straight line. Therefore, the ratio of the length to the center of the first circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. Means for displaying the first straight line at a proportionally divided position,
A two-dimensional display system characterized by being equipped with. The means for analysis further determines the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of classes to be confirmed, and determines the center of the nth circle corresponding to the ratio of the first class. Find the ratio of the proportions of the nth class
Based on the image data, the display means further displays the nth circle corresponding to the nth class having the nth largest ratio among the plurality of classes to be confirmed, and the first circle. And the circle having the same radius as the second circle centered on the position away from the center of the first circle by the distance between the centers of the second circle so as not to overlap with the n-1 circle. The identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class and the nth class is displayed on the center of the first circle and the nth circle. The ratio of the length to the center of the first circle and the length to the center of the nth circle on the n-1 straight line connecting the centers of the second class is the ratio of the second class and the first. The straight line of the n-1 is displayed at a position proportionally divided so as to correspond to the ratio of the ratio of the first class or the ratio of the nth class to the ratio of the first class (n is). 3 ≤ n ≤ (number of classes to be confirmed) integer),
The two-dimensional display system according to claim 1. It is a two-dimensional display system that displays the output result obtained when data is input to a trained training model that has undergone machine learning using the training data prepared for each of a plurality of classes in two dimensions. When data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class.
It is a means for analyzing the output result obtained by inputting unknown data other than the training data into the training model, and a plurality of the classes to be confirmed are ordered according to the magnitude of the ratio and a plurality of the classes to be confirmed are to be confirmed. Corresponds to the first circle corresponding to the first class having the largest proportion among the classes, the second circle corresponding to the second largest class, and the third class having the third largest proportion. The center of the third circle is determined, and the radius is set by dividing the straight line connecting the center of the first circle and the center of the second circle, and further, with the ratio of the first class. The means for determining the ratio of the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class.
A means for forming image data based on the information obtained by the analysis, and a first class having the largest ratio among the plurality of classes to be confirmed based on the image data. For the first circle and the second circle, the first circle and the second circle corresponding to the second largest second class and the third circle corresponding to the third largest third class. As a circle having a radius obtained by dividing a straight line connecting the center of the first circle and the center of the second circle, and for the third circle, the first circle and the second circle. The second circle and the third circle do not overlap each other as a circle having the same radius as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the circles. As described above, the identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the second class, and the third class is attached to the first class. The position where the straight line is proportionally divided so that the ratio of the length to the center of the circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. On an arc concentric with the first circle having a radius of length corresponding to the distance between the center of the first circle and the center of the first circle, and of the center of the second circle and the third circle. The ratio of the length of the arc concentric with the first circle to the center of the second circle and the length of the arc to the center of the third circle, which can be formed between the center, is the third. A means for displaying at a position corresponding to a position proportionally divided so as to correspond to the ratio of the ratio of the second class to the ratio of the second class.
A two-dimensional display system characterized by being equipped with. The means for analysis further determines the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed, and determines the center of the nth circle with the ratio of the nth class. Find the ratio of the proportions of the n-1 class
Based on the image data, the display means further forms the nth circle corresponding to the nth class having the nth largest ratio among the plurality of classes to be confirmed, and the first circle. And the circle having the same radius as the second circle centered on the position away from the center of the first circle by the distance between the centers of the second circle, and does not overlap with the n-1 circle. The identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the n-1 class, and the nth class is attached to the first class. The position where the straight line is proportionally divided so that the ratio of the length to the center of the circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. On an arc concentric with the first circle having a radius of length corresponding to the distance between and the center of the first circle, the center of the n-1 circle and the nth The ratio of the length of the arc concentric with the first circle that can be formed between the center of the circle to the center of the n-1 circle and the length of the arc to the center of the nth circle. Is displayed at the position corresponding to the position proportionally divided so as to correspond to the ratio of the ratio of the nth class to the ratio of the n-1 class (n is 4 ≦ n ≦ (of the class to be confirmed). Quantum)), the two-dimensional display system according to claim 3. In the means for analysis, the radius of the first circle and the radius of the second circle are obtained by dividing the first straight line into two, and the radius of the second circle is the first. The two-dimensional display system according to any one of claims 1 to 4, which is equal to or smaller than the radius of the first circle. In the means for displaying, the first circle and the nth circle are displayed in a state where they are substantially in contact with each other at one point on their circumference (n is an integer of 2 or more). The two-dimensional display system according to any one of Items 1 to 5. In the means for displaying, the nth circle and the n + 1 circle are displayed in a state where they are substantially in contact with each other at one point on their circumference (n is an integer of 2 or more). The two-dimensional display system according to any one of Items 1 to 6. Any of claims 1 to 7, wherein the proportion of the n-1 class is the same as the proportion of the n class (n is an integer of 2 ≦ n ≦ (the number of classes to be confirmed)). Two-dimensional display system described in Crab. The two-dimensional display system according to any one of claims 1 to 8, further comprising display changing means for changing the display method of the identification mark according to the magnitude of the ratio. The two-dimensional display system according to claim 9, wherein the identification mark is displayed at different times in a plurality of the classes selected as display targets by the display changing means. The two-dimensional display system according to claim 9 or 10, wherein the identification mark is displayed in a size corresponding to the size of the ratio by the display changing means. The two-dimensional display system according to any one of claims 9 to 11, wherein the identification mark is blinked and displayed at a speed corresponding to the magnitude of the ratio by the display changing means. The two-dimensional display system according to any one of claims 1 to 12, wherein the analysis means excludes the class whose ratio is equal to or less than a predetermined threshold value from the display target. The two-dimensional display system according to any one of claims 1 to 13, wherein a circle indicating an evaluation criterion for evaluating the identification mark is displayed concentrically with the first circle by the display changing means. The two-dimensional according to any one of claims 1 to 14, wherein a circle other than the first circle is displayed near the center of each circle by a method other than the circle by the display changing means, or nothing is displayed. Display system. The two-dimensional display system according to any one of claims 1 to 15, wherein the learning data includes data for creating the learning model and data for verifying the learning model. It is a two-dimensional display method that displays the output result obtained when data is input to a trained learning model that has undergone machine learning using the training data prepared for each of a plurality of classes in two dimensions. When data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class.
At the stage of analyzing the output result obtained by inputting unknown data other than the training data into the training model, the plurality of the classes to be confirmed are ordered according to the magnitude of the ratio, and the plurality to be confirmed. The center of the first circle corresponding to the first class having the largest proportion among the classes and the second circle corresponding to the second largest class is determined, and the first circle is defined. The radius is set by dividing the first straight line connecting the center of the circle and the center of the second circle, and the ratio of the ratio of the first class to the ratio of the second class is obtained. stage,
The step of forming image data based on the information obtained by the analysis, and the first class having the largest ratio among the plurality of classes to be confirmed based on the image data. The first circle and the second circle corresponding to the second largest second class were obtained by dividing the first straight line connecting the center of the first circle and the center of the second circle. Each is displayed as a circle having a radius, and an identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class and the second class is displayed on the first straight line. Therefore, the ratio of the length to the center of the first circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. At the stage of displaying the first straight line at the proportionally divided position,
A two-dimensional display method characterized by comprising. At the stage of the analysis, the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is further determined, and the ratio of the first class is determined. Find the ratio of the proportions of the nth class
At the stage of displaying, the nth circle corresponding to the nth class having the nth largest ratio among the plurality of classes to be confirmed is further referred to as the first circle based on the image data. And the circle having the same radius as the second circle centered on the position away from the center of the first circle by the distance between the centers of the second circle so as not to overlap with the n-1 circle. The identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class and the nth class is displayed on the center of the first circle and the nth circle. The ratio of the length to the center of the first circle and the length to the center of the nth circle on the n-1 straight line connecting the centers of the second class is the ratio of the second class and the first. The straight line of the n-1 is displayed at a position proportionally divided so as to correspond to the ratio of the ratio of the first class or the ratio of the nth class to the ratio of the first class (n is). 3 ≤ n ≤ (number of classes to be confirmed) integer),
The two-dimensional display method according to claim 17. It is a two-dimensional display method that displays the output result obtained when data is input to a trained learning model that has undergone machine learning using the training data prepared for each of a plurality of classes in two dimensions. When data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class.
At the stage of analyzing the output result obtained by inputting unknown data other than the training data into the training model, the plurality of the classes to be confirmed are ordered according to the magnitude of the ratio, and the plurality to be confirmed. Corresponds to the first circle corresponding to the first class having the largest proportion among the classes, the second circle corresponding to the second largest class, and the third class having the third largest proportion. The center of the third circle is determined, and the radius is set by dividing the straight line connecting the center of the first circle and the center of the second circle, and further, with the ratio of the first class. The step of determining the ratio of the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class.
The stage of forming image data based on the information obtained by the analysis, and the first class having the largest ratio among the plurality of classes to be confirmed based on the image data. For the first circle and the second circle, the first circle and the second circle corresponding to the second largest second class and the third circle corresponding to the third largest third class. As a circle having a radius obtained by dividing a straight line connecting the center of the first circle and the center of the second circle, and for the third circle, the first circle and the second circle. The second circle and the third circle do not overlap each other as a circle having the same radius as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the circles. As described above, the identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the second class, and the third class is attached to the first class. The position where the straight line is proportionally divided so that the ratio of the length to the center of the circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. On an arc concentric with the first circle having a radius of length corresponding to the distance between the center of the first circle and the center of the first circle, and of the center of the second circle and the third circle. The ratio of the length of the arc concentric with the first circle to the center of the second circle and the length of the arc to the center of the third circle, which can be formed between the center, is the third. The stage of displaying at the position corresponding to the position proportionally divided so as to correspond to the ratio of the ratio of the second class to the ratio of the second class.
A two-dimensional display method characterized by comprising. At the stage of the analysis, the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is further determined, and the ratio of the nth class is determined. Find the ratio of the proportions of the n-1 class
At the display stage, further, based on the image data, the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is the first circle. And the circle having the same radius as the second circle centered on the position away from the center of the first circle by the distance between the centers of the second circle, and does not overlap with the n-1 circle. The identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the n-1 class, and the nth class is attached to the first class. The position where the straight line is proportionally divided so that the ratio of the length to the center of the circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. On an arc concentric with the first circle having a radius of length corresponding to the distance between and the center of the first circle, the center of the n-1 circle and the nth The ratio of the length of the arc concentric with the first circle that can be formed between the center of the circle to the center of the n-1 circle and the length of the arc to the center of the nth circle. Is displayed at the position corresponding to the position proportionally divided so as to correspond to the ratio of the ratio of the nth class to the ratio of the n-1 class (n is 4 ≦ n ≦ (of the class to be confirmed). Number) integer),
The two-dimensional display method according to claim 19. It is a two-dimensional display program that displays the output result obtained when data is input to a trained training model that has undergone machine learning using the training data prepared for each of a plurality of classes in two dimensions. When data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class.
On the computer
It is a function to analyze the output result obtained by inputting unknown data other than the training data into the training model, and the plurality of the classes to be confirmed are ordered according to the magnitude of the ratio, and the plurality of to be confirmed. The center of the first circle corresponding to the first class having the largest proportion among the classes and the second circle corresponding to the second largest class is determined, and the first circle is defined. The radius is set by dividing the first straight line connecting the center of the circle and the center of the second circle, and the ratio of the ratio of the first class to the ratio of the second class is obtained. function,
A function of forming image data based on the information obtained by the analysis, and a first class having the largest ratio among the plurality of classes to be confirmed based on the image data. The first circle and the second circle corresponding to the second largest second class were obtained by dividing the first straight line connecting the center of the first circle and the center of the second circle. Each is displayed as a circle having a radius, and an identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class and the second class is displayed on the first straight line. Therefore, the ratio of the length to the center of the first circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. Function to display the first straight line at the proportionally divided position,
A two-dimensional display program to realize. On the computer
In the function to analyze, the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is further determined, and the ratio of the first class is determined. The function of finding the ratio of the proportions of the nth class,
In the function to be displayed, further, based on the image data, the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is the first circle. And the circle having the same radius as the second circle centered on the position away from the center of the first circle by the distance between the centers of the second circle, so as not to overlap with the n-1 circle. The identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class and the nth class is displayed on the center of the first circle and the nth circle. The ratio of the length to the center of the first circle and the length to the center of the nth circle on the n-1 straight line connecting the centers of the second class is the ratio of the second class and the first. A function of displaying the n-1 straight line at a proportionally divided position so as to correspond to the ratio of the ratio of one class or the ratio of the nth class to the ratio of the first class (n is , 3 ≤ n ≤ (number of classes to be confirmed) integer),
21. The two-dimensional display program according to claim 21. It is a two-dimensional display program that displays the output result obtained when data is input to a trained training model that has undergone machine learning using the training data prepared for each of a plurality of classes in two dimensions. When data is input, the learning model outputs the certainty that the data belongs to each of the plurality of classes as a ratio among the plurality of classes for each class.
On the computer
It is a function to analyze the output result obtained by inputting unknown data other than the training data into the training model, and the plurality of the classes to be confirmed are ordered according to the magnitude of the ratio, and the plurality to be confirmed. Corresponds to the first circle corresponding to the first class having the largest proportion among the classes, the second circle corresponding to the second largest class, and the third class having the third largest proportion. The center of the third circle is determined, and the radius is set by dividing the straight line connecting the center of the first circle and the center of the second circle, and further, with the ratio of the first class. The function of obtaining the ratio of the ratio of the second class and the ratio of the ratio of the second class to the ratio of the third class.
The function of forming image data based on the information obtained by the analysis, and the first class corresponding to the first class having the largest ratio among the plurality of classes to be confirmed based on the image data. For the first circle and the second circle, the first circle and the second circle corresponding to the second largest second class and the third circle corresponding to the third largest third class. As a circle having a radius obtained by dividing a straight line connecting the center of the first circle and the center of the second circle, and for the third circle, the first circle and the second circle. The second circle and the third circle do not overlap each other as a circle having the same radius as the second circle centered on a position separated from the center of the first circle by the distance between the centers of the circles. As described above, the identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the second class, and the third class is attached to the first class. The position where the straight line is proportionally divided so that the ratio of the length to the center of the circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. On an arc concentric with the first circle having a radius of length corresponding to the distance between the center of the first circle and the center of the first circle, and of the center of the second circle and the third circle. The ratio of the length of the arc concentric with the first circle to the center of the second circle and the length of the arc to the center of the third circle, which can be formed between the center, is the third. A function to display at the position corresponding to the position proportionally divided so as to correspond to the ratio of the class ratio of the second class to the ratio of the second class.
A two-dimensional display program to realize. On the computer
In the function to analyze, the center of the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is further determined, and the ratio of the nth class is determined. The function of obtaining the ratio of the ratio of the n-1th class,
In the function to be displayed, further, based on the image data, the nth circle corresponding to the nth class having the nth largest ratio among the plurality of the classes to be confirmed is the first circle. And the circle having the same radius as the second circle centered on the position away from the center of the first circle by the distance between the centers of the second circle, and does not overlap with the n-1 circle. The identification mark indicating the certainty that the unknown data belongs to the class in the relationship between the first class, the n-1 class, and the nth class is attached to the first class. The position where the straight line is proportionally divided so that the ratio of the length to the center of the circle to the center of the second circle corresponds to the ratio of the ratio of the second class to the ratio of the first class. On an arc concentric with the first circle having a radius of length corresponding to the distance between and the center of the first circle, the center of the n-1 circle and the nth The ratio of the length of the arc concentric with the first circle that can be formed between the center of the circle to the center of the n-1 circle and the length of the arc to the center of the nth circle. Is a function of displaying at a position corresponding to a position proportionally divided so as to correspond to the ratio of the ratio of the nth class to the ratio of the n-1 class (n is 4 ≦ n ≦ (class to be confirmed). Number of)), integer),
23. The two-dimensional display program according to claim 23. A computer-readable recording medium on which the program according to any one of claims 21 to 24 is recorded.

Images