CN117558410A - Artificial intelligence-based diet management system and method for diabetic nephropathy patients - Google Patents
Artificial intelligence-based diet management system and method for diabetic nephropathy patients Download PDFInfo
- Publication number
- CN117558410A CN117558410A CN202410045683.6A CN202410045683A CN117558410A CN 117558410 A CN117558410 A CN 117558410A CN 202410045683 A CN202410045683 A CN 202410045683A CN 117558410 A CN117558410 A CN 117558410A
- Authority
- CN
- China
- Prior art keywords
- patient
- semantic
- text
- feature vector
- recipe
- 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.)
- Granted
Links
- 235000005911 diet Nutrition 0.000 title claims abstract description 57
- 208000007342 Diabetic Nephropathies Diseases 0.000 title claims abstract description 53
- 208000033679 diabetic kidney disease Diseases 0.000 title claims abstract description 53
- 230000037213 diet Effects 0.000 title claims abstract description 52
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 title claims abstract description 12
- 239000013598 vector Substances 0.000 claims abstract description 296
- 230000002146 bilateral effect Effects 0.000 claims abstract description 108
- 230000000295 complement effect Effects 0.000 claims abstract description 96
- 230000003993 interaction Effects 0.000 claims abstract description 37
- 238000007689 inspection Methods 0.000 claims abstract description 13
- 238000012549 training Methods 0.000 claims description 133
- 238000007726 management method Methods 0.000 claims description 52
- 238000004364 calculation method Methods 0.000 claims description 13
- 230000004927 fusion Effects 0.000 claims description 12
- 238000004458 analytical method Methods 0.000 claims description 10
- 230000000378 dietary effect Effects 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 5
- 230000036541 health Effects 0.000 abstract description 8
- 230000006870 function Effects 0.000 description 23
- 239000008280 blood Substances 0.000 description 6
- 210000004369 blood Anatomy 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- WQZGKKKJIJFFOK-GASJEMHNSA-N Glucose Natural products OC[C@H]1OC(O)[C@H](O)[C@@H](O)[C@@H]1O WQZGKKKJIJFFOK-GASJEMHNSA-N 0.000 description 4
- 230000019771 cognition Effects 0.000 description 4
- 206010012601 diabetes mellitus Diseases 0.000 description 4
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 239000008103 glucose Substances 0.000 description 4
- 235000012054 meals Nutrition 0.000 description 4
- 235000006694 eating habits Nutrition 0.000 description 3
- 230000002452 interceptive effect Effects 0.000 description 3
- 238000005457 optimization Methods 0.000 description 3
- 230000036772 blood pressure Effects 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000001419 dependent effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 235000013305 food Nutrition 0.000 description 2
- 210000003734 kidney Anatomy 0.000 description 2
- 208000017169 kidney disease Diseases 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 206010020772 Hypertension Diseases 0.000 description 1
- 206010030113 Oedema Diseases 0.000 description 1
- 206010062237 Renal impairment Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 210000004204 blood vessel Anatomy 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010411 cooking Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 208000016097 disease of metabolism Diseases 0.000 description 1
- 239000004615 ingredient Substances 0.000 description 1
- 238000012482 interaction analysis Methods 0.000 description 1
- 230000003907 kidney function Effects 0.000 description 1
- 150000002632 lipids Chemical class 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 208000030159 metabolic disease Diseases 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000016709 nutrition Nutrition 0.000 description 1
- 231100000857 poor renal function Toxicity 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 235000021075 protein intake Nutrition 0.000 description 1
- 201000001474 proteinuria Diseases 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 235000021023 sodium intake Nutrition 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H20/00—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance
- G16H20/60—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to nutrition control, e.g. diets
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H10/00—ICT specially adapted for the handling or processing of patient-related medical or healthcare data
- G16H10/60—ICT specially adapted for the handling or processing of patient-related medical or healthcare data for patient-specific data, e.g. for electronic patient records
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- 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
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Physics & Mathematics (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Data Mining & Analysis (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Pathology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Nutrition Science (AREA)
- Probability & Statistics with Applications (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
The invention discloses an artificial intelligence-based diet management system and method for diabetic nephropathy patients, and relates to the technical field of intelligent management, wherein the system and method acquire text description of personal information of patient objects and text description of inspection reports, and perform semantic coding and bilateral complementary feature interaction on the two to obtain patient information-patient inspection text bilateral complementary semantic feature vectors; acquiring text description of an alternative recipe, and carrying out semantic coding and dimension adjustment on the text description to obtain a semantic coding feature vector of the text of the alternative recipe after adjustment; and determining whether the matching degree of the candidate recipe and the patient object meets the preset requirement based on the semantic cross matching characteristics between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector. In this way, the patient can be provided with a recipe that suits his needs and health.
Description
Technical Field
The invention relates to the technical field of intelligent management, in particular to an artificial intelligence-based diet management system and method for diabetic nephropathy patients.
Background
Diabetes is a metabolic disease caused by long-term uncontrolled high blood glucose levels, with diabetic nephropathy being one of the complications of diabetes. In diabetic nephropathy, high blood glucose levels damage the small blood vessels and glomeruli of the kidneys, gradually leading to impaired renal function. Initially, diabetic nephropathy may not cause significant symptoms, but patients may develop symptoms such as proteinuria, hypertension, edema, etc. as the condition progresses.
Treatment of diabetic nephropathy includes controlling blood glucose levels, controlling blood pressure, limiting protein intake, limiting sodium intake, and the like. Diet management is one of the important links in the treatment of diabetic nephropathy, and can control blood sugar and blood pressure through reasonable diet selection and food collocation, so as to reduce kidney burden. Conventional methods of dietary management for diabetic nephropathy patients typically adjust dietary habits based on physician recommendations, but in most cases do not provide specific dietary management regimens for the nephropathy patients, making patient dietary management too dependent on availability of medical resources and time and effort investments for the patient, and not fully covering the patient's daily dietary needs.
Thus, an artificial intelligence-based diet management system and method for diabetic nephropathy patients are desired.
Disclosure of Invention
The invention aims to overcome the defects and provide an artificial intelligence-based diet management system and method for diabetic nephropathy patients.
The embodiment of the invention also provides an artificial intelligence-based diet management system for diabetic nephropathy patients, which comprises the following components:
a patient information acquisition module for acquiring a text description of personal information of a patient object and a text description of an examination report;
the patient information semantic analysis module is used for carrying out semantic coding on the text description of the personal information of the patient object and the text description of the examination report and carrying out feature interaction based on the bilateral information complementary attention so as to obtain a patient information-patient examination text bilateral complementary semantic feature vector;
the alternative recipe acquisition module is used for acquiring text description of the alternative recipes;
the alternative recipe semantic understanding module is used for carrying out semantic coding and dimension adjustment on the text description of the alternative recipe so as to obtain an adjusted alternative recipe text semantic coding feature vector;
the semantic matching measurement module is used for determining whether the matching degree of the candidate recipe and the patient object meets the preset requirement or not based on semantic cross matching characteristics between the patient information-patient check text bilateral complementary semantic feature vector and the adjusted candidate recipe text semantic coding feature vector;
Wherein, the semantic matching measurement module comprises:
the semantic cross matching coefficient calculation unit is used for calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector so as to obtain a patient-recipe semantic matching feature vector;
and the matching degree judging unit is used for enabling the patient-recipe semantic matching feature vector to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the matching degree of the alternative recipe and the patient object meets the preset requirement.
