CN115995301A - Data recovery method and system based on artificial intelligence - Google Patents
Data recovery method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN115995301A CN115995301A CN202211647424.8A CN202211647424A CN115995301A CN 115995301 A CN115995301 A CN 115995301A CN 202211647424 A CN202211647424 A CN 202211647424A CN 115995301 A CN115995301 A CN 115995301A
- Authority
- CN
- China
- Prior art keywords
- prescription
- diagnosis
- treatment
- result
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention provides a data recovery method and system based on artificial intelligence, and relates to the technical field of data recovery. The invention solves the technical problems of slow diagnosis and treatment process and low efficiency of traditional Chinese medicine in the prior art, realizes the rapid traversal of the search library by computer technology, obtains diagnosis and treatment candidate prescriptions, and achieves the effects of providing reference for doctor diagnosis and treatment and further improving doctor diagnosis and treatment efficiency.
Description
Technical Field
The invention relates to the technical field of data restoration, in particular to a data restoration method and system based on artificial intelligence.
Background
The traditional Chinese medicine has been applied to the ancient land of China for thousands of years, and clinical practice for thousands of years proves that the traditional Chinese medicine of China is effective and feasible in terms of treatment, disease prevention and health preservation. Before Western medicine is introduced into China, chinese medicine is used for treating diseases by ancestor, lives of countless people are saved, and the traditional Chinese medicine diagnosis and treatment method still has certain defects, and the traditional Chinese medicine is mainly carried out by a manual recording and sorting mode in an exemplary mode like the traditional Chinese medicine study at present, so that the problems of complicated carrying mode and low carrying efficiency exist. In addition, because the personal time and energy of the traditional Chinese medicine specialists are limited, how to improve the teaching speed and effect of the diagnosis experience of each traditional Chinese medicine specialist by using the scientific technology, thereby enabling more young doctors to finally serve more patients has important significance. Therefore, a certain lifting space exists for the utilization of the high-quality diagnosis and treatment experience of the traditional Chinese medicine.
The traditional Chinese medicine diagnosis and treatment still relies on the traditional sitting and diagnosing mode, and diagnosis and treatment process is slow, and whole diagnosis and treatment efficiency is low, is difficult to make full use of to high-quality diagnosis and treatment experience.
Disclosure of Invention
The embodiment of the application provides a data recovery method and system based on artificial intelligence, which are used for solving the technical problems of slow diagnosis and treatment process and low efficiency of traditional Chinese medicine in the prior art.
In view of the above problems, embodiments of the present application provide a data recovery method and system based on artificial intelligence.
In a first aspect, an embodiment of the present application provides an artificial intelligence-based data restoration method, where the method includes: obtaining expert historical diagnosis and treatment records based on big data, wherein the expert historical diagnosis and treatment records comprise a plurality of historical diagnosis and treatment medical records with expert identifications; extracting the plurality of historical diagnosis and treatment medical cases with expert marks to obtain a target historical medical case, wherein the target historical medical case comprises multiple diagnosis and treatment; each diagnosis and treatment in the multiple diagnosis and treatment comprises a diagnosis and treatment result and a diagnosis and treatment prescription, and the diagnosis and treatment result and the diagnosis and treatment prescription have a corresponding relation; constructing a historical diagnosis and treatment database based on the inquiry result and the diagnosis and treatment prescription, and taking the historical diagnosis and treatment database as a retrieval library; obtaining a target clinical medical case and obtaining a target inquiry result of the target clinical medical case; traversing the target inquiry result in the search library to obtain a target traversing result; analyzing the target traversal result based on a preset prescription library to obtain a prescription reduction result, and taking the prescription reduction result as a diagnosis and treatment candidate prescription of the target clinical medical case.
In a second aspect, embodiments of the present application provide an artificial intelligence based data recovery system, the system comprising: the expert diagnosis and treatment record acquisition module is used for acquiring an expert historical diagnosis and treatment record based on big data, wherein the expert historical diagnosis and treatment record comprises a plurality of historical diagnosis and treatment medical records with expert identifications; the target historical medical records acquisition module is used for extracting the plurality of historical diagnosis and treatment medical records with expert marks to obtain target historical medical records, wherein the target historical medical records comprise multiple diagnosis and treatment; the diagnosis and treatment corresponding relation construction module is used for each diagnosis and treatment in the multiple diagnosis and treatment, wherein the diagnosis and treatment corresponding relation construction module comprises a diagnosis and treatment result and a diagnosis and treatment prescription, and the diagnosis and treatment result and the diagnosis and treatment prescription have a corresponding relation; the diagnosis and treatment database construction module is used for constructing a historical diagnosis and treatment database based on the inquiry result and the diagnosis and treatment prescription, and taking the historical diagnosis and treatment database as a retrieval library; the system comprises a target clinical medical case acquisition module, a target clinical medical case analysis module and a target diagnosis module, wherein the target clinical medical case acquisition module is used for acquiring a target clinical medical case and acquiring a target diagnosis result of the target clinical medical case; the target traversal result acquisition module is used for traversing the target inquiry result in the search library to obtain a target traversal result; the prescription restoration result acquisition module is used for analyzing the target traversal result based on a preset prescription library to obtain a prescription restoration result, and the prescription restoration result is used as a diagnosis and treatment candidate prescription of the target clinical medical case.
