Exercise recommendation method, apparatus, device, and medium

Abstract

The application discloses a sports recommendation method, device, equipment and medium, and the method comprises the following steps: acquiring basic information of a user; querying a pre-constructed sports knowledge graph according to the basic information of the user to obtain a target sports mode list and an evaluation attribute list corresponding to the target sports mode list; determining at least one sports scheme according to the target sports mode list and the evaluation attribute list, wherein the sports scheme comprises at least one sports mode and a recommended sports duration corresponding to the sports mode; performing sorting processing on the at least one sports scheme according to the evaluation attribute list to obtain a sorting result; and recommending the at least one sports scheme from the sorting result. The technical scheme provided in the embodiments of the application can effectively improve the accuracy of the sports recommendation result by querying the pre-constructed sports knowledge graph to obtain a sports mode list and performing optimization processing on the sports mode list.

Description

Technical Field

[0001] This application relates to the technical field of big data processing, and in particular to methods, apparatus, devices and media for motion recommendation. Background Technology

[0002] Sports medicine is receiving increasing attention and importance, especially for people with chronic diseases and those in a sub-healthy state. Regular and scientific exercise helps maintain health and improve quality of life. Scientific research has confirmed that exercise has a positive effect on people with chronic diseases such as diabetes, hypertension, hyperlipidemia, and obesity, helping to improve glucose metabolism, regulate blood pressure, improve blood lipids, and control weight. Summary of the Invention

[0003] Some embodiments of this application provide a method for recommending exercise, the method comprising:

[0004] Obtain basic user information;

[0005] Based on the user's basic information, a pre-constructed motion knowledge graph is queried to obtain a list of target motion methods and a list of evaluation attributes corresponding to the list of target motion methods.

[0006] At least one exercise plan is determined based on the target exercise mode list and the evaluation attribute list, wherein the exercise plan includes at least one exercise mode and a recommended exercise duration corresponding to the exercise mode;

[0007] The at least one motion scheme is sorted according to the evaluation attribute list to obtain the sorting result;

[0008] Based on the ranking results, at least one exercise program should be recommended.

[0009] In some embodiments, a pre-constructed motion knowledge graph is queried based on the user's basic information to obtain a list of target motion methods and a list of evaluation attributes corresponding to the target motion method list, including:

[0010] Construct at least one query statement based on the user's basic information;

[0011] The motion knowledge graph is queried according to each of the at least one query statement to obtain the target motion mode list and the evaluation attribute list corresponding to the target motion mode list.

[0012] In some embodiments, the basic information includes: the user's target disease type, target user attributes, and the user's physical activity level assessment results. Constructing at least one query statement based on the user's basic information includes:

[0013] Obtain the user's physical activity level assessment results;

[0014] When the user's physical activity level assessment result is high, the target user attribute and the target symptom type are used as query conditions to construct a query statement; or...

[0015] When the user's physical activity level assessment result is medium or below, the target user attribute and the target disease type are used as query conditions to construct a first query statement;

[0016] The second query statement is constructed by using the user's physical activity level assessment result as the query condition.

[0017] In some embodiments, the sports knowledge graph includes a first relationship between user attribute entities and sports type entities, and a second relationship between sports type entities and disease type entities. Then, querying the sports knowledge graph according to each of the at least one query statement includes:

[0018] When the user's physical activity level assessment result is high, the first and second relationships in the exercise knowledge graph are queried according to the first query statement to obtain a list of target exercise methods; or...

[0019] When the user's physical activity level assessment result is intermediate or below, the first and second relationships in the exercise knowledge graph are queried according to the first query statement to obtain an initial list of exercise methods.

[0020] Based on the second query statement, the exercise knowledge graph is queried to obtain the exercise methods in the initial exercise method list that meet the user's physical activity level assessment results, which are then used as the target exercise method list.

[0021] Obtain the list of evaluation attributes corresponding to the target motion mode list.

[0022] In some embodiments, determining at least one motion scheme based on the target motion mode list and the evaluation attribute list includes:

[0023] Obtain the user's target calorie consumption;

[0024] The motion combination result is obtained by combining at least one motion mode included in the target motion mode list.

[0025] Based on the exercise combination results, extract the evaluation attributes corresponding to the exercise combination results from the evaluation attribute list;

[0026] Based on the evaluation attributes and the target calorie consumption, a recommended time combination is determined corresponding to the exercise combination result. The recommended time combination includes the recommended exercise duration corresponding to each exercise mode in the exercise combination result.

[0027] The exercise combination result and the recommended time combination corresponding to the exercise combination result are used as the exercise plan.

[0028] In some embodiments, the evaluation attribute includes the calories burned per unit of energy corresponding to the exercise mode. Then, determining the recommended time combination corresponding to the exercise mode combination based on the evaluation attribute and the target calories burned includes:

[0029] Based on the fact that the sum of the products of the unit calorie consumption and the recommended duration for each exercise mode included in the exercise combination result equals the target calorie consumption, the recommended time combination corresponding to the exercise mode combination is determined.

[0030] In some embodiments, determining at least one motion scheme based on the target motion mode list and the evaluation attribute list includes:

[0031] Obtain the user's exercise frequency within the exercise plan cycle;

[0032] The motion combination result is obtained by combining at least one motion mode included in the target motion mode list.

[0033] Determine the maximum and minimum activity levels corresponding to the exercise plan cycle;

[0034] A time array is determined based on the maximum and minimum activity levels, the time array including the exercise frequency and the recommended exercise duration corresponding to the exercise frequency;

[0035] The time array is allocated to the exercise combination result according to the exercise frequency within the exercise planning cycle.

[0036] In some embodiments, combining at least one type of motion included in the target motion mode list includes:

[0037] Based on the user's physical activity level assessment results, a preset number of exercise methods corresponding to the user's physical activity level assessment results are obtained from the target exercise method list, and these are used as the exercise combination results.

[0038] In some embodiments, the evaluation attribute list includes multiple evaluation attributes corresponding to the exercise mode, and sorting the at least one exercise scheme according to the evaluation attribute list includes:

[0039] The user's exercise risk level is determined based on the user's basic information;

[0040] The at least one exercise program is ranked according to the multiple evaluation attributes and exercise risk level attributes.

[0041] In some embodiments, the exercise program includes at least one exercise mode and a recommended exercise duration corresponding to each exercise mode, and the step of ranking the at least one exercise program according to the plurality of evaluation attributes and exercise risk level attributes includes:

[0042] Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode;

[0043] The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated.

[0044] The product is calculated based on the value of the risk level corresponding to each exercise mode and the weighting coefficient corresponding to the risk level.

[0045] The product result is used as an evaluation index for each movement mode in the at least one movement scheme;

[0046] The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode;

[0047] The at least one exercise program is ranked according to the comprehensive evaluation index.

[0048] In some embodiments, the exercise program includes at least one exercise mode and a recommended exercise duration corresponding to each exercise mode, and the step of ranking the at least one exercise program according to the plurality of evaluation attributes and exercise risk level attributes includes:

[0049] Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode;

[0050] The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated.

[0051] The product is calculated by combining the value of the risk level corresponding to each exercise mode with the weighting coefficient corresponding to the risk level.

[0052] The product is calculated based on the recommended exercise duration for each exercise mode and the weighting coefficient corresponding to the recommended exercise duration;

[0053] The product result is used as an evaluation index for each movement mode in the at least one movement scheme;

[0054] The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode;

[0055] The at least one exercise program is ranked according to the comprehensive evaluation index.

[0056] In some embodiments, prior to querying a pre-built motion knowledge graph based on the user's basic information, the method further includes:

[0057] The user's exercise risk level is determined based on the user's basic information;

[0058] The recommended exercise methods are determined based on the aforementioned exercise risk level.

[0059] In some embodiments, determining the user's exercise risk level based on the user's basic information includes:

[0060] The user's basic information is input into a pre-built sports risk classification model to obtain the user's sports risk level.

[0061] In some embodiments, the user's basic information includes the target disease type, target user attributes, and physical activity level assessment results. Obtaining the user's basic information includes:

[0062] The target disease type, target user attributes, and physical activity level assessment results are obtained through electronic questionnaires; or...

[0063] Obtain the user's target disease type and target user attributes through the user's physical examination data;

[0064] The results of the physical activity level assessment of users were obtained through electronic questionnaires.

[0065] In some embodiments, the electronic questionnaire method includes a human-computer interaction display interface or a voice dialogue method.

[0066] Some embodiments of this application also provide an exercise recommendation device, the device comprising:

[0067] The information acquisition unit is used to acquire the user's basic information.

[0068] The graph query unit is used to query a pre-built motion knowledge graph based on the user's basic information to obtain a list of target motion methods and a list of evaluation attributes corresponding to the list of target motion methods.

[0069] The scheme determination unit is used to determine at least one exercise scheme based on the target exercise mode list and the evaluation attribute list, wherein the exercise scheme includes at least one exercise mode and a recommended exercise duration corresponding to the exercise mode;

[0070] The scheme sorting unit is used to sort the at least one motion scheme according to the evaluation list attribute list to obtain a sorting result;

[0071] The scheme recommendation unit is used to recommend at least one motion scheme that ranks highly in the ranking results.

[0072] Some embodiments of this application provide an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor being used to implement the methods described in the embodiments of this application when executing the program.

[0073] In some embodiments, the electronic device further includes: an input device and an output device;

[0074] The input device is used to obtain basic user information;

[0075] The output device is used to recommend at least one motion scheme to the user.

[0076] Some embodiments of this application provide a computer-readable storage medium having a computer program stored thereon for implementing the methods described in the embodiments of this application. Attached Figure Description

[0077] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0078] Figure 1 This illustration shows an application scenario diagram of the exercise recommendation method provided in the embodiments of this application;

[0079] Figure 2 A flowchart illustrating the exercise recommendation method proposed in an embodiment of this application is shown;

[0080] Figure 3 A flowchart illustrating yet another exercise recommendation method provided in an embodiment of this application is shown;

[0081] Figure 4 A schematic diagram illustrating the process of creating a motion knowledge graph according to an embodiment of this application is shown;

[0082] Figure 5 A flowchart illustrating another exercise recommendation method provided in an embodiment of this application is shown;

[0083] Figure 6 This illustration shows a schematic diagram of the basic relationships between entity names stored in a graph database, as provided in an embodiment of this application.

[0084] Figure 7This illustration shows a schematic diagram of the basic relationships between entity names stored in a graph database, as provided in an embodiment of this application.

[0085] Figure 8 This illustration shows a schematic diagram of the basic relationships between entity names stored in a graph database, as provided in an embodiment of this application.

[0086] Figure 9 A flowchart illustrating the synonym processing method for motion names provided in an embodiment of this application is shown.

[0087] Figure 10 A flowchart illustrating the motion risk classification and processing method provided in an embodiment of this application is shown;

[0088] Figure 11 This illustration shows a schematic diagram of the questionnaire data collection method provided in an embodiment of this application;

[0089] Figure 12 A schematic diagram of the query interface provided in an embodiment of this application is shown;

[0090] Figure 13 A schematic diagram of the heat calculation interface provided in an embodiment of this application is shown;

[0091] Figure 14 A schematic diagram of the information collection interface provided in an embodiment of this application is shown;

[0092] Figure 15 A schematic diagram of the motion scheme display interface provided in an embodiment of this application is shown;

[0093] Figure 16 A schematic diagram of the weekly exercise plan display interface provided in an embodiment of this application is shown;

[0094] Figure 17 This illustration shows a schematic diagram of the interface for obtaining exercise recommendation results based on user physical examination data and questionnaire data, provided in an embodiment of this application.

[0095] Figure 18 A flowchart illustrating the presentation method of exercise recommendation results provided in an embodiment of this application is shown;

[0096] Figure 19 A schematic diagram of the structure of the exercise recommendation device provided in an embodiment of this application is shown;

[0097] Figure 20 A schematic diagram of the structure of a computer system suitable for implementing the terminal device or server of the present application is shown. Detailed Implementation

[0098] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the relevant disclosure and not intended to limit the scope of the disclosure. Furthermore, it should be noted that, for ease of description, only the parts relevant to the disclosure are shown in the accompanying drawings.

