Customized report text labeling method and system based on exercise health

Abstract

The embodiment of the invention provides a customized report text labeling method and system based on exercise health, and the method comprises the steps: carrying out the preliminary learning through employing the first sample learning data, generating the first network learning parameter information, and carrying out the further learning through combining with the second sample learning data with the same customized knowledge point. And generating second network learning parameter information. And finally, according to the two groups of network learning parameter information, constructing an efficient and accurate target text labeling network. According to the method, the accuracy and efficiency of text labeling are improved, deep understanding and precise labeling of exercise health customized activities are achieved, and a more intelligent and personalized text processing solution is provided for the field of exercise health.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to a method and system for annotating customized report texts based on sports health. Background Art

[0002] In the field of sports health, with the growing demand for personalization, customized sports health reports have become the mainstream of the market. However, traditional report generation methods mostly rely on manual labeling and sorting, which is not only time-consuming and labor-intensive, but also difficult to ensure the consistency and accuracy of the labeling. Especially when processing a large amount of sports health data, the inefficiency of manual labeling is particularly prominent.

[0003] In recent years, although some machine learning-based text annotation methods have been proposed, these methods often only annotate general texts and are difficult to adapt to the specific customized needs in the field of sports and health. In addition, in the learning and annotation process, existing methods usually only consider current data and ignore the importance of historical prior data, resulting in the need to improve the accuracy and reliability of the annotation results. Summary of the invention

[0004] In view of this, the purpose of this application is to provide a customized report text annotation method and system based on sports health.

[0005] In conjunction with the first aspect of the present application, a method for annotating a customized report text based on sports health is provided, which is applied to a customized report text annotating system based on sports health, and the method comprises:

[0006] Performing text annotation learning on the acquired first sample learning data to generate first network learning parameter information; wherein the first sample learning data is the text data of the first sample sports health customized activity and the prior data recorded by the first sample sports health customized activity within a set time period;

[0007] Acquire second sample learning data; wherein the second sample learning data is text data of the second sample sports health customized activity and current data recorded by the second sample sports health customized activity; the customized knowledge points of the first sample sports health customized activity are the same as the customized knowledge points of the second sample sports health customized activity;

[0008] Performing text annotation learning on the second sample learning data to generate second network learning parameter information;

[0009] A target text annotation network is generated according to the first network learning parameter information and the second network learning parameter information.

[0010] In a possible implementation manner of the first aspect, performing text annotation learning on the acquired first sample learning data to generate first network learning parameter information includes:

[0011] Performing text annotation learning on the first sample learning data to generate a first text annotation network;

[0012] Acquire a first network performance parameter of the first text annotation network; wherein the first network performance parameter represents the text annotation accuracy of the first text annotation network;

[0013] The first network performance parameter is verified, and if the first network performance parameter verification passes, the first network learning parameter information of the first text annotation network is obtained.

[0014] In a possible implementation manner of the first aspect, the verifying the first network performance parameter includes:

[0015] Loading first set sample data into the first text annotation network to generate the first network performance parameter; wherein the first set sample data is data used to verify the first network performance parameter;

[0016] If the first network performance parameter is greater than a first threshold parameter value, it is determined that the first network performance parameter verification has passed.

[0017] In a possible implementation manner of the first aspect, performing text annotation learning on the second sample learning data to generate second network learning parameter information includes:

[0018] Obtaining characteristic distribution parameters of the second sample learning data;

[0019] The characteristic distribution parameters are verified, and if the characteristic distribution parameters are verified, text annotation learning is performed on the second sample learning data to generate the second network learning parameter information.

[0020] In a possible implementation of the first aspect, the method further includes:

[0021] If the characteristic distribution parameter verification fails, obtaining third network learning parameter information associated with the second sample learning data from preset network learning parameter information; wherein the third network learning parameter information is generated by the sample learning data with the shortest interval from the generation node of the second sample learning data;

[0022] The target text annotation network is generated according to the third network learning parameter information and the first network learning parameter information.