One or more technical solutions provided in the embodiments of the present application at least have the following technical effects or advantages:
the embodiment of the application provides a data reduction method based on artificial intelligence, which relates to the technical field of data reduction, and comprises the steps of obtaining expert historical diagnosis and treatment records based on big data, extracting and obtaining a target historical diagnosis and treatment record comprising a plurality of historical diagnosis and treatment medical records with expert marks, wherein each diagnosis and treatment in the plurality of diagnoses comprises a diagnosis and treatment result and a diagnosis and treatment prescription, constructing a historical diagnosis and treatment database based on the diagnosis and treatment result and the diagnosis and treatment prescription and taking the diagnosis and treatment result and the diagnosis and treatment result as a search library to obtain a target clinical medical record and a target diagnosis and treatment result, traversing the target diagnosis and treatment result in the search library to obtain a target traversal result, analyzing the target traversal result based on a preset prescription library to obtain a prescription reduction result, and taking the prescription reduction result as a diagnosis and treatment candidate of the target clinical medical record. The traditional Chinese medicine diagnosis and treatment method solves the technical problems that the traditional sitting and treatment method is slow in diagnosis and treatment process, the overall diagnosis and treatment efficiency is low, and the high-quality diagnosis and treatment experience is difficult to fully utilize, realizes the rapid traversal of a search library through a computer technology, obtains diagnosis and treatment candidate prescriptions, provides a reference for doctor diagnosis and treatment, and further improves the doctor diagnosis and treatment efficiency.
The foregoing description is only an overview of the technical solutions of the present application, and may be implemented according to the content of the specification in order to make the technical means of the present application more clearly understood, and in order to make the above-mentioned and other objects, features and advantages of the present application more clearly understood, the following detailed description of the present application will be given.
Drawings
Fig. 1 is a schematic flow chart of a data recovery method based on artificial intelligence according to an embodiment of the present application;
FIG. 2 is a schematic flow chart of a target traversal result obtained in an artificial intelligence-based data reduction method according to an embodiment of the present application;
FIG. 3 is a schematic flow chart of a prescription traversal result obtained in an artificial intelligence-based data recovery method according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a data recovery system based on artificial intelligence according to an embodiment of the present application.
Reference numerals illustrate: the system comprises an expert diagnosis and treatment record acquisition module 10, a target history medical table acquisition module 20, a diagnosis and treatment corresponding relation construction module 30, a diagnosis and treatment database construction module 40, a target clinical medical table acquisition module 50, a target traversal result acquisition module 60 and a prescription restoration result acquisition module 70.
Detailed Description
According to the data recovery method based on artificial intelligence, the traditional Chinese medicine diagnosis and treatment method based on artificial intelligence is used for solving the technical problems that the traditional Chinese medicine diagnosis and treatment method depends on the traditional sitting and diagnosis mode, the diagnosis and treatment process is slow, the overall diagnosis and treatment efficiency is low, and the high-quality diagnosis and treatment experience is difficult to fully utilize.
Example 1
As shown in fig. 1, an embodiment of the present application provides an artificial intelligence based data recovery method, which is applied to an artificial intelligence based data recovery system, and includes:
step S100: obtaining expert historical diagnosis and treatment records based on big data, wherein the expert historical diagnosis and treatment records comprise a plurality of historical diagnosis and treatment medical records with expert identifications;
specifically, the data restoration method based on the artificial intelligence is applied to a data restoration system based on the artificial intelligence. Firstly, the medical cases are continuous records of syndrome differentiation, legislation and prescription medication when doctors treat diseases, the traditional Chinese medical cases are specific reflection forms of comprehensive application of traditional Chinese medicine theory, law, prescription and medication, the medical cases are true records of medical activities, and reflect clinical experiences and thinking activities of doctors, and the quantity, form, style and style of medical cases are different due to different medical knowledge, hobbies and maintenance of the doctors. The method comprises the steps of acquiring expert historical diagnosis and treatment records through a computer and the Internet, wherein the expert historical diagnosis and treatment records are records containing information such as patient ages, sexes, inquiry results, diagnosis and treatment prescriptions, medical orders and the like, which are issued by a professional traditional Chinese medicine, and the expert historical diagnosis and treatment records comprise a plurality of historical diagnosis and treatment medical records with expert identifications. And a foundation is laid for the subsequent establishment of a retrieval library through the acquisition of expert historical diagnosis and treatment records.
Step S200: extracting the plurality of historical diagnosis and treatment medical cases with expert marks to obtain a target historical medical case, wherein the target historical medical case comprises multiple diagnosis and treatment;
specifically, in the obtained history diagnosis and treatment medical cases with expert marks, screening standards are set by a plurality of professional TCM students, and three aspects of the representativeness of clinical cases, inquiry results and diagnosis and treatment prescriptions are selected according to the screening standards. Judging whether the clinical cases are typical or not according to the historical diagnosis and treatment medical records, and generating a first judgment result as a first grading characteristic; judging whether the inquiry result is typical or not according to the historical diagnosis and treatment medical records, and generating a second judgment result as a second classification characteristic; judging whether the diagnosis and treatment prescription is typical or not according to the historical diagnosis and treatment medical records, and generating a third judgment result as a third grading characteristic; the method comprises the steps of constructing a multi-level medical case decision tree of a history diagnosis and treatment medical case with expert identification based on a first hierarchical feature, a second hierarchical feature and a third hierarchical feature, acquiring a first history medical case according to the history diagnosis and treatment medical case, inputting the first medical case into the multi-level medical case decision tree to obtain a judging result of the first medical case, and taking the history medical case which is satisfied by the first hierarchical feature, the second hierarchical feature and the third hierarchical feature as a target history medical case.
Further, the decision tree is a decision analysis method for evaluating the risk of the project and judging the feasibility of the project by solving the probability that the expected value of the net present value is greater than or equal to zero by forming the decision tree on the basis of the known occurrence probability of various conditions, and the classifier is a graphical method for intuitively applying probability analysis, and can provide correct classification for newly appeared objects and consists of root nodes, internal nodes and leaf nodes. The first grading feature, the second grading feature and the third grading feature can be used as internal nodes of the multi-level medical case decision tree, the features can be classified, the multi-level medical case decision tree is constructed recursively by the method until the last feature leaf node cannot be subdivided, and the description classification is finished, so that the multi-level medical case decision tree of the history medical case with expert identification is formed.