[0099] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0100] Please refer to Figure 1 , Figure 1 A schematic diagram illustrating an application scenario of the exercise recommendation method provided in this application embodiment is shown.

[0101] like Figure 1 As shown, users can use terminal device 101 to query suitable exercise recommendations. Users can obtain exercise recommendations through an exercise recommendation application pre-installed on the terminal device. The exercise recommendation application can be implemented in client mode or web page mode. The system obtains the query conditions input by the user and then sends the query conditions to server 103 via network 102. Query conditions may include, but are not limited to, user physical examination data, basic information input by the user, and physical activity level provided by the user through the human-computer interaction interface.

[0102] The aforementioned terminal device 101 can be a smartphone, tablet computer, smart glasses, smartwatch, other wearable devices, or other mobile devices, or an electronic device such as a desktop computer, but is not limited thereto. For example, the exercise recommendation method provided in this embodiment can also be executed on the terminal device 101. For example, the exercise recommendation method provided in this embodiment can be partially executed on the terminal device 101, and the other part can be sent to the server 103 for execution, that is, it can be jointly executed by the terminal device 101 and the server 103.

[0103] Server 103 executes a query program based on the query conditions sent by terminal device 101 to find motion recommendation results related to the query conditions. Server 103 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0104] The aforementioned network 102 includes, but is not limited to, wireless or wired networks, which use standard communication technologies and / or protocols. The network is typically the Internet, but can also be any network, including but not limited to any combination of Local Area Network (LAN), Metropolitan Area Network (MAN), Wide Area Network (WAN), mobile, wired or wireless networks, private networks, or virtual private networks.

[0105] In related technologies, exercise recommendations are made based on users' fitness needs data. However, this method doesn't comprehensively consider the user's actual physical condition, such as whether the user has a chronic disease. For patients with chronic diseases, appropriate exercise can have a positive effect, helping to improve glucose metabolism. Furthermore, some diseases have contraindications for certain types of exercise; for example, patients with epilepsy should avoid activities with inherent risks such as skiing and diving. Because this approach doesn't comprehensively consider the user's actual physical condition, the recommended exercise programs have limitations.

[0106] This application proposes a technical solution for an exercise recommendation method that comprehensively considers the correlation between exercise volume and disease characteristics, effectively improving the accuracy of exercise recommendation results.

[0107] Please refer to the following. Figure 2 , Figure 2 A flowchart illustrating the exercise recommendation method proposed in an embodiment of this application is shown. Figure 2 As shown, this method can be executed by an exercise recommendation device, which can be configured in a terminal device or a server. The method includes:

[0108] Step 201: Obtain the user's basic information.

[0109] For example, the basic information of the users mentioned above includes the target disease type, target user attributes, and physical activity level assessment results.

[0110] The aforementioned target disease types refer to disease keywords extracted from user physical examination data that correspond to the user's specific condition. For example, disease keywords include, but are not limited to, fasting blood glucose, postprandial blood glucose, liver fat content, fatty liver, dyslipidemia, and diabetes. Extraction methods include, but are not limited to, TF-IDF (Term Frequency-Inverse Document Frequency) algorithms, TextRank algorithms, word2vec algorithms, and BiLSTM-CRF algorithms based on BERT pre-trained models. For example, the BILSTM-CRF algorithm based on a BERT pre-trained model can extract multiple disease keywords from user physical examination data, such as fatty liver and diabetes. When the physical examination report data is structured data, such as structured data stored in key-value format, disease keywords can be directly obtained by reading the value corresponding to the key. For example, the physical examination report might display "Present medical history (key): Hypertension (value)". When the physical examination report data is unstructured data, keyword extraction algorithms can be used to extract disease keywords. For example, a medical examination report might show "Past medical history: Hypertension for over 10 years, currently not taking antihypertensive medication, reports normal blood pressure, and does not regularly monitor blood pressure changes." Using a keyword extraction algorithm, the disease keywords identified are: hypertension, duration: over 10 years. The aforementioned target disease type can also be entered by the user through an interactive interface.

[0111] The aforementioned target user attributes refer to the user's personal information. This personal information includes, but is not limited to, the user's gender, age, height, and weight. For example, user information can be obtained from medical examination data or entered through an interactive interface.

[0112] The physical activity level assessment results described above are indicators used to evaluate a user's physical fitness level. These results can be provided to the user via text or voice assessment reports. The report may include current activity level, activity frequency, activity intensity, sedentary time, exercise safety level, and a comparison chart with the current activity level of other groups (e.g., age, gender, occupation). The physical activity level assessment results include, but are not limited to, high, medium, and low levels. High indicates sufficient physical activity; medium indicates average physical activity; and low indicates insufficient physical activity.

[0113] Basic user information can be obtained through electronic questionnaires to obtain the user's target disease type, target user attributes, and physical activity level assessment results; it can also be obtained through the user's physical examination data to obtain the user's target disease type and user attributes; and it can also be obtained through electronic questionnaires to obtain the user's physical activity level assessment results.

[0114] It can also obtain user physical examination data, determine the content of the electronic questionnaire to be provided to the user based on the user's physical examination data, and obtain the user's physical activity level assessment results based on the content of the electronic questionnaire.

[0115] User health check data can be a comprehensive evaluation of the user's health check results. For example, the comprehensive evaluation result may indicate that the user is in a healthy state, or the user is in a sub-healthy state, or the user has a chronic disease.

[0116] The aforementioned electronic questionnaire methods include, but are not limited to, human-computer interactive display interfaces or voice dialogue methods.

[0117] The content of the electronic questionnaire provided to the user based on the user's physical examination data includes the following: if the user's physical examination data indicates that the user is in a healthy state, the user will be provided with the content of the International Physical Activity Questionnaire; if the user's physical examination data indicates that the user is in a sub-healthy state, or if the user has a chronic disease or other diseases, the user will be provided with the content of the Physical Fitness Adaptability Questionnaire.

[0118] The International Physical Activity Questionnaire (IPAQ) can collect data on the frequency and duration of exercise at different intensities over the past week. A user's current physical activity level can be categorized into three levels: low (insufficient physical activity), moderate (average physical activity), and high (sufficient physical activity).

[0119] For patients with sub-health conditions, chronic diseases, or other illnesses, the Pre-Activity Readiness Questionnaire (PAR-Q) can be used to obtain an assessment of their physical activity level during exercise. This assessment result serves as a parameter for evaluating the safety of the user's exercise.

[0120] like Figure 11 As shown, the IPAQ and PAR-Q questionnaires can be provided to users through a human-computer interaction interface, allowing for the acquisition of basic user information from the completed electronic questionnaires. Alternatively, the IPAQ and PAR-Q questionnaires can be provided to users via voice interaction, enabling the extraction of basic user information from the results of their voice conversations.

[0121] This application's embodiments utilize various electronic questionnaire methods to improve product satisfaction among different users. For example, voice interaction can provide elderly users with a convenient operating method, thereby increasing their satisfaction with the product.

[0122] Step 202: Based on the user's basic information, query the pre-constructed motion knowledge graph to obtain a list of target motion methods and a list of evaluation attributes corresponding to the target motion methods.

[0123] The list of target movement methods mentioned above includes one or more movement methods. Each movement method is obtained by querying a pre-built movement knowledge graph based on the user's basic information.

[0124] The above list of evaluation attributes includes one or more evaluation attributes. Evaluation attributes are attribute indicators used to evaluate the type of exercise. Evaluation attributes include, but are not limited to: exercise intensity value, whether the exercise is common, whether equipment is required, exercise difficulty, and calories burned per unit of energy.

[0125] For example, the list of target exercise methods includes {"brisk walking", "medical gymnastics", "Tai Chi", "Baduanjin", "running", "square dancing", "badminton", "aerobics"}.

[0126] The evaluation attribute list corresponding to the target exercise method list includes {[exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit]}. 快走 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 医疗体操 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 太极拳 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 八段锦 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 跑步 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 广场舞 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 打羽毛球 [Exercise intensity value, whether it is a common exercise, whether equipment is required, exercise difficulty, calories burned per unit] 有氧操}

[0127] The above-mentioned exercise methods include, but are not limited to, brisk walking, running, swimming, Tai Chi, push-ups, medical gymnastics, Baduanjin (Eight Pieces of Brocade), square dancing, badminton, aerobics, basketball, etc.

[0128] Based on the user's basic information, a pre-built sports knowledge graph is queried to obtain a list of sports activities and a list of corresponding evaluation attributes, including:

[0129] Construct at least one query statement based on the user's basic information;

[0130] The exercise knowledge graph is queried according to each query statement in at least one query statement to obtain a list of exercise methods and a list of evaluation attributes corresponding to the list of exercise methods.

[0131] The above-mentioned construction of at least one query statement based on the user's basic information may include:

[0132] Obtain the user's physical activity level assessment results;

[0133] When the user's physical activity level assessment result is high, the target user attributes and target disease type are used as query conditions to construct the query statement;

[0134] When the user's physical activity level assessment result is intermediate or below, the target user attribute and the target disease type are used as query conditions to construct a first query statement;

[0135] The second query statement is constructed by using the user's physical activity level assessment result as the query condition.

[0136] In the embodiments provided in this application, when the physical activity level assessment result is intermediate or below, the execution order of constructing the first query statement and the second query statement is not restricted. For example, the second query statement can be constructed based on the first query statement, that is, the user's physical activity level assessment result is added to the query conditions of the first query statement to construct the second query statement.

[0137] In some embodiments, the pre-constructed exercise knowledge graph includes a first relationship between user attribute entities and exercise type entities, and a second relationship between exercise type entities and disease type entities. When a user's physical activity level assessment result is high, the first and second relationships in the exercise knowledge graph are queried according to a first query statement to obtain a list of target exercise methods.

[0138] When the user's physical activity level assessment result is intermediate or below, the first and second relationships in the exercise knowledge graph are queried according to the first query statement to obtain an initial list of exercise methods.

[0139] Based on the second query statement, the exercise knowledge graph is queried to obtain the exercise methods in the initial exercise method list that meet the user's physical activity level assessment results, which are then used as the target exercise method list.

[0140] Obtain the list of evaluation attributes corresponding to the target motion mode list.

[0141] Step 203: Determine at least one exercise plan based on the target exercise mode list and the evaluation attribute list. The exercise plan includes at least one exercise mode and a recommended exercise duration corresponding to the exercise mode.

[0142] After obtaining the list of target exercise methods and the list of evaluation attributes, at least one exercise program can be determined based on the list of target exercise methods and the list of evaluation attributes, as well as the user's target calorie consumption. Alternatively, at least one exercise program can be determined based on the list of target exercise methods and the list of evaluation attributes, as well as the user's exercise plan cycle.

[0143] The target calorie expenditure mentioned above refers to the target calorie expenditure that the user hopes or needs to consume. The target calorie expenditure can be obtained by the user inputting it in the human-computer interaction interface, or by determining the daily food intake input by the user in the human-computer interaction interface.

[0144] The exercise plan mentioned above refers to the combination of an exercise method and the recommended exercise duration corresponding to that method. For example, an exercise plan could be ["brisk walking", 35 minutes], ["Tai Chi", 45 minutes], etc.

[0145] When determining an exercise plan based on the target calorie consumption, one or more exercise plans can be determined by the target exercise method list, the unit calorie consumption in the evaluation attribute list, and the user's target calorie consumption.

[0146] When determining an exercise program based on a user's exercise plan cycle, one or more exercise programs can be determined using a list of target exercise methods, a list of evaluation attributes, and an exercise plan cycle.

[0147] The exercise plan cycle refers to the length of time a user plans to exercise. For example, a one-week exercise plan cycle means the user plans to exercise for 7 days. The exercise plan cycle can be set by a third party, including but not limited to the user, through the human-computer interaction interface. The exercise plan cycle can also be a default recommended value.

[0148] Step 204: Sort at least one motion scheme according to the evaluation attribute list to obtain the sorting result.

[0149] After obtaining one or more motion schemes, the motion schemes are sorted according to the evaluation attribute list. For example, the motion schemes can be sorted by the combination of some evaluation attributes in the evaluation attribute list to obtain the sorting results. The sorting method can be ascending or descending.

