TWI529652B - Influenza-like illness diagnostic assistance system - Google Patents

Description

Influenza disease diagnosis aid system

The invention relates to a diagnostic aid system for influenza-like diseases, in particular to using various information about influenza-related diseases to establish a diagnostic aid system for influenza-like diseases in a fuzzy inference manner, so that it can be used as an auxiliary for medical personnel to diagnose influenza-like diseases.

Influenza-like and Common Cold have little difference between each other. For example, fever, cough, headache, etc. often occur and are not easy to distinguish. Therefore, when the patient complains of symptoms, it is usually difficult. Directly determine whether the infection is flu-like. At this time, the doctor will collect the flu virus from the throat swab for the patient, and then use the flu rapid screening kit to check whether the virus reaction is positive or negative, but the test may still be somewhat inaccurate. It is still necessary to rely on the doctor's observation and diagnosis experience to judge whether it is an influenza-like disease. Therefore, the whole diagnosis process is quite complicated, and there are many symptom factors to be considered, but the diagnosis result is still misdiagnosed. Therefore, how to establish an influenza-like disease The diagnostic aid system first makes a preliminary judgment on the known symptoms, assists the physician in clinical diagnosis, and improves the efficiency of diagnosis to further improve the diagnosis rate.

With the continuous development of computer technology, various expert-assisted diagnostic systems have also emerged. At present, the expert-assisted diagnostic system mainly uses the objective data obtained by pathological examination, imaging technology, microscopic examination, bacterial culture, or other instruments as the evidence for diagnosis. Correct In terms of the patient's self-reported symptoms or the physician's observation of the patient's external symptoms, because the doctor's professional experience is required to judge the state of the disease, there is a relatively unclear interval in the diagnostic criteria, so it cannot be directly compared by the general comparison criteria. Diagnosis, such as influenza-like illness, has similar or overlapping symptoms to the common cold, but if it can be considered for the symptoms and its severity, the corresponding diagnostic aid system is designed with different degrees of symptoms, and the patient is classified in the initial diagnosis. Carrying out will have a certain effect.

Therefore, the present invention provides an influenza-like disease diagnosis auxiliary system, which can determine whether a patient's symptoms are infected with an influenza-like disease by inputting a symptom of a patient, and using a fuzzy inference method to provide a physician or a nursing staff as a reference to achieve the aforementioned goal. .

In view of the above-mentioned problems of the prior art, the object of the present invention is to provide an influenza-like disease diagnosis auxiliary system, which blurs the symptoms of influenza-like diseases according to different severity levels, cooperates with appropriate fuzzy reasoning principles, and finally defuzzifies the symptoms. The output value is generated and compared with the preset standard value to determine whether the combination of different symptoms and different severity levels is an infectious influenza disease, so that the medical staff can initially screen whether the patient is infected with influenza-like diseases by referring to the judgment result.

According to an aspect of the present invention, an influenza-like disease diagnosis assisting system is provided, which comprises a storage module, a processing module, an input module and an output module. The storage module stores the symptom information of the symptoms of the influenza virus and the plurality of severity values corresponding to the symptom information, and defines the fuzzy attribution relationship between the symptom information and the plurality of severity values, and the fuzzy attribution is to attribute the plurality of severity values to the plural number. a fuzzy set, and determining that each of the plurality of severity values corresponds to a belonging weight value of the fuzzy set; the processing module establishes a judgment rule table according to the fuzzy attribution relationship, and performs all combinations of the plurality of fuzzy sets between different symptom information Establish a judgment weight value separately The weight value and the judgment weight value are used to calculate the output value, and the output value is compared with the preset standard value to determine whether it is an infectious influenza disease; the input module inputs the symptom information and the corresponding actual severity value information, and the processing module calculates the output. The value is used to determine whether the actual severity value is an infectious influenza disease; and the output module outputs the judgment result of the processing module for inspection and as an aid for judging whether it is an infectious influenza disease.

Preferably, the fuzzy attribution relationship between the symptom information and the plurality of severity values may define a plurality of fuzzy sets by the triangle attribution function, and the formula of the triangle attribution function is: Where μ ( x ) is the attribution weight value of the complex fuzzy set, b is the vertex value of the triangle attribution function, and a and c are respectively boundary values of the two sides of the triangle attribution function.

Preferably, the fuzzy attribution relationship between the symptom information and the plurality of severity values can be defined by a single-valued Gaussian type attribution function, such that their respective ownership weight values are all 1.