Through inputting the screening criteria set by doctors in a plurality of professions into the established decision tree, the classical target history medical records in all aspects can be rapidly and accurately screened out, and the technical effect of improving the screening efficiency of the target history medical records is achieved.
Step S300: each diagnosis and treatment in the multiple diagnosis and treatment comprises a diagnosis and treatment result and a diagnosis and treatment prescription, and the diagnosis and treatment result and the diagnosis and treatment prescription have a corresponding relation;
Specifically, the inquiry mainly inquires about the medical history, which is the main part of the medical history, and the occurrence, development and change of the disease and the diagnosis and treatment conditions from the onset to the diagnosis are recorded in detail according to the occurrence sequence of symptoms around the main complaint, and the contents mainly comprise: the time of onset, the urgency, the main symptoms, the characteristics of accompanying symptoms, diagnosis and treatment of diseases and other injuries which are done after the onset, and inquiry of cold and heat, sweat, diet, sleep, emotion, urination and defecation, and inquiry of certain conditions of women and children. If inquiring about cold or heat, i.e. inquiring about the sensation of cold or heat, the patient can provide a basis for determining the deficiency or excess of cold or heat from the exterior or interior of the disease. In each diagnosis and treatment, each inquiry result indicates a plurality of symptoms, and each symptom corresponds to one diagnosis and treatment prescription, so that the inquiry result has one-to-many correspondence with the diagnosis and treatment prescription. And a foundation is laid for the subsequent construction of a retrieval library through the acquisition of the corresponding relation between the inquiry result and the diagnosis and treatment prescription.
Step S400: constructing a historical diagnosis and treatment database based on the inquiry result and the diagnosis and treatment prescription, and taking the historical diagnosis and treatment database as a retrieval library;
specifically, according to the corresponding relation between the inquiry result and the corresponding symptoms and the diagnosis and treatment prescription in the target historical medical records, an n-ary tree is constructed. The n-ary tree is present for realizing convenient and quick searching, the inquiry result is taken as a root node of the tree, the corresponding symptoms are taken as child nodes, each child node corresponds to a diagnosis and treatment prescription, the root node is the node without a father node, and the root node serves as a starting point when traversing the tree, because all other nodes can be reached from the root node, and the constructed n-ary tree is taken as a search library. The search refers to a process of retrieving and finding out required information from a stored information base, and a search base is a database created for searching the required information, and is generally created according to a search relation. When searching, when the inquiry result is input, a plurality of diagnosis and treatment prescriptions can be obtained, and the proper diagnosis and treatment prescriptions are selected according to the severity and the specific conditions of the inquiry. By constructing the search library, the effects of providing treatment references for doctors and further improving diagnosis and treatment efficiency are achieved.
Step S500: obtaining a target clinical medical case and obtaining a target inquiry result of the target clinical medical case;
specifically, the target clinical medical case is the medical case of the target user needing diagnosis, and comprises information such as age, gender, inquiry results and the like of the target user, the target inquiry results are screened out, are inquiry results of main complaints, comprise the characteristics of main symptoms, onset time, illness state urgency, accompanying symptoms and the like of the target user, and the judgment according to the actual situation of the target user is realized through the acquisition of the target inquiry results, so that the diagnosis and treatment efficiency is improved.
Step S600: traversing the target inquiry result in the search library to obtain a target traversing result;
specifically, the Traversal (Traversal) refers to access to all node information in the n-ary tree, that is, each node in the tree is accessed once and only once in turn, and the tree structure has various different Traversal modes, and from the root node of the n-ary tree, the Traversal of the node is divided into three main steps: and operating the current node, traversing the left child node and traversing the right child node. Since there are a plurality of next nodes that can go to, starting from a given node, in the case of sequential computation, access to certain nodes can only be deferred-i.e. saved in some way for later re-access-it is worth noting that the order of traversal in this embodiment does not affect the traversal results, and therefore does not take into account the order, only the traversal results. And traversing to obtain a plurality of diagnosis and treatment groups similar to the target clinical cases in the history diagnosis and treatment, and taking the diagnosis and treatment groups as target traversing results.
Step S700: analyzing the target traversal result based on a preset prescription library to obtain a prescription reduction result, and taking the prescription reduction result as a diagnosis and treatment candidate prescription of the target clinical medical case.
Specifically, the preset prescription library is formed by collecting the existing prescriptions and the types and proportions of traditional Chinese medicines based on big data. Extracting a first prescription in the first prescription traversal results, extracting a second prescription in the second prescription traversal results, and sequentially obtaining a first prescription reduction result of the first prescription and a second prescription reduction result of the second prescription by using a preset prescription library, wherein the first prescription reduction result comprises a first preset prescription and a first addition and subtraction medicine, the second prescription reduction result comprises a second preset prescription and a second addition and subtraction medicine, and combining the first preset prescription, the first addition and subtraction medicine, the second preset prescription and the second addition and subtraction medicine to obtain a prescription reduction result.
Further, a target prescription reduction result in the prescription reduction results is extracted to obtain a judgment instruction, wherein the judgment instruction is used for judging whether the target prescription reduction result is repeated in the prescription reduction results, if yes, a first addition instruction is obtained, the target prescription reduction result is added to a first diagnosis and treatment candidate prescription according to the first addition instruction, the first diagnosis and treatment candidate prescription is ordered according to the number of times of prescription repetition to obtain a first diagnosis and treatment candidate prescription sequence, if not, a second addition instruction is obtained, the target prescription reduction result is added to a second diagnosis and treatment candidate prescription according to the second addition instruction, and the first diagnosis and treatment candidate prescription sequence and the second diagnosis and treatment candidate prescription are combined to obtain the diagnosis and treatment candidate prescription.
The diagnosis and treatment candidate prescription is obtained, so that reference is provided for diagnosis and treatment of doctors, and the effect of improving the treatment efficiency of the doctors is achieved.