[0150] Step 205: Recommend at least one exercise program from the sorting results.

[0151] In the above ranking results, some exercise programs are recommended as exercise recommendations. For example, after evaluating each exercise program according to its rating attributes, the top N exercise programs are selected from the multiple programs with the rating results arranged in descending order (i.e., a sequence of rating results from highest to lowest) as recommendations. The value of N can be determined based on the user's level or the user's physical activity level assessment results. For example, the number of recommended exercise programs can be determined based on the user's physical activity level assessment results. When the physical activity level assessment result is high, N can be 3. When the physical activity level assessment result is medium, N can be 2. When the physical activity level assessment result is low, N can be 1 or 0. When N is zero, the user is advised that exercise is not advisable.

[0152] In some embodiments, this application may further determine the user's exercise risk level based on the user's basic information before obtaining the list of target exercise methods; and determine the range of recommended exercise methods based on the exercise risk level. For example, if the exercise risk level is high, no exercise methods will be recommended to the user. Alternatively, if the exercise risk level is medium, at least one of aerobic exercise, strength training, and flexibility training will be recommended to the user.

[0153] This application embodiment queries a pre-constructed exercise knowledge graph based on the user's basic information to obtain a list of target exercise methods and a list of evaluation attributes corresponding to the target exercise methods. Then, based on the list of target exercise methods, the list of evaluation attributes, and the target calorie consumption, at least one exercise program is determined. Finally, exercise recommendation results are obtained from the sorted exercise programs. Compared with related technologies, it can recommend different exercise programs for each user, which effectively improves the accuracy of recommended exercise.

[0154] This application provides a method for recommending exercise through several embodiments. Please refer to... Figure 3 , Figure 3 This illustration shows a flowchart of yet another exercise recommendation method provided in an embodiment of this application. Figure 3 As shown, this method can be executed by an exercise recommendation device, which can be configured in a terminal device or a server. The method includes:

[0155] Step 301: Obtain the user's basic information;

[0156] After obtaining the user's basic information, the information is preprocessed, for example, by discretizing the input features of the user's physical examination data.

[0157] Gender: Male, Female

[0158] Population segmentation: Adolescents (12-17 years old), Adults (18-64 years old), Seniors (65 years and older)

[0159] Special physiological states of women: menstruation, pregnancy, postpartum period, menopause

[0160] Physical condition: Unknown, underweight, normal weight, overweight, obese, central obesity, pre-central obesity

[0161] Physical activity level: Unknown, Low, Moderate, High

[0162] Fitness environment: Unknown, gym, home, bodyweight training

[0163] Health status: unknown, healthy, sub-healthy, hypertension, diabetes, osteoporosis, hyperlipidemia, etc.

[0164] Fitness goals: All fitness levels, fat loss, body shaping, muscle gain, strength training, blood sugar reduction, blood pressure regulation, and improved blood lipids.

[0165] Health status: healthy people, sub-healthy people, people with chronic diseases, people with limited physical activity, people with developmental disorders, and people with other diseases.

[0166] Types of exercise: aerobic exercise, strength training, stretching exercises, balance exercises, flexibility training

[0167] Exercise intensity: low intensity, moderate intensity, high intensity, low-to-medium intensity, high-to-medium intensity

[0168] Difficulty level: Low, Medium, High

[0169] Exercise heart rate: Calculated using both maximum heart rate method and heart rate reserve method.

[0170] Human resting energy expenditure: calculated using the Mifflin-St Jeor formula based on gender, height, weight, and age, in kcal / day.

[0171] User features are discretized, such as age (three categories), and exercise programs cannot be recommended to users who do not meet the age criteria. Each user feature is then represented as a low-dimensional embedding vector. The embedding pre-training model uses WORD2VEC, and the training data is sourced from medical websites and data compiled from sports medicine books.

[0172] Step 302: Based on the user's basic information, query the pre-constructed motion knowledge graph to obtain a list of target motion methods and a list of evaluation attributes corresponding to the list of target motion methods;

[0173] Step 3031: Obtain the user's target calorie consumption;

[0174] Step 3032: Combine at least one motion mode included in the target motion mode list to obtain a motion combination result, wherein the motion combination result includes at least one motion mode;

[0175] The above-mentioned motion combination result refers to the result of combining one or more motion methods. Combining a list of target motion methods can include:

[0176] Based on the user's physical activity level assessment results, a preset number of exercise methods corresponding to the user's physical activity level assessment results are retrieved from the target exercise method list. The preset number of exercise methods corresponding to the user's physical activity level assessment results is a fixed attribute pre-stored in the exercise knowledge graph.

[0177] In some embodiments, after obtaining the target exercise mode list, a preset number of exercise modes corresponding to the user's physical activity level assessment result can be directly obtained from the target exercise mode list as the exercise combination result. Assume the target exercise mode list is {e1, e2, ... e...} n The user's physical activity level assessment result is intermediate. The preset number of results corresponding to the user's physical activity level assessment result is 2. When retrieving the exercise combination result from the target exercise method list, the first exercise combination result (e1, e2) and the second exercise combination result (e4, e6) can be obtained. This is only for explanation and does not limit the number of exercise combination results.

[0178] When retrieving a preset number of exercise methods corresponding to the user's physical activity level assessment from the target exercise method list as exercise combination results, multiple exercise combination results can be obtained. These multiple exercise combination results are sorted according to the comprehensive evaluation index corresponding to the exercise combination results, and then one or more exercise combination results are determined based on the sorting results.

[0179] Assume the list of target motion modes is {e1, e2, ... e} n If a user's physical activity level assessment result is intermediate, and the preset number of possible combinations corresponding to this assessment result is 2, then multiple exercise combination results will be obtained, such as (e1, e2), (e4, e6). The comprehensive evaluation index K for each exercise combination result will be determined using the following comprehensive evaluation index calculation formula. 综合 :

[0180]

[0181] Then, based on the comprehensive evaluation index results, sort them from largest to smallest or smallest to largest, and obtain the one with the largest comprehensive evaluation index (which can be one of the higher-ranked ones) as the target motion combination result.

[0182] Combining at least one motion mode included in the list of target motion modes may also include:

[0183] The target motion mode list is sorted according to the evaluation attribute list corresponding to the target motion mode list to obtain the sorted target motion mode list.

[0184] Based on the user's physical activity level assessment results, a preset number of exercise methods corresponding to the user's physical activity level assessment results are obtained from the sorted list of target exercise methods, and these are used as the exercise combination results.

[0185] In this embodiment of the application, the list of target motion modes can be sorted to improve the accuracy of the recommendation results.

[0186] When retrieving a preset number of exercise methods corresponding to the user's physical activity level assessment from a sorted list of target exercise methods, and using these as the exercise combination result, the list of target exercise methods is first sorted, and may include:

[0187] Based on the evaluation attribute list corresponding to each movement mode, the evaluation result for each movement mode is calculated. Then, the list of target movement modes is sorted according to the evaluation index K.

[0188] Calculate the evaluation index K for each motion mode according to the evaluation index formula:

[0189] Assume the sorted list of target motion modes is {e3, e1, ... e} n ... e2}, obtain the exercise methods according to the preset number corresponding to the user's physical activity level result. For example, when the user's physical activity level assessment result is high, the top 3 exercise methods in the sorted target exercise method list can be obtained as the exercise combination result, for example, (e3, e1, e4).

[0190] For example, if a user's physical activity level assessment result is intermediate, then the top two ranked exercise methods can be obtained from the sorted list of target exercise methods as the exercise combination result (e3, e1). Step 3033: Extract the evaluation attributes corresponding to the exercise combination result from the evaluation attribute list based on the exercise combination result;

[0191] Step 3034: Determine the recommended time combination corresponding to the exercise combination result based on the evaluation attributes and target calorie consumption. The recommended time combination includes the recommended exercise duration corresponding to each exercise mode in the exercise combination result.

[0192] The evaluation attributes corresponding to the above-mentioned exercise combination results refer to the evaluation attributes corresponding to each exercise mode in the exercise combination results. After obtaining the exercise combination results, recommended exercise durations can be assigned to the exercise modes included in the exercise combination results according to the target calorie consumption or exercise plan cycle.

[0193] When allocating recommended exercise duration based on the target calorie expenditure and corresponding exercise mode, assuming the exercise combination result is (e3, e1), extracting the evaluation attribute corresponding to the exercise combination result from the evaluation attribute list means extracting ([calorie expenditure per unit]) from the evaluation attribute list. e3 [Heat consumption per unit] e1 ).

[0194] Then, based on the evaluation attributes and target calorie consumption, the recommended time combination corresponding to the exercise combination result is determined. This can include determining the recommended time combination corresponding to the exercise combination based on the sum of the products of the unit calorie consumption of each exercise mode included in the exercise combination result and the recommended exercise duration, which equals the target calorie consumption.

[0195] The recommended time combination corresponding to the exercise combination result is determined based on the following recommended time formula.

[0196]

[0197] Indicates the target amount of calories consumed. This indicates the number of motion modes contained in the motion combination vector; This represents the calorie expenditure per unit corresponding to the i-th type of exercise. This represents the recommended exercise duration for the i-th exercise method.

[0198] For example, based on the recommended time formula above, with the target calorie expenditure y being 340 kcal as a constraint, the calorie expenditure per unit corresponding to the exercise combination result (e3, e1) is ( , The recommended exercise duration for each exercise mode in the exercise combination result (e3, e1) is determined to be [(e3, 30), (e1, 50)], that is, when the target calorie expenditure is 340 kcal, the calorie expenditure per unit is ( , Substituting into the above recommended time formula, we can obtain that the recommended exercise duration for exercise mode e3 is 30 minutes, and the recommended exercise duration for exercise mode e1 is 50 minutes.

[0199] Then, according to the pre-set exercise time step (e.g., a time interval of 5 minutes), the recommended exercise duration corresponding to the exercise mode is adjusted in combination with the above recommended time formula, so that multiple exercise schemes to be ranked can be obtained, which can also be called candidate schemes or candidate sets.

[0200] The exercise time step mentioned above refers to the time interval between two recommended exercise durations for the same exercise method. For example, if the exercise method is "brisk walking," and the first recommended exercise duration is 30 minutes and the second is 35 minutes, then the exercise time step is 5 minutes. The exercise time step can be adjusted according to the user's needs. The exercise time step can also be determined based on the user's basic information. For example, if the user's basic information indicates that the user is elderly, then the corresponding exercise time step for the elderly is 5 minutes, meaning that the recommended exercise duration suitable for the elderly is obtained according to a 5-minute time interval. If the user's basic information indicates that the user is young, then the exercise time step for young people can be 10 minutes, meaning that the recommended exercise duration suitable for young people is obtained according to a 10-minute time interval.

[0201] For example, when e3 = "brisk walking" and e1 = "Tai Chi", the calorie expenditure per unit of "brisk walking" is c1 = "228 kcal / hour", and the calorie expenditure per unit of "Tai Chi" is c2 = "240 kcal / hour". According to the recommended time formula above, n = 2, y = 340, then:

[0202] c1*t1+c2*t2=340

[0203] get:

[0204] (228 / 60)*t1+(240 / 60)*t2=340

[0205] That is: 3.8t1 + 4t2 = 340

[0206] The range of values ​​for the regular exercise time is [10, 15, 20, ..., 85, 90].

[0207] For young and middle-aged people, the exercise time t takes values ​​in [10, 15, 20, ..., 85, 90].

[0208] For elderly individuals, the exercise time t can be set between [10, 15, ..., 60]. The exercise time step can be 5 minutes for both young adults and the elderly. Assuming a target calorie expenditure of 340 kcal, assuming t1 is 10 minutes and t2 is 75 minutes, this is unsuitable for this user as they are elderly. Adjusting t1 and t2 according to the exercise time step, assuming t1 is 15 minutes and t2 is 70 minutes, is also unsuitable and should be discarded. This continues until t1 is 30 minutes and t2 is 50 minutes. Therefore, t1=30 and t2=50 is the recommended time combination corresponding to the exercise result. Similarly, t1=35 and t2=50 can be used as the recommended time combination. Finally, t1=40 and t2=45 can be used as the recommended time combination. Continue until all combinations that satisfy the target calorie consumption are found within [10, 15, ... 60].