[0023] In a possible implementation manner of the first aspect, the verifying the characteristic distribution parameter includes:

[0024] Determining the relationship between the characteristic distribution parameters and the target distribution parameters;

[0025] If the characteristic distribution parameter matches the target distribution parameter, it is determined that the characteristic distribution parameter verification has passed.

[0026] In a possible implementation manner of the first aspect, generating a target text annotation network according to the first network learning parameter information and the second network learning parameter information includes:

[0027] Optimizing the first network learning parameter information according to the second network learning parameter information to generate optimized network learning parameter information;

[0028] Generating a second text annotation network according to the optimized network learning parameter information;

[0029] Verifying a second network performance parameter of the second text annotation network; wherein the second network performance parameter represents the text annotation accuracy of the second text annotation network;

[0030] If the second network performance parameter verification passes, the second text annotation network is determined to be the target text annotation network.

[0031] In a possible implementation manner of the first aspect, the verifying the second network performance parameter of the second text annotation network includes:

[0032] Loading second set sample data into the second text annotation network to generate the second network performance parameters; wherein the second set sample data is data used to verify the second network performance parameters;

[0033] If the second network performance parameter is greater than a second threshold parameter value, it is determined that the second network performance parameter verification has passed.

[0034] In a possible implementation of the first aspect, the method further includes:

[0035] Loading the third set sample data into the text annotation network corresponding to the second network learning parameter information to generate a third network performance parameter;

[0036] If the third network performance parameter is greater than a third threshold parameter value, the network learning parameter information corresponding to the second network learning parameter information in the preset network learning parameter information is optimized with the second network learning parameter information and saved.

[0037] In combination with the second aspect of the present application, a customized report text annotation system based on sports health is provided, wherein the customized report text annotation system based on sports health includes a machine-readable storage medium and a processor, wherein the machine-readable storage medium stores machine-executable instructions, and when the processor executes the machine-executable instructions, the customized report text annotation system based on sports health implements the aforementioned customized report text annotation method based on sports health.

[0038] In conjunction with the third aspect of the present application, a computer-readable storage medium is provided, in which computer-executable instructions are stored. When the computer-executable instructions are executed, the aforementioned customized report text annotation method based on sports health is implemented.

[0039] In combination with any of the above aspects, the embodiment of the present application performs preliminary learning using the first sample learning data and generates the first network learning parameter information, and then further learns in combination with the second sample learning data with the same customized knowledge point to generate the second network learning parameter information. Finally, based on these two sets of network learning parameter information, an efficient and accurate target text annotation network is constructed. This method not only improves the accuracy and efficiency of text annotation, but also achieves a deep understanding and accurate annotation of customized sports and health activities, providing a more intelligent and personalized text processing solution for the field of sports and health. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained by combining these drawings without paying creative work.

[0041] Figure 1 A flowchart of a customized report text annotation method based on sports health provided in an embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all of the embodiments. According to the embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or end that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or ends.

[0044] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0045] Figure 1 The flowchart of the customized report text annotation method based on sports health provided by the embodiment of the present application is shown. It should be understood that in other embodiments, the order of some steps in the customized report text annotation method based on sports health of this embodiment can be shared with each other according to actual needs, or some steps can be omitted or maintained. The details of the customized report text annotation method based on sports health include:

[0046] Step S110, performing text annotation learning on the acquired first sample learning data to generate first network learning parameter information, wherein the first sample learning data is the text data of the first sample sports health customized activity and the prior data recorded in the set time period of the first sample sports health customized activity.

[0047] In this embodiment, it is assumed that the server is located in a data center of a large sports health management company, which provides various personalized sports health customized services.

[0048] The first example of customized sports and health activities is a customized "fat loss and fitness" activity for the middle-aged population.

[0049] Its text data includes descriptions of activities, such as "This is a 12-week fat-loss fitness plan, including aerobic exercise, strength training and diet control. Aerobic exercise is mainly jogging three times a week, each time for more than 30 minutes; strength training focuses on core muscles and limb muscles twice a week; in terms of diet control, reduce the intake of high-calorie foods and increase the intake of vegetables and high-quality protein."

[0050] Prior data records some information from similar fat-reducing fitness activities in the past. For example, in the data of the 50 people who participated in the previous fat-reducing fitness activity, their age (average age is 40 years old), gender (60% male, 40% female), initial weight (average initial weight is 75 kg), initial body fat percentage (average initial body fat percentage is 25%), heart rate changes during exercise (average heart rate is 130-140 beats / minute when jogging), gradual increase in weight of strength training (increase of 5%-10% training weight per week), and actual changes in calorie intake (average reduction of 500-800 kcal per week), etc. are recorded.

[0051] The server first annotates the key information in the text data. For example, "fat loss and fitness" is annotated as the activity goal, "12 weeks" is annotated as the activity cycle, and "jogging", "strength training", and "diet control" are annotated as specific intervention methods. For prior data, age, gender, etc. are annotated as basic characteristics of the participants, initial weight, body fat percentage, etc. are annotated as initial state data, and changes in heart rate during exercise, increase in training weight, and changes in calorie intake are annotated as process data.

[0052] The server uses a deep learning framework (such as TensorFlow or PyTorch) to build an initial text annotation network model. This model can be a multi-layer neural network, for example, including an input layer, a hidden layer, and an output layer. The input layer receives the first sample learning data after annotation processing, the hidden layer extracts and transforms the data, and the output layer outputs the results related to the annotation task, such as predicting the trend of body fat percentage changes in the next time period or evaluating the effect of exercise.

[0053] During the learning process, the server uses the labeled data to train the network. For example, the back propagation algorithm is used to adjust the weights and biases of the network. For each input labeled data point, the error between the network output and the actual labeled result is calculated, and then the network parameters are updated based on this error, so that the network output gradually approaches the correct labeled result.

[0054] When the network training reaches a certain stage, the server obtains the first network performance parameter of the first text annotation network. For example, the server uses a validation set containing 100 new first sample sports health customized activity data (these data have the same structure as the data used for training before, but different data content) to test the network performance.

[0055] For this validation set, the server inputs each data point into the trained first text annotation network, and the network outputs the corresponding annotation result. The server then calculates the accuracy between the predicted annotation result and the actual annotation result. This accuracy is the first network performance parameter. Assuming that in this validation set, the network correctly annotates 80 data points, then the first network performance parameter (accuracy) is 80%.

[0056] The server loads the first set sample data into the first text annotation network to generate a first network performance parameter. Assume that the first threshold parameter value is set to 75%.

[0057] Since the first network performance parameter calculated previously is 80%, and 80%>75%, it is determined that the first network performance parameter verification has passed.

[0058] Once the verification is passed, the server obtains the first network learning parameter information of the first text annotation network. These parameter information include various parameters such as network weights and biases, which are obtained after the network is trained and can be used for subsequent network optimization or fusion with other networks.

[0059] Step S120, obtaining second sample learning data. The second sample learning data is the text data of the second sample sports health customized activity and the current data recorded by the second sample sports health customized activity. The customized knowledge points of the first sample sports health customized activity are the same as the customized knowledge points of the second sample sports health customized activity.

[0060] Continuing with the example of the sports and health management company, the second example of customized sports and health activities is also a customized "fat loss and fitness" activity for the middle-aged population.

[0061] The text data may be as follows: "This is a new fat-reducing fitness activity, which also lasts for 12 weeks and includes aerobic exercise such as alternating brisk walking and jogging four times a week, 25-35 minutes each time; strength training focuses on balanced training of upper and lower body muscles three times a week; the diet follows the principle of low carbohydrates, high fiber and moderate protein."

[0062] The current data records the relevant information of the participants who are currently doing this activity. For example, there are currently 30 participants, their ages are between 38 and 45 years old, the gender ratio is 55% male and 45% female, the average initial weight is 72 kg, the average initial body fat rate is 23%, the average heart rate during brisk walking and jogging is 125-135 beats / minute, the number of repetitions and sets of each movement in strength training, and the actual daily intake of various nutrients in the diet.

[0063] The server queries and obtains text data and current data related to the second example sports health customized activity from the company's database. This database stores relevant information of all sports health customized activities. The server accurately obtains the required data based on the activity type (fat loss and fitness) and target population (middle-aged group).

[0064] Step S130: Perform text annotation learning on the second sample learning data to generate second network learning parameter information.

[0065] In this embodiment, the server first analyzes the second sample learning data to obtain its characteristic distribution parameters. For example, for the age feature of the participant, the server calculates the mean, variance and other statistics of the age. Assume that the mean of age is 42 years old and the variance is 5. For the weight feature, the mean is 72 kg and the variance is 3 kg. For the heart rate feature during exercise, the mean is 130 beats / minute and the variance is 5 beats / minute, etc. These statistics constitute the characteristic distribution parameters.

[0066] Next, we can determine the relationship between the characteristic distribution parameters and the target distribution parameters. Assume that the target distribution parameters are set based on the ideal data distribution of fat loss and fitness activities for middle-aged people that the company has accumulated over a long period of time. For example, the ideal mean age is 40 years old and the variance is 4; the ideal mean weight is 70 kg and the variance is 4 kg; the ideal mean heart rate is 128 beats / minute and the variance is 4 beats / minute.

[0067] The server compares the actual feature distribution parameters with the target distribution parameters. For age, although the mean of 42 years old deviates from the target mean of 40 years old, considering factors such as variance, if the overall deviation is within an acceptable range (for example, by setting a comprehensive deviation calculation method and calculating that the total deviation is less than a certain threshold), the feature distribution parameter verification is determined to have passed.

[0068] If the feature distribution parameter verification passes, the server performs a text annotation learning process similar to step S110 on the second sample learning data.

[0069] The server builds a new text annotation network (it can also be initialized and adjusted based on the previous network structure) and inputs the annotated second sample learning data into this network for training. The back propagation algorithm and other methods are also used to adjust the weights and biases of the network. As the training progresses, the network gradually learns the characteristics and rules in the second sample learning data, and finally obtains the second network learning parameter information of the trained network, including the network weights, biases and other parameters.

[0070] Step S140: generating a target text annotation network according to the first network learning parameter information and the second network learning parameter information.

[0071] In this embodiment, the server integrates and optimizes the first network learning parameter information and the second network learning parameter information. For example, assuming that a certain weight in the first network learning parameter information is [0.3, 0.5, 0.2], and the corresponding weight in the second network learning parameter information is [0.4, 0.4, 0.2]. The server can use a weighted average method to optimize this weight, such as the new weight is [(0.3+0.4) / 2, (0.5+0.4) / 2, (0.2+0.2) / 2] = [0.35, 0.45, 0.2]. In this way, other parameters in the first network learning parameter information (such as bias, etc.) are also optimized in a similar manner to generate optimized network learning parameter information.

[0072] The server uses the optimized network learning parameter information to initialize a new text annotation network, which is the second text annotation network. The network structure can be the same as the previous network structure, but because it is initialized with optimized parameters, its initial performance will be closer to expectations.

[0073] The server loads the second set sample data into the second text annotation network to generate a second network performance parameter. Assume that the second set sample data is another 100 data points related to the fat loss fitness activity.

[0074] The server inputs these data points into the second text annotation network, and calculates the accuracy between the annotation result output by the network and the actual annotation result. Assuming that the accuracy is 85%, the second threshold parameter value is set to 80%.

[0075] Since 85%>80%, it is determined that the second network performance parameter verification has passed.

[0076] If the second network performance parameter is verified, the second text annotation network is determined as the target text annotation network. This target text annotation network can be used to annotate text data of new sports health customized activities, thereby providing sports health management companies with more accurate activity analysis, participant status prediction and other services.

[0077] The server may also load the third set sample data into the text annotation network corresponding to the second network learning parameter information to generate the third network performance parameter. For example, the third set sample data is another batch of data from different sources but related to fat loss fitness activities.

[0078] If the third network performance parameter is greater than the third threshold parameter value, the network learning parameter information corresponding to the second network learning parameter information in the preset network learning parameter information is optimized with the second network learning parameter information and saved. Assuming that the third threshold parameter value is 70%, if the calculated third network performance parameter is 75%, then the optimization and saving operations are performed in accordance with the regulations so that more optimized network learning parameter information can be used in future learning and applications.

[0079] Based on the above steps, the embodiment of the present application performs preliminary learning using the first sample learning data and generates the first network learning parameter information, and then further learns in combination with the second sample learning data with the same customized knowledge points to generate the second network learning parameter information. Finally, based on these two sets of network learning parameter information, an efficient and accurate target text annotation network is constructed. This method not only improves the accuracy and efficiency of text annotation, but also achieves a deep understanding and accurate annotation of customized sports and health activities, providing a more intelligent and personalized text processing solution for the field of sports and health.

[0080] In a possible implementation manner, performing text annotation learning on the acquired first sample learning data to generate first network learning parameter information includes:

[0081] Perform text annotation learning on the first sample learning data to generate a first text annotation network.

[0082] A first network performance parameter of the first text annotation network is obtained, wherein the first network performance parameter represents the text annotation accuracy of the first text annotation network.

[0083] The first network performance parameter is verified, and if the first network performance parameter verification passes, the first network learning parameter information of the first text annotation network is obtained.

[0084] In a possible implementation manner, the verifying the first network performance parameter includes:

[0085] Loading first set sample data into the first text annotation network to generate the first network performance parameter, wherein the first set sample data is data used to verify the first network performance parameter.

[0086] If the first network performance parameter is greater than a first threshold parameter value, it is determined that the first network performance parameter verification has passed.

[0087] In a possible implementation manner, performing text annotation learning on the second sample learning data to generate second network learning parameter information includes:

[0088] Obtain characteristic distribution parameters of the second sample learning data.

[0089] The characteristic distribution parameters are verified, and if the characteristic distribution parameters are verified, text annotation learning is performed on the second sample learning data to generate the second network learning parameter information.

[0090] In a possible implementation, the method further includes:

[0091] If the characteristic distribution parameter verification fails, third network learning parameter information associated with the second sample learning data is obtained from preset network learning parameter information, wherein the third network learning parameter information is generated by the sample learning data with the shortest interval from the generation node generating the second sample learning data.

[0092] The target text annotation network is generated according to the third network learning parameter information and the first network learning parameter information.

[0093] In a possible implementation, the verifying the characteristic distribution parameter includes:

[0094] A relationship between the characteristic distribution parameters and the target distribution parameters is determined.

[0095] If the characteristic distribution parameter matches the target distribution parameter, it is determined that the characteristic distribution parameter verification has passed.

[0096] In a possible implementation, generating a target text annotation network according to the first network learning parameter information and the second network learning parameter information includes:

[0097] The first network learning parameter information is optimized according to the second network learning parameter information to generate optimized network learning parameter information.

[0098] A second text annotation network is generated according to the optimized network learning parameter information.

[0099] A second network performance parameter of the second text annotation network is verified, wherein the second network performance parameter represents the text annotation accuracy of the second text annotation network.

[0100] If the second network performance parameter verification passes, the second text annotation network is determined to be the target text annotation network.

[0101] In a possible implementation manner, the verifying the second network performance parameter of the second text annotation network includes:

[0102] The second set sample data is loaded into the second text annotation network to generate the second network performance parameter, wherein the second set sample data is data used to verify the second network performance parameter.

[0103] If the second network performance parameter is greater than a second threshold parameter value, it is determined that the second network performance parameter verification has passed.

[0104] In a possible implementation, the method further includes:

[0105] The third set sample data is loaded into the text annotation network corresponding to the second network learning parameter information to generate third network performance parameters.

[0106] If the third network performance parameter is greater than a third threshold parameter value, the network learning parameter information corresponding to the second network learning parameter information in the preset network learning parameter information is optimized with the second network learning parameter information and saved.

[0107] In the above embodiment, the sports health-based customized report text annotation system for executing the above method embodiment has at least one processor, a control module (chip set) coupled to at least one of the (at least one) processors, a memory coupled to the control module, a non-volatile memory (NVM) / storage device coupled to the control module, at least one input / output device coupled to the control module, and a network interface coupled to the control module.

[0108] The processor may include at least one single-core or multi-core processor, and the processor may include any combination of a general-purpose processor or a dedicated processor (e.g., a graphics processor, an application processor, a baseband processor, etc.). For some alternative implementations, the customized report text annotation system based on sports health can be used as a customized report text annotation system device based on sports health such as the gateway described in the embodiments of the present application.

[0109] For some alternative embodiments, a customized report text annotation system based on sports health may include at least one computer-readable medium (e.g., a memory or NVM / storage device) having instructions and at least one processor integrated with the at least one computer-readable medium and configured to execute instructions to implement a module to perform the actions described in the present disclosure.

[0110] For one embodiment, the control module may include any suitable interface controller to provide any suitable interface to at least one of the processor(s) and / or any suitable device or component in communication with the control module.

[0111] The control module may include a memory controller module to provide an interface to the memory. The memory controller module may be a hardware module, a software module, and / or a firmware module.

[0112] The memory may be used, for example, to load and store data and / or instructions for a sports health-based customized report text annotation system. For one embodiment, the memory may include any suitable volatile memory, for example, a suitable DRAM.

[0113] For one embodiment, the control module may include at least one input / output controller to provide an interface to the NVM / storage device and the (at least one) input / output device.

[0114] For example, the NVM / storage device may be used to store data and / or instructions. The NVM / storage device may include any suitable non-volatile memory (e.g., flash memory) and / or may include any suitable (at least one) non-volatile storage device (e.g., at least one hard disk drive (HDD), at least one compact disk (CD) drive, and / or at least one digital versatile disk (DVD) drive).

[0115] The NVM / storage device may include a storage resource that is physically part of a device on which the sports health-based customized report text annotation system is installed, or it may be accessible to the device without being part of the device. For example, the NVM / storage device may be accessed via (at least one) input / output device based on a network.

[0116] (At least one) input / output device may provide an interface for the sports health-based customized report text annotation system to communicate with any other appropriate device, and the input / output device may include a communication component, a pinyin component, a sensor component, etc. The network interface may provide an interface for the sports health-based customized report text annotation system to predict network communication based on at least one user preference, and the sports health-based customized report text annotation system may wirelessly communicate with at least one component of a wireless network based on any standard and / or protocol of at least one wireless network standard and / or protocol, such as accessing a wireless network based on a communication standard.

[0117] For one embodiment, at least one of the (at least one) processors may be loaded together with the logic of at least one controller of a control module (e.g., a memory controller module). For one embodiment, at least one of the (at least one) processors may be loaded together with the logic of at least one controller of a control module to form a system level load. For one embodiment, at least one of the (at least one) processors may be fused on the same die with the logic of at least one controller of a control module. For one embodiment, at least one of the (at least one) processors may be fused on the same die with the logic of at least one controller of a control module to form a system on chip (SoC).

[0118] The embodiments of the present application are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core idea of ​​the present application. At the same time, for general technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

[0119] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the method for annotating a customized report text based on sports health described in the aforementioned embodiment.

[0120] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps in the customized report text annotation method based on sports health described in the above embodiment.

[0121] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple user preference prediction network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0122] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, a magnetic disk storage, a magnetic tape storage, or any other medium that can be used to be computer-readable with or store data.

[0123] Finally, it should be noted that what is disclosed above is only a preferred embodiment of the present invention, which is only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it should be understood by those skilled in the art that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A customized report text annotation method based on sports health, characterized in that: The method comprises: Performing text annotation learning on the acquired first sample learning data to generate first network learning parameter information; wherein the first sample learning data is the text data of the first sample sports health customized activity and the prior data recorded by the first sample sports health customized activity within a set time period; Acquire second sample learning data; wherein the second sample learning data is text data of the second sample sports health customized activity and current data recorded by the second sample sports health customized activity; the customized knowledge points of the first sample sports health customized activity are the same as the customized knowledge points of the second sample sports health customized activity; Performing text annotation learning on the second sample learning data to generate second network learning parameter information; A target text annotation network is generated according to the first network learning parameter information and the second network learning parameter information.

2. The customized report text annotation method based on sports health according to claim 1 is characterized in that: The step of performing text annotation learning on the acquired first sample learning data to generate first network learning parameter information includes: Performing text annotation learning on the first sample learning data to generate a first text annotation network; Acquire a first network performance parameter of the first text annotation network; wherein the first network performance parameter represents the text annotation accuracy of the first text annotation network; The first network performance parameter is verified, and if the first network performance parameter verification passes, the first network learning parameter information of the first text annotation network is obtained.

3. The customized report text annotation method based on sports health according to claim 2 is characterized in that: The verifying the first network performance parameter includes: Loading first set sample data into the first text annotation network to generate the first network performance parameter; wherein the first set sample data is data used to verify the first network performance parameter; If the first network performance parameter is greater than a first threshold parameter value, it is determined that the first network performance parameter verification has passed.

4. The customized report text annotation method based on sports health according to claim 1 is characterized in that: The performing text annotation learning on the second sample learning data to generate second network learning parameter information includes: Obtaining characteristic distribution parameters of the second sample learning data; The characteristic distribution parameters are verified, and if the characteristic distribution parameters are verified, text annotation learning is performed on the second sample learning data to generate the second network learning parameter information.

5. The customized report text annotation method based on sports health according to claim 4 is characterized in that: The method further comprises: If the characteristic distribution parameter verification fails, obtaining third network learning parameter information associated with the second sample learning data from preset network learning parameter information; wherein the third network learning parameter information is generated by the sample learning data with the shortest interval from the generation node of the second sample learning data; The target text annotation network is generated according to the third network learning parameter information and the first network learning parameter information.

6. The customized report text annotation method based on sports health according to claim 4 is characterized in that: The verifying of the characteristic distribution parameters includes: Determining the relationship between the characteristic distribution parameters and the target distribution parameters; If the characteristic distribution parameter matches the target distribution parameter, it is determined that the characteristic distribution parameter verification has passed.

7. The customized report text annotation method based on sports health according to claim 1 is characterized in that: The step of generating a target text annotation network according to the first network learning parameter information and the second network learning parameter information includes: Optimizing the first network learning parameter information according to the second network learning parameter information to generate optimized network learning parameter information; Generating a second text annotation network according to the optimized network learning parameter information; Verifying a second network performance parameter of the second text annotation network; wherein the second network performance parameter represents the text annotation accuracy of the second text annotation network; If the second network performance parameter verification passes, the second text annotation network is determined to be the target text annotation network.

8. The customized report text annotation method based on sports health according to claim 7 is characterized in that: The verifying the second network performance parameter of the second text annotation network includes: Loading second set sample data into the second text annotation network to generate the second network performance parameters; wherein the second set sample data is data used to verify the second network performance parameters; If the second network performance parameter is greater than a second threshold parameter value, it is determined that the second network performance parameter verification has passed.

9. The customized report text annotation method based on sports health according to claim 4 is characterized in that: The method further comprises: Loading the third set sample data into the text annotation network corresponding to the second network learning parameter information to generate a third network performance parameter; If the third network performance parameter is greater than a third threshold parameter value, the network learning parameter information corresponding to the second network learning parameter information in the preset network learning parameter information is optimized with the second network learning parameter information and saved.

10. A customized report text annotation system based on sports health, characterized in that: It includes a processor and a computer-readable storage medium, wherein the computer-readable storage medium stores machine-executable instructions, and when the machine-executable instructions are executed by a computer, the customized report text annotation method based on sports health described in any one of claims 1-9 is implemented.

Images