Further, as shown in fig. 2, step S600 of the present application further includes:
step S610: extracting a first inquiry result in the search library;
step S620: performing feature analysis on the first inquiry result to form a first historical inquiry feature set;
step S630: performing feature analysis on the target inquiry result to form a target inquiry feature set;
step S640: analyzing the first historical inquiry feature set and the target inquiry feature set, and calculating to obtain a first inquiry similarity index according to an analysis result;
step S650: the first inquiry similarity index is used for representing the similarity degree between the first inquiry result and the target inquiry result;
step S660: obtaining a preset similarity threshold, and judging whether the first inquiry similarity index meets the preset similarity threshold or not;
step S670: if yes, an adding instruction is obtained, wherein the adding instruction is used for adding the first inquiry result to the target traversal result.
Specifically, the first inquiry result is any inquiry result in the search library, the total number of symptoms in the first inquiry result, such as fever and vomiting, the existence of certain features, the nonexistence of certain features, the chi-square value of each feature, which is used as a first historical inquiry feature set, is counted, and the target inquiry feature set is obtained in the same method. Setting a preset label scheme according to the total number of symptoms, if the symptoms are fever and vomiting, labeling the first historical inquiry feature set according to the preset label scheme to obtain a first label vector of the first inquiry result, labeling the target inquiry feature set according to the preset label scheme to obtain a target label vector of the target inquiry result, comparing the first label vector with the target label vector, and calculating by utilizing a similarity coefficient algorithm principle to obtain the first inquiry similarity index.
Setting a preset similarity threshold, wherein the preset similarity threshold is used for defining the highest value and the lowest value which can be achieved by the similarity index between the target inquiry result and the first inquiry result, namely the value range of the similarity index, and the similarity index exceeding the preset similarity threshold indicates that the similarity between the target inquiry result and the first inquiry result is too low and does not meet the requirements. If the similarity index meets a preset similarity threshold, the similarity between the target inquiry result and the first inquiry result is indicated to meet the requirement, and an adding instruction is obtained and used for adding the first inquiry result to the target traversal result. The screening of diagnosis and treatment similar to the target clinical medical records in the history diagnosis and treatment is realized, the effect of rapidly screening the inquiry result is achieved, and the screening efficiency is further improved.
Further, step S600 of the present application further includes:
step S680: if not, obtaining a skip instruction;
step S690: and carrying out skip processing on the first inquiry result according to the skip instruction.
Specifically, if the first query similarity index does not meet the preset similarity threshold, that is, the similarity between the first query result and the target query result is too low, a skip instruction is generated to skip the first query result. By skipping the inquiry results which do not meet the similar threshold, the effect of rapidly removing the inquiry results which do not meet the requirements is achieved, and the screening efficiency is further improved.
Further, step S640 of the present application further includes:
step S641: obtaining a preset label scheme;
step S642: labeling the first historical inquiry feature set according to the preset labeling scheme to obtain a first label vector of the first inquiry result;
step S643: labeling the target inquiry feature set according to the preset label scheme to obtain a target label vector of the target inquiry result;
step S644: comparing the first tag vector with the target tag vector, and calculating to obtain the first inquiry similarity index by using a similarity coefficient algorithm principle, wherein the calculation formula of the first inquiry similarity index is as follows:
wherein the P (I i I) refers to the first query similarity index, I refers to the target query result, I i Refers to the first inquiry result, the Q yy +Q nn Refers to the number of matching pairs of the first tag vector and the target tag vector that agree, the Q Total (S) Refers to the total number of matching pairs of the first tag vector and the target tag vector, and Q Total (S) ＝Q yy +Q yn +Q ny +Q nn 。
Specifically, a preset label scheme is established according to common symptoms of traditional Chinese medicine, such as wind cold, fever, cough and the like, for one symptom feature, the feature is represented By y, the feature is not represented By n, and the first historical inquiry feature set is labeled according to the feature, if the first historical inquiry feature set is wind cold A and fever B, the label vectors of the first historical inquiry feature set are Ay and Bn, and therefore the first label vectors Ay and By of the first inquiry result are obtained. And obtaining the target label vector of the target inquiry result by the same method. Superposing the first label vector and the target label vector to obtain Ayy and Bny, when the first label vector and the target label vector are identical to each other for the same disease feature, indicating that the matching is consistent, and when one of the first label vector and the target label vector is identical to each other for y, indicating that the matching is inconsistent, counting the quantity of the matching is consistent and the quantity of the matching is inconsistent, and obtaining a first inquiry similarity index through calculation, wherein the calculation formula of the first inquiry similarity index is as follows:
Wherein the P (I i I) refers to the first query similarity index, I refers to the target query result, I i Refers to the first inquiry result, the Q yy +Q nn Refers to the number of matching pairs of the first tag vector and the target tag vector that agree, the Q Total (S) Refers to the total number of matching pairs of the first tag vector and the target tag vector, and Q Total (S) ＝Q yy +Q yn +Q ny +Q nn 。
As can be derived from the formula, the first query similarity index is the ratio of the number of matching pairs of the first tag vector and the target tag vector to the total number of matching pairs of the first tag vector and the target tag vector, i.e., the ratio of the number of matching pairs to the total number.
Further, as shown in fig. 3, before step S700 of the present application, the method further includes:
step S710: extracting a plurality of inquiry results in the target traversal results, and reversely matching a plurality of diagnosis and treatment prescriptions of the inquiry results;
step S720: sequentially extracting a first diagnosis and treatment prescription and a second diagnosis and treatment prescription in the plurality of diagnosis and treatment prescriptions;
step S730: wherein the first diagnosis and treatment prescription comprises a first traditional Chinese medicine combination, and the second diagnosis and treatment prescription comprises a second traditional Chinese medicine combination;
step S740: traversing the first traditional Chinese medicine combination in the preset prescription library to obtain a first prescription traversing result;
Step S750: traversing the second traditional Chinese medicine combination in the preset prescription library to obtain a second prescription traversing result;
step S760: and combining the first prescription traversal result with the second prescription traversal result to obtain a prescription traversal result.
Specifically, the target traversal result is a set of multiple diagnosis and treatment compositions similar to the target clinical medical cases in the history diagnosis and treatment cases, in which the patient conditions of multiple medical cases are the same as the target patient conditions, and the diagnosis and treatment methods adopted by each patient are different historically, namely, multiple prescriptions exist. The first diagnosis and treatment prescription is any one of a plurality of diagnosis and treatment prescriptions, the second diagnosis and treatment prescription is any one of a plurality of diagnosis and treatment prescriptions which are different from the first diagnosis and treatment prescriptions, each diagnosis and treatment prescription contains a traditional Chinese medicine combination, namely, according to the actual symptoms of patients, corresponding traditional Chinese medicines are added for each symptom, and all the traditional Chinese medicines added for the symptoms form the diagnosis and treatment prescriptions for the diagnosis and treatment result. The preset prescription library is a prescription which is prepared by acquiring the prior prescription based on big data and the types and proportions of traditional Chinese medicines, and is prepared by acquiring the big data, wherein the prescription comprises 9 g of bupleurum, 9 g of scutellaria baicalensis, 6 g of rhizoma pinelliae preparata, 3 g of honey-fried licorice root, 3 pieces of ginger, 3 jujubes and 6 g of codonopsis pilosula, and is mainly used for treating the symptoms of shaoyang syndrome such as cold and heat going on, chest and hypochondrium fullness, no desire to eat, vexation, vomiting, bitter taste, dry throat, deafness, graying eyes, thin and white tongue coating and wiry and rapid pulse.
Comparing the composition of each traditional Chinese medicine in the first traditional Chinese medicine combination with each traditional Chinese medicine composition in a preset prescription library, namely accessing all traditional Chinese medicine compositions in the preset prescription library, and only accessing each traditional Chinese medicine composition according to the accessed traditional Chinese medicine compositions, sorting according to similarity obtained in the traversal process, taking the highest similarity as a first prescription traversal result, acquiring a second prescription traversal result by the same method, merging the first prescription traversal result with the second prescription traversal result, namely reserving the same traditional Chinese medicine combination in the prescription traversal result, and merging different traditional Chinese medicine compositions in the first prescription traversal result and the second prescription traversal result.
Further, step S700 of the present application further includes:
step S770: extracting a first prescription in the first prescription traversal result;
step S780: extracting a second prescription in the second prescription traversal result;
step S790: sequentially obtaining a first prescription reduction result of the first prescription and a second prescription reduction result of the second prescription by using the preset prescription library;
step S7100: the first prescription recovery result comprises a first preset prescription and a first addition and subtraction medicine, and the second prescription recovery result comprises a second preset prescription and a second addition and subtraction medicine;
Step S7200: and combining the first preset prescription and the first add-subtract medicine with the second preset prescription and the second add-subtract medicine to obtain the prescription restoration result.
Specifically, exemplarily, the obtained first prescription traversal result is a first prescription for treating the shaoyang syndrome, and the first prescription traversal result is obtained; the second prescription traversal result is also a second prescription for treating the shaoyang syndrome, and the second prescription traversal result is acquired. The prescription of the small bupleurum decoction for treating the shaoyang disease is obtained by presetting a prescription library, and comprises 9 g of bupleurum, 9 g of baical skullcap root, 6 g of prepared pinellia tuber, 3 g of honey-fried licorice root, 3 pieces of ginger, 3 jujubes and 6 g of pilose asiabell root, and the prescription is mainly used for treating the cold and heat of the shaoyang disease, chest and hypochondrium fullness, no desire for eating, vexation and vomiting, bitter taste and dry throat, deafness and eye, thin and white tongue fur and wiry and rapid pulse.
The first prescription traversing result can be obtained, and the symptom aimed by the first prescription traversing result has cough symptom on the basis of shaoyang syndrome, so that the first prescription in the first prescription traversing result is added with Chinese magnoliavine on the basis of xiaochaihu decoction and is used for astringing lung to relieve cough, the first preset prescription is the prescription in a preset prescription library matched with the first prescription traversing result, and the first add-subtract medicine is the traditional Chinese medicine which is added and reduced after being adjusted according to actual conditions as compared with the prescription in the preset prescription library; the first prescription traversing result can be obtained, and the symptoms aimed by the second prescription traversing result do not have vomiting symptoms on the basis of shaoyang symptoms, so that the second prescription in the second prescription traversing result also removes pinellia ternate and ginger on the basis of the small bupleurum decoction, the pinellia ternate and the ginger are used for stopping vomiting, and similarly, the second preset prescription is the prescription in a preset prescription library matched with the second prescription traversing result, and the second add-subtract medicine is the traditional Chinese medicine which is added and reduced after the second prescription traversing result is adjusted according to actual conditions compared with the prescription in the preset prescription library. And combining the first preset prescription and the first add-subtract medicine with the second preset prescription and the second add-subtract medicine to obtain the prescription restoration result.
Further, step S700 of the present application further includes:
step S7300: extracting a target prescription reduction result in the prescription reduction results;
step S7400: obtaining a judging instruction, wherein the judging instruction is used for judging whether the target prescription reduction result is repeated in the prescription reduction result;
step S7500: if yes, a first adding instruction is obtained, and the target prescription reduction result is added to a first diagnosis and treatment candidate prescription according to the first adding instruction;
step S7600: ordering the first diagnosis and treatment candidate prescriptions according to the prescription repetition times to obtain a first diagnosis and treatment candidate prescription sequence;
step S7700: if not, a second adding instruction is obtained, and the target prescription reduction result is added to a second diagnosis and treatment candidate prescription according to the second adding instruction;
step S7800: and merging the first diagnosis and treatment candidate prescription sequence and the second diagnosis and treatment candidate prescription to obtain the diagnosis and treatment candidate prescription.