[0209] Based on the recommended time formula above, multiple exercise plans can be obtained, such as {[(e3,30), (e1,50)], [(e3,40), (e1,45)], [(e3,50), (e1,40)]}. These exercise plans contain the same exercise mode, (e3, e1). The recommended exercise duration for each exercise mode (e3, e1) is different in each exercise plan. For example, in the first exercise plan [(e3,30), (e1,50)], the recommended exercise durations for exercise modes (e3, e1) are 30 and 50 minutes, respectively. In the second exercise plan [(e3,40), (e1,45)], the recommended exercise durations for exercise modes (e3, e1) are 40 and 45 minutes, respectively.

[0210] Based on the list of target exercise methods, the list of evaluation attributes, and the target calorie consumption, the exercise plan is determined, resulting in a set of exercise plans: {[(e3,30), (e1,50)], [(e3,40), (e1,45)], [(e3,50), (e1,40)]}, where [(e3,30), (e1,50)] represents one exercise plan.

[0211] Step 304: Sort the at least one motion scheme according to the evaluation attribute list to obtain the sorting result;

[0212] In the steps described above, an exercise plan may include one or more exercise methods and a recommended duration for each method. The exercise plan may also include heart rate, expected calorie expenditure, and exercise grouping. Exercise grouping refers to a detailed breakdown of the exercise method. Exercise methods are subdivided based on exercise grouping. For example, when the exercise category is strength training, exercise grouping might include 2-3 sets of supine knee bends (8-12 repetitions per set), 2-3 sets of leg presses (8-12 repetitions per set), and 2-3 sets of bench presses (8-12 repetitions per set). Subdividing exercise methods through exercise grouping provides users with more accurate exercise recommendations.

[0213] The evaluation attribute list can include multiple evaluation attributes corresponding to the mode of motion.

[0214] In some embodiments, the user's exercise risk level can be determined based on the user's basic information, and at least one exercise program can be ranked according to multiple evaluation attributes and exercise risk level attributes.

[0215] In some embodiments, ranking the at least one exercise program based on the plurality of evaluation attributes and the exercise risk level attribute may include:

[0216] Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode;

[0217] The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated.

[0218] The product is calculated based on the value of the risk level corresponding to each exercise mode and the weighting coefficient corresponding to the risk level.

[0219] The product result is used as an evaluation index for each movement mode in the at least one movement scheme;

[0220] The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode;

[0221] The at least one exercise program is ranked according to the comprehensive evaluation index.

[0222] In the above steps, the weight coefficient corresponding to the evaluation attribute refers to the weight value used to characterize the evaluation attribute in the calculation of the evaluation index. For example, if the evaluation attribute is the exercise intensity value, its proportion in the calculation of the evaluation index is 0.3.

[0223] The evaluation index K for each movement mode is obtained according to the following evaluation index formula:

[0224]

[0225] In the above formula, arrive These are the weighting coefficients corresponding to exercise intensity, whether the exercise is common, whether exercise equipment is needed, exercise difficulty, and user's exercise risk level. k1, ..., k5 can be set empirically. Alternatively, parameter values ​​can be obtained through training on labeled data. For example, the training process can use the training feature X as X, along with the weighting coefficients for exercise intensity, whether the exercise is common, whether exercise equipment is needed, exercise difficulty, user's exercise risk level, and exercise duration, etc., and train the training feature X according to a preset training algorithm to obtain the label Y.

[0226] After obtaining the evaluation index K corresponding to each movement mode, the comprehensive evaluation index K for each movement scheme can be determined using the following comprehensive evaluation index calculation formula. 综合 :

[0227]

[0228] in, , , ..., This represents the weighting coefficient for each motion mode, with a value ranging from 0 to 1. , ..., This represents the evaluation index corresponding to each movement mode, which can be determined using the evaluation index calculation formula mentioned above. N is the number of evaluation modes.

[0229] Finally, based on the comprehensive evaluation index K 综合 Sort the exercise plans.

[0230] The above embodiments rank the exercise programs by obtaining the comprehensive evaluation index corresponding to each exercise program, which can improve the accuracy of the recommendation results.

[0231] In some embodiments, ranking the at least one exercise program based on the plurality of evaluation attributes and the exercise risk level attribute may include:

[0232] Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode;

[0233] The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated.

[0234] The product is calculated by combining the value of the risk level corresponding to each exercise mode with the weighting coefficient corresponding to the risk level.

[0235] The product is calculated based on the recommended duration of each exercise mode and the weight coefficient corresponding to the recommended duration;

[0236] The product result is used as an evaluation index for each movement mode in the at least one movement scheme;

[0237] The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode;

[0238] The at least one exercise program is ranked according to the comprehensive evaluation index.

[0239] In some embodiments, recommended exercise duration may also be considered in the calculation of evaluation indicators, and the evaluation indicator K for each exercise mode can be obtained according to the following evaluation indicator formula:

[0240]

[0241] in, arrive These are the weighting coefficients corresponding to exercise intensity, whether the exercise is common, whether exercise equipment is needed, exercise difficulty, and the user's exercise risk level. This refers to the weighted coefficient value of the exercise duration corresponding to the exercise mode. This value can be determined based on the user's demographic attributes. Normalization time refers to the value used to normalize the exercise duration. k0, k1, ..., k5 can be set empirically. Alternatively, parameter values ​​can be obtained through training on labeled data. For example, the training process can use training features as X, including exercise intensity, whether the exercise is common, whether exercise equipment is needed, exercise difficulty, user exercise risk level, and the weighted coefficient value of exercise duration, etc., and train the training features X according to a preset training algorithm to obtain the label Y. The normalization time can be, for example, 100.

[0242] After obtaining the evaluation index K corresponding to each movement mode, the comprehensive evaluation index K for each movement scheme can be determined using the aforementioned comprehensive evaluation index calculation formula. 综合 Finally, based on the comprehensive evaluation index K 综合 Sort the exercise plans.

[0243] Step 305: Recommend at least one exercise program from the sorting results.

[0244] In some embodiments provided in this application, the exercise programs are ranked by obtaining a comprehensive evaluation index for each exercise program, thereby improving the accuracy of recommending exercise to individual users.

[0245] Some embodiments of this application also propose a method for constructing a motion knowledge graph. Please refer to... Figure 4 , Figure 4 A schematic flowchart of a method for constructing a motion knowledge graph according to an embodiment of this application is shown. The method includes:

[0246] Step 401: Perform named entity recognition on the original medical data, the original medical motion combined data, and the original motion data to obtain an entity set.

[0247] The aforementioned entity set includes user attribute entities, exercise mode entities, disease type entities, evaluation attribute entities corresponding to exercise mode entities, and contraindication attribute entities corresponding to disease type feature entities. User attribute entities include, but are not limited to, adults and the elderly, such as... Figure 6 As shown. Entities related to exercise include, but are not limited to, brisk walking and calf raises, such as... Figure 6-8 As shown. Disease type entities include, but are not limited to, hypertension, diabetes, etc., such as... Figure 6-7 As shown. Evaluation attribute instances corresponding to the exercise mode entity include, but are not limited to, exercise intensity, whether it is machine-based exercise, and exercise category, such as... Figure 6-7 As shown. The contraindication attribute entities corresponding to the disease type feature entities include, but are not limited to, exercise contraindications and indications, such as... Figure 6-7 As shown.

[0248] In the above steps, data is obtained from clinical medical guidelines, sports medicine books, literature, and medical websites. The BILSTM-CRF algorithm, based on a BERT-trained model, performs motion-related named entity recognition on the data to be extracted. The Bootstrapping algorithm, combined with rule templates, is used for relation extraction. The Word2vec model can also be used to obtain the similarity of synonymous expressions of different exercise types, and a first candidate set is obtained based on this similarity. Finally, entity disambiguation is performed on the second candidate set based on the estimated calorie data of the exercise type. For example, ... Figure 9 As shown, this embodiment of the application can also train a model using the Word2vec algorithm to obtain entity names with semantic similarity greater than a preset threshold as a first candidate synonym set; calculate the average calorie count of the first candidate synonym set, and query from the sports knowledge graph sports names whose absolute value of the difference from the average calorie count is less than 10 kcal / hour as a second candidate synonym set; combine the first and second candidate synonym sets to obtain the final candidate synonym set; and manually extract, check, and evaluate the final candidate synonym set. Through the above processing of the synonym set, multiple synonymous expressions for sports names corresponding to sports methods can be obtained, eliminating ambiguity between multiple synonyms.

[0249] Data sources include, but are not limited to, clinical medical guidelines, sports medicine books, literature, medical websites, and national fitness guidelines. Multi-threaded web crawling can also be used for data crawling and parsing, as well as data cleaning, integration, and consolidation.

[0250] Step 402: Construct a set of semantic relation triples based on user attribute entities, movement type entities, disease type entities, evaluation attributes corresponding to movement mode entities, and taboo attributes corresponding to disease type entities.

[0251] The set of semantic relation triples includes a first relation between user attribute entities and movement type entities, a second relation between movement type entities and disease type entities, a third relation between disease type entities and the taboo attributes corresponding to disease type entities, and a fourth relation between movement mode entities and the evaluation attributes corresponding to movement mode entities.

[0252] The sports knowledge graph constructed in this application can contain 2684 entities and 10578 triples. Entities can include sports form, sports category (upper limb, core, stretching, lower limb, strength, yoga, Pilates, etc.), sports level, muscle group, equipment requirements, calories, contraindicated sports, recommended sports, suitable population, movement techniques, fitness effects, etc. Entity recognition can use the BILSTM-CRF algorithm based on the BERT pre-trained model. During the entity recognition process, relevant dictionary data such as diseases, symptoms, sports methods, movements, human skeleton, muscles, tissues, diseases, and exercise parts are fused. Relation extraction can use the Bootstrapping algorithm combined with rule templates.

[0253] In some embodiments, the entities included in the motion knowledge graph, the relationships between entities, and the attribute types are shown in Tables 1, 2, and 3, respectively. Table 1 shows the entity types of the motion knowledge graph; Table 2 shows the relationships between entities in the motion knowledge graph; and Table 3 shows the attribute types of the motion knowledge graph.

[0254]

[0255] Table 1. Entity Types in the Motion Knowledge Graph

[0256] As shown in Table 1 above, the sports knowledge graph contains multiple entities, such as diseases, sports subcategories, sports levels, sports types, suitable populations, sports intensity, and the role of sports. This application extracts multiple sports-related attributes as entities, providing rich queryable fields for the query interface and effectively improving the query efficiency of the sports knowledge graph.

[0257]

[0258] Table 2. Entity Relationship Types in the Motion Knowledge Graph

[0259] As shown in Table 2 above, the sports knowledge graph contains various entity relationship attributes, such as subclasses of sports, categories to which sports belong, equipment requirements, major muscle groups, effects of sports, etc. These entity relationships connect multiple entities into a graph, and graph matching query methods can quickly and accurately find results related to the target search object.

[0260]

[0261] Table 3. Attribute Types of the Motion Knowledge Graph

[0262] As shown in Table 3 above, the exercise knowledge graph also provides attribute types. Different attribute types are used to explain the relevant attributes of an exercise method. For example, the exercise name, calories burned, and key points of the movement can all provide a brief description of the exercise method, making it easier for users to understand the content related to the exercise method. Even reminders and precautions are pre-created as attributes of the exercise method in the exercise knowledge graph.

[0263] In the process of constructing a motion knowledge graph, such as Figure 6 As shown, synonyms for "fitness walk" include "aerobic fitness walk", "brisk walk", "fast walk", and "healthy walk"; synonyms for "jogging (6-8 km / h)" include "running, 6-8 km / h".

[0264] The knowledge graph of movement can also contain taboo information. Extraction of this taboo information includes, but is not limited to, named entity recognition using the BILSTM-CRF algorithm based on a BERT pre-trained model. The associations between this taboo information and movement patterns and diseases can be extracted using the Bootstrapping algorithm combined with rule templates.