Preferably, the fuzzy attribution relationship between the symptom information and the plurality of severity values may define a plurality of fuzzy sets by a Gaussian type attribution function, and the formula of the Gaussian type attribution function is: Where μ ( x ) is the attribution weight value of the complex fuzzy set, b is the severity value of the complex fuzzy set, and c is the range of the severity value.

Preferably, the processing module calculates the output value by a product synthesis method, and the calculation formula is: Where Y is the output value, r is the attribution weight value, w is the judgment weight value, and n is the combination number of the plurality of fuzzy sets.

Preferably, the processing module compares the judgment result with the actual diagnosis result, and corrects the weight value according to the error between the judgment result and the actual diagnosis result.

Preferably, the input module can include a physiological measuring device, and directly measures the physiological signal and transmits the signal to the processing module.

Preferably, the physiological measuring device can be a Bluetooth ear thermometer, and the measured body temperature information is transmitted to the processing module by wireless transmission.

Preferably, the output unit can include a display that displays the result of the processing module determination.

Preferably, the input module and the output module can be a handheld device including a touch screen, and the touch screen is used to input the symptom information and the actual severity value, and the result is judged by the touch screen display.

As described above, according to the influenza-like disease diagnostic aid system of the present invention, it may have one or more of the following advantages:

(1) The diagnostic aid system for influenza diseases can provide a preliminary diagnosis of whether or not an influenza-like illness can be diagnosed by the patient's symptom information and severity value, and provide medical staff with auxiliary information for diagnosis and improve the efficiency of initial diagnosis of the patient.

(2) The diagnostic aid system for such influenza diseases can be diagnosed by actual diagnosis. If the comparison is to correct the judgment weight value in the system, the judgment result can be closer to the actual diagnosis result, and the diagnosis rate of the diagnosis is improved.

(3) This type of influenza disease diagnosis aid system can transmit data to the system through handheld devices and wireless transmission methods, reducing human error and improving the convenience of use.

10‧‧‧Storage module

11‧‧‧Symptom Information

12‧‧‧Severity value

13‧‧‧ Fuzzy attribution

20‧‧‧Processing module

21‧‧‧Fuzzy diagnostic unit

211‧‧‧Fuzzification

212‧‧‧Fuzzy reasoning

213‧‧‧Unfuzzification

30‧‧‧Input module

40‧‧‧Output module

50‧‧‧ class neural network architecture

51‧‧‧Input layer

52‧‧‧ attribution layer

53‧‧‧ rule layer

54‧‧‧Inference layer

55‧‧‧Output layer

56‧‧‧ Actual diagnosis results

Figure 1 is a block diagram of a diagnostic aid system for influenza diseases such as the present invention.

Figure 2 is a schematic diagram of the fuzzy attribution relationship of the present invention.

Figure 3 is a schematic diagram of another fuzzy attribution relationship of the present invention.

Figure 4 is a schematic diagram of still another fuzzy attribution relationship of the present invention.

Figure 5 is a schematic diagram of the fuzzy inference mode of the present invention.

Figure 6 is a schematic diagram of a neural network architecture such as the present invention.

The technical features, contents, advantages and advantages of the present invention will be understood by the reviewing committee, and the present invention will be described in detail with reference to the accompanying drawings. The subject matter is only for the purpose of illustration and description. It is not intended to be a true proportion and precise configuration after the implementation of the present invention. Therefore, the scope and configuration relationship of the attached drawings should not be interpreted or limited. First described.

Please refer to FIG. 1 , which is a block diagram of the influenza disease diagnosis auxiliary system of the present invention. As shown in the figure, the influenza-like disease diagnosis auxiliary system includes a storage module 10 , a processing module 20 , an input module 30 , and an output . Module 40. Storage module 10 Interview with experts to understand the clinical symptoms of the diagnosis of influenza-related diseases, or through medical literature for influenza-like illnesses, such as fever, cough, headache, muscle aches, fatigue, stuffy nose, sore throat, runny nose, fight Sneezing or chills, dizziness, diarrhea, vomiting, joint pain and other symptoms, as a symptom of the symptoms of influenza-like diseases,11, because these symptoms may be partly the same or similar to the symptoms of the common cold, can not directly related to whether it is related Symptoms to determine whether it is a flu-like illness or a general cold, you must increase the severity of the various symptoms of the information 11 to make further judgments, so set a corresponding severity value of 12 for different symptoms, the severity of the symptoms, Numerically, these severity values 12 can be filled in by questionnaires in advance by the patient, or obtained through preliminary interviews by medical staff.

As described above, the symptom information 11 and the severity value 12 of the patient are usually judged by the medical knowledge and accumulated experience of the physician to determine whether the patient is infected with the influenza-like disease. This method makes the judgment standard have some vague space, so the class The flu disease diagnosis aid system uses the characteristics of fuzzy theory to design the system. Each symptom information 11 is established with a fuzzy attribution relationship 13 between different severity values 12, where the fuzzy attribution relationship 13 refers to dividing the severity value 12 into a plurality of fuzzy sets according to the symptom type, such as an asymptomatic collection, The symptoms are mildly aggregated, the symptoms are severely aggregated, and the severity value 12 is assigned to the belonging weight values of different fuzzy sets. For example, when the body temperature is 38 ° C, 80% of the total is attributed to the fever. The value of the attribution weight is 0~1, which can be defined by the attribution function. After having the information of the symptom information 11 and the fuzzy attribution relationship 13 corresponding to the severity value 12, the processing module 20 includes a fuzzy diagnosis unit 21, and the fuzzy diagnosis unit 21 utilizes different symptom information 11 and its corresponding severity value 12 The inference method of fuzzy theory judges whether it is an infectious influenza disease, according to historical data or expert opinions. The judgment weight value in the fuzzy diagnosis unit 21 is set to be close to the judgment standard of the actual diagnosis.

After the establishment of the ownership weight value and the judgment weight value in the influenza-like disease diagnosis auxiliary system, the user can input the symptom information 11 of the patient and the severity value of the symptom 12 to the processing module by using the appropriate input module 30. 20, the input module 30 includes various input interfaces, such as keyboard, mouse, touch screen, etc., and the input module can be connected to various physiological measuring devices, such as Bluetooth ear temperature. The gun transmits the measured body temperature information directly to the processing module 20 in a wireless manner, further improving input efficiency and reducing human input errors. The input information is calculated by the processing module 20 according to the fuzzy diagnosis unit 21, and the output value is compared with the preset standard value to determine whether the symptom information 11 and the severity value 12 of the patient are infectious influenza diseases, and The result is output via the output module 40. Here, the output module 40 can include a computer or a handheld device used by the medical staff, so that the medical staff can refer to the judgment result to assist the actual diagnosis.

The fuzzy attribution relationship for different symptom information and its severity value will be further explained below. In Fig. 2, the fuzzy attribution relationship of symptom information is defined by a single-valued Gaussian type attribution function, and the third and fourth figures are triangle attribution functions. To define the fuzzy attribution relationship of the symptom information, but the invention is not limited thereto. For example, the third and fourth pictures may also define a fuzzy attribution relationship by a Gaussian type attribution function, or may be defined by a trapezoid and other related attribution functions, depending on the symptom. The characteristics of the information are determined. Please refer to FIG. 2, which is a schematic diagram of the fuzzy attribution relationship of the present invention. As shown in the figure, a fuzzy attribution relationship is defined for the episode of fever in the symptom information, wherein the episode of fever is divided into no according to the severity value, the occurrence time is less than 6 hours, the occurrence time is 6 to 12 hours, and the occurrence time is greater than 12 hours. 4 kinds of fuzzy sets, with 0, 1, 2, 3 as their input values, and using single-valued Gaussian type attribution function The four fuzzy sets are defined separately, so that the attribution weights are all 1, that is, in the symptom information of the patient, the time of the attack period is completely attributed to the relevant fuzzy set according to the above range, for example, suddenly within 6 to 12 hours. A fever may be diagnosed as being infected with an influenza-like disease. Therefore, if the symptom of the patient is a 10-hour episode, enter 2 a fuzzy set belonging to the occurrence time of 6 to 12 hours, the attribution weight value is 1, and the subsequent judgment inference has A higher judgment weight value.

Please refer to FIG. 3, which is a schematic diagram of another fuzzy attribution relationship of the present invention. As shown in the figure, the fuzzy relationship is defined for the body temperature in the symptom information. The interval between the measured body temperature is set between 35 ° C and 39 ° C, but the body temperature is not as clear as the time of occurrence to distinguish the fever or the body temperature is normal. When the fuzzy relationship is established for body temperature, the two fuzzy sets for fever and normal body temperature are respectively defined by the triangle attribution function. As shown in the figure, when the input body temperature is less than 37 °C, the body temperature is normal, and when it is greater than 36 °C. It belongs to the fever, so that the measured body temperature may meet the two kinds of fuzzy sets at the same time. According to the function value, there are corresponding attribution weight values in the two fuzzy sets, and the attribution weight value is between 0 and 1. For example, if the measured body temperature is At 38 ° C, the weight of the normal body temperature is close to 0, and the weight of the fever is 0.8.

Please refer to FIG. 4, which is a schematic diagram of still another fuzzy attribution relationship of the present invention. As shown in the figure, as shown in the figure, if the symptoms are cough, headache, muscle pain, fatigue, nasal congestion or sore throat, the diagnosis is more blurred, so When defining the fuzzy attribution relationship for this type of symptom, the patient thinks that the severity of the symptom is divided into 1~9 level input. The greater the severity value, the more severe the symptom is, and the fuzzy set is divided into the asymptomatic Slight, normal and severe, respectively defined by the triangle attribution function, as shown in the figure, when the severity of the input is 1~4, it belongs to the slightest, and when it is 3~7, it belongs to the ordinary, at 6~9. Attributable to serious, according to the corresponding function value is not There is a corresponding attribution weight value in the same fuzzy set, and the attribution weight value is between 0 and 1. For example, when the input severity value is 8, the weight values attributed to the slight and ordinary fuzzy sets are close to 0, and the attribution is The weight value for a severe fuzzy set is 0.5.

Please refer to FIG. 5, which is a schematic diagram of the fuzzy diagnosis method of the present invention. As shown in the figure, the fuzzy diagnosis method is that the user inputs the symptom information of the patient to be determined by the input module 30 and the severity value of the actual collection, and the result is output to the output module 40 after being processed by the fuzzy diagnosis unit 21. The fuzzy diagnosis unit 21 includes three parts: blurring 211, fuzzy reasoning 212, and defuzzification 213. The blurring 211 is an episode of fever, body temperature, cough, headache, muscle pain, fatigue, nasal congestion, and sore throat. According to the fuzzy attribution relationship in the storage module, the severity input values of the individual symptoms of the patient are respectively assigned to the set fuzzy set, so that they have corresponding to the fuzzy set of belongings. The attribution weight value; the fuzzy reasoning 212 is to establish a judgment rule table for the relationship between the fuzzy sets of all the symptoms, for example, to select two kinds of symptom information in Table 1, the cough X ₁ or the headache X _{2 have} two kinds of symptom information respectively. Fuzzy sets (none, slight, normal, severe), which have corresponding attribution weight values μ ₁₁ ~ μ ₁₄ , μ ₂₁ ~ μ ₂₄ , and the relationship between the symptoms of the fuzzy sets has a total of 16 combinations, which represent a total of 16 kinds of symptoms determination rule r ₁ ~ r _16, each provided with a determination rule determining weights corresponding to the weight value w ₁ ~ w _16, the weight values determined weights w ₁ ~ w ₁₆ can utilize Or historical data generated through subsequent comparison with the results to correct the actual diagnosis, determines weight values w ₁ ~ w ₁₆ between 0 and 1.

The part of the defuzzification 213 is calculated by using the judgment weight values w ₁ to w ₁₆ and the attribution weight values μ ₁₁ ~ μ ₁₄ and μ ₂₁ ~ μ ₂₄ as the output values of the fuzzy inference. The output value is obtained by the calculation of the product synthesis method, and the calculation formula is:

In the above case, the number of n is 16, and the output value Y is calculated according to the above formula, and compared with the preset standard value to diagnose whether it is an infectious influenza disease. The above is only an example of two symptoms. However, the present invention is not limited to this, and the information on other related symptoms can also be obtained by using the above calculation method to obtain an output value, thereby determining whether it is an infectious influenza disease.

Please refer to FIG. 6, which is a schematic diagram of a neural network architecture such as the present invention. As shown, the influenza disease diagnostic aid system of the present invention can be configured in conjunction with a neural network architecture 50 that includes an input layer 51, a home layer 52, a rule layer 53, an inference layer 54, and an output layer 55, each layer Between the plurality of neurons, the neuron has an initial weight value, and the final output values are calculated by using the initial weight values, and the initial weight values may be set first through historical data, or generated by using a random method. After the learning and correction process, the results of the diagnosis of the entire influenza-like diagnostic aid system can be brought closer to the actual diagnosis of the physician. Wherein, the input layer 51 is a type of influenza symptoms diseases i X _i, X _i with each of j different types of symptoms fuzzy sets corresponding to the severity of each symptom X _i value can be attributed to the home of the layer 52 corresponding to the fuzzy set and Corresponding attribute weight value μ _ij is formed, and the judgment rule r _k , k in the rule layer 53 is formed according to the mutual relationship between the fuzzy sets, and k is the number of all combinations, and finally inferred according to the information of the home layer 52 and the rule layer 53. The calculation of the layer 54 obtains the determined output value y . After comparing the output value with the preset standard value, the determination result Y is outputted by the output layer 55, and the user can obtain the result of the judgment via the output layer 55 as an auxiliary diagnosis. in accordance with. In addition, the neural network architecture 50 can also correct the input data via the physician D actual diagnosis result 56, and compare the error between the judgment result Y and the actual diagnosis result 56, thereby correcting the aforementioned weight value to make the entire influenza-like disease. The diagnostic aid system can be improved. In such a neural network architecture 50, the number of levels and nodes to be calculated is not limited to the number shown in this embodiment. To obtain a more accurate judgment, the number of layers of the network architecture and the number of judgment rules can be increased, but the relative calculation is performed. The time may be lengthened, causing the output to take a long time. Therefore, the number of layers and the number of judgment rules of the above architecture can be adjusted according to actual operation conditions to meet the best benefit.

The above is intended to be illustrative only and not limiting. Any equivalent modifications or alterations to the spirit and scope of the invention are intended to be included in the scope of the appended claims.

10‧‧‧Storage module

11‧‧‧Symptom Information

12‧‧‧Severity value

13‧‧‧ Fuzzy attribution

20‧‧‧Processing module

21‧‧‧Fuzzy diagnostic unit

30‧‧‧Input module

40‧‧‧Output module

Claims (9)

The invention relates to a diagnosis system for influenza-like diseases, which comprises: a storage module, which is a symptom information of a symptom of storing influenza-like diseases and a plurality of severity values corresponding to the symptom information, and defines the symptom information and the plurality of severity values. a fuzzy attribution relationship, the fuzzy attribution relationship assigning the plurality of severity values to the plurality of fuzzy sets, and determining that each of the plurality of severity values corresponds to one of the fuzzy sets of belonging weight values; And establishing a judgment rule table according to the fuzzy attribution relationship, respectively establishing a judgment weight value for all combinations of the plurality of fuzzy sets between the symptom information, and calculating an output value by using the home weight value and the judgment weight value And comparing the output value with a preset standard value to determine whether it is an infectious influenza disease; an input module is configured to input the symptom information and information corresponding to one actual severity value, and the processing module calculates the output value Determining whether the actual severity value is an infectious influenza disease; and an output module is the processing The output determination for the group of view and determines whether the auxiliary as influenza infection diseases; wherein the processing module is calculated based synthesis to a product of the output value, which is calculated as: Where Y is the output value, r is the attribution weight value, w is the judgment weight value, and n is the combination number of the plurality of fuzzy sets. The influenza disease diagnosis auxiliary system as described in claim 1, wherein the fuzzy attribution relationship between the symptom information and the plurality of severity values is a plurality of fuzzy sets defined by a triangle attribution function, and the triangle attribution function The formula is: Where μ ( x ) is the attribution weight value of the complex fuzzy set, b is the vertex value of the triangle attribution function, and a and c are respectively boundary values of the two sides of the triangle attribution function. The influenza disease diagnosis auxiliary system as described in claim 1, wherein the fuzzy attribution relationship between the symptom information and the plurality of severity values is defined by a single-valued Gaussian type attribution function, so that each of the The attribution weights are all 1. The influenza disease diagnosis auxiliary system as described in claim 1, wherein the fuzzy attribution relationship between the symptom information and the plurality of severity values is defined by a Gaussian type attribution function, the Gaussian type attribution The formula for the function is: Where μ ( x ) is the attribution weight value of the complex fuzzy set, b is the curve center of the Gaussian type attribution function, c is the curve width of the Gaussian type attribution function, and e is a mathematical constant. Assessing the diagnosis of influenza diseases as described in item 1 of the patent application The system, wherein the processing module compares the judgment result with the actual diagnosis result, and corrects the judgment weight value according to an error between the judgment result and the actual diagnosis result. For example, the influenza disease diagnosis assistance system described in claim 1 , wherein the input module comprises a physiological measurement device, and directly measures a physiological signal and transmits the signal to the processing module. For example, the influenza disease diagnosis assisting system described in claim 6 is wherein the physiological measuring device is a Bluetooth ear thermometer, and the measured body temperature information is transmitted to the processing module by wireless transmission. The influenza disease diagnosis assistance system as described in claim 1, wherein the output unit comprises a display, and the result of the determination by the processing module is displayed. For example, the input module and the output module are handheld devices including a touch screen, and the symptom information and the actual use are input by using the touch screen. The severity value is displayed by the touch screen display.