In the artificial intelligence-based diet management system for diabetic nephropathy patients, the semantic cross matching coefficient calculating unit is used for:
calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and each position feature value of the adjustment candidate recipe text semantic coding feature vector in the following semantic cross metric formula to obtain the patient-recipe semantic matching feature vector; wherein, the semantic cross metric formula is:
;
wherein,checking text for the patient information-patient for the +.th in the text bilateral complementary semantic feature vector >Personal characteristic value->Encoding semantically the feature vector for the adjusted alternative recipe text +.>Personal characteristic value->Checking the dimensions of text bilateral complementary semantic feature vectors for the patient information-patient, ++>Is->Semantic cross-matching coefficients, i.e. the +.th in the patient-recipe semantic matching feature vector>Personal characteristic value->A logarithmic function operation with a base of 2 is represented.
In the artificial intelligence-based diet management system for diabetic nephropathy patients, the patient information semantic analysis module comprises:
the patient personal information semantic understanding unit is used for carrying out semantic encoding on the text description of the personal information of the patient object through a patient personal information semantic encoder based on a transducer so as to obtain a patient information text semantic encoding feature vector;
a patient examination report semantic understanding unit, configured to perform semantic encoding on a text description of an examination report of the patient object by using a patient examination report semantic encoder based on a transducer to obtain a patient examination text semantic encoding feature vector;
and the characteristic interaction unit is used for carrying out characteristic interaction on the patient information text semantic coding characteristic vector and the patient examination text semantic coding characteristic vector so as to obtain the patient information-patient examination text bilateral complementary semantic characteristic vector.
In the above artificial intelligence based diet management system for diabetic nephropathy patients, the feature interaction unit comprises:
and the bilateral information complementation subunit is used for obtaining the patient information-patient check text bilateral complementation semantic feature vector through the patient information text semantic coding feature vector and the patient check text semantic coding feature vector by a feature interaction module based on bilateral information complementation attention.
In the artificial intelligence-based diet management system for diabetic nephropathy patients, the bilateral information complementation subunit is used for:
concatenating the patient information text semantic coding feature vector and the patient exam text semantic coding feature vector to obtain a first concatenated vector;
passing the first series of vectors through a Softmax layer to obtain a first attention weight vector;
vector multiplication is carried out on the patient information text semantic coding feature vector and the first attention weight vector to obtain a first fusion vector;
passing the first fusion vector through a full connection layer to obtain a patient information text semantic coding feature vector containing patient examination text semantic information;
passing the patient information text semantically encoded feature vector through a Softmax layer to obtain a second attention weight vector;
Vector multiplication is carried out on the second attention weight vector and the patient examination text semantic coding feature vector to obtain a second fusion vector;
passing the second fusion vector through a full connection layer to obtain a patient check text semantic coding feature vector containing personal information of a patient;
and merging the patient information text semantic coding feature vector containing patient examination text semantic information and the patient examination text semantic coding feature vector containing patient personal information to obtain the patient information-patient examination text bilateral complementary semantic feature vector.
In the artificial intelligence-based diet management system for diabetic nephropathy patients, the alternative recipe semantic understanding module is used for:
the alternative recipe semantic coding unit is used for carrying out semantic coding on the text description of the alternative recipe so as to obtain an alternative recipe text semantic coding feature vector;
and the dimension adjustment unit is used for carrying out dimension adjustment on the semantic coding feature vector of the alternative recipe text so as to obtain the semantic coding feature vector of the alternative recipe text after adjustment.
In the artificial intelligence-based diabetic nephropathy patient diet management system described above, the dimension adjusting unit is configured to:
And the candidate recipe text semantic coding feature vector is subjected to a dimension adjustment module based on a full connection layer to obtain the adjusted candidate recipe text semantic coding feature vector.
The artificial intelligence-based diabetic nephropathy patient diet management system further comprises a training module for training the patient personal information semantic encoder based on the Transformer, the patient examination report semantic encoder based on the Transformer, the feature interaction module based on the bilateral information complementary attention, the dimension adjustment module based on the full-connection layer and the classifier; wherein, training module includes:
the training information acquisition unit is used for acquiring training data, wherein the training data comprises text description of training personal information of a patient object, text description of training inspection report, text description of training alternative recipes and whether the matching degree of the training alternative recipes and the patient object reaches a true value of a preset requirement;
the training patient personal information semantic coding unit is used for carrying out semantic coding on the text description of the training personal information of the patient object through a patient personal information semantic coder based on a transducer so as to obtain training patient information text semantic coding feature vectors;
A training patient examination report semantic coding unit, configured to perform semantic coding on a text description of a training examination report of the patient object by using a patient examination report semantic coder based on a transducer to obtain a training patient examination text semantic coding feature vector;
the training bilateral information complementary feature interaction unit is used for enabling the training patient information text semantic coding feature vector and the training patient checking text semantic coding feature vector to pass through a feature interaction module based on bilateral information complementary attention so as to obtain training patient information-patient checking text bilateral complementary semantic feature vector;
the training alternative recipe semantic coding unit is used for carrying out semantic coding on the text description of the training alternative recipe so as to obtain a training alternative recipe text semantic coding feature vector;
the training dimension adjustment unit is used for enabling the training alternative recipe text semantic coding feature vector to pass through a dimension adjustment module based on a full connection layer to obtain a training adjustment alternative recipe text semantic coding feature vector;
the training semantic cross matching coefficient calculation unit is used for calculating semantic cross matching coefficients between the training patient information-patient check text bilateral complementary semantic feature vectors and the training adjustment candidate recipe text semantic coding feature vectors so as to obtain training patient-recipe semantic matching feature vectors;
A classification loss function value calculation unit, configured to pass the training patient-recipe semantic matching feature vector through a classifier to obtain a classification loss function value;
a specific loss function value calculation unit, configured to calculate a specific loss function value of the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment candidate recipe text semantic coding feature vector;
the training unit is used for training the patient personal information semantic encoder based on the Transformer, the patient examination report semantic encoder based on the Transformer, the feature interaction module based on the bilateral information complementary attention, the dimension adjustment module based on the full-connection layer and the classifier by taking the weighted sum of the classified loss function value and the specific loss function value as the loss function value.
The embodiment of the invention also provides an artificial intelligence-based diet management method for diabetic nephropathy patients, which comprises the following steps:
acquiring a text description of personal information of a patient object and a text description of an examination report;
semantic coding and bilateral complementary feature interaction are carried out on the text description of the personal information of the patient object and the text description of the examination report so as to obtain a patient information-patient examination text bilateral complementary semantic feature vector;
Acquiring a text description of an alternative recipe;
performing semantic coding and dimension adjustment on the text description of the alternative recipe to obtain an adjusted alternative recipe text semantic coding feature vector;
determining whether the matching degree of the candidate recipe and the patient object meets a preset requirement based on semantic cross matching features between the patient information-patient check text bilateral complementary semantic feature vector and the adjusted candidate recipe text semantic coding feature vector;
wherein determining whether the matching degree of the candidate recipe and the patient object meets a predetermined requirement based on semantic cross matching features between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector comprises:
calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vectors and the adjustment candidate recipe text semantic coding feature vectors to obtain patient-recipe semantic matching feature vectors;
and passing the patient-recipe semantic matching feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the matching degree of the alternative recipe and the patient object meets the preset requirement.
In the above artificial intelligence based diet management method for diabetic nephropathy patients, calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector to obtain a patient-recipe semantic matching feature vector, comprising:
calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and each position feature value of the adjustment candidate recipe text semantic coding feature vector in the following semantic cross metric formula to obtain the patient-recipe semantic matching feature vector; wherein, the semantic cross metric formula is:
;
wherein,checking text for the patient information-patient for the +.th in the text bilateral complementary semantic feature vector>Personal characteristic value->Encoding semantically the feature vector for the adjusted alternative recipe text +.>Personal characteristic value->Checking the dimensions of text bilateral complementary semantic feature vectors for the patient information-patient, ++>Is->Semantic cross-matching coefficients, i.e. the +.th in the patient-recipe semantic matching feature vector>Personal characteristic value->A logarithmic function operation with a base of 2 is represented.
The method comprises the steps of obtaining text description of personal information of a patient object and text description of an inspection report, and carrying out semantic coding and bilateral complementary feature interaction on the text description and the text description of the inspection report to obtain a patient information-patient inspection text bilateral complementary semantic feature vector; acquiring text description of an alternative recipe, and carrying out semantic coding and dimension adjustment on the text description to obtain a semantic coding feature vector of the text of the alternative recipe after adjustment; and determining whether the matching degree of the candidate recipe and the patient object meets the preset requirement based on the semantic cross matching characteristics between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector. In this way, the patient can be provided with a recipe that suits his needs and health.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art. In the drawings:
Fig. 1 is a block diagram of an artificial intelligence-based diet management system for diabetic nephropathy patients in an embodiment of the present invention.
Fig. 2 is a flowchart of a method for diet management of diabetic nephropathy patients based on artificial intelligence according to an embodiment of the present invention.
Fig. 3 is a schematic diagram of a system architecture of an artificial intelligence-based diet management method for diabetic nephropathy patients according to an embodiment of the present invention.
Fig. 4 is an application scenario diagram of an artificial intelligence-based diet management system for diabetic nephropathy patients provided in an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the embodiments of the present invention will be described in further detail with reference to the accompanying drawings. The exemplary embodiments of the present invention and their descriptions herein are for the purpose of explaining the present invention, but are not to be construed as limiting the invention.
Unless defined otherwise, all technical and scientific terms used in the examples of this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to limit the scope of the present application.
In the description of the embodiments of the present application, unless otherwise indicated and defined, the term "connected" should be construed broadly, and for example, may be an electrical connection, may be a communication between two elements, may be a direct connection, or may be an indirect connection via an intermediary, and it will be understood by those skilled in the art that the specific meaning of the term may be understood according to the specific circumstances.
It should be noted that, the term "first\second\third" in the embodiments of the present application merely distinguishes similar objects, and does not represent a specific order for the objects, and it is understood that "first\second\third" may interchange a specific order or sequence where allowed. It is to be understood that the "first\second\third" distinguishing objects may be interchanged where appropriate such that the embodiments of the present application described herein may be implemented in sequences other than those illustrated or described herein.
In one embodiment of the present invention, FIG. 1 is a block diagram of an artificial intelligence based diet management system for diabetic nephropathy patients provided in an embodiment of the present invention. As shown in fig. 1, an artificial intelligence based diet management system 100 for diabetic nephropathy patients according to an embodiment of the present invention includes: a patient information acquisition module 110 for acquiring a text description of personal information of a patient object and a text description of an examination report; the patient information semantic analysis module 120 is configured to perform semantic coding on the text description of the personal information of the patient object and the text description of the examination report, and perform feature interaction based on the complementary attention of the bilateral information to obtain a patient information-patient examination text bilateral complementary semantic feature vector; an alternative recipe acquisition module 130 for acquiring a text description of an alternative recipe; an alternative recipe semantic understanding module 140, configured to perform semantic encoding and dimension adjustment on the text description of the alternative recipe to obtain an adjusted alternative recipe text semantic encoding feature vector; the semantic matching measurement module 150 is configured to determine whether the matching degree between the candidate recipe and the patient object meets a predetermined requirement based on the semantic cross-matching feature between the patient information-patient check text bilateral complementary semantic feature vector and the adjusted candidate recipe text semantic coding feature vector.
As described above in the background art, conventional methods for managing the diet of diabetic nephropathy patients generally adjust the eating habits based on the advice of doctors, but in most cases, specific diet management schemes are not provided for the nephropathy patients, so that the diet management of patients is too dependent on the availability of medical resources and the time and effort of patients, and cannot fully cover the daily diet requirements of patients.
Aiming at the technical problems, the technical concept of the application is to perform semantic interaction analysis on personal information and an inspection report of a patient by adopting an artificial intelligence technology based on deep learning so as to form condition cognition of the patient, and perform semantic cross matching on the condition cognition information of the patient and text description of alternative recipes so as to adapt personalized recipes for the patient. In this way, the patient can be provided with a recipe that suits his needs and health.
Based on this, in the technical solution of the present application, first, a text description of personal information of a patient object and a text description of an inspection report are acquired. Here, the personal information of the patient's subject includes information of the patient's age, sex, height, weight, lifestyle, diet preference, etc., and provides clues about the basic characteristics and lifestyle of the patient, helping to understand the patient's diet requirements and restrictions. The examination report includes various examination results, such as blood glucose level, kidney function index, blood lipid level, etc., of the patient, and can provide information about the current health condition of the patient, particularly an index related to diabetic nephropathy. By analyzing the personal information and the examination report of the patient, the disease progress condition, risk factors and treatment requirements of the patient can be known, so that recipe recommendation can be better performed.
In one embodiment of the present application, the patient information semantic analysis module includes: the patient personal information semantic understanding unit is used for carrying out semantic encoding on the text description of the personal information of the patient object through a patient personal information semantic encoder based on a transducer so as to obtain a patient information text semantic encoding feature vector; a patient examination report semantic understanding unit, configured to perform semantic encoding on a text description of an examination report of the patient object by using a patient examination report semantic encoder based on a transducer to obtain a patient examination text semantic encoding feature vector; and the characteristic interaction unit is used for carrying out characteristic interaction on the patient information text semantic coding characteristic vector and the patient examination text semantic coding characteristic vector so as to obtain the patient information-patient examination text bilateral complementary semantic characteristic vector.
