CN113035319A - Automatic recommendation method and system for lifestyle management path of diabetic patient - Google Patents
Automatic recommendation method and system for lifestyle management path of diabetic patient Download PDFInfo
- Publication number
- CN113035319A CN113035319A CN202110451694.0A CN202110451694A CN113035319A CN 113035319 A CN113035319 A CN 113035319A CN 202110451694 A CN202110451694 A CN 202110451694A CN 113035319 A CN113035319 A CN 113035319A
- Authority
- CN
- China
- Prior art keywords
- diabetic
- knowledge graph
- management
- feature points
- data
- 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.)
- Withdrawn
Links
- 206010012601 diabetes mellitus Diseases 0.000 title claims abstract description 221
- 238000000034 method Methods 0.000 title claims abstract description 27
- 230000001575 pathological effect Effects 0.000 claims abstract description 59
- 238000013136 deep learning model Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 13
- 235000005911 diet Nutrition 0.000 claims description 41
- 230000037213 diet Effects 0.000 claims description 36
- 238000005516 engineering process Methods 0.000 claims description 24
- 238000000605 extraction Methods 0.000 claims description 18
- 238000010276 construction Methods 0.000 claims description 14
- 238000005295 random walk Methods 0.000 claims description 13
- 238000004891 communication Methods 0.000 claims description 11
- 238000012545 processing Methods 0.000 claims description 7
- 230000000378 dietary effect Effects 0.000 claims description 5
- 230000037361 pathway Effects 0.000 claims 6
- 238000004422 calculation algorithm Methods 0.000 description 18
- 239000008280 blood Substances 0.000 description 13
- 210000004369 blood Anatomy 0.000 description 13
- 238000001914 filtration Methods 0.000 description 9
- 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 6
- 239000008103 glucose Substances 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 235000020880 diabetic diet Nutrition 0.000 description 4
- 235000013305 food Nutrition 0.000 description 4
- 235000021395 porridge Nutrition 0.000 description 4
- 229940079593 drug Drugs 0.000 description 3
- 239000003814 drug Substances 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 235000015872 dietary supplement Nutrition 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 208000017667 Chronic Disease Diseases 0.000 description 1
- 208000002249 Diabetes Complications Diseases 0.000 description 1
- 208000003443 Unconsciousness Diseases 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 210000004153 islets of langerhan Anatomy 0.000 description 1
- 235000013372 meat Nutrition 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
- 208000001072 type 2 diabetes mellitus Diseases 0.000 description 1
Images
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
- G16H20/30—ICT specially adapted for therapies or health-improving plans, e.g. for handling prescriptions, for steering therapy or for monitoring patient compliance relating to physical therapies or activities, e.g. physiotherapy, acupressure or exercising
-
- 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
- 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
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- Epidemiology (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Physical Education & Sports Medicine (AREA)
- Nutrition Science (AREA)
- Biomedical Technology (AREA)
- Databases & Information Systems (AREA)
- Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention discloses a method and a system for automatically recommending a lifestyle management path of a diabetic, wherein the method comprises the following steps: obtaining data related to lifestyle management of a diabetic patient to form multi-source heterogeneous data; constructing a multi-modal knowledge graph for the lifestyle management of the diabetic patient based on the multi-source heterogeneous data; performing relation completion on the multi-mode knowledge graph; inputting the semantic information of the multi-modal knowledge graph after the relationship is completed into a deep learning model for training and learning; the method comprises the steps of constructing a sub-knowledge graph based on pathological information and target life style information of a certain diabetic, inputting semantic information of the sub-knowledge graph into a trained deep learning model for graph similarity matching prediction, and extracting the part with the highest similarity with the sub-knowledge graph in the multi-mode knowledge graph to serve as a recommended life style management path of the diabetic.
Description
Technical Field
The invention relates to the technical field of diabetic lifestyle management path recommendation, in particular to a method and a system for automatically recommending a diabetic lifestyle management path based on a knowledge graph.
Background
Diabetes is a chronic disease with prevalence rates as high as 10%, and complications can affect various organs of the body. Diet control is one of the main means for managing diabetes, and the main function of the diet control is to reduce the load of the pancreatic islets of the patient, maintain the blood sugar and the blood fat at the level of normal people, and effectively improve the health condition of the patient by being matched with the use of medicines. Traditionally, the diet of a diabetic patient is prepared by a specific doctor or a dietician, and one or more patients are difficult to manage in procedure, and the types of food are single, so that the conditions of not conforming to the taste of the patient and the like occur, and the workload of the dietician is greatly increased. In addition, exercise is also performed by a professional who instructs a patient to take a prescribed exercise after performing a caloric calculation. Studies have also found whether the combination of dietary supplementation with physical exercise may be more beneficial to patients newly diagnosed with type 2 diabetes than if only dietary supplementation were undertaken, which provides data support for personalized management of diabetic patients. Although the effect of managing diabetes can be achieved through lifestyle intervention control of patients, compared with the vast population of diabetes patients, the resources of diabetes medical care personnel are insufficient, and the problem of personalized recommendation of diet and exercise schemes of diabetes patients is urgently needed to be solved.
With the rise and rapid development of the internet, users are unconscious in the face of diabetes patients with huge data volume and unconsciously submerged in massive data. Despite the existing dietary and exercise clinical guidelines of diabetic patients, there are many new intervention modes with the penetration of big data studies. There is a technical report for developing diabetes personalized management application based on multi-source heterogeneous data, and this technology performs mining analysis on basic information, acquired physiological signal data, living habits, medication conditions and the like of a diabetes patient, and uses a dual clustering recommendation method to push information such as personalized living styles, exercise management, medication schemes and the like for the user, guide the user to select better living habits, actively perform behavior intervention, control the onset of diabetes complications, reduce the expenditure of families on medical health, and improve the quality of life of the diabetes patient.
The generation of the recommendation algorithm meets the requirements of the diabetic patients on available information, the dilemma that the users cannot obtain the data really needed by the users is eliminated, and the utilization rate of the information is improved. For example, by means of a speed sensor, a wireless internet technology and a global positioning system, the recommendation system can analyze the exercise state (sleeping, exercising, working, eating, sitting idle, walking and the like) of a patient by acquiring the exercise speed and heart rate information of the patient, and then provides a diversified and detailed exercise prescription according to the result of a blood glucose meter, so that only increasing or decreasing of exercise can be recommended. And the software combined exercise equipment can give exercise prescriptions through comprehensive human body component analysis reports, aerobic capacity reports and muscle strength reports. The recommendation algorithm is used for recommending other information which is unknown by the user and is interesting to the user according to the attributes of the user such as interest and demand information. And these attributes are used to establish an association between users. The classified users have close relationship with each other due to the interest points generated by the recommendation algorithm analysis. The personalized service generates dependence for users. Recommendation algorithms have been applied in many fields, most typically the field of e-commerce, and are rapidly and increasingly mature in this field. In the field of intelligent medical treatment, recommendation algorithms are also concerned to be applied. The personalized diet recommendation algorithm for the diabetic is one of the important research subjects in the field of intelligent medical treatment. Traditional diabetic diet recommendation algorithms are faced with increasing patient data and fail to accurately match the relationship between patient body indices and recommended foods. On the basis of diabetes food exchange shares, the existing diabetes diet recommendation algorithms mostly adopt a recommendation algorithm based on association rules, a recommendation algorithm based on content, a recommendation algorithm based on hierarchical analysis, a collaborative filtering recommendation algorithm, a recommendation algorithm based on constraint, or a plurality of methods which are jointly completed under the control of diagnosis of an attending doctor and a dietician. Diabetes management programs are scored by a paradigm inference algorithm (CBR), then user unscored items are inferentially scored, and finally Top-N recommendations are made with user-based collaborative filtering. By deep knowledge of a case reasoning algorithm (CBR), the characteristic that case reasoning can be used for prediction is applied to the cold start problem of collaborative filtering, the case library is corrected through a forgetting model, then the user unscored items are subjected to reasoning and scoring, and finally Top-N recommendation is carried out through the collaborative filtering based on the user. Collaborative filtering combined with paradigm reasoning has certain auxiliary effect on relieving the sparsity problem of the collaborative filtering matrix. In addition, a diet recommendation platform is created between the patient and the doctor by means of the improved recommendation engine, the mobile platform and the information support system. Some researchers use collaborative filtering algorithms to realize diet recommendation, including user-based collaborative filtering algorithms and project-based collaborative filtering algorithms. The method comprises the steps of constructing a user-item score matrix through the score information of a historical dish ordering template of a user, calculating the similarity between users or between items through a cosine similarity formula, finding out a neighbor set with the highest similarity with the current user or item, predicting the unscored items of a target user by using the history score data of the neighbor user on the items by adopting a weighted average method, finding the highest score of a diet template according to the ranking of the scores of the generated recommended item set, recommending the diet template to the target user finally, calculating the calorie required by the patient for one day, and then giving quantitative recommendation to food. The method comprises the steps of establishing a model framework of a diabetes diet recommendation system by utilizing the idea of recommending diet for the diabetic through a heterogeneous information network by sukifeng and the like, and explaining a model flow of the recommendation system through an actual case. The diabetic diet recommendation process comprises four data types of a patient type P, a patient symptom A, a recommendation principle T and a recommendation scheme F, the four data types are used as heterogeneous information networks, attribute selection contained in the data types is explained, meanwhile, potential relations among the data types are analyzed, the diabetic diet heterogeneous information network with the patient as the center is constructed, patient clustering is achieved based on a method combining traditional clustering and heterogeneous information network sequencing, a diet recommendation scheme is obtained, a diabetic diet recommendation system comprising a patient-side module, a recommendation module, a database management module and a medical-side module is built, and the system achieves functions of diet recommendation and diet recording.
Disclosure of Invention
Aiming at the problems and the defects in the prior art, the invention provides a novel method and a system for automatically recommending a lifestyle management path of a diabetic patient based on a knowledge graph.
The invention solves the technical problems through the following technical scheme:
the invention provides an automatic recommendation method for a lifestyle management path of a diabetic, which is characterized by comprising the following steps of:
s1, acquiring data related to the diabetes patient life style management to form multi-source heterogeneous data, wherein the multi-source heterogeneous data comprises pathological data and life style management data of the diabetes patient, and the life style management data comprises diet data and exercise data;
s2, constructing a multi-modal knowledge graph for the life style management of the diabetic patient based on multi-source heterogeneous data: extracting different feature points of the diabetic patient and the relation between the different feature points from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in a knowledge graph, the feature points with the superior and inferior conceptual relation belong to the nodes corresponding to the feature points with the superior concept, and the connection edge between the two corresponding nodes is constructed based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
s3, performing relation completion on the multi-modal knowledge graph: calculating the link similarity between the nodes in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without an edge relation in the multi-modal knowledge graph, and constructing an edge between the two nodes without the edge relation when the link similarity reaches a set value;
s4, inputting the semantic information of the multi-modal knowledge map after the relationship is completed into a deep learning model for training and learning;
s5, constructing a sub-knowledge graph based on pathological information and target life style information of a certain diabetic, inputting semantic information of the sub-knowledge graph into a trained deep learning model for graph similarity matching prediction, and extracting the part with the highest similarity with the sub-knowledge graph in the multi-mode knowledge graph to serve as a recommended life style management path of the diabetic.
Preferably, in step S1, the data related to the lifestyle management of the diabetic is crawled through web crawler technology, and the data related to the lifestyle management of the diabetic is crawled from a diet exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guide, and the obtained data related to the lifestyle management of the diabetic is entity identified and encoded by natural language technology.
Preferably, in step S2, assuming the null knowledge-map N, the node I having an initial value of 0 and the connecting edge J having an initial value of 0, for each diabetic:
s21, extracting characteristic points of the diabetic;
s22, judging whether the feature point exists in the knowledge graph N, if so, repeating the step S21, otherwise, entering the step S23;
s23, adding the feature point to the knowledge graph N as an independent node, and repeating step S21, where I is I + 1;
s24, extracting the relationship between two different characteristic points of the diabetic;
s25, judging whether two different feature points have a relation, if so, entering a step S26, otherwise, entering a step S27;
s26, constructing connecting edges between the nodes corresponding to the two different feature points, and repeating the step S24;
s27, no connecting edge is constructed between the nodes corresponding to the two different feature points, and the step S24 is repeated.
The invention also provides an automatic recommendation method for the lifestyle management path of the diabetic, which is characterized by comprising the following steps:
s1, acquiring data related to the diabetes patient life style management to form multi-source heterogeneous data, wherein the multi-source heterogeneous data comprises pathological data and life style management data of the diabetes patient, and the life style management data comprises diet data and exercise data;
s2, constructing a multi-modal knowledge graph for the life style management of the diabetic patient based on multi-source heterogeneous data: extracting different feature points of the diabetic patient and the relation between the different feature points from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in a knowledge graph, the feature points with the superior and inferior conceptual relation belong to the nodes corresponding to the feature points with the superior concept, and the connection edge between the two corresponding nodes is constructed based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
s3, performing relation completion on the multi-modal knowledge graph: calculating the link similarity between the nodes in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without an edge relation in the multi-modal knowledge graph, and constructing an edge between the two nodes without the edge relation when the link similarity reaches a set value;
s4, acquiring pathological information and life style management target information of a certain diabetic, and finding out a node corresponding to the pathological information in the multi-modal knowledge map as a starting point and a node corresponding to the life style management target information as an end point;
s5, judging whether the relation value between the starting point and any one of other nodes with the connection edge relation in the multi-modal knowledge graph is larger than a first set value or not, and if so, adding the node into the starting point set;
s6, judging whether the relation value between the end point and any one of other nodes with connecting edge relation in the multi-modal knowledge graph is larger than a second set value or not, and if so, adding the node into the end point set;
s7, merging the starting point set and the end point set, and removing nodes which do not appear in the merged set in the multi-modal knowledge graph;
and S8, calculating the sum of the joint relation values of the starting point and the end point in the multi-modal knowledge graph after the removing operation, which pass through other nodes except the starting point and the end point, and extracting the part with the highest sum of the joint relation values in the multi-modal knowledge graph to be used as the recommended optimal path for the life style management of the diabetic patient.
Preferably, in step S1, the data related to the lifestyle management of the diabetic is crawled through web crawler technology, and the data related to the lifestyle management of the diabetic is crawled from a diet exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guide, and the obtained data related to the lifestyle management of the diabetic is entity identified and encoded by natural language technology.
The invention also provides an automatic recommendation system for the lifestyle management path of the diabetic, which is characterized by comprising an acquisition module, a construction module, a completion module, a training module and a prediction module;
the acquisition module is used for acquiring data related to the life style management of the diabetic patient to form multi-source heterogeneous data, the multi-source heterogeneous data comprises pathological data and life style management data of the diabetic patient, and the life style management data comprises diet data and motion data;
the building module is used for building a multi-modal knowledge graph for managing the life style of the diabetic patient based on multi-source heterogeneous data, extracting the relation between different feature points and different feature points of the diabetic patient from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in the knowledge graph, the feature points with the upper and lower concept relation belong to the nodes corresponding to the feature points with the upper concept, and the connecting edges between the two corresponding nodes are built based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
the completion module is used for performing relationship completion on the multi-modal knowledge graph, calculating the link similarity between node pairs in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without a connection relationship in the multi-modal knowledge graph, and constructing a connection edge between the two nodes without the connection relationship when the link similarity reaches a set value;
the training module is used for inputting the semantic information of the multi-modal knowledge map after the relationship completion into the deep learning model for training and learning;
the prediction module is used for constructing a sub-knowledge map based on pathological information and target life style information of a certain diabetic, inputting semantic information of the sub-knowledge map into a trained deep learning model for map similarity matching prediction, and extracting the part with the highest similarity with the sub-knowledge map in the multi-mode knowledge map to serve as a recommended life style management path of the diabetic.
Preferably, the acquiring module is configured to crawl data related to lifestyle management of the diabetic through a web crawler technology, capture the data related to lifestyle management of the diabetic from a dietary exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guide, and perform entity identification and encoding on each acquired data related to lifestyle management of the diabetic by using a natural language technology.
Preferably, the building module comprises a first extraction submodule, a first judgment submodule, an addition submodule, a second extraction submodule, a second judgment submodule and a building submodule;
setting an empty knowledge graph N, setting the initial value of a node I to be 0, and setting the initial value of a connecting edge J to be 0, aiming at each diabetic:
the first extraction submodule is used for extracting the characteristic points of the diabetic patient;
the first judgment submodule is used for judging whether the characteristic point exists in the knowledge graph N or not, if so, the first extraction submodule is repeatedly called, and if not, the adding submodule is called;
the adding submodule is used for adding the characteristic point into the knowledge graph N to serve as an independent node, and the first extracting submodule is repeatedly called, wherein I is I + 1;
the second extraction submodule is used for extracting the relation between two different feature points of the diabetic patient;
the second judgment submodule is used for judging whether two different feature points have a relationship, and if so, the construction submodule is called;
the construction submodule is used for constructing a connecting edge between nodes corresponding to two different feature points and repeatedly calling the second extraction submodule.
The invention also provides an automatic recommendation system for the lifestyle management path of the diabetic, which is characterized by comprising a first acquisition module, a construction module, a completion module, a second acquisition module, a first judgment module, a second judgment module, a processing module and a recommendation module;
the first acquisition module is used for acquiring data related to life style management of the diabetic patient to form multi-source heterogeneous data, the multi-source heterogeneous data comprises pathological data and life style management data of the diabetic patient, and the life style management data comprises diet data and motion data;
the building module is used for building a multi-modal knowledge graph for managing the life style of the diabetic patient based on multi-source heterogeneous data, extracting the relation between different feature points and different feature points of the diabetic patient from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in the knowledge graph, the feature points with the upper and lower concept relation belong to the nodes corresponding to the feature points with the upper concept, and the connecting edges between the two corresponding nodes are built based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
the completion module is used for performing relationship completion on the multi-modal knowledge graph, calculating the link similarity between node pairs in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without a connection relationship in the multi-modal knowledge graph, and constructing a connection edge between the two nodes without the connection relationship when the link similarity reaches a set value;
the second acquisition module is used for acquiring pathological information and life style management target information of a certain diabetic patient, and finding out a node corresponding to the pathological information in the multi-mode knowledge map as a starting point and a node corresponding to the life style management target information as an end point;
the first judging module is used for judging whether a relation value between the starting point and any one of other nodes with a connecting edge relation in the multi-modal knowledge graph is larger than a first set value or not, and if so, adding the node into the starting point set;
the second judging module is used for judging whether a relation value between the endpoint and any node in other nodes with connection edge relations in the multi-modal knowledge graph is larger than a second set value or not, and if so, adding the node into the endpoint set;
the processing module is used for merging the starting point set and the end point set, and removing nodes which do not appear in the merged set in the multi-modal knowledge graph;
the recommending module is used for calculating the sum of the connection relationship values of other nodes except the starting point and the end point between the starting point and the end point in the multi-modal knowledge map after the removing operation, and extracting the part with the highest sum of the connection relationship values in the multi-modal knowledge map to be used as the recommended optimal path for the life style management of the diabetic patient.
Preferably, the acquiring module is configured to crawl data related to lifestyle management of the diabetic through a web crawler technology, capture the data related to lifestyle management of the diabetic from a dietary exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guide, and perform entity identification and encoding on each acquired data related to lifestyle management of the diabetic by using a natural language technology.
On the basis of the common knowledge in the field, the above preferred conditions can be combined randomly to obtain the preferred embodiments of the invention.
The positive progress effects of the invention are as follows:
the method comprises the steps of obtaining multi-source heterogeneous data of the diabetes lifestyle management, constructing a multi-mode knowledge map of the diabetes lifestyle management, dynamically increasing the multi-mode knowledge map and completing relations, realizing link prediction by restarting a random walk model, introducing characteristic discovery and matching of attention model depth map learning, and realizing personalized automatic recommendation of diabetes lifestyle management paths according to lifestyle habits of users and set diabetes management targets.
Drawings
Fig. 1 is a flowchart of a method for automatically recommending lifestyle management routes for diabetic patients according to embodiment 1 of the present invention.
Fig. 2 is a block diagram of a system for automatically recommending a lifestyle management route for a diabetic according to embodiment 1 of the present invention.
Fig. 3 is a flowchart of a method for automatically recommending lifestyle management routes for diabetic patients according to embodiment 2 of the present invention.
Fig. 4 is a block diagram of a system for automatically recommending a lifestyle management route for a diabetic according to embodiment 2 of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, are within the scope of the present invention.
Example 1
As shown in fig. 1, the present embodiment provides an automatic recommendation method for lifestyle management routes of diabetic patients, which includes the following steps:
In step 101, data related to lifestyle management of the diabetic patient is crawled through a web crawler technology, the data related to lifestyle management of the diabetic patient is captured from a diet exercise management database, a diabetic patient communication forum, an existing diabetic literature database and a diabetic patient lifestyle management guide, and entity identification and encoding are performed on each acquired data related to lifestyle management of the diabetic patient by using a natural language technology.
102, constructing a multi-modal knowledge graph for the lifestyle management of the diabetic patient based on multi-source heterogeneous data: extracting different feature points of the diabetic patient and the relation between the different feature points from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in a knowledge graph, the feature points with the superior and inferior conceptual relation belong to the nodes corresponding to the feature points with the superior concept, and the connecting edge between the two corresponding nodes is constructed based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data.
For example: the pathological characteristic points comprise types of diabetes patients, blood sugar values, blood fat values and the like, and the life style management characteristic points comprise diet characteristic points, motion characteristic points and the like. Extracting different pathological feature points (such as type feature points, blood sugar value feature points and blood fat value feature points of the diabetes patient), different diet feature points (porridge feature points, steamed bun feature points and steamed bun feature points) and different movement feature points (running feature points and slow walking feature points) of the diabetes patient from multi-source heterogeneous data, wherein meat buns and three-piece buns belong to the steamed stuffed buns, have upper and lower conceptual relations and belong to the steamed bun feature points, and different feature points are respectively used as independent nodes in a knowledge graph. If the diabetes patient eats porridge and steamed stuffed buns in the morning, the porridge characteristic points and the steamed stuffed buns characteristic points have a relationship, and the connecting edges between the two nodes corresponding to the porridge characteristic points and the steamed stuffed buns characteristic points are constructed.
In step 102, a null knowledge map N is set, the initial value of the node I is 0, the initial value of the connecting edge J is 0, and for each diabetic patient:
s21, extracting characteristic points of the diabetic;
s22, judging whether the feature point exists in the knowledge graph N, if so, repeating the step S21, otherwise, entering the step S23;
s23, adding the feature point to the knowledge graph N as an independent node, and repeating step S21, where I is I + 1;
s24, extracting the relationship between two different characteristic points of the diabetic;
s25, judging whether two different feature points have a relation, if so, entering a step S26, otherwise, entering a step S27;
s26, constructing connecting edges between the nodes corresponding to the two different feature points, and repeating the step S24;
s27, no connecting edge is constructed between the nodes corresponding to the two different feature points, and the step S24 is repeated.
103, performing relation completion on the multi-modal knowledge graph: and calculating the link similarity between the nodes in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without an edge relation in the multi-modal knowledge graph, and constructing an edge between the two nodes without the edge relation when the link similarity reaches a set value.
And 104, inputting the semantic information of the multi-modal knowledge graph after the relationship is completed into a deep learning model (attention model) for training and learning.
And 105, constructing a sub-knowledge graph based on the pathological information and the target life style information of a certain diabetic, inputting the semantic information of the sub-knowledge graph into a trained deep learning model for graph similarity matching prediction, and extracting the part with the highest similarity with the sub-knowledge graph in the multi-mode knowledge graph to serve as the recommended life style management path of the diabetic.
For example: the pathological information of a certain diabetic patient is information such as diabetes type, blood sugar and the like, the target life style information is information such as multi-diet and multi-exercise life style hoped to be adopted, a sub-knowledge map is constructed based on the information, the semantic information of the sub-knowledge map is input into a trained deep learning model for map similarity matching prediction, and the part with the highest similarity with the sub-knowledge map in the multi-modal knowledge map is extracted to be used as a recommended life style management path of the diabetic patient.
As shown in fig. 2, the embodiment further provides an automatic recommendation system for lifestyle management routes of diabetic patients, which includes an obtaining module 11, a building module 12, a completion module 13, a training module 14, and a prediction module 15.
The acquisition module 11 is used for acquiring data related to the lifestyle management of the diabetic patient to form multi-source heterogeneous data, wherein the multi-source heterogeneous data comprises pathological data and lifestyle management data of the diabetic patient, and the lifestyle management data comprises diet data and exercise data.
The acquiring module 11 is configured to crawl data related to lifestyle management of the diabetic through a web crawler technology, capture the data related to lifestyle management of the diabetic from a diet exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guide, and perform entity identification and encoding on each acquired data related to lifestyle management of the diabetic by using a natural language technology.
The building module 12 is configured to build a multi-modal knowledge graph for managing lifestyle of a diabetic patient based on multi-source heterogeneous data, extract relationships between different feature points and different feature points of the diabetic patient from the multi-source heterogeneous data, where each different feature point is used as an independent node in the knowledge graph, the feature points having a top-bottom conceptual relationship belong to nodes corresponding to feature points having a top concept, and a connecting edge between two corresponding nodes is built based on the relationships between different feature points, where the feature points include pathological feature points in pathological data and lifestyle management feature points in lifestyle management data.
The construction module 12 comprises a first extraction submodule, a first judgment submodule, an addition submodule, a second extraction submodule, a second judgment submodule and a construction submodule;
setting an empty knowledge graph N, setting the initial value of a node I to be 0, and setting the initial value of a connecting edge J to be 0, aiming at each diabetic:
the first extraction submodule is used for extracting the characteristic points of the diabetic patient;
the first judgment submodule is used for judging whether the characteristic point exists in the knowledge graph N or not, if so, the first extraction submodule is repeatedly called, and if not, the adding submodule is called;
the adding submodule is used for adding the characteristic point into the knowledge graph N to serve as an independent node, and the first extracting submodule is repeatedly called, wherein I is I + 1;
the second extraction submodule is used for extracting the relation between two different feature points of the diabetic patient;
the second judgment submodule is used for judging whether two different feature points have a relationship, and if so, the construction submodule is called;
the construction submodule is used for constructing a connecting edge between nodes corresponding to two different feature points and repeatedly calling the second extraction submodule.
The completion module 13 is configured to perform relationship completion on the multimodal knowledge graph, calculate link similarity between node pairs in the multimodal knowledge graph by using the restart random walk model based on the feature points of the nodes, thereby calculating link similarity between two nodes without a connection relationship in the multimodal knowledge graph, and construct a connection edge between two nodes without a connection relationship when the link similarity reaches a set value.
The training module 14 is configured to input the semantic information of the multi-modal knowledge graph after the relationship completion into the deep learning model for training and learning.
The prediction module 15 is configured to construct a sub-knowledge graph based on pathological information and target lifestyle information of a certain diabetic, input semantic information of the sub-knowledge graph into a trained deep learning model to perform graph similarity matching prediction, and extract a part of the multi-modal knowledge graph with the highest similarity to the sub-knowledge graph to serve as a recommended lifestyle management path of the diabetic.
Example 2
As shown in fig. 3, the present embodiment provides an automatic recommendation method for lifestyle management routes of diabetic patients, which includes the following steps:
In step 201, data related to lifestyle management of the diabetic patient is crawled through a web crawler technology, and the data related to lifestyle management of the diabetic patient is captured from a diet exercise management database, a diabetic patient communication forum, an existing diabetic literature database and a diabetic patient lifestyle management guide, and entity identification and encoding are performed on each acquired data related to lifestyle management of the diabetic patient by using a natural language technology.
And step 204, acquiring pathological information and life style management target information of a certain diabetic, and finding out a node corresponding to the pathological information in the multi-modal knowledge graph as a starting point and a node corresponding to the life style management target information as an end point.
For example: if the pathological information of a certain diabetic is blood glucose high information and the lifestyle management target information is blood glucose value reduction to target value information, the node corresponding to the blood glucose high information in the multimodal knowledge map is found as a starting point and the node corresponding to the blood glucose value reduction to the target value information in the multimodal knowledge map is found as an end point based on the blood glucose high information.
And step 205, judging whether a relation value between the starting point and any one of other nodes with a connection edge relation in the multi-modal knowledge graph is larger than a first set value, and if so, adding the node into the starting point set.
And step 206, judging whether the relation value between the end point and any one of other nodes with connecting edge relation in the multi-modal knowledge graph is larger than a second set value, and if so, adding the node into the end point set.
And step 207, merging the starting point set and the end point set, and removing nodes which do not appear in the merged set in the multi-modal knowledge graph.
And step 208, calculating the sum of the values of the continuous edge relationship of the nodes with the continuous edge relationship except the starting point and the end point between the starting point and the end point in the multi-modal knowledge graph after the removing operation, and extracting the part with the highest sum of the values of the continuous edge relationship in the multi-modal knowledge graph to be used as the recommended optimal path for the life style management of the diabetic patient.
For example: and concentrating the starting point to be A and the end point to be D, and further storing a node B and a node C, wherein the starting point A and the end point D both have a connecting edge relationship with the node B and the node C, calculating the sum of a connecting edge relationship value between the starting point and the node B and a connecting edge relationship value between the node B and the end point D, calculating the sum of a connecting edge relationship value between the starting point and the node C and a connecting edge relationship value between the node C and the end point D, and finding out the part with the highest sum value from the sum value to extract the part as the recommended life style management optimal path of the diabetic patient.
As shown in fig. 4, the present embodiment further provides an automatic recommendation system for lifestyle management routes of diabetic patients, which includes a first obtaining module 21, a building module 22, a completing module 23, a second obtaining module 24, a first determining module 25, a second determining module 26, a processing module 27, and a recommending module 28.
The first obtaining module 21 is configured to obtain data related to lifestyle management of a diabetic patient to form multi-source heterogeneous data, where the multi-source heterogeneous data includes pathological data and lifestyle management data of the diabetic patient, and the lifestyle management data includes diet data and exercise data.
The building module 22 is configured to build a multi-modal knowledge graph for managing the lifestyle of the diabetic patient based on multi-source heterogeneous data, extract relationships between different feature points and different feature points of the diabetic patient from the multi-source heterogeneous data, where each different feature point is used as an independent node in the knowledge graph, the feature points having a top-bottom conceptual relationship belong to nodes corresponding to feature points having a top concept, and a connecting edge between two corresponding nodes is built based on the relationships between different feature points, where the feature points include pathological feature points in pathological data and lifestyle management feature points in lifestyle management data.
The completion module 23 is configured to perform relationship completion on the multimodal knowledge graph, calculate a link similarity between node pairs in the multimodal knowledge graph by using the restart random walk model based on the feature points of the nodes, thereby calculating a link similarity between two nodes without a connection relationship in the multimodal knowledge graph, and construct a connection edge between two nodes without a connection relationship when the link similarity reaches a set value.
The second obtaining module 24 is configured to obtain pathological information and lifestyle management target information of a certain diabetic patient, and find a node corresponding to the pathological information in the multimodal knowledge graph as a starting point and a node corresponding to the lifestyle management target information as an ending point.
The first judging module 25 is configured to judge whether a relationship value between the starting point and any node in other nodes having a connection edge relationship in the multimodal knowledge graph is greater than a first set value, and if so, add the node to the starting point set.
The second determining module 26 is configured to determine whether a relationship value between the endpoint and any node in other nodes having a connection relationship in the multimodal knowledge graph is greater than a second set value, and if so, add the node to the endpoint set.
The processing module 27 is configured to perform union processing on the starting point set and the end point set, and remove nodes that do not appear in the union set in the multimodal knowledge graph.
The recommending module 28 is configured to calculate a sum of the link relationship values between the starting point and the end point in the multi-modal knowledge graph after the removing operation and passing through other nodes except the starting point and the end point, and extract a portion of the multi-modal knowledge graph with the highest sum of the link relationship values to serve as a recommended optimal path for the lifestyle management of the diabetic patient.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that these are by way of example only, and that the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the spirit and scope of the invention, and these changes and modifications are within the scope of the invention.
Claims (10)
1. A method for automatically recommending a lifestyle management path of a diabetic patient is characterized by comprising the following steps:
s1, acquiring data related to the diabetes patient life style management to form multi-source heterogeneous data, wherein the multi-source heterogeneous data comprises pathological data and life style management data of the diabetes patient, and the life style management data comprises diet data and exercise data;
s2, constructing a multi-modal knowledge graph for the life style management of the diabetic patient based on multi-source heterogeneous data: extracting different feature points of the diabetic patient and the relation between the different feature points from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in a knowledge graph, the feature points with the superior and inferior conceptual relation belong to the nodes corresponding to the feature points with the superior concept, and the connection edge between the two corresponding nodes is constructed based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
s3, performing relation completion on the multi-modal knowledge graph: calculating the link similarity between the nodes in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without an edge relation in the multi-modal knowledge graph, and constructing an edge between the two nodes without the edge relation when the link similarity reaches a set value;
s4, inputting the semantic information of the multi-modal knowledge map after the relationship is completed into a deep learning model for training and learning;
s5, constructing a sub-knowledge graph based on pathological information and target life style information of a certain diabetic, inputting semantic information of the sub-knowledge graph into a trained deep learning model for graph similarity matching prediction, and extracting the part with the highest similarity with the sub-knowledge graph in the multi-mode knowledge graph to serve as a recommended life style management path of the diabetic.
2. The automated diabetic lifestyle management pathway recommendation method according to claim 1, wherein in step S1, data related to diabetic lifestyle management is crawled through web crawler technology, and data related to diabetic lifestyle management is crawled from a diet exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guideline, and each of the obtained data related to diabetic lifestyle management is entity-identified and encoded by natural language technology.
3. The automated diabetic lifestyle management pathway recommendation method of claim 1 wherein, in step S2, assuming an empty knowledge map N, node I having an initial value of 0, and a connecting edge J having an initial value of 0, for each diabetic:
s21, extracting characteristic points of the diabetic;
s22, judging whether the feature point exists in the knowledge graph N, if so, repeating the step S21, otherwise, entering the step S23;
s23, adding the feature point to the knowledge graph N as an independent node, and repeating step S21, where I is I + 1;
s24, extracting the relationship between two different characteristic points of the diabetic;
s25, judging whether two different feature points have a relation, if so, entering a step S26, otherwise, entering a step S27;
s26, constructing connecting edges between the nodes corresponding to the two different feature points, and repeating the step S24;
s27, no connecting edge is constructed between the nodes corresponding to the two different feature points, and the step S24 is repeated.
4. A method for automatically recommending a lifestyle management path of a diabetic patient is characterized by comprising the following steps:
s1, acquiring data related to the diabetes patient life style management to form multi-source heterogeneous data, wherein the multi-source heterogeneous data comprises pathological data and life style management data of the diabetes patient, and the life style management data comprises diet data and exercise data;
s2, constructing a multi-modal knowledge graph for the life style management of the diabetic patient based on multi-source heterogeneous data: extracting different feature points of the diabetic patient and the relation between the different feature points from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in a knowledge graph, the feature points with the superior and inferior conceptual relation belong to the nodes corresponding to the feature points with the superior concept, and the connection edge between the two corresponding nodes is constructed based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
s3, performing relation completion on the multi-modal knowledge graph: calculating the link similarity between the nodes in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without an edge relation in the multi-modal knowledge graph, and constructing an edge between the two nodes without the edge relation when the link similarity reaches a set value;
s4, acquiring pathological information and life style management target information of a certain diabetic, and finding out a node corresponding to the pathological information in the multi-modal knowledge map as a starting point and a node corresponding to the life style management target information as an end point;
s5, judging whether the relation value between the starting point and any one of other nodes with the connection edge relation in the multi-modal knowledge graph is larger than a first set value or not, and if so, adding the node into the starting point set;
s6, judging whether the relation value between the end point and any one of other nodes with connecting edge relation in the multi-modal knowledge graph is larger than a second set value or not, and if so, adding the node into the end point set;
s7, merging the starting point set and the end point set, and removing nodes which do not appear in the merged set in the multi-modal knowledge graph;
and S8, calculating the sum of the joint relation values of the starting point and the end point in the multi-modal knowledge graph after the removing operation, which pass through other nodes except the starting point and the end point, and extracting the part with the highest sum of the joint relation values in the multi-modal knowledge graph to be used as the recommended optimal path for the life style management of the diabetic patient.
5. The automated diabetic lifestyle management pathway recommendation method according to claim 1, wherein in step S1, data related to diabetic lifestyle management is crawled through web crawler technology, and data related to diabetic lifestyle management is crawled from a diet exercise management database, a diabetic communication forum, an existing diabetic literature database, and a diabetic lifestyle management guideline, and each of the obtained data related to diabetic lifestyle management is entity-identified and encoded by natural language technology.
6. An automatic recommendation system for a lifestyle management path of a diabetic patient is characterized by comprising an acquisition module, a construction module, a completion module, a training module and a prediction module;
the acquisition module is used for acquiring data related to the life style management of the diabetic patient to form multi-source heterogeneous data, the multi-source heterogeneous data comprises pathological data and life style management data of the diabetic patient, and the life style management data comprises diet data and motion data;
the building module is used for building a multi-modal knowledge graph for managing the life style of the diabetic patient based on multi-source heterogeneous data, extracting the relation between different feature points and different feature points of the diabetic patient from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in the knowledge graph, the feature points with the upper and lower concept relation belong to the nodes corresponding to the feature points with the upper concept, and the connecting edges between the two corresponding nodes are built based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
the completion module is used for performing relationship completion on the multi-modal knowledge graph, calculating the link similarity between node pairs in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without a connection relationship in the multi-modal knowledge graph, and constructing a connection edge between the two nodes without the connection relationship when the link similarity reaches a set value;
the training module is used for inputting the semantic information of the multi-modal knowledge map after the relationship completion into the deep learning model for training and learning;
the prediction module is used for constructing a sub-knowledge map based on pathological information and target life style information of a certain diabetic, inputting semantic information of the sub-knowledge map into a trained deep learning model for map similarity matching prediction, and extracting the part with the highest similarity with the sub-knowledge map in the multi-mode knowledge map to serve as a recommended life style management path of the diabetic.
7. The automated diabetic lifestyle management pathway recommendation system of claim 6 wherein the retrieval module is configured to crawl data related to diabetic lifestyle management via web crawler technology, and to capture data related to diabetic lifestyle management from a dietary exercise management database, a diabetic communication forum, an existing diabetic literature database, a diabetic lifestyle management guideline, and to perform entity identification and encoding of each retrieved data related to diabetic lifestyle management using natural language technology.
8. The automated diabetic lifestyle management pathway recommendation system of claim 6 wherein the construction module comprises a first extraction sub-module, a first judgment sub-module, an addition sub-module, a second extraction sub-module, a second judgment sub-module, and a construction sub-module;
setting an empty knowledge graph N, setting the initial value of a node I to be 0, and setting the initial value of a connecting edge J to be 0, aiming at each diabetic:
the first extraction submodule is used for extracting the characteristic points of the diabetic patient;
the first judgment submodule is used for judging whether the characteristic point exists in the knowledge graph N or not, if so, the first extraction submodule is repeatedly called, and if not, the adding submodule is called;
the adding submodule is used for adding the characteristic point into the knowledge graph N to serve as an independent node, and the first extracting submodule is repeatedly called, wherein I is I + 1;
the second extraction submodule is used for extracting the relation between two different feature points of the diabetic patient;
the second judgment submodule is used for judging whether two different feature points have a relationship, and if so, the construction submodule is called;
the construction submodule is used for constructing a connecting edge between nodes corresponding to two different feature points and repeatedly calling the second extraction submodule.
9. An automatic recommendation system for a lifestyle management path of a diabetic patient is characterized by comprising a first acquisition module, a construction module, a completion module, a second acquisition module, a first judgment module, a second judgment module, a processing module and a recommendation module;
the first acquisition module is used for acquiring data related to life style management of the diabetic patient to form multi-source heterogeneous data, the multi-source heterogeneous data comprises pathological data and life style management data of the diabetic patient, and the life style management data comprises diet data and motion data;
the building module is used for building a multi-modal knowledge graph for managing the life style of the diabetic patient based on multi-source heterogeneous data, extracting the relation between different feature points and different feature points of the diabetic patient from the multi-source heterogeneous data, wherein each different feature point is used as an independent node in the knowledge graph, the feature points with the upper and lower concept relation belong to the nodes corresponding to the feature points with the upper concept, and the connecting edges between the two corresponding nodes are built based on the relation between the different feature points, wherein the feature points comprise pathological feature points in pathological data and life style management feature points in life style management data;
the completion module is used for performing relationship completion on the multi-modal knowledge graph, calculating the link similarity between node pairs in the multi-modal knowledge graph by using the restarting random walk model based on the characteristic points of the nodes, thereby calculating the link similarity between two nodes without a connection relationship in the multi-modal knowledge graph, and constructing a connection edge between the two nodes without the connection relationship when the link similarity reaches a set value;
the second acquisition module is used for acquiring pathological information and life style management target information of a certain diabetic patient, and finding out a node corresponding to the pathological information in the multi-mode knowledge map as a starting point and a node corresponding to the life style management target information as an end point;
the first judging module is used for judging whether a relation value between the starting point and any one of other nodes with a connecting edge relation in the multi-modal knowledge graph is larger than a first set value or not, and if so, adding the node into the starting point set;
the second judging module is used for judging whether a relation value between the endpoint and any node in other nodes with connection edge relations in the multi-modal knowledge graph is larger than a second set value or not, and if so, adding the node into the endpoint set;
the processing module is used for merging the starting point set and the end point set, and removing nodes which do not appear in the merged set in the multi-modal knowledge graph;
the recommending module is used for calculating the sum of the connection relationship values of other nodes except the starting point and the end point between the starting point and the end point in the multi-modal knowledge map after the removing operation, and extracting the part with the highest sum of the connection relationship values in the multi-modal knowledge map to be used as the recommended optimal path for the life style management of the diabetic patient.
10. The automated diabetic lifestyle management pathway recommendation system of claim 6 wherein the retrieval module is configured to crawl data related to diabetic lifestyle management via web crawler technology, and to capture data related to diabetic lifestyle management from a dietary exercise management database, a diabetic communication forum, an existing diabetic literature database, a diabetic lifestyle management guideline, and to perform entity identification and encoding of each retrieved data related to diabetic lifestyle management using natural language technology.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110451694.0A CN113035319A (en) | 2021-04-26 | 2021-04-26 | Automatic recommendation method and system for lifestyle management path of diabetic patient |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110451694.0A CN113035319A (en) | 2021-04-26 | 2021-04-26 | Automatic recommendation method and system for lifestyle management path of diabetic patient |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113035319A true CN113035319A (en) | 2021-06-25 |
Family
ID=76454563
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110451694.0A Withdrawn CN113035319A (en) | 2021-04-26 | 2021-04-26 | Automatic recommendation method and system for lifestyle management path of diabetic patient |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113035319A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113377967A (en) * | 2021-08-12 | 2021-09-10 | 明品云(北京)数据科技有限公司 | Target scheme acquisition method and system, electronic equipment and medium |
| CN114023418A (en) * | 2022-01-06 | 2022-02-08 | 苏州百孝医疗科技有限公司 | Insulin recommendation method and device and system for monitoring blood sugar level |
| CN117558410A (en) * | 2024-01-12 | 2024-02-13 | 吉林大学 | Artificial intelligence-based diet management system and method for diabetic nephropathy patients |
-
2021
- 2021-04-26 CN CN202110451694.0A patent/CN113035319A/en not_active Withdrawn
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113377967A (en) * | 2021-08-12 | 2021-09-10 | 明品云(北京)数据科技有限公司 | Target scheme acquisition method and system, electronic equipment and medium |
| CN113377967B (en) * | 2021-08-12 | 2021-12-21 | 明品云(北京)数据科技有限公司 | Target scheme acquisition method and system, electronic equipment and medium |
| CN114023418A (en) * | 2022-01-06 | 2022-02-08 | 苏州百孝医疗科技有限公司 | Insulin recommendation method and device and system for monitoring blood sugar level |
| CN114023418B (en) * | 2022-01-06 | 2022-04-01 | 苏州百孝医疗科技有限公司 | Insulin recommendation method and device and system for monitoring blood sugar level |
| CN117558410A (en) * | 2024-01-12 | 2024-02-13 | 吉林大学 | Artificial intelligence-based diet management system and method for diabetic nephropathy patients |
| CN117558410B (en) * | 2024-01-12 | 2024-03-22 | 吉林大学 | Artificial intelligence-based diet management system and method for diabetic nephropathy patients |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Rostami et al. | A novel time-aware food recommender-system based on deep learning and graph clustering | |
| CN113035319A (en) | Automatic recommendation method and system for lifestyle management path of diabetic patient | |
| US20170249434A1 (en) | Multi-format, multi-domain and multi-algorithm metalearner system and method for monitoring human health, and deriving health status and trajectory | |
| KR20200038628A (en) | Apparatus and method for providing personalized medication information | |
| KR102217307B1 (en) | Machine Learning and Semantic Knowledge-based Big Data Analysis: A Novel Healthcare Monitoring Method and Apparatus Using Wearable Sensors and Social Networking Data | |
| CN108511057A (en) | Transfusion volume model foundation and prediction technique, device, equipment and its storage medium | |
| CN104318351A (en) | Traditional Chinese medicine health management system and method | |
| CN107358014A (en) | The clinical pre-treating method and system of a kind of physiological data | |
| CN114072782A (en) | Question recommendation method, device and system, electronic equipment and readable storage medium | |
| WO2021158796A1 (en) | Forecasting and explaining user health metrics | |
| CN115116612A (en) | Intelligent risk assessment system and method for child patient state of illness | |
| KR20200113954A (en) | System and method for providing user-customized health information service | |
| KR20210052122A (en) | System and method for providing user-customized food information service | |
| Kaur et al. | Food image-based nutritional management system to overcome polycystic Ovary Syndrome using DeepLearning: A systematic review | |
| Hassan et al. | Early predictive analytics in healthcare for diabetes prediction using machine learning approach | |
| Habib et al. | A neutrosophic clinical decision-making system for cardiovascular diseases risk analysis | |
| Chou et al. | Extracting drug utilization knowledge using self-organizing map and rough set theory | |
| KR102360651B1 (en) | System for providing personalized meal plan service for improving medical test result using medical record | |
| Kasemthaweesab et al. | Association analysis of diabetes mellitus (DM) with complication states based on association rules | |
| CN117672451A (en) | Medicine recommendation method for type 2 diabetics | |
| CN116721730B (en) | Whole-course patient management system based on digital therapy | |
| Bhattacharyya et al. | Delirium prediction in the ICU: designing a screening tool for preventive interventions | |
| JP2009031900A (en) | Medical checkup data processor | |
| CN117637139A (en) | Machine learning-based type 2diabetes auxiliary decision making system | |
| Abu-Issa et al. | Design and implementation of proactive multi-type context-aware recommender system for patients suffering diabetes |
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 | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20210625 |
|
| WW01 | Invention patent application withdrawn after publication |