[0265] For example, the text states, "Those with epilepsy should avoid activities with inherent risks, such as skiing and diving. Swimming in the sea or rivers is prohibited. Working at heights and operating machinery should be avoided."

[0266] Named entity recognition can identify the disease "epilepsy" and the types of exercise "skiing", "diving", and "swimming".

[0267] Relation extraction can yield semantic relation triples such as "epilepsy-forbidden sports-skiing", "epilepsy-forbidden sports-diving", and "epilepsy-forbidden sports-swimming".

[0268] The text states, "Lower back pain is divided into acute and chronic. For acute lower back pain, strenuous exercise should be avoided, and daily life is fine. For chronic lower back pain, exercise can help relieve the pain. Swimming is recommended as it puts the least strain on the lumbar spine. Basketball, badminton, and other sports are not recommended because they may cause injury, especially for heavier individuals. If swimming is not an option, plank exercises can be performed at home to strengthen the abdominal and back muscles, i.e., the core muscles. Tai Chi can also strengthen the core muscles and improve balance."

[0269] Named entity recognition identifies the following terms: "lower back pain", "acute lower back pain", and "chronic lower back pain".

[0270] Relation extraction: "Chronic lower back pain - Recommended exercise - Swimming", "Chronic lower back pain - Recommended exercise - Plank", "Chronic lower back pain - Recommended exercise - Tai Chi", "Chronic lower back pain - Contraindicated exercise - Basketball", "Chronic lower back pain - Contraindicated exercise - Badminton".

[0271] Step 403: Store the entity set and the semantic relation triple set into the graph database to obtain the motion knowledge graph.

[0272] The aforementioned entity set includes, but is not limited to, user attribute entities, exercise mode entities, disease type entities, evaluation attribute entities corresponding to exercise mode entities, and prohibition attribute entities corresponding to disease type entities.

[0273] The aforementioned semantic relationships include, but are not limited to, the first relationship between user attribute entities and motion type entities, the second relationship between motion type entities and disease type entities, the third relationship between disease type entities and the taboo attribute entities corresponding to disease type entities, and the fourth relationship between motion mode entities and the evaluation attributes corresponding to motion mode entities.

[0274] The named entity recognition results and relation extraction results are stored in a graph database to obtain a motion knowledge graph. For example, Neo4j graph database can be used for storage.

[0275] like Figure 6-7As shown, taking fitness walking as an example, the aliases, suitable population, exercise category, and suitable ailments associated with fitness walking can be used to construct multiple semantic relation triples. Each semantic relation triple can represent a relation attribute between two entities. For example, the first relation between the user attribute entity and the exercise type entity can be represented as <fitness walking, best population, elderly>. The second relation between the exercise type entity and the disease type entity can be represented as <fitness walking, suitable ailment, diabetes>. The third relation between the disease type entity and the corresponding contraindication attribute entity can be represented as <diabetes, exercise contraindication, recurrent hypoglycemia>. The fourth relation between the exercise method entity and the corresponding evaluation attribute can be represented as <fitness walking, exercise intensity, moderate intensity>, <fitness walking, default, calories burned per unit>, etc.

[0276] The motion mode and attribute relationship are stored in the graph database as named entity recognition results and relationship type extraction results, thereby constructing the motion knowledge graph proposed in the embodiments of this application.

[0277] The exercise knowledge graph proposed in this application can filter exercise methods associated with target disease types based on contraindication attributes of the target disease type, and transform clinical exercise guidance into an achievable recommendation strategy through contraindication attributes, effectively improving the accuracy of recommended exercise methods.

[0278] To better understand the exercise recommendation method proposed in this application, the following section combines... Figure 5-17 Using a male's physical examination results and questionnaire survey results as an example, this paper describes the methods for recommending exercise.

[0279] like Figure 5 As shown, the overall exercise recommendation method can be divided into offline and online components. The offline component is used to construct an exercise knowledge graph, while the online component is the application of the exercise knowledge graph constructed in the offline component.

[0280] Online component: User data can be obtained first, including but not limited to user physical examination data and the content of the electronic questionnaire to be provided based on the user's physical examination data. Then, the user's physical activity level assessment results can be obtained based on the content of the electronic questionnaire.

[0281] Suppose we obtain the following user's medical examination report data:

[0282] Gender: Male

[0283] Age: 66

[0284] Height: 176cm

[0285] Weight: 70kg

[0286] Contraindications for exercise: None

[0287] The electronic questionnaire, including the PAR-Q questionnaire, is provided to users based on their age. The PAR-Q questionnaire can be delivered via voice interaction. After receiving the voice response to the PAR-Q questionnaire, the user's physical activity level is assessed as moderate by analyzing the voice response.

[0288] Based on the user's physical examination report data and physical activity level assessment results, the user's exercise risk assessment result was obtained through a sports risk classification model: Medium.

[0289] The exercise risk classification model includes, but is not limited to, models trained using the AdaBoost algorithm. A user's exercise risk level can be represented by a risk level label, which includes, but is not limited to, four categories: no risk, low risk, medium risk, and high risk.

[0290] The AdaBoost algorithm trains multiple weak classifiers and then linearly combines them to obtain a strong classifier. Here, the CART classification and regression tree is used as the weak classifier.

[0291] When training the exercise risk classification model, the input data includes, but is not limited to: basic information: gender, age, occupation; health history: family history, present illness history, allergy history, medication history, surgical history, menstrual and reproductive history (female), physical symptoms; physical examination information: height, weight, waist circumference, hip circumference, systolic blood pressure, diastolic blood pressure, heart rate, fasting plasma glucose (FPG), postprandial blood glucose (2HPG), glycated hemoglobin, total cholesterol, triglycerides, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, liver fat content, blood uric acid, bone mineral density; lifestyle habits: diet, exercise, sleep, smoking, alcohol consumption; environment: humidity, PM2.5, etc., and physical activity level (IPAQ) obtained through questionnaire assessment.

[0292] Body state can be calculated using height, weight, and waist circumference. The specific calculation method is as follows:

[0293]

[0294] BMI calculations for other age groups refer to international standards. Among these, the relationship between fasting blood glucose, glycated hemoglobin, and diabetes provides important indicators for understanding a user's blood glucose status and determining whether blood glucose levels are abnormal.

[0295] The aforementioned systolic and diastolic blood pressure are important indicators for determining whether a user has hypertension or elevated blood pressure. Similarly, blood uric acid is an important indicator for determining whether a user has gout; bone mineral density is an important indicator for determining whether a user has osteoporosis, and this indicator can also determine the severity of osteoporosis. Liver fat content is an important indicator for determining whether a user has fatty liver or dyslipidemia. If bone mineral density is unavailable, the OSTA (Osteoporosis Self-Screening Tool for Asians) can be used to predict the risk level of osteoporosis based on age and weight.

[0296] The aforementioned exercise risk classification model can also incorporate expert knowledge during the training phase. For example, clinical guidelines state that for individuals aged 18 <= age <= 40, without any medical history or medication history, exercise capacity assessment is unnecessary, and the PAR-Q questionnaire should not be sent to the user. The PAR-Q questionnaire results do not need to be referenced during exercise capacity assessment. For individuals aged > 40, exercise capacity assessment is required, and the PAR-Q questionnaire can be sent to the user, with the PAR-Q questionnaire results incorporated into the exercise capacity assessment. In some embodiments, for individuals of any age who are generally inactive, have a certain disease, or have potential health conditions, an exercise risk assessment should be conducted before starting exercise and increasing activity levels.

[0297] Preprocessing user data yields query variables corresponding to the exercise knowledge graph. For example, the user's age in their physical examination data determines that they belong to the elderly population. Specifically, based on age 66, the user is identified as elderly, represented as the query variable "elderly." The user's height, weight, and waist circumference determine their body mass level as obese. Optionally, waist circumference can also be used to determine the presence of central obesity. Specifically, based on height (176cm) and weight (70kg), the user's BMI is determined to be obese, represented as the query variable "obese." Fasting blood glucose is used to determine if there is significant hyperglycemia. Diastolic and systolic blood pressure are used to determine if there is significant hypertension. If either "significant hyperglycemia" or "significant hypertension" is true, the user is deemed temporarily not to need exercise, and a suggestion is made to resume exercise after blood glucose and blood pressure have stabilized. By preprocessing the above user physical examination data, the following query variables can be obtained:

[0298] Gender: Male

[0299] Target audience: Seniors

[0300] BMI status: Obese

[0301] Medical condition: Diabetes

[0302] Physical activity level: Moderate

[0303] Contraindications for exercise: None

[0304] Exercise risk assessment results: Medium

[0305] Significantly high blood sugar: No

[0306] Does the patient have significant hypertension? No

[0307] After obtaining the aforementioned query variables, a query statement is constructed based on these variables. This query statement is then used to query the pre-created motion knowledge graph, yielding query results that match the query category variables. For example, the Cypher command can be used to query the neo4j knowledge graph.

[0308] like Figure 7 As shown, the query retrieves a list of exercise methods from the knowledge graph that satisfy the disease or symptom "obesity" and "diabetes" and are suitable for the population "elderly". Using the target user attributes and target disease type as query conditions, the query statement is constructed as follows:

[0309] MATCH (a:Action)-[r:s2AdapSymptoms]->(s:Symptoms) where s.name="obesity" and s.name="diabetes"WITH a,r,s MATCH (a:Action)-[r1:s2population]->(p:population) where p.name="elderly" return a,r,s,r1,pLIMIT 10

[0310] The query result 'a' (let's assume action_list=a for clarity) is:

[0311] action_list = ["brisk walking", "medical gymnastics", "Tai Chi", "Baduanjin", "running", "swimming", "square dancing", "basketball", "badminton", "aerobics"]

[0312] The above query retrieves a list of recommended exercise methods for the user. These include "brisk walking," "medical gymnastics," and "Tai Chi."

[0313] The action_list is filtered based on "Physical Activity Level Assessment Result: Moderate" to obtain exercise types with an Intensity of "Low Intensity" or "Moderate Intensity". The user's physical activity level assessment result is then used as the query condition to construct the following query statement:

[0314] MATCH (a:Action)-[r:s2intensity]->(s:Intensity) where s.name="low intensity" or s.name="medium intensity" where a.name in action_list return a.name, a.calorie, r,s

[0315] Where a.name is the name of the sport, and a.calorie is the calories burned per hour (in kcal).

[0316] After the above filtering process, the filtered list of exercise types and the corresponding list of calorie consumption per unit of exercise type are obtained as follows:

[0317] a.name list = ["brisk walking", "Tai Chi", "medical gymnastics", "Baduanjin", "running", "square dancing", "badminton", "aerobics"]

[0318] a.calorie list = ["228 kcal / hour", "180 kcal / hour", "240 kcal / hour", "270 kcal / hour", "670 kcal / hour", "180 kcal / hour", "270 kcal / hour", "354 kcal / hour"]

[0319] For example, using target user attributes and target disease types as query conditions, the query statement can be constructed as follows:

[0320] MATCH (a:Action)-[r:s2AdapSymptoms]->(s:Symptoms) where s.name="obesity" and s.name="diabetes" WITH a,r,s MATCH (a:Action)-[r1:s2population]->(p:population) where p.name="elderly" return a.name,a.calorie,,r,s,r1,pLIMIT 10

[0321] Where a.name is the name of the sport, and a.calorie is the calories burned per hour (in kcal).

[0322] The system retrieves the user's target calorie expenditure of 340 kcal input through the user interface. Based on the target exercise mode list, evaluation attribute list, and target calorie expenditure, it determines the user's exercise plan. Given the user's "Physical Activity Level Assessment Result: Moderate," it retrieves two exercise modes from the filtered exercise mode list. Then, it determines the recommended exercise duration combination corresponding to each exercise mode based on the recommended exercise duration formula.

[0323] Suppose we retrieve "brisk walking" and "Tai Chi" as the exercise combination result from the list a.name (i.e., the list of exercise methods), where... This indicates the calories burned per unit of time during a brisk walk. This indicates the unit of heat consumption corresponding to Tai Chi.

[0324]

[0325] Based on the above formula, the user's exercise combination result is as follows:

[0326] =30; =50

[0327] =40; =45

[0328] The exercise time ranges from [10, 15, 20, ..., 85, 90], and the corresponding exercise time for each exercise method can be determined within this range. For example, a time step of 10 minutes is ideal for "brisk walking." When the recommended exercise duration for "brisk walking" is 30 minutes, then according to the above formula, with a target calorie expenditure of 340 kcal as the constraint, the recommended exercise duration for "Tai Chi" can be determined to be 50 minutes. Then, increasing the recommended exercise duration for "brisk walking" by 10 minutes, we get a recommended exercise duration of 40 minutes. Then, according to the above formula, with a target calorie expenditure of 340 kcal as the constraint, the recommended exercise duration for "Tai Chi" can be determined to be 45 minutes. By using the above method with a target calorie expenditure of 340 kcal as the constraint, we can determine the recommended time combinations corresponding to multiple exercise combinations, thus obtaining multiple exercise plans.

[0329] Based on the list of target exercise methods, the list of evaluation attributes, and the target calorie expenditure, multiple exercise programs Z are determined as follows:

[0330] Z = {[("Brisk Walking", 30), ("Tai Chi" 50)], [("Brisk Walking", 40), ("Tai Chi" 45)], ...}

[0331] By sorting the above-mentioned multiple exercise schemes and then obtaining exercise recommendation results from the sorted exercise schemes, the accuracy of personalized recommendation results can be effectively improved.

[0332] Suppose that the set of evaluation attributes corresponding to "brisk walking" includes {exercise intensity value: 3.8; whether it is a common exercise: 1; whether exercise equipment is needed: 0; exercise difficulty: 0; user exercise risk level: 0}.

[0333] Among them, whether it is a common sport: if yes, the value is 0, if no, the value is 1. Whether equipment is required: if yes, the value is 0, if no, the value is 1. Exercise difficulty: low, the value is 0; medium, the value is 1; high, the value is 2. User exercise risk level: no risk, the value is 0; low risk, the value is 1; medium risk, the value is 2; high risk, the value is 3.

[0334] The exercise intensity value for "brisk walking" can be determined by querying a knowledge graph of exercise knowledge. For example, the following query can be used to retrieve the exercise intensity value for "brisk walking":

[0335] MATCH (a:Action{name:"walk"}) RETURN a.mets;

[0336] The exercise intensity of "brisk walking" was found to be mets=3.8.

[0337] Then, determine the evaluation index corresponding to "brisk walking" according to the following evaluation index formula. for

[0338]

[0339] The evaluation index corresponding to each movement mode in the exercise combination result can be calculated through a similar process described above, such as "Tai Chi" in the exercise combination result ("brisk walking", "Tai Chi").

[0340] Then, determine the comprehensive evaluation index corresponding to the exercise combination result ("brisk walking", "Tai Chi") according to the following comprehensive evaluation index formula. ,

[0341]

[0342] In the above formula This represents the weighting coefficient corresponding to brisk walking, and its value is, for example, 0.3; This represents the weighting coefficient corresponding to Tai Chi, and its value is, for example, 0.7.

[0343] The exercise program set {[("Brisk Walking", 30), ("Tai Chi" 50)], [("Brisk Walking", 40), ("Tai Chi" 45)], ...} is sorted in descending order based on the comprehensive evaluation index (i.e., sorted from highest to lowest comprehensive evaluation index). Here, [("Brisk Walking", 30), ("Tai Chi" 50)] represents one exercise program, and the set of exercise programs includes multiple programs. According to the comprehensive evaluation index formula mentioned above, a comprehensive evaluation index is calculated for each exercise program, and the programs are sorted from highest to lowest comprehensive evaluation index to obtain the sorted result {[[("Brisk Walking", 40), ("Tai Chi" 45)], ("Brisk Walking", 30), ("Tai Chi" 50)], ...}.

[0344] The top N results from the ranking are selected as the final exercise recommendations. The value of N is related to the user's physical activity level assessment result, and N can be a natural number such as 1, 2, or 3. For example, in the above embodiment, if the user is an elderly person, the top two exercise programs in the ranking results can be selected as the exercise recommendations.

[0345] After the above processing, the exercise combination results are optimized and sorted according to the user's personality characteristics to obtain more personalized recommendation results that are more suitable for the individual user. Compared with the way related technologies recommend exercise to users, the embodiments of this application can accurately provide users with highly applicable exercise combination results, thereby improving the accuracy of exercise recommendation results.

[0346] In some embodiments, such as Figure 16 As shown, exercise plans can also be recommended to users. These plans are a set of exercise programs recommended to the user within a specific exercise plan period, following the recommended exercise frequency. The exercise plan period refers to the length of time the user engages in periodic exercise. For example, an exercise plan period of one week is called a weekly exercise plan. Below is an example of a user with the following health check data: age: 66, gender: male, height: 176cm, weight: 70kg, waist circumference: 88cm. Assume a weekly exercise plan is recommended to this user, with a plan period of 7 days.

[0347] Obtain the exercise plan cycle and exercise frequency. The exercise plan cycle can be user-inputted or set by system default. Exercise frequency refers to the number of times exercise is performed within the exercise plan cycle. Exercise frequency can be determined based on the user's physical examination data and questionnaire data, representing the optimal number of exercises within the exercise plan cycle. For example, if the exercise plan cycle is 7 days, and the user's age is 66 years old based on the physical examination data and the questionnaire data (i.e., the physical activity level assessment result is intermediate), the user's exercise frequency can be determined using a pre-built table showing the relationship between user age, physical activity level assessment result, and exercise frequency. This would determine that the exercise frequency for an elderly person with a moderate physical activity level assessment result is 4, meaning the exercise frequency within the 7-day exercise plan cycle would be 4 times.

[0348] In some embodiments, the exercise frequency can also be determined based on a pre-built table relating user age, physical activity level assessment results, exercise type, and exercise frequency. For example, if a user is a young adult with an intermediate physical activity level assessment result and exercises aerobic and strength training, then the exercise frequency for aerobic exercise is determined to be 6 times and the exercise frequency for strength training is 4 times within a 7-day exercise plan period.

[0349] When allocating recommended exercise duration corresponding to the exercise mode based on the exercise plan cycle, the exercise plan cycle and exercise frequency are obtained, and the time array is determined based on the exercise plan cycle and exercise frequency.

[0350] A time array refers to a set of exercise durations corresponding to exercise frequency. The time array can be expressed in various ways, including but not limited to [exercise duration, exercise frequency], [recommended exercise duration], etc. The following formula for determining the time array yields a time array that meets the specified conditions:

[0351]

[0352] As shown in the formula above, assuming the time array is T=[t i [f], where the exercise frequency is f times, the minimum activity level v1 corresponding to the exercise plan cycle, the maximum activity level v2 corresponding to the exercise plan cycle, and t. i This indicates the suggested duration for each exercise session.

[0353] The maximum and minimum activity levels corresponding to the exercise plan cycle can be obtained by querying a pre-built exercise knowledge graph based on user physical examination data, questionnaire data, and exercise risk assessment results.

[0354] Among them, t i The time units for v1 and v2 are minutes, while the unit for f is days or times.

[0355] Assuming the preset time selection range is T 范围=[10,15,20,……,90]. t i The recommended exercise duration can be repeatedly obtained from the selected time range to obtain a time array.

[0356] For example, a time array of [35, 4] indicates that four 35-minute intervals are selected from the time range, and the recommended exercise duration for each exercise type is 35 minutes. "Four" means the user needs to exercise four times within the exercise plan cycle.

[0357] For example, the time array [15,25,35,45] means that 15 minutes, 25 minutes, 35 minutes, and 45 minutes are selected from the time range as the recommended exercise duration corresponding to the exercise method.

[0358] Assuming that the time array is determined according to the above formula, the user's time array T=[35,4] is obtained.

[0359] Assign target dates corresponding to exercise combination results within the exercise planning cycle according to the time array T, including:

[0360] Determine the recommended exercise duration for each exercise mode in the exercise combination results;

[0361] Determine the target date corresponding to the exercise combination results in the exercise plan cycle based on the exercise frequency.

[0362] For example, based on the user's basic information, a pre-built exercise knowledge graph is queried to obtain a list of target exercise methods and a list of evaluation attributes corresponding to the target exercise methods. The target exercise methods are then sorted according to their corresponding evaluation attributes to obtain a sorted list of target exercise methods. Then, based on the user's age of 66, the user is determined to be an elderly person, and the number of exercise methods corresponding to the elderly person is either one or two, based on a preset limit.

[0363] In some embodiments, it is assumed that one movement mode is obtained from the sorted list of target movement modes as the movement combination result, namely "brisk walking";

[0364] When the number of exercise combination results n is 1, the recommended exercise duration for each exercise mode in the exercise combination results is determined as follows:

[0365] If the exercise combination result is "brisk walking", it is recommended to exercise for 35 minutes each time, 4 times a week;

[0366] If the exercise combination result is "Tai Chi", it is recommended to exercise for 40 minutes each time, 4 times a week;

[0367] Then, based on the exercise frequency, determine the target dates corresponding to the exercise combination results in the exercise plan cycle, that is, determine the target dates corresponding to "brisk walking" as follows:

[0368] Monday: 35-minute brisk walk

[0369] Wednesday: 35-minute brisk walk

[0370] Thursday: 35-minute brisk walk

[0371] Saturday: 35-minute brisk walk

[0372] The target dates for "Tai Chi Chuan" are as follows:

[0373] Tuesday: 40 minutes of Tai Chi

[0374] Wednesday: 40 minutes of Tai Chi

[0375] Friday: 40 minutes of Tai Chi

[0376] Sunday: 40 minutes of Tai Chi

[0377] In some embodiments, it is assumed that two movement modes are obtained from the sorted list of target movement modes as the movement combination result, that is, the movement combination result is ("brisk walking", "Tai Chi").

[0378] The number of exercise combination results, n, is 2. The exercise combination result is ("brisk walking", "Tai Chi"). Assuming the recommended total exercise time corresponding to the exercise combination result is 45 minutes, the recommended exercise time corresponding to each exercise method is determined according to the number of exercise methods in the exercise combination result. Then the recommended exercise time corresponding to ("brisk walking", "Tai Chi") is [("brisk walking", 25 minutes; ("Tai Chi", 20 minutes))].

[0379] For this user, the target dates corresponding to the exercise combinations within the exercise plan cycle based on exercise frequency are determined as follows: Specifically, the target dates for ("brisk walking", "Tai Chi") are determined as follows:

[0380] Monday: 25 minutes of brisk walking; 20 minutes of Tai Chi.

[0381] Wednesday: 25 minutes of brisk walking; 20 minutes of Tai Chi.

[0382] Thursday: 25 minutes of brisk walking; 20 minutes of Tai Chi.

[0383] Saturday: 25 minutes of brisk walking; 20 minutes of Tai Chi.

[0384] In some embodiments, the exercise combination results can also be categorized according to the type of exercise, and then the exercise frequency corresponding to the type of exercise can be determined. For example, for aerobic exercise f=3, [1, 3, 5] can be selected, that is, exercise can be performed on Monday, Wednesday, and Friday. When strength exercise f=1, strength training is performed on a randomly selected day when no aerobic exercise was performed. When strength exercise f=2, strength training is performed on the day with the least aerobic exercise time, and another strength training is performed on a randomly selected day when no aerobic exercise was performed. When strength exercise f=3, strength training is performed on 1-2 days with the least aerobic exercise time, and the remaining strength training is performed on a randomly selected day when no aerobic exercise was performed.

[0385] like Figure 15 The interface shown displays recommended exercise programs for the user. These recommendations include the exercise type, the corresponding exercise method, and the exercise duration. Users can determine the appropriate exercise duration, frequency, and intensity based on the exercise category. This interface can be a web-based version or the interface of a pre-installed application on a mobile device.

[0386] For male users aged 66, the target exercise plan date can be selected based on the type of exercise and the exercise frequency f within a preset duration. For example, if the aerobic exercise frequency f=4, Monday, Tuesday, Thursday, and Friday can be selected for exercise.

[0387] The embodiments of this application can effectively improve the accuracy of exercise recommendation results through the above-described methods.

[0388] In some embodiments, when a user uses the product, a query interface may also be provided to the user based on the user's needs, such as... Figure 12 The visual interface shown allows users to search for information related to a particular type of exercise by directly entering the name of the exercise as a keyword. For example, entering "Tai Chi" will retrieve information such as the exercise category, calorie consumption, exercise intensity, metabolic equivalent (METs) value, and activity equivalent per thousand steps.

[0389] In some embodiments, when a user uses the product, other query interfaces may also be provided to the user according to their needs, such as calculation of estimated calorie consumption during exercise, etc. Figure 13 The interactive interface shown allows users to select information displayed on the interface or input their weight, exercise duration, and exercise type through the input interface provided within the interface to determine the estimated calories burned for the corresponding exercise type.

[0390] Upon receiving input query parameters (such as weight, exercise duration, and exercise type);

[0391] Fill the query parameters into the exercise estimated calorie consumption calculation template to obtain the exercise estimated calorie consumption corresponding to the query parameters.

[0392] For example, the template for calculating estimated calorie expenditure during exercise is: weight * exercise duration * metabolic equivalent * 0.0167.

[0393] The metabolic equivalent (METs) can be obtained by querying a pre-built exercise knowledge graph based on the exercise mode. For example, if the exercise mode is cycling, the metabolic equivalent (METs) corresponding to cycling can be obtained by querying the fourth relation of the exercise knowledge graph. For example, the metabolic equivalent for cycling (slow speed, 16-19.2 km / h) is 6.0.

[0394] Then, the query parameters obtained from the interface are filled into the exercise prediction calorie consumption calculation template to obtain, as shown below. Figure 13 The query results are shown above. In the calculation template above, * represents a multiplication sign, weight is in kilograms, exercise duration is in minutes, and metabolic equivalent (METs) is a floating-point number. For example, if a person weighs 60 kilograms, exercises for 40 minutes, and has a metabolic equivalent of 6.0, the estimated calorie expenditure is 240 kcal.

[0395] To meet the needs of different users, some embodiments of this application also provide a voice interaction method. For example, when a user requests to "play the exercise plan" (e.g., a voice request), the system will automatically play the exercise plan for the day. The exercise plan is displayed on the screen in real time, along with a video of the movements, guiding the user to perform the exercise plan according to the standard movements. The screen can be a home television, projector, etc.

[0396] In some embodiments of this application, the exercise heart rate can also be queried, and the maximum heart rate can be calculated based on the exercise heart rate of healthy individuals using an exercise heart rate calculation template. For example, when the exercise level is low, the exercise heart rate = (220 - age) * (50%~60%); when the exercise level is moderate, the exercise heart rate = (220 - age) * (60%~70%); and when the exercise level is high, the exercise heart rate = (220 - age) * (70%~80%).

[0397] For patients with chronic diseases, sub-health conditions, and other illnesses, the exercise frequency range recommended by the exercise prescription compiled from the guidelines should be used for calculation.

[0398] In some embodiments of this application, the exercise frequency can also be queried and determined using an exercise frequency calculation template based on exercise level. When the exercise level is low, the aerobic exercise frequency f=3 and the strength exercise frequency f=1; when the exercise level is moderate, the aerobic exercise frequency f=4 and the strength exercise frequency f=2; when the exercise level is high, the aerobic exercise frequency f=5 and the strength exercise frequency f=2~3. Stretching exercises are required before and after exercise.

[0399] like Figure 8 As shown, in some embodiments of this application, a method for querying and recommending strength training exercises can also be provided to the user:

[0400] Each session is divided into three body parts ("back", "abs", "thighs"], "shoulder", "neck", "arms"], and "hips", "chest", and "calves"], with level = "beginner" and equipment = "bodyweight training". The system queries a health knowledge graph based on the body part, level, and equipment fields. From the results, three exercise names are selected sequentially for each body part, resulting in nine exercises per session (3*3=9). The exercise intensity is then mapped according to the user's fitness level (low-low intensity, moderate-moderate intensity, high-high intensity). The exercise knowledge graph is then queried to return information such as exercise type, muscle groups, key points of the movement, and equipment requirements.

[0401] Figure 17 This application illustrates some embodiments of exercise plans categorized by exercise type, derived from user physical examination data and physical activity level assessments. Under each exercise type, exercise methods corresponding to the exercise frequency are assigned. For example, aerobic exercise is recommended for 6 days a week, with a recommended duration of 75 minutes per day. Similarly, strength training provides detailed information on the target muscle groups, number of repetitions, and intensity of the exercise program.

[0402] For patients with illnesses, it is necessary to consult the knowledge graph for information on exercise contraindications and precautions. For example, asthma patients cannot exercise in dry or pollen-filled environments. Therefore, the recommendation model should recommend exercise in a relatively humid indoor environment, accurately calculate the exercise heart rate, emphasize that patients should not do high-intensity exercise, and remind patients to stop exercising immediately and provide reasonable treatment suggestions if the intensity is detected to be greater than a certain threshold or if they feel difficulty breathing.

[0403] During the recommendation process, big data analysis is conducted based on online data and exercise reports to extract the most popular live-streamed fitness courses, such as high-efficiency fat burning, rhythmic street dance, and full-body shaping. The top 10 global fitness trends of 2019 are also considered: wearable technology, group training, high-intensity interval training (HIIT), fitness for seniors, bodyweight training, hiring certified coaches, yoga, personal training, functional training, and health-oriented exercise. This knowledge is incorporated into the recommendation model to increase the trendiness and novelty of the recommended exercises.

[0404] The system obtains the weather conditions (humidity, temperature, season, etc.) of the user's location, allows the user to choose the exercise environment (such as indoor or outdoor, exercise equipment or bodyweight), reorders the exercise plan, and adjusts the exercise plan in real time.

[0405] Recommendations include exercise type, intensity, duration, frequency, heart rate (hr), and total weekly exercise time (v).

[0406] If some fields of information about the user in the discrete variables are unknown or unrestricted, exercise plans can be recommended to the user. In this case, for example, if only the user's gender and age are known, big data analysis can be used to identify suitable exercises for men and women of different ages, and then these exercises can be recommended to the user.

[0407] Example of recommendation results:

[0408] User 1: Age = 75, Exercise Type = Aerobic Exercise, Health Status = Diabetes, Other Information Unknown. Recommendations are as follows:

[0409] {

[0410] 'type': 'Aerobic exercise'

[0411] 'action': ['Walk slowly (60-70 steps / minute)'],

[0412] 'intensity': 'low intensity'

[0413] 'hr': '87~122 times / minute'

[0414] 'duration': '30min / time'

[0415] 'frequency': '2 times / day',

[0416] }

[0417] {

[0418] 'type': 'strength training'

[0419] Actions: [Wall push-ups, Standing resistance band curls, Seated resistance band presses, Supine crunches, Standing calf raises, Half squats, Standing high knees with chair back support]

[0420] 'intensity': 'moderate intensity'

[0421] 'hr': ['', '50%~70% of maximum muscle strength'],

[0422] 'frequency': '2-3 sets per day, 2-3 days per week'

[0423] 'part': '6~10',

[0424] 'repeat': '8~12',

[0425] }

[0426] User 2: Given the following data: age = 13, exercise type = aerobic exercise, health status = obese, other information is unknown, then the following recommendations are made:

[0427] {

[0428] 'type': 'Aerobic exercise'

[0429] 'action': ['long walk', 'jogging', 'swimming'],

[0430] 'intensity': 'moderate to medium intensity'

[0431] It should be noted that although the operations of the methods of this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all of the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart may be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0432] Please refer to the following. Figure 19 , Figure 19 A schematic diagram of the structure of the exercise recommendation device provided in an embodiment of this application is shown. The device may include:

[0433] Information acquisition unit 801 is used to acquire basic information of the user;

[0434] The graph query unit 802 is used to query a pre-built motion knowledge graph based on the user's basic information to obtain a list of target motion methods and a list of evaluation attributes corresponding to the list of target motion methods.

[0435] The scheme determination unit 803 is used to determine at least one exercise scheme based on the target exercise mode list and the evaluation attribute list, wherein the exercise scheme includes at least one exercise mode and a recommended exercise duration corresponding to the exercise mode;

[0436] The scheme sorting unit 804 is used to sort the at least one motion scheme according to the evaluation list attribute list to obtain a sorting result;

[0437] The scheme recommendation unit 805 is used to recommend at least one motion scheme that ranks highly in the ranking results.

[0438] The map query unit is used for:

[0439] Construct at least one query statement based on the user's basic information;

[0440] The motion knowledge graph is queried according to each of the at least one query statement to obtain the target motion mode list and the evaluation attribute list corresponding to the target motion mode list.

[0441] The basic information includes: the user's target disease type, target user attributes, and the user's physical activity level assessment results. The atlas query unit is also used for:

[0442] Obtain the user's physical activity level assessment results;

[0443] When the user's physical activity level assessment result is high, the target user attribute and the target symptom type are used as query conditions to construct a query statement; or...

[0444] When the user's physical activity level assessment result is medium or below, the target user attribute and the target disease type are used as query conditions to construct a first query statement;

[0445] The second query statement is constructed by using the user's physical activity level assessment result as the query condition.

[0446] The exercise knowledge graph includes a first relationship between user attribute entities and exercise type entities, and a second relationship between exercise type entities and disease type entities. The graph query unit is also used for:

[0447] When the user's physical activity level assessment result is high, the first and second relationships in the exercise knowledge graph are queried according to the first query statement to obtain the list of target exercise methods; or...

[0448] When the user's physical activity level assessment result is intermediate or below, the first and second relationships in the exercise knowledge graph are queried according to the first query statement to obtain an initial list of exercise methods.

[0449] Based on the second query statement, the exercise knowledge graph is queried to obtain the exercise methods in the initial exercise method list that meet the user's physical activity level assessment results, which are then used as the target exercise method list.

[0450] Obtain the list of evaluation attributes corresponding to the target motion mode list.

[0451] The scheme determination unit is used for:

[0452] Obtain the user's target calorie consumption;

[0453] The motion combination result is obtained by combining at least one motion mode included in the target motion mode list.

[0454] Based on the exercise combination results, extract the evaluation attributes corresponding to the exercise combination results from the evaluation attribute list;

[0455] Based on the evaluation attributes and the target calorie consumption, a recommended time combination is determined corresponding to the exercise combination result. The recommended time combination includes the recommended exercise duration corresponding to each exercise mode in the exercise combination result.

[0456] The exercise combination result and the recommended time combination corresponding to the exercise combination result are used as the exercise plan.

[0457] The scheme determination unit is also used for:

[0458] Obtain the user's exercise frequency within the exercise plan cycle;

[0459] The motion combination result is obtained by combining at least one motion mode included in the target motion mode list.

[0460] Determine the maximum and minimum activity levels corresponding to the exercise plan cycle;

[0461] A time array is determined based on the maximum and minimum activity levels, the time array including the exercise frequency and the recommended exercise duration corresponding to the exercise frequency;

[0462] The time array is allocated to the exercise combination result according to the exercise frequency within the exercise planning cycle.

[0463] The scheme determination unit is also used for:

[0464] Based on the user's physical activity level assessment results, a preset number of exercise methods corresponding to the user's physical activity level assessment results are obtained from the target exercise method list, and these are used as the exercise combination results.

[0465] The evaluation attributes include the unit calorie consumption corresponding to the exercise mode, and the scheme ranking unit is used for:

[0466] Based on the fact that the sum of the products of the unit calorie consumption and the recommended duration for each exercise mode included in the exercise combination result equals the target calorie consumption, the recommended time combination corresponding to the exercise mode combination is determined.

[0467] The evaluation attribute list includes multiple evaluation attributes corresponding to the motion mode, and the scheme sorting unit is used for:

[0468] The user's exercise risk level is determined based on the user's basic information;

[0469] The at least one exercise program is ranked according to the multiple evaluation attributes and exercise risk level attributes.

[0470] The exercise plan includes at least one exercise mode and a recommended exercise duration for each exercise mode. The plan sorting unit is used for:

[0471] Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode;

[0472] The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated.

[0473] The product is calculated based on the value of the risk level corresponding to each exercise mode and the weighting coefficient corresponding to the risk level.

[0474] The product result is used as an evaluation index for each movement mode in the at least one movement scheme;

[0475] The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode;

[0476] The at least one exercise program is ranked according to the comprehensive evaluation index.

[0477] The exercise plan includes at least one exercise mode and a recommended exercise duration for each exercise mode. The plan sorting unit is used for:

[0478] Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode;

[0479] The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated.

[0480] The product is calculated by combining the value of the risk level corresponding to each exercise mode with the weighting coefficient corresponding to the risk level.

[0481] The product is calculated based on the recommended exercise duration for each exercise mode and the weighting coefficient corresponding to the recommended exercise duration;

[0482] The product result is used as an evaluation index for each movement mode in the at least one movement scheme;

[0483] The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode;

[0484] The at least one exercise program is ranked according to the comprehensive evaluation index.

[0485] The device also includes:

[0486] A sports risk determination unit is used to determine the user's sports risk level based on the user's basic information;

[0487] The recommended range determination unit is used to determine the range of recommended exercise methods based on the exercise risk level.

[0488] The exercise risk determination unit is used to input the user's basic information into a pre-built exercise risk classification model to obtain the user's exercise risk level.

[0489] The user's basic information includes the target disease type, target user attributes, and physical activity level assessment results. The basic information acquisition unit is used for:

[0490] The target disease type, target user attributes, and physical activity level assessment results are obtained through electronic questionnaires; or...

[0491] Obtain the user's target disease type and target user attributes through the user's physical examination data;

[0492] The results of the physical activity level assessment of users were obtained through electronic questionnaires.

[0493] The electronic questionnaire can be presented through a human-computer interactive display interface or a voice dialogue method.

[0494] Compared to related technologies that recommend exercise to users, the embodiments of this application can accurately provide users with highly applicable exercise combinations, thereby improving the accuracy of exercise recommendation results.

[0495] It should be understood that the units or modules described in the above-mentioned device are the same as those in the reference. Figure 2 The steps in the described method correspond to each other. Therefore, the operations and features described above for the method also apply to the above-described device and the units contained therein, and will not be repeated here. The above-described device can be pre-implemented in the browser or other security applications of an electronic device, or it can be loaded into the browser or other security applications of an electronic device through download or other means. The corresponding units in the above-described device can cooperate with the units in the electronic device to implement the solutions of the embodiments of this application.

[0496] The division of modules or units mentioned in the detailed description above is not mandatory. In fact, according to the embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0497] The following is for reference. Figure 20 , Figure 20 A schematic diagram of the structure of a computer system suitable for implementing the terminal device or server of the present application is shown.

[0498] like Figure 20 As shown, the computer system includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 902 or programs loaded from storage section 908 into random access memory (RAM) 903. RAM 903 also stores various programs and data required for the operation of system 900. CPU 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0499] The following components are connected to I / O interface 905: an input section 906 including a keyboard, mouse, etc.; an output section 907 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN card, modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to I / O interface 905 as needed. A removable medium 911, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 910 as needed so that computer programs read from it can be installed into storage section 908 as needed.

[0500] Specifically, according to embodiments of this disclosure, the flowcharts above refer to... Figure 2 The described process can be implemented as a computer software program. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a machine-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication section 909, and / or installed from removable medium 911. When the computer program is executed by central processing unit (CPU) 901, it performs the functions defined in the system of this application.

[0501] It should be noted that the computer-readable medium disclosed herein may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0502] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0503] The units or modules described in the embodiments of this application can be implemented in software or hardware. The described units or modules can also be housed in a processor; for example, a processor can be described as including an information acquisition unit, a map query unit, a scheme determination unit, a scheme ranking unit, and a scheme recommendation unit. The names of these units or modules do not necessarily limit the unit or module itself; for example, the information acquisition unit can also be described as "a unit for acquiring basic user information."

[0504] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium stores one or more programs that are used by one or more processors to execute the motion recommendation method described in this application.

[0505] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for recommending exercise, characterized in that, The method includes: Obtain basic user information; Based on the user's basic information, a pre-constructed motion knowledge graph is queried to obtain a list of target motion methods and a list of evaluation attributes corresponding to the list of target motion methods. At least one exercise plan is determined based on the target exercise mode list and the evaluation attribute list, wherein the exercise plan includes at least one exercise mode and a recommended exercise duration corresponding to the exercise mode; The at least one motion scheme is sorted according to the evaluation attribute list to obtain the sorting result; Recommend at least one exercise program based on the ranking results; Prior to querying the pre-built motion knowledge graph based on the user's basic information, the method further includes: The user's exercise risk level is determined based on the user's basic information; The recommended exercise methods are determined based on the aforementioned exercise risk level.

2. The method according to claim 1, characterized in that, The step of querying a pre-constructed motion knowledge graph based on the user's basic information to obtain a list of target motion methods and a list of evaluation attributes corresponding to the target motion method list includes: Construct at least one query statement based on the user's basic information; The motion knowledge graph is queried according to each of the at least one query statement to obtain the target motion mode list and the evaluation attribute list corresponding to the target motion mode list.

3. The method according to claim 2, characterized in that, The basic information includes: the user's target disease type, target user attributes, and the user's physical activity level assessment results. Constructing at least one query statement based on the user's basic information includes: Obtain the user's physical activity level assessment results; When the user's physical activity level assessment result is high, the target user attribute and the target symptom type are used as query conditions to construct a query statement; or... When the user's physical activity level assessment result is medium or below, the target user attribute and the target disease type are used as query conditions to construct a first query statement; The second query statement is constructed by using the user's physical activity level assessment result as the query condition.

4. The method according to claim 3, characterized in that, The exercise knowledge graph includes a first relationship between user attribute entities and exercise type entities, and a second relationship between exercise type entities and disease type entities. The step of querying the exercise knowledge graph according to each of the at least one query statement includes: When the user's physical activity level assessment result is high, the first relationship and the second relationship in the exercise knowledge graph are queried according to the first query statement to obtain the list of target exercise methods; or... When the user's physical activity level assessment result is intermediate or below, the first relationship and the second relationship in the exercise knowledge graph are queried according to the first query statement to obtain an initial list of exercise methods. Based on the second query statement, the exercise knowledge graph is queried to obtain the exercise methods in the initial exercise method list that meet the user's physical activity level assessment results, and these are used as the target exercise method list. Obtain the list of evaluation attributes corresponding to the target motion mode list.

5. The method according to any one of claims 1-4, characterized in that, Determining at least one motion scheme based on the target motion mode list and the evaluation attribute list includes: Obtain the user's target calorie consumption; The motion combination result is obtained by combining at least one motion mode included in the target motion mode list. Based on the exercise combination results, extract the evaluation attributes corresponding to the exercise combination results from the evaluation attribute list; Based on the evaluation attributes and the target calorie consumption, a recommended time combination is determined corresponding to the exercise combination result. The recommended time combination includes the recommended exercise duration corresponding to each exercise mode in the exercise combination result. The exercise combination result and the recommended time combination corresponding to the exercise combination result are used as the exercise plan.

6. The method according to claim 5, characterized in that, The evaluation attributes include the unit calorie expenditure corresponding to the exercise mode. Based on the evaluation attributes and the target calorie expenditure, a recommended time combination corresponding to the exercise mode combination is determined, including: Based on the fact that the sum of the products of the unit calorie consumption and the recommended duration for each exercise mode included in the exercise combination result equals the target calorie consumption, the recommended time combination corresponding to the exercise mode combination is determined.

7. The method according to any one of claims 1-4, characterized in that, Determining at least one motion scheme based on the target motion mode list and the evaluation attribute list includes: Obtain the user's exercise frequency within the exercise plan cycle; The motion combination result is obtained by combining at least one motion mode included in the target motion mode list. Determine the maximum and minimum activity levels corresponding to the exercise plan cycle; A time array is determined based on the maximum and minimum activity levels, the time array including the exercise frequency and the recommended exercise duration corresponding to the exercise frequency; The time array is allocated to the exercise combination result according to the exercise frequency within the exercise planning cycle.

8. The method according to claim 7, characterized in that, The combination of at least one motion mode included in the target motion mode list includes: Based on the user's physical activity level assessment results, a preset number of exercise methods corresponding to the user's physical activity level assessment results are obtained from the target exercise method list, and these are used as the exercise combination results.

9. The method according to claim 8, characterized in that, The evaluation attribute list includes multiple evaluation attributes corresponding to the movement mode. The process of sorting the at least one movement scheme according to the evaluation attribute list includes: The user's exercise risk level is determined based on the user's basic information; The at least one exercise program is ranked according to the multiple evaluation attributes and exercise risk level attributes.

10. The method according to claim 9, characterized in that, The exercise plan includes at least one exercise mode and a recommended exercise duration for each exercise mode. The step of ranking the at least one exercise mode based on the multiple evaluation attributes and exercise risk level attributes includes: Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode; The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated. The product is calculated based on the value of the risk level corresponding to each exercise mode and the weighting coefficient corresponding to the risk level. The product result is used as an evaluation index for each movement mode in the at least one movement scheme; The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode; The at least one exercise program is ranked according to the comprehensive evaluation index.

11. The method according to claim 9, characterized in that, The exercise plan includes at least one exercise mode and a recommended exercise duration for each exercise mode. The step of ranking the at least one exercise mode based on the multiple evaluation attributes and exercise risk level attributes includes: Obtain the weight coefficients corresponding to the multiple evaluation attributes for each motion mode; The product of the weight coefficient corresponding to each evaluation attribute and the value corresponding to each evaluation attribute is calculated. The product is calculated by combining the value of the risk level corresponding to each exercise mode with the weighting coefficient corresponding to the risk level. The product is calculated based on the recommended exercise duration for each exercise mode and the weighting coefficient corresponding to the recommended exercise duration; The product result is used as an evaluation index for each movement mode in the at least one movement scheme; The comprehensive evaluation index for each exercise scheme is determined based on the evaluation index and the weighting coefficient corresponding to each exercise mode; The at least one exercise program is ranked according to the comprehensive evaluation index.

12. The method according to claim 1, characterized in that, Determining the user's exercise risk level based on the user's basic information includes: The user's basic information is input into a pre-built sports risk classification model to obtain the user's sports risk level.

13. The method according to claim 1, characterized in that, The user's basic information includes the target disease type, target user attributes, and physical activity level assessment results. Obtaining the user's basic information includes: The target disease type, target user attributes, and physical activity level assessment results are obtained through electronic questionnaires; or... Obtain the user's target disease type and target user attributes through the user's physical examination data; The results of the physical activity level assessment of users were obtained through electronic questionnaires.

14. The method according to claim 13, characterized in that, The electronic questionnaire can be presented through a human-computer interactive display interface or a voice dialogue method.

15. A sports recommendation device, characterized in that, The device includes: The information acquisition unit is used to acquire the user's basic information. The graph query unit is used to query a pre-built motion knowledge graph based on the user's basic information to obtain a list of target motion methods and a list of evaluation attributes corresponding to the list of target motion methods. The scheme determination unit is used to determine at least one exercise scheme based on the target exercise mode list and the evaluation attribute list, wherein the exercise scheme includes at least one exercise mode and a recommended exercise duration corresponding to the exercise mode; The scheme sorting unit is used to sort the at least one motion scheme according to the evaluation list attribute list to obtain a sorting result; The scheme recommendation unit is used to recommend at least one motion scheme that ranks first in the ranking results. The device further includes: The determining unit is configured to determine the user's exercise risk level based on the user's basic information before querying the pre-constructed exercise knowledge graph based on the user's basic information; and to determine the range of recommended exercise methods based on the exercise risk level.

16. An electronic device, characterized in that, The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor being configured to implement the method as described in any one of claims 1-14 when executing the program.

17. The electronic device according to claim 16, characterized in that, The electronic device further includes: an input device and an output device; The input device is used to obtain basic user information; The output device is used to recommend at least one motion scheme to the user.

18. A computer-readable storage medium having a computer program stored thereon, the computer program being used to implement the method as claimed in any one of claims 1-14.

Images