And then, respectively carrying out semantic coding on the text description of the personal information of the patient object and the text description of the examination report to obtain a patient information text semantic coding feature vector and a patient examination text semantic coding feature vector. It should be understood that the personal information of the patient object contains basic information of the patient, while the examination report contains health indicators and disease states of the patient. Semantic characteristic information about physical conditions, disease indexes, eating habits and the like of a patient can be obtained by respectively carrying out semantic coding on the two, and personalized diet recipe recommendation can be carried out based on the information.
And then, the patient information text semantic coding feature vector and the patient examination text semantic coding feature vector pass through a feature interaction module based on bilateral information complementary attention so as to obtain a patient information-patient examination text bilateral complementary semantic feature vector. That is, the semantic coding feature vector of the patient information text and the semantic coding feature vector of the patient examination text are subjected to feature interaction so as to fuse multi-source information, and information complementarity between the patient information and the examination result is fully utilized to obtain more comprehensive and accurate semantic features. Meanwhile, the feature interaction module based on the bilateral information complementary attention can learn the association weight between the text semantic coding feature vector of the patient information and the text semantic coding feature vector of the patient check through the bilateral information complementary attention mechanism, highlight important semantic features and perform deeper semantic characterization, so that personal features and various disease indexes of the patient are comprehensively considered, and more targeted feature representation is provided for subsequent matching analysis.
In a specific embodiment of the present application, the feature interaction unit includes: and the bilateral information complementation subunit is used for obtaining the patient information-patient check text bilateral complementation semantic feature vector through the patient information text semantic coding feature vector and the patient check text semantic coding feature vector by a feature interaction module based on bilateral information complementation attention.
Further, the bilateral information complementation subunit is configured to: concatenating the patient information text semantic coding feature vector and the patient exam text semantic coding feature vector to obtain a first concatenated vector; passing the first series of vectors through a Softmax layer to obtain a first attention weight vector; vector multiplication is carried out on the patient information text semantic coding feature vector and the first attention weight vector to obtain a first fusion vector; passing the first fusion vector through a full connection layer to obtain a patient information text semantic coding feature vector containing patient examination text semantic information; passing the patient information text semantically encoded feature vector through a Softmax layer to obtain a second attention weight vector; vector multiplication is carried out on the second attention weight vector and the patient examination text semantic coding feature vector to obtain a second fusion vector; passing the second fusion vector through a full connection layer to obtain a patient check text semantic coding feature vector containing personal information of a patient; and merging the patient information text semantic coding feature vector containing patient examination text semantic information and the patient examination text semantic coding feature vector containing patient personal information to obtain the patient information-patient examination text bilateral complementary semantic feature vector.
At the same time, a textual description of the alternative recipe is obtained. And then, carrying out semantic coding on the text description of the alternative recipe to obtain a text semantic coding feature vector of the alternative recipe. It should be appreciated that the textual description of the alternative recipe provides information about the food material, the cooking method, the nutritional ingredients, etc. By semantically encoding the alternative recipes, these recipe features can be incorporated into a matching analysis that more fully evaluates the matching of the alternative recipes to the patient, thereby selecting the appropriate alternative recipe based on the patient's health and diet requirements.
In a specific embodiment of the present application, the alternative recipe semantic understanding module is configured to: the alternative recipe semantic coding unit is used for carrying out semantic coding on the text description of the alternative recipe so as to obtain an alternative recipe text semantic coding feature vector; and the dimension adjustment unit is used for carrying out dimension adjustment on the semantic coding feature vector of the alternative recipe text so as to obtain the semantic coding feature vector of the alternative recipe text after adjustment.
Further, it is contemplated that the alternative recipe text semantic coding feature vector and the patient information-patient check text bilateral complementary semantic feature vector may have different dimensions. For subsequent feature interactions and matching analysis, their dimensions need to be unified. Thus, the candidate recipe text semantic coding feature vector is passed through a full connection layer based dimension adjustment module to obtain an adjusted candidate recipe text semantic coding feature vector. That is, by adjusting the candidate recipe text semantic coding feature vector to the same dimension as the patient information-patient check text bilateral complementary semantic feature vector based on the dimension adjustment module of the fully connected layer, the candidate recipe text semantic coding feature vector is better aligned with other features in the feature space so as to perform subsequent feature interaction operation, thereby more accurately evaluating the matching degree of the candidate recipe and the patient.
In a specific embodiment of the present application, the dimension adjustment unit is configured to: and the candidate recipe text semantic coding feature vector is subjected to a dimension adjustment module based on a full connection layer to obtain the adjusted candidate recipe text semantic coding feature vector.
In one embodiment of the present application, the semantic matching metric module includes: the semantic cross matching coefficient calculation unit is used for calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector so as to obtain a patient-recipe semantic matching feature vector; and the matching degree judging unit is used for enabling the patient-recipe semantic matching feature vector to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the matching degree of the alternative recipe and the patient object meets the preset requirement.
Then, calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vectors and the adjustment candidate recipe text semantic coding feature vectors to obtain patient-recipe semantic matching feature vectors. Here, the degree of semantic matching between the patient information and the candidate recipe is quantified by calculating a semantic cross-matching coefficient between the patient information-patient check text bilateral complementary semantic feature vector and the adjusted candidate recipe text semantic coding feature vector. That is, the value of the semantic cross-matching coefficient reflects the degree of semantic similarity or matching between the two.
Specifically, by calculating the semantic cross matching coefficient, a cross association relationship between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector can be embodied, namely, the association of the patient information-patient check text bilateral complementary semantic feature vector relative to the adjustment candidate recipe text semantic coding feature vector and the association of the adjustment candidate recipe text semantic coding feature vector relative to the patient information-patient check text bilateral complementary semantic feature vector can be realized, and the cross association relationship can carry out two-way comparison on semantic features expressed by the patient information-patient check text bilateral complementary semantic feature vector and can further represent the matching degree of semantic features expressed between the candidate recipe and the physical condition of a patient.
In a specific embodiment of the present application, the semantic cross matching coefficient calculating unit is configured to: calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and each position feature value of the adjustment candidate recipe text semantic coding feature vector in the following semantic cross metric formula to obtain the patient-recipe semantic matching feature vector; wherein, the semantic cross metric formula is:
;
Wherein,checking text for the patient information-patient for the +.th in the text bilateral complementary semantic feature vector>Personal characteristic value->Encoding semantically the feature vector for the adjusted alternative recipe text +.>Personal characteristic value->Checking the dimensions of text bilateral complementary semantic feature vectors for the patient information-patient, ++>Is->Semantic cross-matching coefficients, i.e. the +.th in the patient-recipe semantic matching feature vector>Personal characteristic value->A logarithmic function operation with a base of 2 is represented.
The patient-recipe semantic matching feature vector is then passed through a classifier to obtain a classification result that is used to indicate whether the matching degree of the candidate recipe to the patient object meets a predetermined requirement. Here, determining whether the candidate recipe meets the predetermined requirement, i.e. whether the degree of matching with the patient object meets the set criterion, based on the output result of the classifier, helps to provide the patient with a recipe recommendation suited to his needs and health condition.
In one embodiment of the present application, the artificial intelligence based diabetic nephropathy patient diet management system further comprises a training module for training the Transformer based patient personal information semantic encoder, the Transformer based patient examination report semantic encoder, the bilateral information complementary attention based feature interaction module, the full-connection layer based dimension adjustment module and the classifier; wherein, training module includes: the training information acquisition unit is used for acquiring training data, wherein the training data comprises text description of training personal information of a patient object, text description of training inspection report, text description of training alternative recipes and whether the matching degree of the training alternative recipes and the patient object reaches a true value of a preset requirement; the training patient personal information semantic coding unit is used for carrying out semantic coding on the text description of the training personal information of the patient object through a patient personal information semantic coder based on a transducer so as to obtain training patient information text semantic coding feature vectors; a training patient examination report semantic coding unit, configured to perform semantic coding on a text description of a training examination report of the patient object by using a patient examination report semantic coder based on a transducer to obtain a training patient examination text semantic coding feature vector; the training bilateral information complementary feature interaction unit is used for enabling the training patient information text semantic coding feature vector and the training patient checking text semantic coding feature vector to pass through a feature interaction module based on bilateral information complementary attention so as to obtain training patient information-patient checking text bilateral complementary semantic feature vector; the training alternative recipe semantic coding unit is used for carrying out semantic coding on the text description of the training alternative recipe so as to obtain a training alternative recipe text semantic coding feature vector; the training dimension adjustment unit is used for enabling the training alternative recipe text semantic coding feature vector to pass through a dimension adjustment module based on a full connection layer to obtain a training adjustment alternative recipe text semantic coding feature vector; the training semantic cross matching coefficient calculation unit is used for calculating semantic cross matching coefficients between the training patient information-patient check text bilateral complementary semantic feature vectors and the training adjustment candidate recipe text semantic coding feature vectors so as to obtain training patient-recipe semantic matching feature vectors; a classification loss function value calculation unit, configured to pass the training patient-recipe semantic matching feature vector through a classifier to obtain a classification loss function value; a specific loss function value calculation unit, configured to calculate a specific loss function value of the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment candidate recipe text semantic coding feature vector; the training unit is used for training the patient personal information semantic encoder based on the Transformer, the patient examination report semantic encoder based on the Transformer, the feature interaction module based on the bilateral information complementary attention, the dimension adjustment module based on the full-connection layer and the classifier by taking the weighted sum of the classified loss function value and the specific loss function value as the loss function value.
In the above technical solution, the training patient information-patient check text bilateral complementary semantic feature vector expresses text semantic feature interactive coding features of the text description of the training personal information of the patient object and the text description of the training check report, and the training adjustment back-up recipe text semantic coding feature vector expresses text semantic full-connection associated coding features of the text description of the training back-up recipe, where, considering a source text semantic density difference between the text description of the training personal information of the patient object and the text description of the training check report and the text description of the training back-up recipe, the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment back-up recipe text semantic coding feature vector may also have significant feature group density representation differences in feature dimensions, so that when the model is trained as a whole, there may be iteration imbalance between the training patient information text semantic feature coding and the training back-up recipe text semantic feature coding, affecting the overall training efficiency of the model.
Thus, the applicant of the present application considers promoting feature group density representation consistency of the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment back-up recipe text semantic coding feature vector, thereby further introducing a specific penalty function for the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment back-up recipe text semantic coding feature vector expressed as: calculating the training patient information-patient check text bilateral complementary semantic feature vectors and the specific loss function values of the training adjustment candidate recipe text semantic coding feature vectors according to the following optimization formula; wherein, the optimization formula is:
;
Wherein,is the training patient information-patient check text bilateral complementary semantic feature vector, < >>Is the feature vector obtained by linear interpolation of the text semantic coding feature vector of the candidate recipe after training and adjustment, and the feature vectorAnd the feature vector->Having the same length->And->Representing the square of the two norms of the vector, +.>Andis the feature vector +.>And the feature vector->Characteristic value of>Representing the specific loss function value,>representing the calculation of a value of the natural exponent function raised to a power of a value, ">Representing per-position subtraction.
Here, the penalty function performs group count attention based on feature group density by recursively mapping the group count as output feature group density to perform adaptive attention of different density representation patterns between the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment alternate recipe text semantic coding feature vector. By taking the model as a loss function to train the model, the model can avoid overestimation and underestimation aiming at different density modes under the characteristic distribution of the training patient information-patient checking text bilateral complementary semantic characteristic vector and the training adjustment candidate recipe text semantic coding characteristic vector, and learn the corresponding relation between the characteristic value distribution and the group density distribution, thereby realizing the consistency optimization of the characteristic group density representation between the training patient information-patient checking text bilateral complementary semantic characteristic vector and the training adjustment candidate recipe text semantic coding characteristic vector with different characteristic densities, and improving the overall training efficiency of the model.
In summary, the artificial intelligence based diet management system 100 for diabetic nephropathy patients according to the embodiment of the present invention is illustrated, which performs semantic interactive analysis on personal information and inspection reports of patients by using artificial intelligence technology based on deep learning to form condition cognition for the patients, and performs semantic cross matching on condition cognition information of the patients and text description of alternative recipes, so as to adapt personalized recipes for the patients. In this way, the patient can be provided with a recipe that suits his needs and health.
As described above, the artificial intelligence based diet management system 100 for diabetic nephropathy patients according to the embodiment of the present invention can be implemented in various terminal devices, such as a server for artificial intelligence based diet management for diabetic nephropathy patients, etc. In one example, the artificial intelligence based diabetic nephropathy patient meal management system 100 according to embodiments of the present invention may be integrated into the terminal device as one software module and/or hardware module. For example, the artificial intelligence based diabetic nephropathy patient meal management system 100 can be a software module in the operating system of the terminal device or can be an application developed for the terminal device; of course, the artificial intelligence based diabetic nephropathy patient diet management system 100 can also be one of the numerous hardware modules of the terminal device.
Alternatively, in another example, the artificial intelligence based diabetic nephropathy patient meal management system 100 and the terminal device may also be separate devices, and the artificial intelligence based diabetic nephropathy patient meal management system 100 may be connected to the terminal device through a wired and/or wireless network and transmit interactive information in a agreed data format.
Fig. 2 is a flowchart of a method for diet management of diabetic nephropathy patients based on artificial intelligence according to an embodiment of the present invention. Fig. 3 is a schematic diagram of a system architecture of an artificial intelligence-based diet management method for diabetic nephropathy patients according to an embodiment of the present invention. As shown in fig. 2 and 3, an artificial intelligence-based diet management method for diabetic nephropathy patients includes: 210, obtaining a text description of personal information of the patient object and a text description of the examination report; 220, performing semantic coding and bilateral complementary feature interaction on the text description of the personal information of the patient object and the text description of the examination report to obtain a patient information-patient examination text bilateral complementary semantic feature vector; 230, obtaining a text description of the alternative recipe; 240, carrying out semantic coding and dimension adjustment on the text description of the alternative recipe to obtain an adjusted alternative recipe text semantic coding feature vector; 250, determining whether the matching degree of the candidate recipe and the patient object meets the preset requirement based on the semantic cross matching feature between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector.
It will be appreciated by those skilled in the art that the specific operation of the steps in the above-described artificial intelligence-based diet management method for diabetic patients has been described in detail above with reference to the description of the artificial intelligence-based diet management system for diabetic patients of fig. 1, and thus, repetitive descriptions thereof will be omitted.
Fig. 4 is an application scenario diagram of an artificial intelligence-based diet management system for diabetic nephropathy patients provided in an embodiment of the present invention. As shown in fig. 4, in this application scenario, first, a text description of personal information of a patient object (C1 as illustrated in fig. 4) and a text description of an examination report (C2 as illustrated in fig. 4) are acquired; then, the obtained text description of the personal information and the text description of the inspection report are input into a server (S as illustrated in fig. 4) deployed with an artificial intelligence based diabetic nephropathy patient diet management algorithm, wherein the server is capable of processing the text description of the personal information and the text description of the inspection report based on the artificial intelligence based diabetic nephropathy patient diet management algorithm to determine whether the matching degree of the candidate recipe and the patient object meets a predetermined requirement.
The foregoing description of the embodiments has been provided for the purpose of illustrating the general principles of the invention, and is not meant to limit the scope of the invention, but to limit the invention to the particular embodiments, and any modifications, equivalents, improvements, etc. that fall within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (8)
1. An artificial intelligence based diet management system for diabetic nephropathy patients, comprising:
a patient information acquisition module for acquiring a text description of personal information of a patient object and a text description of an examination report;
the patient information semantic analysis module is used for carrying out semantic coding on the text description of the personal information of the patient object and the text description of the examination report and carrying out feature interaction based on the bilateral information complementary attention so as to obtain a patient information-patient examination text bilateral complementary semantic feature vector;
the alternative recipe acquisition module is used for acquiring text description of the alternative recipes;
the alternative recipe semantic understanding module is used for carrying out semantic coding and dimension adjustment on the text description of the alternative recipe so as to obtain an adjusted alternative recipe text semantic coding feature vector;
The semantic matching measurement module is used for determining whether the matching degree of the candidate recipe and the patient object meets the preset requirement or not based on semantic cross matching characteristics between the patient information-patient check text bilateral complementary semantic feature vector and the adjusted candidate recipe text semantic coding feature vector;
wherein, the semantic matching measurement module comprises:
the semantic cross matching coefficient calculation unit is used for calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector so as to obtain a patient-recipe semantic matching feature vector;
the matching degree judging unit is used for enabling the patient-recipe semantic matching feature vector to pass through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the matching degree of the alternative recipe and the patient object meets a preset requirement;
the semantic cross matching coefficient calculating unit is used for:
calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and each position feature value of the adjustment candidate recipe text semantic coding feature vector in the following semantic cross metric formula to obtain the patient-recipe semantic matching feature vector; wherein, the semantic cross metric formula is:
;
Wherein,checking text for the patient information-patient for the +.th in the text bilateral complementary semantic feature vector>The value of the characteristic is a value of,encoding semantically the feature vector for the adjusted alternative recipe text +.>Personal characteristic value->Checking the dimensions of text bilateral complementary semantic feature vectors for the patient information-patient, ++>Is->Semantic cross-matching coefficients, i.e. the +.th in the patient-recipe semantic matching feature vector>Personal characteristic value->A logarithmic function operation with a base of 2 is represented.
2. The artificial intelligence based diabetic nephropathy patient diet management system of claim 1, wherein the patient information semantic analysis module comprises:
the patient personal information semantic understanding unit is used for carrying out semantic encoding on the text description of the personal information of the patient object through a patient personal information semantic encoder based on a transducer so as to obtain a patient information text semantic encoding feature vector;
a patient examination report semantic understanding unit, configured to perform semantic encoding on a text description of an examination report of the patient object by using a patient examination report semantic encoder based on a transducer to obtain a patient examination text semantic encoding feature vector;
And the characteristic interaction unit is used for carrying out characteristic interaction on the patient information text semantic coding characteristic vector and the patient examination text semantic coding characteristic vector so as to obtain the patient information-patient examination text bilateral complementary semantic characteristic vector.
3. The artificial intelligence based diabetic nephropathy patient diet management system of claim 2, wherein the feature interaction unit comprises:
and the bilateral information complementation subunit is used for obtaining the patient information-patient check text bilateral complementation semantic feature vector through the patient information text semantic coding feature vector and the patient check text semantic coding feature vector by a feature interaction module based on bilateral information complementation attention.
4. The artificial intelligence based diabetic nephropathy patient diet management system of claim 3, wherein the bilateral information complementation subunit is configured to:
concatenating the patient information text semantic coding feature vector and the patient exam text semantic coding feature vector to obtain a first concatenated vector;
passing the first series of vectors through a Softmax layer to obtain a first attention weight vector;
Vector multiplication is carried out on the patient information text semantic coding feature vector and the first attention weight vector to obtain a first fusion vector;
passing the first fusion vector through a full connection layer to obtain a patient information text semantic coding feature vector containing patient examination text semantic information;
passing the patient information text semantically encoded feature vector through a Softmax layer to obtain a second attention weight vector;
vector multiplication is carried out on the second attention weight vector and the patient examination text semantic coding feature vector to obtain a second fusion vector;
passing the second fusion vector through a full connection layer to obtain a patient check text semantic coding feature vector containing personal information of a patient;
and merging the patient information text semantic coding feature vector containing patient examination text semantic information and the patient examination text semantic coding feature vector containing patient personal information to obtain the patient information-patient examination text bilateral complementary semantic feature vector.
5. The artificial intelligence based diabetic nephropathy patient diet management system of claim 4, wherein the alternative recipe semantic understanding module is configured to:
The alternative recipe semantic coding unit is used for carrying out semantic coding on the text description of the alternative recipe so as to obtain an alternative recipe text semantic coding feature vector;
and the dimension adjustment unit is used for carrying out dimension adjustment on the semantic coding feature vector of the alternative recipe text so as to obtain the semantic coding feature vector of the alternative recipe text after adjustment.
6. The artificial intelligence based diabetic nephropathy patient diet management system of claim 5, wherein the dimension adjustment unit is configured to:
and the candidate recipe text semantic coding feature vector is subjected to a dimension adjustment module based on a full connection layer to obtain the adjusted candidate recipe text semantic coding feature vector.
7. The artificial intelligence based diabetic nephropathy patient dietary management system of claim 6, further comprising a training module for training the Transformer based patient personal information semantic encoder, the Transformer based patient exam report semantic encoder, the bilateral information complementary attention based feature interaction module, the full connectivity layer based dimension adjustment module, and the classifier; wherein, training module includes:
The training information acquisition unit is used for acquiring training data, wherein the training data comprises text description of training personal information of a patient object, text description of training inspection report, text description of training alternative recipes and whether the matching degree of the training alternative recipes and the patient object reaches a true value of a preset requirement;
the training patient personal information semantic coding unit is used for carrying out semantic coding on the text description of the training personal information of the patient object through a patient personal information semantic coder based on a transducer so as to obtain training patient information text semantic coding feature vectors;
a training patient examination report semantic coding unit, configured to perform semantic coding on a text description of a training examination report of the patient object by using a patient examination report semantic coder based on a transducer to obtain a training patient examination text semantic coding feature vector;
the training bilateral information complementary feature interaction unit is used for enabling the training patient information text semantic coding feature vector and the training patient checking text semantic coding feature vector to pass through a feature interaction module based on bilateral information complementary attention so as to obtain training patient information-patient checking text bilateral complementary semantic feature vector;
The training alternative recipe semantic coding unit is used for carrying out semantic coding on the text description of the training alternative recipe so as to obtain a training alternative recipe text semantic coding feature vector;
the training dimension adjustment unit is used for enabling the training alternative recipe text semantic coding feature vector to pass through a dimension adjustment module based on a full connection layer to obtain a training adjustment alternative recipe text semantic coding feature vector;
the training semantic cross matching coefficient calculation unit is used for calculating semantic cross matching coefficients between the training patient information-patient check text bilateral complementary semantic feature vectors and the training adjustment candidate recipe text semantic coding feature vectors so as to obtain training patient-recipe semantic matching feature vectors;
a classification loss function value calculation unit, configured to pass the training patient-recipe semantic matching feature vector through a classifier to obtain a classification loss function value;
a specific loss function value calculation unit, configured to calculate a specific loss function value of the training patient information-patient check text bilateral complementary semantic feature vector and the training adjustment candidate recipe text semantic coding feature vector;
the training unit is used for training the patient personal information semantic encoder based on the Transformer, the patient examination report semantic encoder based on the Transformer, the feature interaction module based on the bilateral information complementary attention, the dimension adjustment module based on the full-connection layer and the classifier by taking the weighted sum of the classified loss function value and the specific loss function value as the loss function value.
8. An artificial intelligence-based diet management method for a diabetic nephropathy patient, comprising the following steps:
acquiring a text description of personal information of a patient object and a text description of an examination report;
semantic coding and bilateral complementary feature interaction are carried out on the text description of the personal information of the patient object and the text description of the examination report so as to obtain a patient information-patient examination text bilateral complementary semantic feature vector;
acquiring a text description of an alternative recipe;
performing semantic coding and dimension adjustment on the text description of the alternative recipe to obtain an adjusted alternative recipe text semantic coding feature vector;
determining whether the matching degree of the candidate recipe and the patient object meets a preset requirement based on semantic cross matching features between the patient information-patient check text bilateral complementary semantic feature vector and the adjusted candidate recipe text semantic coding feature vector;
wherein determining whether the matching degree of the candidate recipe and the patient object meets a predetermined requirement based on semantic cross matching features between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector comprises:
Calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vectors and the adjustment candidate recipe text semantic coding feature vectors to obtain patient-recipe semantic matching feature vectors;
passing the patient-recipe semantic matching feature vector through a classifier to obtain a classification result, wherein the classification result is used for indicating whether the matching degree of the alternative recipe and the patient object meets a preset requirement;
the method for calculating the semantic cross matching coefficient between the patient information-patient check text bilateral complementary semantic feature vector and the adjustment candidate recipe text semantic coding feature vector to obtain a patient-recipe semantic matching feature vector comprises the following steps:
calculating semantic cross matching coefficients between the patient information-patient check text bilateral complementary semantic feature vector and each position feature value of the adjustment candidate recipe text semantic coding feature vector in the following semantic cross metric formula to obtain the patient-recipe semantic matching feature vector; wherein, the semantic cross metric formula is:
;
wherein,checking text for the patient information-patient for the +.th in the text bilateral complementary semantic feature vector >The value of the characteristic is a value of,encoding semantically the feature vector for the adjusted alternative recipe text +.>Personal characteristic value->Checking the dimensions of text bilateral complementary semantic feature vectors for the patient information-patient, ++>Is->Semantic cross-matching coefficients, i.e. the +.th in the patient-recipe semantic matching feature vector>Personal characteristic value->A logarithmic function operation with a base of 2 is represented.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410045683.6A CN117558410B (en) | 2024-01-12 | 2024-01-12 | Artificial intelligence-based diet management system and method for diabetic nephropathy patients |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202410045683.6A CN117558410B (en) | 2024-01-12 | 2024-01-12 | Artificial intelligence-based diet management system and method for diabetic nephropathy patients |
Publications (2)
Publication Number | Publication Date |
---|---|
CN117558410A true CN117558410A (en) | 2024-02-13 |
CN117558410B CN117558410B (en) | 2024-03-22 |
Family
ID=89821850
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202410045683.6A Active CN117558410B (en) | 2024-01-12 | 2024-01-12 | Artificial intelligence-based diet management system and method for diabetic nephropathy patients |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN117558410B (en) |
Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108335731A (en) * | 2018-02-09 | 2018-07-27 | 辽宁工程技术大学 | A kind of invalid diet's recommendation method based on computer vision |
CN109903836A (en) * | 2019-03-31 | 2019-06-18 | 山西慧虎健康科技有限公司 | A kind of diet intelligent recommendation and matching system and method based on constitution and big data |
CN111599438A (en) * | 2020-04-02 | 2020-08-28 | 浙江工业大学 | Real-time diet health monitoring method for diabetic patient based on multi-modal data |
CN112233772A (en) * | 2020-10-13 | 2021-01-15 | 杭州电子科技大学 | Healthy recipe recommendation system based on machine learning |
CN113035319A (en) * | 2021-04-26 | 2021-06-25 | 南通大学 | Automatic recommendation method and system for lifestyle management path of diabetic patient |
CN113314198A (en) * | 2021-04-29 | 2021-08-27 | 天津市第五中心医院 | Method and system for automatically generating recipes according to nutritional ingredients required by dialysis patient |
KR102360651B1 (en) * | 2021-06-15 | 2022-02-10 | 메디프레쉬주식회사 | System for providing personalized meal plan service for improving medical test result using medical record |
CN114864047A (en) * | 2022-03-29 | 2022-08-05 | 青岛海尔科技有限公司 | Recipe recommendation method, storage medium and electronic device |
CN116168804A (en) * | 2023-04-21 | 2023-05-26 | 杭州青橄榄网络技术有限公司 | Patient diet recommendation system and method based on HIS system |
CN116246743A (en) * | 2022-12-19 | 2023-06-09 | 万达信息股份有限公司 | Precise nutrition comprehensive evaluation system based on portrait label |
CN116417115A (en) * | 2023-06-07 | 2023-07-11 | 北京四海汇智科技有限公司 | Personalized nutrition scheme recommendation method and system for gestational diabetes patients |
CN116525067A (en) * | 2023-06-21 | 2023-08-01 | 安徽宏元聚康医疗科技有限公司 | Nutrient recipe recommendation system and method |
KR20230116355A (en) * | 2022-01-28 | 2023-08-04 | 씨제이올리브네트웍스 주식회사 | Method and apparatus for recommending personalized recipe |
CN116682532A (en) * | 2023-06-12 | 2023-09-01 | 中日友好医院(中日友好临床医学研究所) | Method and device for managing health of pre-diabetes patient |
CN116978513A (en) * | 2022-04-24 | 2023-10-31 | 北京左医科技有限公司 | Diet information reminding method and device, electronic equipment and storage medium |
CN117198466A (en) * | 2023-11-08 | 2023-12-08 | 北京四海汇智科技有限公司 | Diet management method and system for kidney disease patients |
-
2024
- 2024-01-12 CN CN202410045683.6A patent/CN117558410B/en active Active
Patent Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108335731A (en) * | 2018-02-09 | 2018-07-27 | 辽宁工程技术大学 | A kind of invalid diet's recommendation method based on computer vision |
CN109903836A (en) * | 2019-03-31 | 2019-06-18 | 山西慧虎健康科技有限公司 | A kind of diet intelligent recommendation and matching system and method based on constitution and big data |
CN111599438A (en) * | 2020-04-02 | 2020-08-28 | 浙江工业大学 | Real-time diet health monitoring method for diabetic patient based on multi-modal data |
CN112233772A (en) * | 2020-10-13 | 2021-01-15 | 杭州电子科技大学 | Healthy recipe recommendation system based on machine learning |
CN113035319A (en) * | 2021-04-26 | 2021-06-25 | 南通大学 | Automatic recommendation method and system for lifestyle management path of diabetic patient |
CN113314198A (en) * | 2021-04-29 | 2021-08-27 | 天津市第五中心医院 | Method and system for automatically generating recipes according to nutritional ingredients required by dialysis patient |
KR102360651B1 (en) * | 2021-06-15 | 2022-02-10 | 메디프레쉬주식회사 | System for providing personalized meal plan service for improving medical test result using medical record |
KR20230116355A (en) * | 2022-01-28 | 2023-08-04 | 씨제이올리브네트웍스 주식회사 | Method and apparatus for recommending personalized recipe |
CN114864047A (en) * | 2022-03-29 | 2022-08-05 | 青岛海尔科技有限公司 | Recipe recommendation method, storage medium and electronic device |
CN116978513A (en) * | 2022-04-24 | 2023-10-31 | 北京左医科技有限公司 | Diet information reminding method and device, electronic equipment and storage medium |
CN116246743A (en) * | 2022-12-19 | 2023-06-09 | 万达信息股份有限公司 | Precise nutrition comprehensive evaluation system based on portrait label |
CN116168804A (en) * | 2023-04-21 | 2023-05-26 | 杭州青橄榄网络技术有限公司 | Patient diet recommendation system and method based on HIS system |
CN116417115A (en) * | 2023-06-07 | 2023-07-11 | 北京四海汇智科技有限公司 | Personalized nutrition scheme recommendation method and system for gestational diabetes patients |
CN116682532A (en) * | 2023-06-12 | 2023-09-01 | 中日友好医院(中日友好临床医学研究所) | Method and device for managing health of pre-diabetes patient |
CN116525067A (en) * | 2023-06-21 | 2023-08-01 | 安徽宏元聚康医疗科技有限公司 | Nutrient recipe recommendation system and method |
CN117198466A (en) * | 2023-11-08 | 2023-12-08 | 北京四海汇智科技有限公司 | Diet management method and system for kidney disease patients |
Non-Patent Citations (3)
Title |
---|
李冬梅;李云凤;董霞;奥日瀚;陈晔;白晓光;: "成人糖尿病患者饮食配给方法的探讨――基本食物匹配法", 实用糖尿病杂志, no. 01, 20 February 2010 (2010-02-20) * |
汪名凤;任红霞;朱乐凤;: "个体化饮食护理在1型糖尿病患者治疗中的作用", 当代护士(下旬刊), no. 03, 1 March 2020 (2020-03-01) * |
邓丽丽;张广清;林静霞;刘惠;傅立哲;吴一凡;: "个体化健康食谱的建立与实施对早期糖尿病肾病患者的影响", 世界科学技术-中医药现代化, no. 06, 20 September 2013 (2013-09-20) * |
Also Published As
Publication number | Publication date |
---|---|
CN117558410B (en) | 2024-03-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
KR102024375B1 (en) | Apparatus and method for predicting disease risk of chronic kidney disease | |
CN109411082A (en) | A kind of Evaluation of Medical Quality and medical recommended method | |
CN109378066A (en) | A kind of control method and control device for realizing disease forecasting based on feature vector | |
CN110427486B (en) | Body condition text classification method, device and equipment | |
CN116417115B (en) | Personalized nutrition scheme recommendation method and system for gestational diabetes patients | |
CN115293128A (en) | Model training method and system based on multi-modal contrast learning radiology report generation | |
Mittal et al. | DFR-HL: Diabetic Food Recommendation Using Hybrid Learning Methods | |
CN113488168A (en) | Personalized physical examination item recommendation method based on user health information characteristic set | |
CN114496243A (en) | Data processing method, data processing device, storage medium and electronic equipment | |
CN116797572A (en) | Rheumatoid arthritis activity grading device based on multi-mode data | |
CN114864090A (en) | Data intelligence-based intervention effect evaluation system for overweight and obese children | |
CN111863247A (en) | Brain age cascade refining prediction method and system based on structural magnetic resonance image | |
CN117558410B (en) | Artificial intelligence-based diet management system and method for diabetic nephropathy patients | |
CN113220895A (en) | Information processing method and device based on reinforcement learning and terminal equipment | |
CN115547502A (en) | Hemodialysis patient risk prediction device based on time sequence data | |
Roy et al. | OBESEYE: Interpretable Diet Recommender for Obesity Management using Machine Learning and Explainable AI | |
KR20220158158A (en) | A deep learning based health care service platform by vitro diagnosis and the recommendation method for health care using the same | |
Ispirova et al. | Exploring Knowledge Domain Bias on a Prediction Task for Food and Nutrition Data | |
CN109119159B (en) | Deep learning medical diagnosis system based on rapid weight mechanism | |
CN117594195B (en) | Kidney disease patient diet analysis system and method based on big data | |
CN117373619B (en) | Recipe generation method and generation system based on intestinal ostomy bag excrement monitoring result | |
CN112233797A (en) | Health management method and device and computer readable storage medium | |
CN118098482B (en) | Intelligent medical management system and method based on 5G technology | |
CN111354456A (en) | Traditional Chinese medicine constitution type detection method and device | |
CN114708965B (en) | Diagnosis recommendation method and device, electronic equipment and storage medium |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |