TWM589863U - System for constructing prediction models with big data technology - Google Patents

Abstract

A system for constructing a predictive model using big data technology, including a communication module connected to a user terminal through the communication network, a diagnostic case for storing multiple related diseases and corresponding to multiple patients, and multiple cases A storage module related to the characteristic attributes of the disease, each diagnosed patient case includes at least one target characteristic attribute selected from the characteristic attributes, and a processing module, receiving a After the disease includes at least one undiagnosed patient case to be compared with the characteristic attribute, according to the at least one characteristic feature to be compared and the at least one target characteristic attribute in each confirmed patient case, a case-based reasoning is used to obtain a correspondence The incidence probability score of the undiagnosed patient case, and the incidence probability score is transmitted to the user end through the communication module.

Description

A system for constructing prediction models using big data technology

The present invention relates to a prediction system, especially a system related to a disease, which uses big data technology to construct a prediction model.

The attitude of modern medicine to diseases focuses on early prevention, early diagnosis, and early treatment, so that the current medical science is also full of various disease prediction systems, such as various biochemical test data obtained based on health examinations, or According to the living habits of the examinee, after the analysis, the examinee is informed that the body may be at risk of suffering from a certain disease. Therefore, this case proposes a brand new prediction system to provide the choice or reference for the examinee or doctor.

Therefore, the purpose of the present invention is to provide a system capable of predicting the incidence of a disease.

Therefore, the system for constructing a predictive model using big data technology includes a communication module, a storage module, and a processing module electrically connected to the communication module and the storage module.

The communication module is connected to the user terminal through the communication network.

The storage module is used to store multiple diagnosed patient cases related to a disease and corresponding to multiple patients, and multiple feature attributes related to the disease. Each diagnosed patient case includes at least one selected from such The target characteristic attribute of the characteristic attribute.

The processing module receives an undiagnosed patient case related to the disease from the user terminal and including at least one characteristic attribute to be compared through the communication module, according to the at least one characteristic attribute to be compared, and each The at least one target feature attribute in a confirmed patient case, using a case-based reasoning to obtain an incidence probability score corresponding to the undiagnosed patient case, and transmitting the incidence probability score to the user end through the communication module .

The effect of the present invention lies in that the processing module uses the at least one feature attribute to be compared in the undiagnosed patient case and the at least one target feature attribute in each confirmed patient case, and utilizes the case Formula reasoning to obtain the morbidity score corresponding to the undiagnosed patient case, and then transmit the morbidity score to the user end through the communication module, so that the examinee or doctor holding the user end knows the disease The probability score is used as a reference for future treatment.

Referring to FIG. 1, an embodiment of the new system 1 for constructing a prediction model using big data technology is connected to a user terminal 2 through a communication network 100 and used to execute a disease incidence prediction method.

The system 1 for constructing prediction models using big data technology includes a communication module 11 connected to the communication network 100, a storage module 12, and a processing module 13 electrically connected to the communication module 11 and the storage module 12 . The storage module 12 is used to store multiple pieces of diagnosed patient information related to a disease and corresponding to multiple patients, multiple characteristic attributes related to the disease, multiple weights corresponding to these characteristic attributes, and multiple strokes respectively Diagnosed patient cases that are related to the disease and correspond to multiple patients, and each confirmed patient case includes at least one target feature attribute selected from the feature attributes.

Each diagnosed patient's information includes a complaint record related to the patient's subjective statement of his or her own condition, a separate complaint record related to information obtained by a medical staff objectively observing the patient directly or indirectly, and a related The patient's assessment of the patient's condition and a medical plan related to the patient.

The user terminal 2 includes a user terminal communication module 21 connected to the communication network 100, a user terminal display module 22, and a user terminal processing electrically connected to the user terminal communication module 21 and the user terminal display module 22 Module 23.

In this embodiment, the implementation of the system 1 that uses big data technology to construct a predictive model is, for example, a personal computer, a server, or a cloud host, but not limited to this.

In this embodiment, the implementation of the user terminal 2 is, for example, a personal computer or a handheld electronic device, but it is not limited thereto.

The following will explain the operation details of each component of the new system 1 and the user terminal 2 using the big data technology to construct a prediction model by using the disease incidence prediction method. The disease incidence prediction method includes a feature attribute acquisition procedure and an incidence probability Score earning procedure.

Referring to FIG. 2, the feature attribute obtaining process obtains the feature attributes based on the diagnosed patient information, and includes a step 51, a step 52, and a step 53.

In step 51, the processing module 13 obtains the total number of times each keyword appears in the diagnosed patient information according to the multiple keywords in each diagnosed patient information. In particular, the value is that in this embodiment, the processing module 13 uses Wen Ziyun to obtain the complaint record related to the patient’s subjective statement of his condition, and the objective objective of the medical staff All the keywords in the other complaint records of the information directly or indirectly observed by the patient. In addition, the processing module 13 uses the R language to obtain the evaluation status related to the patient's condition and related to the patient. All the keywords in this medical plan.

In step 52, the processing module 13 obtains a plurality of target keywords greater than a preset number of times and corresponding keywords from the keywords based on the total number of times each keyword appears in the diagnosed patient information Weights.

In step 53, the processing module 13 uses the target keywords as the characteristic attributes, and then stores the characteristic attributes and their corresponding weights in the storage module 12. In particular, the value is that in this embodiment, for each feature attribute, the greater the total number of times the feature attribute corresponds, the greater the weight corresponding to the feature attribute.

Referring to FIG. 3, the morbidity score obtaining program uses a case-based reasoning to obtain an morbidity score, and includes a step 61, a step 62, and a step 63.

In step 61, the user terminal 2 transmits an undiagnosed patient case related to the disease and including at least one feature attribute to be compared to the system for constructing a prediction model using big data technology through the user terminal communication module 21 1.

In step 62, after receiving the undiagnosed patient case from the user terminal 2 through the communication module 11, the processing module 13 according to the at least one characteristic feature to be compared, and in each confirmed patient case The at least one target feature attribute, using the case-based reasoning (CBR) to obtain the incidence probability score corresponding to the undiagnosed patient case, and then transmitting the incidence probability score through the communication module 11 To the end of use 2.

Referring to FIG. 4, it is worth noting that the step 62 further includes a sub-step 621 and a sub-step 622.

In sub-step 621, the processing module 13 calculates the undiagnosed patient according to the at least one feature attribute to be compared in the undiagnosed patient case and the at least one target feature attribute in each confirmed patient case The degree of similarity between the case and each confirmed patient case, and from these confirmed patient cases, a target confirmed patient case with the highest degree of similarity to the undiagnosed patient case is obtained.

In sub-step 622, the processing module 13 obtains the morbidity score corresponding to the undiagnosed patient case according to the similarity between the target confirmed patient case and the undiagnosed patient case, and then through the communication module 11 Send the incidence probability score to the user terminal 2.

…(1) It is worth noting that in sub-step 621, for each confirmed patient case, the processing module 13 uses the following formula (1) to calculate the degree of similarity between the undiagnosed patient case and the confirmed patient case To obtain the target confirmed patient case with the highest degree of similarity to the undiagnosed patient case.

…(1)

Among them, n represents the number of these characteristic attributes, i represents each characteristic attribute from 1 to n, Wi represents the weight of each characteristic attribute, sim(Xi, Yi) represents the i-th characteristic attribute, the undiagnosed patient case A similar value between Xi and the confirmed patient case Yi. Among them, the similarity value sim(Xi, Yi) is Bollinger's value. When the undiagnosed patient case Xi and the confirmed patient case Yi both have the i-th feature attribute, the similarity value sim(Xi, Yi) is true.

In addition, in this embodiment, the higher the degree of similarity between the target confirmed patient case and the undiagnosed patient case, the higher the incidence probability score corresponding to the undiagnosed patient case.

In step 63, the user-side processing module 23 receives the morbidity score from the system 1 that constructs the prediction model using big data technology through the user-side communication module 21, and then displays the morbidity score on the usage端Displaymodule 22.

In summary, in the new system of constructing a predictive model using big data technology, the processing module 13 is based on the at least one feature attribute to be compared in the undiagnosed patient case received through the communication module 11 , And the at least one target feature attribute in each confirmed patient case, and using the case-based reasoning (formula (1)) to obtain the incidence probability score corresponding to the undiagnosed patient case, and then through the communication module 11 Transmit the morbidity probability score to the user terminal 2, so that the subject or doctor holding the operative terminal 2 knows the morbidity probability score and serves as a reference basis for future treatment. Therefore, it can indeed achieve the purpose of new cost.

However, the above are only examples of the new model. When the scope of the new model cannot be limited by this, any simple equivalent changes and modifications made according to the patent application scope and patent specification content of the new model are still regarded as Within the scope of this new patent.

100‧‧‧Communication network 1‧‧‧ Big data technology to build predictive model system 11‧‧‧Communication module 12‧‧‧storage module 13‧‧‧Processing module 2‧‧‧Use end 21‧‧‧Use end communication module 22‧‧‧Use end display module 23‧‧‧Using end processing module 51~53‧‧‧Step 61~63‧‧‧Step 621, 622‧‧‧ Substep

Other features and functions of the new model will be clearly presented in the embodiment with reference to the drawings, in which: FIG. 1 is a block diagram illustrating an embodiment of the new model system for constructing prediction models using big data technology; FIG. 2 FIG. 3 is a flowchart illustrating a procedure for obtaining a characteristic attribute of a disease incidence prediction method performed by the embodiment; FIG. 3 is a flowchart illustrating a procedure for obtaining an incidence probability score of the disease incidence prediction method performed by the embodiment; and FIG. 4 It is a flowchart illustrating the detailed flow of how the morbidity score obtaining program executed in the embodiment obtains an morbidity score.

100‧‧‧Communication network

1‧‧‧ Big data technology to build predictive model system

11‧‧‧Communication module

12‧‧‧storage module

13‧‧‧Processing module

2‧‧‧Use end

21‧‧‧Use end communication module

22‧‧‧Use end display module

23‧‧‧Using end processing module

Claims (10)

A system for constructing a prediction model using big data technology is connected to a user terminal through a communication network. The system for constructing a prediction model using big data technology includes: a communication module connected to the user terminal through the communication network; and a storage module , Used to store multiple confirmed patient cases related to a disease and corresponding to multiple patients, and multiple characteristic attributes related to the disease, each confirmed patient case includes at least one selected from the characteristic attributes Target characteristic attribute; and a processing module, which is electrically connected to the communication module and the storage module, and receives, through the communication module, an unrelated feature related to the disease and including at least one characteristic attribute to be compared After confirming the patient case, according to the at least one feature attribute to be compared and the at least one target feature attribute in each diagnosed patient case, a case-based reasoning is used to obtain an incidence probability corresponding to the undiagnosed patient case Score, and send the incidence probability score to the user end through the communication module. The system for constructing a predictive model using big data technology as described in claim 1, wherein the processing module is based on the at least one feature feature to be compared in the undiagnosed patient case and the At least one target feature attribute, calculate the similarity between the undiagnosed patient case and each confirmed patient case, and from these confirmed patient cases, obtain a highest degree of similarity to the undiagnosed patient case The target confirmed patient case, and then according to the similarity between the target confirmed patient case and the undiagnosed patient case, obtain the probability score corresponding to the undiagnosed patient case. The system for constructing predictive models using big data technology as described in claim 2, wherein the higher the degree of similarity between the target confirmed patient case and the undiagnosed patient case, the corresponding incidence of the undiagnosed patient case The higher the score. The system for constructing prediction models using big data technology as described in claim 3, wherein the storage module also stores a plurality of weights corresponding to the characteristic attributes respectively. For each confirmed patient case, the processing module uses the following Formula to calculate the similarity between the undiagnosed patient case and the confirmed patient case:

, Where n represents the number of such feature attributes, i represents each feature attribute from 1 to n, Wi represents the weight of each feature attribute, and sim(Xi,Yi) represents the i-th feature attribute, the undiagnosed patient case A similar value between Xi and the confirmed patient case Yi. The system for constructing prediction models using big data technology as described in claim 4, wherein the similarity value sim(Xi, Yi) is the Bollinger value, when both the undiagnosed patient case Xi and the confirmed patient case Yi exist For the i-th feature attribute, the similarity value sim(Xi, Yi) is true. The system for constructing a predictive model using big data technology as described in claim 5, wherein the storage module also stores multiple pieces of confirmed patient information related to the disease and corresponding to multiple patients, respectively, and the processing module is also based on Multiple keywords in each diagnosed patient information to obtain the total number of times each keyword appears in the diagnosed patient information, and from these keywords, obtain multiple target keywords greater than a preset number of times These characteristic attributes are stored in the storage module. The system for constructing a predictive model using big data technology as described in claim 6, wherein the processing module also obtains and stores in the storage according to the total number of occurrences of the information of the diagnosed patients corresponding to each feature attribute Each feature attribute of the module has a corresponding weight. The system for constructing a prediction model using big data technology as described in claim 7, wherein for each feature attribute, the more the total number of times the feature attribute corresponds, the greater the weight corresponding to the feature attribute. The system for constructing a predictive model using big data technology as described in claim 6, wherein each diagnosed patient information stored in the storage module includes a complaint record related to the patient's subjective statement of his condition, a Relevant records related to information directly or indirectly observed by a medical staff on the patient, an assessment status related to the patient's condition, and a medical plan related to the patient. The system for constructing a predictive model using big data technology as described in claim 9, wherein the processing module uses Wen Ziyun to obtain the relevant main complaint record and all keywords in the other complaint records, the processing module It is to use R language to obtain the relevant assessment status and all keywords related to the medical plan.

Images