Specifically, the judging instruction is configured to judge whether the target prescription recovery result is repeated in the prescription recovery result, and when the target prescription recovery result is repeated, it is stated that a plurality of patients use the same prescription historically, and the probability of being used by the target patient is high, that is, the prescription has general applicability, so that the target prescription recovery result is added to the first diagnosis and treat candidate prescription as the judging result, the first diagnosis and treat candidate prescription is ordered according to the repetition number, and the more the repetition number is, the more the prescription is used, the more the universality is, and the more the prescription is suitable for most people; when there is no repetition, only one patient in the description history uses the prescription, so the prescription is not commonly applicable, and may be used only for a specific situation, and the prescription is used as a replacement candidate, the target prescription reduction result is added to a second diagnosis candidate prescription as a judgment result, and the first diagnosis candidate prescription sequence and the second diagnosis candidate prescription are combined to obtain the diagnosis candidate prescription.
Example two
Based on the same inventive concept as the data restoring method based on artificial intelligence in the foregoing embodiments, as shown in fig. 4, the present application provides a data restoring system based on artificial intelligence, including:
the expert diagnosis and treatment record acquisition module 10 is used for acquiring an expert historical diagnosis and treatment record based on big data, wherein the expert historical diagnosis and treatment record comprises a plurality of historical diagnosis and treatment medical records with expert identifications;
the target historical medical case acquisition module 20 is used for extracting the plurality of historical diagnosis and treatment medical cases with expert marks to obtain a target historical medical case, wherein the target historical medical case comprises multiple diagnosis and treatment;
the diagnosis and treatment correspondence construction module 30, wherein the diagnosis and treatment correspondence construction module 30 is used for each diagnosis and treatment in the multiple diagnosis and treatment, each diagnosis and treatment comprises a diagnosis and treatment result and a diagnosis and treatment prescription, and the diagnosis and treatment result and the diagnosis and treatment prescription have a correspondence;
a diagnosis and treat database construction module 40, wherein the diagnosis and treat database construction module 40 is used for constructing a historical diagnosis and treat database based on the inquiry result and the diagnosis and treat prescription, and taking the historical diagnosis and treat database as a retrieval library;
The target clinical medical case acquisition module 50 is used for acquiring a target clinical medical case and acquiring a target inquiry result of the target clinical medical case;
the target traversal result acquisition module 60, where the target traversal result acquisition module 60 is configured to traverse the target inquiry result in the search library to obtain a target traversal result;
the prescription restoration result obtaining module 70 is configured to analyze the target traversal result based on a preset prescription library, obtain a prescription restoration result, and use the prescription restoration result as a diagnosis candidate prescription of the target clinical medical case.
Further, the system further comprises:
the first inquiry result extraction module is used for extracting a first inquiry result in the search library;
the feature analysis module is used for carrying out feature analysis on the first inquiry result to form a first historical inquiry feature set;
the target inquiry feature set acquisition module is used for carrying out feature analysis on the target inquiry result to form a target inquiry feature set;
the first inquiry similarity index acquisition module is used for analyzing the first historical inquiry feature set and the target inquiry feature set and calculating to obtain a first inquiry similarity index according to an analysis result;
The first inquiry similarity index is used for representing the similarity degree between the first inquiry result and the target inquiry result;
the preset similarity threshold obtaining module is used for obtaining a preset similarity threshold and judging whether the first inquiry similarity index meets the preset similarity threshold or not;
and the adding instruction acquisition module is used for acquiring an adding instruction if yes, wherein the adding instruction is used for adding the first inquiry result to the target traversal result.
Further, the system further comprises:
the skip instruction acquisition module is used for acquiring a skip instruction if not;
and the skipping processing module is used for skipping the first inquiry result according to the skipping instruction.
Further, the system further comprises:
the preset label scheme acquisition module is used for acquiring a preset label scheme;
the label marking module is used for marking the first historical inquiry feature set according to the preset label scheme to obtain a first label vector of the first inquiry result;
the target label vector acquisition module is used for labeling the target inquiry feature set according to the preset label scheme to obtain a target label vector of the target inquiry result;
The first inquiry similarity index obtaining module is used for comparing the first tag vector with the target tag vector and calculating to obtain the first inquiry similarity index by utilizing a similarity coefficient algorithm principle, wherein the calculation formula of the first inquiry similarity index is as follows:
wherein the P (I i I) refers to the first query similarity index, I refers to the target query result, I i Refers to the first inquiry result, the Q yy +Q nn Refers to the number of matching pairs of the first tag vector and the target tag vector that agree, the Q Total (S) Refers to the total number of matching pairs of the first tag vector and the target tag vector, and Q Total (S) ＝Q yy +Q yn +Q ny +Q nn 。
Further, the system further comprises:
the multiple inquiry result extraction modules are used for extracting multiple inquiry results in the target traversal results and reversely matching multiple diagnosis and treatment prescriptions of the multiple inquiry results;
the diagnosis and treatment prescription extraction module is used for sequentially extracting a first diagnosis and treatment prescription and a second diagnosis and treatment prescription in the plurality of diagnosis and treatment prescriptions;
wherein the first diagnosis and treatment prescription comprises a first traditional Chinese medicine combination, and the second diagnosis and treatment prescription comprises a second traditional Chinese medicine combination;
the first prescription traversal result acquisition module is used for traversing the first traditional Chinese medicine combination in the preset prescription library to obtain a first prescription traversal result;
The second prescription traversal result acquisition module is used for traversing the second traditional Chinese medicine combination in the preset prescription library to obtain a second prescription traversal result;
and the traversal result merging module is used for merging the first prescription traversal result with the second prescription traversal result to obtain a prescription traversal result.
Further, the system further comprises:
the first prescription extraction module is used for extracting a first prescription in the first prescription traversal result;
a second prescription extraction module for extracting a second prescription in the second prescription traversal result;
the prescription reduction result acquisition module is used for sequentially acquiring a first prescription reduction result of the first prescription and a second prescription reduction result of the second prescription by utilizing the preset prescription library;
the first prescription recovery result comprises a first preset prescription and a first addition and subtraction medicine, and the second prescription recovery result comprises a second preset prescription and a second addition and subtraction medicine;
and the merging module is used for merging the first preset prescription and the first addition and subtraction medicine with the second preset prescription and the second addition and subtraction medicine to obtain the prescription reduction result.
Further, the system further comprises:
Extracting a target prescription reduction result, and extracting a target prescription reduction result in the prescription reduction result;
the judging instruction acquisition module is used for acquiring a judging instruction, wherein the judging instruction is used for judging whether the target prescription reduction result is repeated in the prescription reduction result or not;
the first adding instruction acquisition module is used for acquiring a first adding instruction if yes, and adding the target prescription reduction result to a first diagnosis and treatment candidate prescription according to the first adding instruction;
the ordering module is used for ordering the first diagnosis and treatment candidate prescriptions according to the prescription repetition times to obtain a first diagnosis and treatment candidate prescription sequence;
the second adding instruction acquisition module is used for acquiring a second adding instruction if not, and adding the target prescription reduction result to a second diagnosis and treatment candidate prescription according to the second adding instruction;
the diagnosis and treatment candidate prescription acquisition module is used for combining the first diagnosis and treatment candidate prescription sequence and the second diagnosis and treatment candidate prescription to obtain the diagnosis and treatment candidate prescription.
In the foregoing description of an artificial intelligence based data recovery method and system, those skilled in the art can clearly understand that, for the device disclosed in this embodiment, the description is relatively simple, and relevant places refer to the description of the method section, because the device corresponds to the method disclosed in the embodiment.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (8)
1. An artificial intelligence based data recovery method, comprising:
obtaining expert historical diagnosis and treatment records based on big data, wherein the expert historical diagnosis and treatment records comprise a plurality of historical diagnosis and treatment medical records with expert identifications;
extracting the plurality of historical diagnosis and treatment medical cases with expert marks to obtain a target historical medical case, wherein the target historical medical case comprises multiple diagnosis and treatment;
each diagnosis and treatment in the multiple diagnosis and treatment comprises a diagnosis and treatment result and a diagnosis and treatment prescription, and the diagnosis and treatment result and the diagnosis and treatment prescription have a corresponding relation;
constructing a historical diagnosis and treatment database based on the inquiry result and the diagnosis and treatment prescription, and taking the historical diagnosis and treatment database as a retrieval library;
Obtaining a target clinical medical case and obtaining a target inquiry result of the target clinical medical case;
traversing the target inquiry result in the search library to obtain a target traversing result;
analyzing the target traversal result based on a preset prescription library to obtain a prescription reduction result, and taking the prescription reduction result as a diagnosis and treatment candidate prescription of the target clinical medical case.
2. The data reduction method of claim 1, wherein traversing the target query results in the search pool to obtain target traversal results comprises:
extracting a first inquiry result in the search library;
performing feature analysis on the first inquiry result to form a first historical inquiry feature set;
performing feature analysis on the target inquiry result to form a target inquiry feature set;
analyzing the first historical inquiry feature set and the target inquiry feature set, and calculating to obtain a first inquiry similarity index according to an analysis result;
the first inquiry similarity index is used for representing the similarity degree between the first inquiry result and the target inquiry result;
obtaining a preset similarity threshold, and judging whether the first inquiry similarity index meets the preset similarity threshold or not;
If yes, an adding instruction is obtained, wherein the adding instruction is used for adding the first inquiry result to the target traversal result.
3. The data reduction method according to claim 2, wherein after the obtaining a preset similarity threshold and determining whether the first query similarity index satisfies the preset similarity threshold, further comprising:
if not, obtaining a skip instruction;
and carrying out skip processing on the first inquiry result according to the skip instruction.
4. The data reduction method according to claim 2, wherein analyzing the first historical query feature set and the target query feature set, and calculating a first query similarity index according to the analysis result, includes:
obtaining a preset label scheme;
labeling the first historical inquiry feature set according to the preset labeling scheme to obtain a first label vector of the first inquiry result;
labeling the target inquiry feature set according to the preset label scheme to obtain a target label vector of the target inquiry result;
comparing the first tag vector with the target tag vector, and calculating to obtain the first inquiry similarity index by using a similarity coefficient algorithm principle, wherein the calculation formula of the first inquiry similarity index is as follows:
Wherein the P (I i I) refers to the first query similarity index, I refers to the target query result, I i Refers to the first inquiry result, the Q yy +Q nn Refers to the number of matching pairs of the first tag vector and the target tag vector that agree, the Q Total (S) Refers to the total number of matching pairs of the first tag vector and the target tag vector, and Q Total (S) ＝Q yy +Q yn +Q ny +Q nn 。
5. The data reduction method according to claim 1, further comprising, before the analyzing the target traversal result based on the preset recipe library to obtain a recipe reduction result:
extracting a plurality of inquiry results in the target traversal results, and reversely matching a plurality of diagnosis and treatment prescriptions of the inquiry results;
sequentially extracting a first diagnosis and treatment prescription and a second diagnosis and treatment prescription in the plurality of diagnosis and treatment prescriptions;
wherein the first diagnosis and treatment prescription comprises a first traditional Chinese medicine combination, and the second diagnosis and treatment prescription comprises a second traditional Chinese medicine combination;
traversing the first traditional Chinese medicine combination in the preset prescription library to obtain a first prescription traversing result;
traversing the second traditional Chinese medicine combination in the preset prescription library to obtain a second prescription traversing result;
And combining the first prescription traversal result with the second prescription traversal result to obtain a prescription traversal result.
6. The data reduction method of claim 5, further comprising:
extracting a first prescription in the first prescription traversal result;
extracting a second prescription in the second prescription traversal result;
sequentially obtaining a first prescription reduction result of the first prescription and a second prescription reduction result of the second prescription by using the preset prescription library;
the first prescription recovery result comprises a first preset prescription and a first addition and subtraction medicine, and the second prescription recovery result comprises a second preset prescription and a second addition and subtraction medicine;
and combining the first preset prescription and the first add-subtract medicine with the second preset prescription and the second add-subtract medicine to obtain the prescription restoration result.
7. The data reduction method of claim 6, further comprising:
extracting a target prescription reduction result in the prescription reduction results;
obtaining a judging instruction, wherein the judging instruction is used for judging whether the target prescription reduction result is repeated in the prescription reduction result;
If yes, a first adding instruction is obtained, and the target prescription reduction result is added to a first diagnosis and treatment candidate prescription according to the first adding instruction;
ordering the first diagnosis and treatment candidate prescriptions according to the prescription repetition times to obtain a first diagnosis and treatment candidate prescription sequence;
if not, a second adding instruction is obtained, and the target prescription reduction result is added to a second diagnosis and treatment candidate prescription according to the second adding instruction;
and merging the first diagnosis and treatment candidate prescription sequence and the second diagnosis and treatment candidate prescription to obtain the diagnosis and treatment candidate prescription.
8. An artificial intelligence based data retrieval system comprising:
the expert diagnosis and treatment record acquisition module is used for acquiring an expert historical diagnosis and treatment record based on big data, wherein the expert historical diagnosis and treatment record comprises a plurality of historical diagnosis and treatment medical records with expert identifications;
the target historical medical records acquisition module is used for extracting the plurality of historical diagnosis and treatment medical records with expert marks to obtain target historical medical records, wherein the target historical medical records comprise multiple diagnosis and treatment;
The diagnosis and treatment corresponding relation construction module is used for each diagnosis and treatment in the multiple diagnosis and treatment, wherein the diagnosis and treatment corresponding relation construction module comprises a diagnosis and treatment result and a diagnosis and treatment prescription, and the diagnosis and treatment result and the diagnosis and treatment prescription have a corresponding relation;
the diagnosis and treatment database construction module is used for constructing a historical diagnosis and treatment database based on the inquiry result and the diagnosis and treatment prescription, and taking the historical diagnosis and treatment database as a retrieval library;
the system comprises a target clinical medical case acquisition module, a target clinical medical case analysis module and a target diagnosis module, wherein the target clinical medical case acquisition module is used for acquiring a target clinical medical case and acquiring a target diagnosis result of the target clinical medical case;
the target traversal result acquisition module is used for traversing the target inquiry result in the search library to obtain a target traversal result;
the prescription restoration result acquisition module is used for analyzing the target traversal result based on a preset prescription library to obtain a prescription restoration result, and the prescription restoration result is used as a diagnosis and treatment candidate prescription of the target clinical medical case.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211647424.8A CN115995301A (en) | 2022-12-21 | 2022-12-21 | Data recovery method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211647424.8A CN115995301A (en) | 2022-12-21 | 2022-12-21 | Data recovery method and system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115995301A true CN115995301A (en) | 2023-04-21 |
Family
ID=85989834
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211647424.8A Pending CN115995301A (en) | 2022-12-21 | 2022-12-21 | Data recovery method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115995301A (en) |
-
2022
- 2022-12-21 CN CN202211647424.8A patent/CN115995301A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhuang et al. | An Effective WSSENet-Based Similarity Retrieval Method of Large Lung CT Image Databases. | |
| Zhang et al. | Stroke lesion detection and analysis in MRI images based on deep learning | |
| Hu et al. | Automatic construction of Chinese herbal prescriptions from tongue images using CNNs and auxiliary latent therapy topics | |
| Grossi et al. | Detection of an autism EEG signature from only two EEG channels through features extraction and advanced machine learning analysis | |
| Yang et al. | Histopathology-based diagnosis of oral squamous cell carcinoma using deep learning | |
| CN116910172B (en) | Follow-up table generation method and system based on artificial intelligence | |
| CN109074858A (en) | There is no hospital's matching in the health care data library for going identification of obvious standard identifier | |
| Huan et al. | Multilevel and multiscale feature aggregation in deep networks for facial constitution classification | |
| Agarwal et al. | Survival prediction based on histopathology imaging and clinical data: A novel, whole slide cnn approach | |
| Pendyala et al. | Automated medical diagnosis from clinical data | |
| CN112381598A (en) | Product service information pushing method and device | |
| CN115438040A (en) | Pathological archive information management method and system | |
| Gansky | Dental data mining: potential pitfalls and practical issues | |
| Shi et al. | An automatic classification method on chronic venous insufficiency images | |
| Huang et al. | Assistant diagnosis of basal cell carcinoma and seborrheic keratosis in Chinese population using convolutional neural network | |
| AlZu’Bi et al. | transfer learning enabled CAD system for monkey pox classification | |
| Boulahia et al. | Att2ResNet: A deep attention‐based approach for melanoma skin cancer classification | |
| Zhang et al. | A time-sensitive hybrid learning model for patient subgrouping | |
| CN115083550B (en) | Patient similarity classification method based on multi-source information | |
| Tobias et al. | Android Application for Chest X-ray Health Classification From a CNN Deep Learning TensorFlow Model | |
| CN115995301A (en) | Data recovery method and system based on artificial intelligence | |
| Wang et al. | Imbalanced learning for hospital readmission prediction using national readmission database | |
| Tang et al. | RTC_TongueNet: An improved tongue image segmentation model based on DeepLabV3 | |
| CN110010231A (en) | A kind of data processing system and computer readable storage medium | |
| Rout et al. | Identification of similar gastrointestinal images through content based image retrieval system based on analytical hierarchical process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |