JPH0488429A - Approximate inference device - Google Patents

Abstract

PURPOSE:To obtain a conclusion which has high articulation efficiently by calculating static information amounts of respective events by using a stored relation and outputting the calculated static information amounts in decreasing order in relation with the events. CONSTITUTION:The device is equipped with a knowledge storage device 11, a knowledge composition device 12, an after-composition knowledge storage device 13, an event value input device 14, a consistency arithmetic unit 15, a dynamic information amount arithmetic unit 16, a possibility arithmetic unit 17, a possibility display device 18, a static information amount arithmetic unit 19, an articulation arithmetic unit 20, an articulation storage device 21, an articulation addition device 22, and an articulation display device 23. Expert's knowledge showing the relation between events and conclusions is stored previously and input data is applied to this knowledge to calculate the possibility of a conclusion by inference; and articulation addition values are calculated by the articulation and conclusions of respective events and input events which are effective for the improvement of the articulation addition value is guided. Further, event data is inputted by an operator or automatically according to the guidance. Consequently, the conclusion which is high in articulation is obtained efficiently.

Description

[Detailed Description of the Invention] Expert knowledge (rules) representing the relationship between summary events and conclusions of the invention is memorized in advance, and inferences are made by applying human data to this knowledge to determine the possibility of the conclusion. In an approximate reasoning device that calculates the clarity of each event regarding the possibility of each conclusion obtained and the clarity addition value for each conclusion, it provides guidance on effective events to be input to further increase the clarity addition value. . This guidance allows event data to be entered manually or automatically. This allows for efficient clarity (reliability)
A high conclusion can be obtained.

BACKGROUND OF THE INVENTION Technical Field A second invention relates to an approximate inference device that performs inference using expert knowledge representing the relationship between events and conclusions.

Conventional Approximate inference is known in which the inference result is modified or changed depending on the amount of information of the event used to derive the inference result (for example, ZhangHongmi
n AN EXPERT SYSTEM MITI
(T)IINKING INIMAGES”, Pr
eprints of 5econd IPsA Co
ngress.

Tokyo, July 20-25.1987. P
, 765).

The second approximate inference method calculates the amount of information for each event (i.e., information discrimination ability of the event) using the membership function given for each conclusion for the event, and calculates the inference result (i.e., the probability that the conclusion holds true). ) by modifying or changing the amount of information of the event used to draw the conclusion (by multiplying the probability and the amount of information), it attempts to improve the discriminative ability of the inference result.

However, in conventional approximate inference methods such as this, when drawing the possibility of a conclusion based on input event data, if data regarding all events are not input, the probability and clarity of the conclusion will be low. humans have not been able to accurately judge.

Also, if you input data for all events, the clarity will be higher, but in cases where speedy processing is required, such as trouble diagnosis, it will take too much time, so it is not recommended to input all data. It's impossible.

Furthermore, when handled by experts, it is possible to select and input only the most efficient event data based on experience, but for beginners, due to their lack of experience, it is necessary to input all event data to ensure clarity. The problem was that it was not high enough.

Summary of the Invention Purpose of the Invention The present invention is to efficiently input event data that has a large influence on the conclusion, and to enable a conclusion with high clarity (reliability) to be obtained with as little event data as possible. With the goal.

Structure 9 of the Invention Operation and Effect The approximate inference device according to the present invention includes a storage means for storing the relationship between an event and a conclusion in advance, and a static information amount of each event using the relationship stored in the storage means. The present invention is characterized by comprising a calculation means and a means for outputting the amount of static information calculated by the calculation means in descending order of magnitude in relation to an event.

By providing a means for instructing to input event data of events corresponding to the static information amount of each event calculated by the static information amount calculation means in descending order of the static information amount of each event, a human (worker) can input event data.

Further, by providing a means for inputting event data of events corresponding to the static information amount of each event calculated by the static information amount calculation means in descending order of magnitude, event data can be inputted without any operation by a single person. Ru.

The clarity of each event is calculated for each conclusion using the static information amount obtained from the static information amount calculating means.

The clarity of each conclusion can be determined by providing a clarity calculation means that calculates the clarity of each conclusion by adding up the clarity of events for which data is actually input using the calculated clarity.

The reliability of the conclusion is increased by comprising control means for repeating event data input and clarity addition until the clarity addition value calculated by the clarity calculation means reaches a predetermined value for at least one conclusion. It can be improved.

The approximate inference device includes approximate inference means that calculates the probability of a conclusion by applying the input data of each event to the relationship stored in the storage means.

The approximate inference means preferably includes a goodness-of-fit calculation means that converts the input data into goodness of fit using a membership function expressed by the above relationship, and a goodness-of-fitness calculation means that uses this goodness of fit to calculate the amount of dynamic information for each event. and a probability calculation means for calculating the probability of a conclusion using the degree of fitness and the amount of dynamic information.

According to this invention, since the static information amount of each event is arranged and output in descending order of static information amount, it is possible to input event data of each event in descending order of static information amount. Since the amount of static information of an event indicates the ability of the membership function of that event to identify a conclusion,
By inputting event value data in order of increasing amount of static information, a conclusion with higher clarity can be obtained by inputting less event data (or fewer input event data). Approximate inference can be performed efficiently and processing time can be shortened. Especially for beginners who are unable or have difficulty deciding which event data to enter first, it will tell you the order of event data, allowing you to draw conclusions efficiently without hesitation. .

The approximate inference device according to the present invention includes a storage means for storing the relationship between an event and a conclusion in advance, and a static information amount calculation for calculating the amount of static information of each event using the relationship stored in the storage means. a means for calculating the clarity of each event for each conclusion using the amount of static information obtained from the static information amount calculating means; and a means for calculating the clarity of each event for each conclusion obtained from the above-mentioned clarity calculating means. The present invention is characterized by comprising means for displaying the clarity in descending order of clarity.

The above approximate reasoning device is also A clarity adding means is provided for calculating the clarity of each conclusion by adding up the clarity of events for which data is actually input.

The approximate reasoning device further includes means for controlling the event data input and the clarity addition until the clarity addition value calculated by the clarity addition means reaches a predetermined value for at least one conclusion.

The approximate inference device further includes approximate inference means that calculates the probability of a conclusion by applying the input data of each event to the relationship stored in the storage means.

The approximate inference means preferably includes a goodness-of-fit calculation means that converts the input data into goodness of fit using a membership function expressed by the above relationship, and a goodness-of-fitness calculation means that uses this goodness of fit to calculate the amount of dynamic information for each event. and a probability calculation means for calculating the probability of a conclusion using the degree of fitness and the amount of dynamic information.

According to the approximate inference device of the present invention, since the order of magnitude of clarity of each event for each conclusion is specified, event value data can be input in accordance with this order of magnitude,
As in the case described above, it is possible to efficiently obtain highly clear conclusions in a short time with a small number of inputs of one event data, and even a beginner can operate it easily and efficiently. .

The approximate inference device according to the present invention includes a storage means for storing in advance a relationship between an event and a conclusion, an approximate inference means for calculating the possibility of a conclusion by applying input data to the relationship stored in the storage means, Calculation means that calculates the clarity of each event for each conclusion using the relationship stored in the storage means, and adds the clarity of the events for which data is actually input to obtain a clarity addition value for each conclusion. summing means and means for selecting the event that maximizes the clarity add value for a given conclusion.

Preferably, the selection means selects the event based on the most probable conclusion obtained by the approximate inference means.

The approximate inference means preferably includes a goodness-of-fit calculation means that converts the input data into goodness of fit using a membership function expressed by the above relationship, and a goodness-of-fitness calculation means that uses this goodness of fit to calculate the amount of dynamic information for each event. It is composed of a dynamic information amount calculation means 2 and a possibility calculation means for calculating the possibility of a conclusion using the above-mentioned fitness degree and dynamic information amount.

By providing display means for instructing input of event data of the event selected by the selection means, one person (worker, user) can input event data in accordance with this instruction.

By providing means for inputting event data of the event selected by the selection means, the event data can be automatically inputted without any operation by a single person.

The clarity of the predetermined conclusion is increased by comprising control means for repeating event data input and clarity addition until the clarity addition value calculated by the addition means reaches a predetermined value for the predetermined conclusion. It can be increased up to a predetermined value.

By providing means for controlling the selection means to repeatedly select events for a predetermined number of conclusions, it becomes possible to additionally input event data for many conclusions.

By providing means for controlling the approximate inference means to recalculate the probability of a conclusion by applying all input data, including event data input according to the selection by the selection means, to the storage means, Re-inference is performed using the input event data according to the selection made by the selection means, and information about the possibility of a highly clear conclusion is obtained.

As described above, according to the approximate inference device of the present invention, the event that maximizes the added clarity value is selected, and by inputting the event data of this selected event, the data related to all events can be Because it is possible to quickly obtain a clarity addition value sufficient for determining a conclusion without inputting information, a conclusion can be drawn efficiently in a short period of time.

When the number of input event data is so small that it is considered difficult or dangerous to evaluate the conclusion.

When there are several highly possible/probable conclusions □ 1. In cases where it is considered that the data for an event that is considered important has not yet been entered, by activating the above selection method, the most effective event can be selected. This makes it possible to input data, allowing even non-experts to draw conclusions efficiently. Particularly for beginners, the system instructs the events for which data should be input one by one, allowing efficient input operations.

If the knowledge set at the time of design is inappropriate, the clear addition value will not become high unless a lot of event data is input. According to this invention, it becomes possible to evaluate the intelligence created at the design stage, and it becomes possible to support optimal design.

Description of Embodiments (1) Overall configuration of approximate inference device An example of the overall configuration of the approximate inference device is shown in FIGS. 1 and 2. FIG. 1 shows a structure according to a first embodiment of the invention, and FIG. 2 shows a structure according to a second embodiment of the invention.

In these figures, the approximate inference device is a knowledge storage device 1
1. Knowledge synthesis device 121, post-synthesis knowledge storage device 13, event value input device 14. Fitness calculation device 15. Dynamic information amount calculation device 1B, possibility calculation device 17. Possibility display device 18.
Static information amount calculation device 19. Clarity calculation device 20. Clarity storage 21. It includes a clarity adder 22 and a clarity display 23. These devices 11 to 23 are common to the first embodiment (FIG. 1) and the second embodiment (FIG. 2).

The approximate inference device of the first embodiment shown in FIG. 1 further includes a static information amount display/input instruction device 31.

The approximate inference device of the second embodiment shown in FIG. 2 further includes a one-event selection calculation device 32 and an event input guidance display device 33.

Hereinafter, first, the configuration common to both embodiments will be explained, and then the configuration and operation of each embodiment will be explained separately.

(2) Knowledge Storage Device The knowledge storage device 11 is a device that stores knowledge input by experts and the like in a format that shows the relationship between events and conclusions. This device can store the knowledge of multiple experts.

Two experts ext stored in the knowledge storage device
, ex2 knowledge is shown below in the form of a rule.

Expert exl: jf20≦f1≦60, 0≦f2≦40゜””ait≦
fi≦bit”” 40≦rl≦80. 6(1; f2≦100”, ai
2≦N≦bi2””a, ≦fi≦b,.

IJ IJ Expert ex2 if30≦f1≦50.10≦f2≦30°・= ,
a≦fi≦b11゜50≦fl≦70. 70≦12≦90゜""a'+
2≦f i≦b I 2- =hen hen hen hen hen hen...(1)...ku2)...(3)...(4) ir=-, a--≦fi≦b,, ,-then c
jIJ IJ...(6) Expert exl fl f2 fi are events, and these are respectively event 1. Event 2. Sometimes called event l. The event index l takes a value in the range 1 to m. cl, c2. cj is the conclusion, and these are respectively concluded 1. It is sometimes called conclusion 2 conclusion j. The subscript j of the conclusion takes a value in the range of 1 to n.

Also, a i expressed as al, ≦fi≦blj
j.

I.J. bl are called the minimum value and maximum value, respectively.

IJ The above rules are expressed in the form of a table for each expert as follows.

Expert ex2 (blank below) (3) Knowledge synthesis device The knowledge synthesis device 12 is a device that synthesizes the knowledge of a plurality of experts stored in the knowledge storage device 11 and combines it into one piece of knowledge.

There are various ways to synthesize knowledge, but here we calculate the average value and standard deviation of multiple experts for the maximum and minimum values of each event involved in each conclusion.

The knowledge synthesis process will be described below, taking as an example the knowledge of the two experts described above that leads to the conclusion c1 from the event f1.

From the above rules (Equations (1) and (4)), event 1
The rule for determining conclusion 1 (cl) from (fl) is extracted as follows.

Expert ext: if 20≦fl≦60 th
en cl・(7) Expert ex2: if 30
≦f1≦50 then cl−=(8) The average value m of the minimum values and the average value calm In m of the maximum values are calculated.

IaK 20+30 □. 1□--25 river (9) tori ax H+ 50 bite 55 ...(10) Standard deviation of the minimum value σ and standard deviation of the maximum value in The deviation σ is calculated.

The above-mentioned rule (
For equations (1) to (6), if all the minimum and maximum values of each event involved in each conclusion are examined, a table as shown below is obtained.

(Left below) Table 3 Generally, in approximate reasoning, a membership function is given to an event. Here, as an example, a method for determining membership functions using Gaussian distribution using expert knowledge synthesized as described above will be described.

Average value m of minimum values,, average value m of maximum values.

1111n
Standard deviation of IaX minimum value σ1. Using the standard deviation a of the maximum value nσ, the membership function is expressed as 1aX by the following equation.

...(13) however.

X: Value of input data to the event Φ(X) Degree to which the two-person data X fits the event (degree of fit) Gauss(x): Value of the Gaussian distribution in the input X.

FIG. 3 shows an example of a Gaussian distribution. In the 20 Gaussian distribution only the left half is used to create the membership function. The position of X at Φ(x)-0,5 is m,
or m determined by ff1ln
IaK, the slope is determined by σ, or σ11n 1lla
X Ru.

As an example, the membership function for determining the conclusion c1 from the event f1 is created as shown in FIGS. 4a to 4c using the values calculated by equations (9) to (12). In this case, equation (13) becomes as follows.

...(14) Figure 4a is the first right-hand side of equation (13) or equation (14).
Figure 4b shows the second term on the right side of equation (13) or (14), and Figure 4C shows the result of subtracting the second term from the first term, that is, equation (13) or equation (14). It represents the membership function expressed by equation (14).

Each event rl, f'2. created based on the synthesized knowledge shown in Table 3. Regarding fi, conclusion c1. e2
．． Examples of membership functions for determining cj are shown in Figures 5a, 5b and 5c.

(4) Post-synthesis knowledge storage device The post-synthesis knowledge storage device 13 stores the average value and standard deviation calculated by the knowledge synthesis device 12 in the format shown in Table 3. Since it is not necessary to synthesize knowledge every time an inference is made, the results calculated in advance in this way are memorized. By reading out and using the values in the storage device 13 each time an inference is made, it is possible to speed up the inference process.

(5) Event value input device The event value input device 14 is a device that reads input data input for each event from a keyboard, communication interface device, memory, file, etc. The input data is given to the fitness computing device 15, and information as to whether data for each event has been input is given to the clarity addition device 22.

The event value may be given not only as a definite value but also as a linguistic value or a membership function.

(6) Fitness calculation device The fitness calculation device 15 calculates the fitness of the data input from the event value input device 14 for each membership function (or conclusion). Specifically, the goodness of fit is determined by substituting the input data as the variable X on the right side of equation (13) when the input data is a definite value.
It is required as. Of course, such an arithmetic expression does not necessarily have to be used. If the input data is a linguistic value or a membership function, the degree of fitness may be calculated using MIN-MAX, for example.

(7) Dynamic information calculation device and static information calculation device event r
The event value (input data) of event f2 is xl, and the event value of event f2 is x2. Let the event value of event fi be xi. These data are input from the event value input device 14.

As shown in Figures 6a and 6b, each degree of fitness 11
12 1j 21°"22"'2j is determined as follows.

m11: Degree of suitability of input data x1 to the membership function of conclusion c1 m12: Degree of suitability of input data xi to the membership function of conclusion c2 m1j: Degree of suitability of input data Xi to the membership function of conclusion cj m21: Input data x2 The goodness of fit of the conclusion cl to the membership function m22: The goodness of fit of the input data x2 to the membership function of the conclusion c2 m2j: The goodness of fit of the input data x2 to the membership function of the conclusion cj Generally, as shown in FIG. 6c, Let the goodness of fit m, j be the goodness of fit of the input data (event value) χ1 to the membership function of the conclusion cj.

These degrees of fitness are calculated based on the input data X1. X2. It is calculated by the fitness calculating device 15 when xi is given.

Here we consider the concept of fuzzy pot entropy.

Fuzzy entropy El when input xi is given
'l is defined as follows.

Efj--main ((m, / Mi) l og (m, j
/M, ) 1...(15) Here, M , -, E, m ij,, (16) This fuzzy entropy is a type of index of information discrimination ability, and when input data x1 is given, , the value is small enough that the conclusion can be clearly identified, and the value is large enough that the conclusion can only be vaguely identified.

The range of values that the fuzzy entropy Ef can take is as shown below.

0≦Ef≦log(n) n: number of conclusions on the event Next, using this fuzzy entropy Eft, dynamic information amount Iff (
Find xi). Here, the amount of dynamic information 1f'1o(xj
) is the ability to identify events in order to establish a conclusion when making inferences.

Therefore, the dynamic information amount IfD(xi
) from the maximum fuzzy entropy, the input data xi
is defined as minus the fuzzy entropy Eft when is given.

lfip(xi)-log(n)+mi1(mij/M,
) log (m1j/M, ) 1...(17
) The dynamic information amount of an event fi to which no input data xi is given is zero.

The dynamic information amount calculating device 16 uses the fitness obtained by the fitness calculating device 15 to calculate the dynamic information amount for each event according to equation (17).

The amount of dynamic information depends on the input data xi as described above.

On the other hand, the amount of static information does not depend on the input data, and the static information of the entire event is calculated by subtracting the average fuzzy entropy within the range of the event from the maximum fuzzy entropy. Quantity. The amount of static information about event i is expressed by the following formula 7% miJ (x'k'' input data xik about event fi
When calculating the fitness static information amount for the membership function of the conclusion cj, as shown in Figure 5C,
The range width xi1 to xi of the event ri is set at intervals of δ (p-1
). (1/p) [Σ
, 1. ] represents the operation of calculating the fuzzy entropy for each xik and finding the average of them.

Equation (18) and ji! As can be seen from equation (19), the greater the overlap between the event membership functions.

The amount of static information of an event is small, and the smaller the overlap between the membership functions of an event, the larger the amount of static information of an event.

In other words, the amount of static information indicates the ability of an event's membership function to identify a conclusion.

The static information amount calculating device 19 calculates and stores the amount of static information for each event from the membership function obtained by the combined wisdom according to the above-mentioned equations (18) and (19). do. Since the static information amount does not depend on input data, it only needs to be calculated once.

(8) Possibility calculation device For each conclusion, calculates the amount of information of events such that the sum of the amount of information of events involved in that conclusion is 1, and the relative strength of the amount of information of those events does not change. . This calculated amount of information is called a weight.

For example, if the amount of dynamic information described above is used, the weight vi for each event fi will be as follows.

vi −1ri (xi)/ ,X, IN, (xi)
-(20) Next, calculate the product of these weights and fitness.

The total sum for each conclusion is calculated as the probability Pr of each conclusion cj.

For example, in the above example, Pr, -Σ (wixmij)... (21> The possibility calculation unit 17 performs the above-mentioned calculation to calculate the probability for each conclusion.

(9) Possibility Display Device This possibility display device 18 displays the possibilities calculated by the possibility calculation device 17 for each conclusion. The possibility may be displayed for all conclusions, or one or more conclusions with high probability may be displayed. Furthermore, the possibility may be transmitted to another device through one communication and then stored in a memory or a file.

(10) Clarity calculation device The clarity calculation device 20 is a device that calculates the clarity of each event for each conclusion. Here, the clarity of each event for each conclusion indicates the relative discernment ability of each event when evaluating the possibility of a certain conclusion. Therefore, this clarity makes it possible to compare the discriminative ability of multiple events in order to establish a certain conclusion, and to understand which event has a high discriminative ability (contains a large amount of information). Ru.

The clarity of each event for each conclusion is determined by the following formula.

Tatashi.

A, giant Σ jk-1 Namely.

ml j (X l k) (23) If the membership function of conclusion cj exists in event fi, A1. >0 J If the event f i is the membership function of the conclusion cj, then A, -0 J The denominator of equation (22), which expresses the clarity of each event for each conclusion, is the membership function of the conclusion cj existing event ri
The total sum of the static information amount 1ris is calculated only for .

Since clarity is obtained by normalizing the amount of static information using equation (22), it is easy to intuitively understand whether it is a relative discrimination ability.

In this way, the clarity calculation device 20 calculates the clarity of each event for each conclusion.

(11) Clarity Storage Device The clarity storage device 21 is a device that stores the clarity of each event for each conclusion calculated by the clarity calculation device 20.

Clarity calculations do not need to be performed every time an inference is made. Therefore, the clarity calculated when knowledge is synthesized is stored in the clarity storage device 21.

Each time an inference is made, the values stored in the clarity storage device 21 are compared. This makes it possible to speed up the inference processing.

(12) Clarity Addition Device The clarity addition device 22 is a device that calculates the clarity of an event to which data is actually input for each conclusion. here,
for the inferences that are actually made.

Take the summation for each conclusion of the clarity of the events for which data has been entered. The sum of this clarity indicates the clarity of the inference result. It can be said that the higher the clarity, the greater the amount of information for deriving the inference result (conclusion). Therefore, clarity can be used as an indicator to judge the reliability of the inference result itself.

The clarity for the inference results (that is, for each conclusion) is calculated according to the following formula.

Cp, -Σ C11...(24> j +-+ lj However, in equation (24), the clarity CL of the event f1 for which no input data xi is given is treated as 0. In other words, the input data xi The addition of equation (24) only for clarity CI1... of a given event is row 1.
It is broken.

The possible range of the clarity B of the inference result is 0.0≦CL≦1.0. In other words, if you make an inference by inputting data about all the events that can be used to draw a certain conclusion within the knowledge given before making the inference,
The clarity of the conclusion is 10. Also, only some of the events that can be used to draw a certain conclusion.

If the data is entered, the clarity will be a value between 0.0 and 1.0. At this time, it can be said that if a large number of events with high clarity are used among the events that can be used, the clarity of the conclusion will be increased, and a highly reliable inference result will be obtained.

(13) Clarity Display Device The clarity display device 23 is a device that displays the clarity of the inference result (-the possibility mentioned above as an example) calculated by the clarity addition device 22. The clarity may be displayed along with the inference results, or the clarity may be transmitted to another device or stored in memory or a file.

This clarity display displays all conclusions of the inference results. Therefore, if there are multiple conclusions, the clarity corresponding to each conclusion is displayed.

In this way, every time data is entered.

By calculating the information amount of the event to which the input data belongs and displaying the clarity of the inference result, the user can judge the reliability of the inference result.

(14) Static information amount display/input instruction device (Fig. 1) As mentioned above, the static information amount for each event is calculated by the calculation device 19, and the results are shown for each conclusion as shown in Fig. 7. The information is stored in the arithmetic unit 19 or the static information amount display/input instruction device 31 in the form of a table. , in the table of FIG. 7, the order is in the order of the magnitude of the amount of static information.

The static information amount 1f+s indicates the ability of the event membership function to identify the conclusion, as described above.

The static information amount display/input instruction device 31 instructs to display the order of events for which data should be input in descending order of static information amount. This display may be displayed on the screen by adding a number representing the corresponding event to a table as shown in Figure 7, or by displaying and instructing the event of the data to be input each time event data is input on the screen. It can be anything. A human (operator) looks at this display and enters the event value input device 14.
Input event data from 1 to 2 in the order of events with the largest amount of static information. Or, in a case where the event data is represented by a detection signal from a sensor, the indicating device 31 inputs the detection signal from the sensor for each event in the order of the event value input device 14 in the order of the event with the largest amount of static information. Control as follows.

The approximate inference device of the first embodiment displays the amount of static information on the device 31 before performing approximate inference, and controls the input device 14 to have a human input event data or to input event data accordingly. to secure the data necessary for inference. After this, the fitness calculation device 15. The possibility calculation device 17 and the like are activated to perform approximate inference.

FIG. 8 shows the processing procedure by the three static information amount display and input instruction devices.

First, the static information amount of each event calculated by the static information amount calculation device 19 is displayed in order of magnitude, or an input instruction is given to the input device 14 to prompt or control the input of event data. (Step 41). When event data is entered (step 42), the data is entered for the event.

The clarity is added for each conclusion in the adding device 22 (step 43). The process ends when the added value of clarity for a predetermined conclusion reaches a predetermined value (a value sufficient to make the conclusion reliable, and generally a value close to 1) (step 44). Steps 42 and 43 may be repeated until the clarity addition value reaches a predetermined value for each of a plurality of predetermined conclusions.

When a human (operator) determines that a conclusion can be determined even if the added value of clarity does not reach a predetermined value, event data input may be stopped.

In the first embodiment, event data is input based on the amount of static information, but the level of clarity of each event for each conclusion calculated using the amount of static information is used as the standard. Then, for each conclusion, event data of corresponding events may be input in order of increasing clarity. In this case as well, the process ends when the added value of clarity reaches a predetermined value for each conclusion.

As described above, according to the first embodiment, since the static information amount of each event is arranged and output in descending order, the event data of each event is input in descending order of the static information amount. It becomes possible to do so. The amount of static information for an event indicates the membership function of that event or the ability to devise a conclusion. (with fewer data inputs) conclusions with greater clarity can be obtained. Approximate inference can be performed efficiently and processing time can be shortened. Especially for beginners who cannot or have difficulty deciding whether to enter the event data for the event first, it will tell you the order of the event data, so you can draw conclusions efficiently without hesitation. can.

(15) Event selection calculation device and event input guidance display device (Fig. 2) The second embodiment calculates the probability of a conclusion and adds clarity based on the data of each event input from the input device 14. It is useful for inputting additional event data when a conclusion with a sufficiently high degree of clarity has not yet been obtained from the already input event data once an approximate inference operation including the above is executed. For example, if the number of event data inputs is considered too small to evaluate a conclusion (the clarity of likely conclusions is low), the clarity value for each is reasonably high, but This is the case when the final conclusion cannot be determined accurately because the probability values of the two are almost the same.

In general, both probability and clarity are low when the number of event data input is small or when data on events that are particularly important for the conclusion are not input. Event selection calculation device 3
2 refers to the data about the clarity of each event for each conclusion stored in the 5 clarity storage device 21, and selects an event that maximizes the clarity for each conclusion (for which data has not yet been input). It is the one that selects the one that does not exist.

This selected event is provided to the event input instruction display 33, and the display 33 displays that data should be entered regarding the event. Seeing this, a human (worker) inputs data regarding the displayed event from the input device fii14. Alternatively, the event selected by the event selection arithmetic unit 32 and which maximizes the clarity is provided to the input device 14. In response to this, the input device 14 reads and takes in a detection signal of each event from a sensor or the like related to the event. It is preferable to perform the above processing starting from the one with the highest probability of conclusion.

9 to 11 show approximate inference results calculated based on given event data. FIG. 9 shows the values of clarity CD of each event for each conclusion stored in the clarity storage device 21. FIG. 10 shows a single event data input flag table, in which an input flag is provided for each event for each conclusion. When event data is input from the input device 14, the input flag is set to 1. If the input flag is 0, the event data has not yet been input. This input flag table is created by the input device 14. FIG. 11 shows the probability and clarity addition value C11 for each conclusion. In this figure, the conclusions are arranged in descending order of probability. The probability of the conclusion is obtained by the possibility calculation device 17, and the clarity addition value C1, is obtained by the clarity addition device 22.

The various data shown in FIGS. 9 to 11 described above are provided to the event selection calculation device 32 when the approximate inference is once completed. The event selection calculation device 32 uses these data to execute the processing shown in FIG.

Referring to FIG. 12, first, the most likely conclusion is selected from the data shown in FIG. 11 (step 51). Regarding this conclusion, it is checked whether the clarity addition value Cg is greater than or equal to a predetermined value (a value large enough to evaluate or confirm the conclusion) (step 52). If the additional clarity value CD is greater than or equal to the predetermined value, the next most likely conclusion is selected (step 60).

If the clarity addition value Cg, for the selected conclusion is less than the predetermined value, use the data shown in Figure 9 to select the event with the highest clarity CD, among the events related to the selected conclusion. Selectively step 53).

Referring to the input flag table shown in FIG. 10, it is checked whether event data has been input for the selected event (step 54). If no input has been made, the number of the event fi is given to the event input guidance display device 33 or the input device 14 (step 55).

Upon receiving the event number, the event input guidance display device 33 displays a command to input the event data of that event number. A person looks at this display and inputs event data of the displayed event number from the input device 14. When the input device 14 receives an event number, it reads and takes in event data from a sensor or the like regarding the event with that event number. It is preferable that event data for which there is no or unknown event data does not necessarily need to be input.

When event data is input, the clarity addition device 22 adds the clarity value of the event in its conclusion to the previous clarity addition value to obtain a clarity addition value Cp.

is updated (step 56). Then, it is checked whether the new added clarity value is greater than or equal to a predetermined value (step 57). The predetermined value in step 57 may be the same as or different from the predetermined value in step 52 described above.

If the added clarity value does not reach the predetermined value, referring to the data in FIG. 9, select the next highest clarity C11, for the same conclusion (step J58) and return to step 52. Repeat steps 53 to 57.

When the additional clarity value CL becomes equal to or greater than a predetermined value, the processing for that conclusion is finished.

Steps 52 to 5B, EiO processing is repeated for a predetermined number of conclusions in order from the most probable conclusion (Step 5
9).

When a human determines that a conclusion can be determined even if the added clarity value does not reach a predetermined value, the above processing for that conclusion may be stopped.

The event input guidance display device 33 may not only display the information on the screen, but may also print it out using a printer or store it in a memory or file using one communication.

When the above-described processing is completed for a predetermined number of conclusions, the goodness-of-fit calculation device 15. [Dynamic information calculation device] 6 Restart the possibility calculation device 17, etc., and approximate inference is executed again using the event data with newly input event data added, and the obtained new data (possibility of conclusion clarity, clarity addition value, etc.) are displayed on the display device ill, 23.

As described above, according to the second embodiment, it is not necessary to input all the event data in the first (initial) approximate inference, and the first (initial) inference is based on the tentatively input event data. Approximate inference is performed. Then, using the data obtained through this approximate inference, an event that maximizes the additional clarity value is selected. By inputting the event data of this selected event, it is possible to quickly obtain a sufficient clarity value for determining a conclusion without inputting data regarding all events, so a conclusion can be drawn efficiently in a short period of time. You will be able to do this.

The second example is particularly useful when the number of input event data is small and it is considered difficult or dangerous to evaluate the conclusion, or when there are several likely conclusions. In cases where it is thought that no data has been input, it becomes possible to input the data of the most effective phenomenon, and even people who are not necessarily experts can draw conclusions efficiently. Particularly for beginners, the system instructs the events for which data should be input one by one, allowing efficient input operations.

It goes without saying that each of the devices 11 to 23 and 31 to 33 described above can be realized by a combination unit including a memory and a display device. For example, knowledge synthesis device 12゜various calculation devices 15.1B, 17.19.20.22.31.32
is suitably realized by a CPU that operates according to a program.

[Brief explanation of the drawing]

FIG. 1 shows a first embodiment, and is a block diagram showing the overall configuration of an approximate inference device. FIG. 2 shows a second embodiment, and is a block diagram showing the overall configuration of an approximate inference device. FIGS. 3 to 6C are for explaining matters common to both embodiments. FIG. 3 is a graph showing a Gaussian distribution. Figures 4a to 4c are graphs showing how membership functions are formed. Figures 5a, 5b and 5c are graphs showing the membership functions obtained for each of the two events. FIGS. 6a, 6b, and 6c are graphs showing how the goodness of fit is determined. FIGS. 7 and 8 are for explaining the first embodiment. FIG. 7 is a diagram showing a static information amount table. FIG. 8 is a flow chart showing the processing procedure of the static information amount display/input instruction device. 9 to 12 are for explaining the second embodiment. FIG. 9 is a table showing the clarity of each event for each conclusion. FIG. 10 is a diagram showing an input flag table. FIG. 11 shows the probability and clarity addition value for each conclusion. FIG. 12 is a flow chart mainly showing the processing procedure of the event selection calculation device. 11...Knowledge storage device. 12... Wisdom synthesis device 13... Post-synthesis knowledge storage device 14... Event value input device 15... Fitness calculation device. 16...Dynamic information amount calculation device. 17... Possibility calculation device 18... Possibility display device 19... Static information amount calculation device. 20... Clarity calculation device. 21... Clarity memory. 22... Clarity addition device. 23... Clarity display device. 31...Static information amount display/input instruction device 32...Event selection calculation device 33...Event input guidance display device. Below

Claims (20)

[Claims] (1) A storage means for storing in advance the relationship between an event and a conclusion, a means for calculating the static information amount of each event using the relationship stored in the storage means, and a means for calculating the static information amount of each event using the relationship stored in the storage means, and a An approximate inference device comprising means for outputting static information amounts in descending order of magnitude in relation to events. (2) Means for instructing to input event data of events corresponding to the static information amount of each event calculated by the static information amount calculation means in descending order of the amount of static information calculated by the static information amount calculating means, according to claim (1). Approximate reasoning device. (3) Claim (1) comprising means for inputting event data of events corresponding to the static information amount of each event calculated by the static information amount calculating means in descending order of magnitude.
The approximate inference device described in . (4) Calculate the clarity of each event for each conclusion using the static information amount obtained from the static information amount calculation means, and use this calculated clarity to calculate the clarity of the event for which data was actually input. The approximate inference device according to claim 2 or 3, further comprising: clarity calculation means for calculating clarity for each conclusion by adding clarity. (5) Claim (4) comprising: means for controlling to repeat event data input and clarity addition until the clarity addition value calculated by the clarity calculation means reaches a predetermined value for at least one conclusion. The approximate inference device described in . (6) Approximate inference means for calculating the possibility of a conclusion by applying input data of each event to the relationship stored in the storage means, any one of claims (2) to (5). The approximate inference device described in . (7) The approximation inference means converts the input data into a degree of fitness using the membership function represented by the relationship, and the degree of fitness is used to calculate the amount of dynamic information for each event. 7. The approximate inference device according to claim 6, comprising: a dynamic information amount calculation means; and a possibility calculation means for calculating the possibility of a conclusion using the fitness degree and the dynamic information amount. (8) Storage means for storing the relationship between events and conclusions in advance; static information amount calculation means for calculating the static information amount of each event using the relationship stored in the storage means; A clarity calculation means that calculates the clarity of each event for each conclusion using the amount of static information obtained from the information amount calculation means, and a method that increases the clarity of each event for each conclusion obtained from the clarity calculation means. A means of displaying in order from Approximate reasoning device with (9) The approximate inference device according to claim (8), further comprising: clarity adding means for calculating the clarity of each conclusion by adding up the clarity of events for which data is actually input. (10) Claim (9) comprising: means for controlling to repeat event data input and clarity addition until the clarity addition value calculated by the clarity addition means reaches a predetermined value for at least one conclusion. The approximate inference device described in . (11) The approximate inference device according to claim (10), further comprising: approximate inference means for calculating the possibility of a conclusion by applying input data of each event to the relationship stored in the storage means. (12) The approximation inference means converts the input data into a degree of fitness using the membership function represented by the relationship, and the degree of fitness is used to calculate the amount of dynamic information for each event. The approximate inference device according to claim 11, comprising: a dynamic information amount calculation means; and a possibility calculation means for calculating the possibility of a conclusion using the fitness degree and the dynamic information amount. (13) Storage means for storing the relationship between events and conclusions in advance; approximate inference means for calculating the possibility of a conclusion by applying input data to the relationship stored in the storage means; and storage in the storage means. a calculation means for calculating the clarity of each event for each conclusion using the relationship that is given, an addition means for calculating the clarity of each event for each conclusion by adding the clarity of the events for which data is actually input, and a predetermined An approximate inference device comprising: means for selecting an event that maximizes the added value of clarity for the conclusion of . (14) Claim (1) wherein the predetermined conclusion is the most likely conclusion obtained by the approximate reasoning means.
The approximate inference device described in 3). (15) The approximate inference means converts the input data into a degree of fitness using the membership function expressed by the relationship, and the degree of fitness is used to calculate the amount of dynamic information for each event. 14. The approximate inference device according to claim 13, comprising: a dynamic information amount calculation means; and a possibility calculation means for calculating the possibility of a conclusion using the fitness degree and the dynamic information amount. (16) The approximate inference device according to claim (13), further comprising: display means for instructing input of event data of the event selected by the selection means. (17) The approximate inference device according to claim (13), further comprising: means for inputting event data of the event selected by the selection means. (18) Control means for repeating event data input and clarity addition until the clarity addition value calculated by the addition means reaches a predetermined value for the predetermined conclusion. 17). (19) The approximate inference device according to claim (13), further comprising means for controlling the selection means to repeatedly select events for a predetermined number of conclusions. (20) means for controlling the approximate inference means to recalculate the probability of a conclusion by applying all input data, including event data input according to the selection by the selection means, to the storage means; The approximate inference device according to any one of claims (16) to (19), comprising: