Intelligent recommendation system based on personalized archives of old people

Abstract

The invention belongs to the technical field of intelligent recommendation. The invention provides an intelligent recommendation system based on personalized archives of old people, and the system comprises an archive obtaining module which is used for calling the personalized archives of target old people and transmitting the personalized archives to a push information screening module; the push information screening module is used for receiving the personalized archive, performing semantic understanding on the personalized archive by using a semantic analysis plug-in, and extracting a plurality of pieces of interest information from the personalized archive; screening according to each piece of interest information to obtain a plurality of pieces of corresponding recommendation information, and sorting each piece of recommendation information; and the information pushing module is used for pushing each piece of recommendation information to terminal equipment associated with the target old people based on the sorting processing result, so that the terminal equipment outputs each piece of recommendation information. According to the method, the most suitable recommendation information can be decided according to the personalized archives of the old people, the pushing accuracy is higher, and richer mental life can be brought to the old people.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent recommendation, and more particularly to an intelligent recommendation system based on the personalized profile of the elderly. Background Art

[0002] Automatically recommending information for the elderly can achieve the purpose of enriching their spiritual life, enhancing their physical health, etc., thus significantly improving the quality of life of the elderly. The recommended information includes but is not limited to entertainment information, health care information, current affairs news, preferential promotion information, etc.

[0003] The existing information recommendation systems for the elderly mainly classify the elderly into multiple categories according to preset criteria, such as classification by personality, classification by occupation, etc., and screen and push information for each category of the elderly according to unified rules. However, even for the elderly of the same type, their actual situations are also very different. Therefore, the existing information push method based on category classification cannot meet the actual needs of the elderly, resulting in poor push effects. Summary of the Invention

[0004] In view of the above technical problems, the present invention provides an intelligent recommendation system, an electronic device, a computer storage medium, and a computer program product based on the personalized profile of the elderly.

[0005] The present invention discloses an intelligent recommendation system based on the personalized profile of the elderly. The intelligent recommendation system includes a profile acquisition module, a push information screening module, and an information push module; wherein, The profile acquisition module is configured to retrieve the personalized profile of the target elderly and transmit it to the push information screening module; The push information screening module is configured to receive the personalized profile, use a semantic analysis plugin to perform semantic understanding on the personalized profile, extract several pieces of interest information therefrom; screen out corresponding multiple pieces of recommended information according to each piece of interest information, and perform sorting processing on each piece of recommended information; The information push module is configured to push each piece of recommended information to the terminal device associated with the target elderly based on the sorting result, so that the terminal device outputs each piece of recommended information; Wherein, the profile acquisition module is further configured to obtain personalized profile update information related to the target elderly, determine an update probability according to the personalized profile update information, generate a pause push signal when the update probability is higher than a probability threshold, and transmit the pause push signal to the push information screening module, so that the push information screening module pauses the screening of the recommended information.

[0006] As an optional embodiment, the semantic analysis plugin is used to semantically understand the personalized profile, and several pieces of interest information are obtained therefrom, including: Use the semantic analysis plugin to semantically understand the personalized profile, obtain several pieces of first interest information therefrom, determine the preference probability of each piece of first interest information, and sort each piece of first interest information from largest to smallest according to the preference probability; Identify the first quantity of the first interest information that does not match each historical interest information in the historical interest information set of the target elderly person, and calculate the ratio of the first quantity to the total number of each historical interest information in the historical interest information set; Determine a second quantity according to the ratio and the preset total number of recommended information, and screen out the second quantity of the first interest information at the head of the sorted queue; In addition, screen out a third quantity of the historical interest information from the historical interest information set, and the historical interest information is different from the screened first interest information; wherein, the third quantity is equal to the preset total number of recommended information minus the second quantity.

[0007] As an optional embodiment, the semantic analysis plugin includes a semantic analysis model constructed based on Transformer and a multi-domain knowledge graph; wherein, the multi-domain knowledge graph is used to be input into the semantic analysis model together with the personalized profile to assist it in semantic understanding.

[0008] As an optional embodiment, the profile acquisition module is further configured to acquire personalized profile update information related to the target elderly person, and make a decision on the update probability according to the personalized profile update information, including: Acquire personalized profile update information related to the target elderly person, and the personalized profile update information at least includes the medical and health information of the target elderly person; Determine a medical start time according to the medical and health information, predict the probability that the current period belongs to the rehabilitation period according to the time difference between the medical start time and the current period, and use this probability as the update probability.

[0009] As an optional embodiment, the probability that the current period belongs to the rehabilitation period is predicted according to the time difference between the medical start time and the current period, and this probability is used as the update probability, including: Predict a first probability that the current period belongs to the rehabilitation period according to the time difference between the medical start time and the current period; Acquire the medical environment data of the target elderly person between the medical start time and the current period, and evaluate an optimization value according to the medical environment data; The second probability is calculated based on the optimized value and the first probability, and the second probability is used as the updated probability.

[0010] As an alternative embodiment, obtaining the medical environment data of the target elderly person between the medical start time and the current period, and evaluating the optimized value according to the medical environment data includes: Obtain the medical environment data of the target elderly person between the medical start time and the current period, where the medical environment data includes the type of medical environment and the corresponding duration. Use an evaluation model based on a large model to evaluate the medical environment data to obtain the optimized value.

[0011] In this embodiment, the target elderly person may be in multiple medical environments during the rehabilitation period. For example, after surgery, the elderly person stays in the hospital for one month first, and then stays in a nursing home / home for two months. Using an evaluation model based on a large model to evaluate this medical environment data can obtain the optimized value.

[0012] The above-mentioned large model refers to general large models such as GPT and BERT. These large models are machine learning models with huge scales and powerful functions. Usually, they far exceed conventional machine learning models in terms of training data, the number of model parameters, and computing resources, and do not require local model construction. However, it needs to be fine-tuned first to obtain an evaluation model that can be used for the above-mentioned evaluation purpose of the present invention.

[0013] As an alternative embodiment, the file acquisition module is further configured to: After transmitting the pause push signal to the push information screening module, periodically obtain the personalized file of the target elderly person, and compare it with the personalized file obtained last time. If there is a difference, generate a restart signal and transmit the restart signal to the push information screening module.

[0014] The present invention also discloses an electronic device, which is applied to the intelligent recommendation system based on the elderly person's personalized file; it includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

[0015] The present invention also discloses a computer storage medium, which is applied to the intelligent recommendation system based on the elderly person's personalized file; the computer-readable storage medium stores a computer program.

[0016] The present invention also discloses a computer program product, which is applied to the intelligent recommendation system based on the elderly person's personalized file; the computer program product contains computer code, and the computer code can be executed by the processor of the electronic device.

[0017] The beneficial effects of the present invention are as follows: The present invention can determine the most suitable recommended information based on the personalized profiles of the elderly. Compared with the traditional method of uniformly pushing information according to the classification of the elderly, the accuracy of the push is higher, which is more conducive to bringing a richer spiritual life to the elderly. In addition, the present invention can also determine whether to push recommended information according to the update probability of the personalized profiles of the elderly, reducing the probability of the situation where the recommended information does not match the latest profiles of the elderly, and further ensuring the accuracy of the push. Brief Description of the Drawings

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 is a schematic structural diagram of an intelligent recommendation system based on the personalized profiles of the elderly disclosed in the embodiments of the present invention; Figure 2 is a schematic structural diagram of the semantic analysis plug-in disclosed in the embodiments of the present invention. Detailed Embodiments

[0020] The following specific embodiments illustrate the implementation manners of the present application. Those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope protected by the present application.

[0021] In addition, the technical features involved in different implementation manners of the present application described below can be combined with each other as long as they do not conflict with each other.

[0022] The existing information recommendation systems for the elderly mainly divide the elderly into multiple categories according to preset criteria, such as classification by personality, classification by occupation, etc., and screen and push information for each category of the elderly according to unified rules. However, even for the elderly of the same type, their actual situations are very different. Therefore, the existing information push method based on category classification cannot meet the actual needs of the elderly, resulting in poor push effects.

[0023] For the above technical problems, as Figure 1As shown in the figure, an intelligent recommendation system based on the personalized profile of the elderly is disclosed in an embodiment of the present invention. The intelligent recommendation system includes a profile acquisition module, a push information screening module, and an information push module. Among them, The profile acquisition module is used to retrieve the personalized profile of the target elderly person and transmit it to the push information screening module; The push information screening module is used to receive the personalized profile, use a semantic analysis plug-in to perform semantic understanding on the personalized profile, extract several pieces of interest information therefrom, screen out corresponding multiple pieces of recommended information according to each piece of interest information, and perform sorting processing on each piece of recommended information; The information push module is used to push each piece of recommended information to the terminal device associated with the target elderly person based on the sorting result, so that the terminal device outputs each piece of recommended information; Among them, the profile acquisition module is further used to obtain the personalized profile update information related to the target elderly person, determine the update probability according to the personalized profile update information, generate a pause push signal when the update probability is higher than the probability threshold, and transmit the pause push signal to the push information screening module, so that the push information screening module pauses the screening of the recommended information.

[0024] Compared with the prior art, the present invention does not decide which type of information should be pushed according to the category to which the elderly belong, but specifically decides which type of information should be pushed based on the personalized profile of the elderly themselves, which can significantly improve the accuracy of the pushed information. Specifically: First, the profile acquisition module retrieves the personalized profile of the target elderly person. The personalized profile of the target elderly person can be formulated by medical institutions, elderly care institutions, guardians, etc., and contains the interests and hobbies, health and medical conditions, etc. of the target elderly person, and transmits it to the push information screening module.

[0025] Then, the push information screening module calls the semantic analysis plug-in to perform semantic understanding on the personalized profile written in natural language, extracts multiple pieces of interest information of the target elderly person therefrom, such as fishing, current news, movies, etc. Correspondingly, multiple pieces of recommended information can be screened out according to these pieces of interest information, and then sorting processing is performed on each piece of recommended information. Among them, the recommended information can be screened out from the local database (such as the local database of the elderly care institution), or can be screened out from Internet resources. The present invention does not make specific limitations. The information push module can push each piece of recommended information to the terminal device associated with the target elderly person (such as a smart phone, a tablet computer, a smart TV, etc.) based on the sorting result, so that the terminal device outputs each piece of recommended information.

[0026] In addition, the file acquisition module of the present invention also acquires personalized file update information related to the target elderly person, and based on this, determines the update probability of the personalized file of the target elderly person in the current period. When the update probability is higher than the probability threshold, a pause push signal is generated to trigger the push information screening module and the information push module to pause the screening and pushing of recommended information, so as to avoid the situation that the recommended information cannot meet the latest needs of the target elderly person.

[0027] Therefore, the present invention can decide the most suitable recommended information according to the personalized file of the elderly. Compared with the traditional unified push method classified by the elderly, the push accuracy is higher, which is more conducive to bringing a richer spiritual life to the elderly. In addition, the present invention can also decide whether to push the recommended information according to the update probability of the personalized file of the elderly, reducing the probability of the situation that the recommended information does not match the latest file of the elderly, and further ensuring the push accuracy.

[0028] It should be noted that the personalized files and the medical and health information contained therein involved in the present invention are all obtained through legal channels, and need to be authorized and confirmed by the owner of the information and the directly involved subject of the information, and unnecessary identity information needs to be erased, or the identity information is represented by an internal code and encrypted during information transmission.

[0029] As an optional embodiment, the semantic analysis plug-in is used to semantically understand the personalized file, and several pieces of interest information are obtained therefrom, including: Use the semantic analysis plug-in to semantically understand the personalized file, obtain several pieces of first interest information therefrom, and determine the preference probability of each piece of the first interest information, and sort each piece of the first interest information from largest to smallest according to the preference probability; Identify the first quantity of the first interest information that does not match each historical interest information in the historical interest information set of the target elderly person, and calculate the ratio of the first quantity to the total number of each historical interest information in the historical interest information set; Determine the second quantity according to the ratio and the preset total number of recommended information, and screen out the second quantity of the first interest information located at the head of the sorted queue; And, screen out the third quantity of the historical interest information from the historical interest information set, and the historical interest information is different from the screened first interest information; wherein, the third quantity is equal to the preset total number of recommended information minus the second quantity.

[0030] In this embodiment, the semantic analysis plugin can extract multiple pieces of interest information from the personalized profile, and can infer the preference probability for the first interest information based on the description content of each first interest information about the target elderly person in the personalized profile. Sort the first interest information in descending order according to the preference probability.

[0031] Meanwhile, a historical interest information set is also constructed for the target elderly person, which contains multiple pieces of historical interest information. These historical interest information are the interest information identified during previous information recommendations and have been de-duplicated. Each identified first interest information is matched one by one with the historical interest information contained in the historical interest information set of the target elderly person, and the first quantity of the first interest information with matching failures (i.e., no similar historical interest information) is counted. Further calculate the ratio of the first quantity to the total number of all historical interest information.

[0032] Thus, according to this ratio, the proportion and quantity of the new interest information (i.e., the first interest information with matching failures) and the historical interest information in the final interest information are determined. Then, the second quantity of "new" first interest information is selected from the head of the sorted queue, and the third quantity of "old" historical interest information is selected from the historical interest information set. These interest information are combined into the final interest information, and subsequent multiple recommended information are screened according to the final interest information. With such a setting, the ratio of new and old interests can be dynamically adjusted, thereby improving the diversity of the recommended information and avoiding the emergence of information cocoons.

[0033] It should be noted that the ratio can also be added to the preset ratio and then operated with the preset total number of recommended information to obtain the second quantity. The preset ratio is, for example, 15% or 20%, which can increase the proportion of the "new" first interest information and adapt to the new habits of users. There is no limitation on this.

[0034] As an alternative embodiment, as Figure 2 shown, the semantic analysis plugin includes a semantic analysis model constructed based on Transformer, and a multi-domain knowledge graph; wherein, the multi-domain knowledge graph is used to be input into the semantic analysis model together with the personalized profile to assist it in semantic understanding.

[0035] In this embodiment, the semantic analysis plug-in of the present invention includes a semantic analysis model and a multi-domain knowledge graph. The multi-domain knowledge graph is a composite knowledge graph, and multiple knowledge graphs can be determined in advance according to the occupation, education level, hobbies, etc. of the target elderly person, and then these knowledge graphs are combined to obtain a composite multi-domain knowledge graph. This multi-domain knowledge graph can play a prompting role in the semantic understanding of the personalized profile of the target elderly person by the semantic analysis model to assist in more accurate semantic analysis and understanding.

[0036] Transformer is a deep learning model based on the self-attention mechanism, which can efficiently process sequence data and capture long-range dependencies, making it perform excellently in semantic analysis tasks. Therefore, the above-mentioned semantic analysis model of the present invention is constructed based on Transformer, and the above characteristics of Transformer are used to efficiently analyze and understand the natural language in the personalized profile.

[0037] As an optional embodiment, the profile acquisition module is further configured to obtain personalized profile update information related to the target elderly person, and determine an update probability according to the personalized profile update information, including: Obtain personalized profile update information related to the target elderly person, where the personalized profile update information at least includes the medical and health information of the target elderly person; Determine a medical starting time according to the medical and health information, predict the probability that the current period belongs to the rehabilitation period according to the time difference between the medical starting time and the current period, and use this probability as the update probability.

[0038] In this embodiment, the personalized profile of the target elderly person will include the medical and health information of the target elderly person, and the first interest information obtained from the foregoing analysis will include this medical and health information. For example, this medical and health information indicates that the target elderly person underwent cervical spine treatment three months ago, and the doctor's advice is to perform cervical spine activities periodically. Therefore, the recommended information determined will include the corresponding cervical spine activity video, and the target elderly person can perform cervical spine activities by referring to this cervical spine activity video, thereby improving the speed of their cervical spine rehabilitation. However, after the target elderly person's condition has recovered, there is no longer a need to determine the recommended information based on this, and therefore, it is necessary to delete this first interest information from the personalized profile. Therefore, the present invention further sets to determine the medical starting time (i.e., the specific time of examination and treatment) based on the medical and health information of the target elderly person, and predicts the probability that the current period belongs to the recovery period based on the time difference between the medical starting time and the current period. For example, the recovery period after the treatment of a certain disease is three months. If the time difference between the medical starting time and the current period is close to or equal to three months, it is determined that the target elderly person has recovered. At this time, the probability that medical staff, service personnel in elderly care institutions, and the guardians of the elderly modify their personalized profiles is greater; if the time difference between the medical starting time and the current period is less than three months, it is determined that the target elderly person has not recovered. At this time, the probability that medical staff, service personnel in elderly care institutions, and the guardians of the elderly modify their personalized profiles is smaller. In view of this, the present invention sets the probability that the current period belongs to the recovery period as the update probability. When this update probability is higher than the probability threshold, a pause push signal can be generated to wait until the personalized profile is updated, and then the recommended information pushed for the target elderly person is determined based on the updated personalized profile.

[0039] As an optional embodiment, the predicting the probability that the current period belongs to the recovery period based on the time difference between the medical starting time and the current period, and using this probability as the update probability includes: Predicting a first probability that the current period belongs to the recovery period based on the time difference between the medical starting time and the current period; Obtaining the medical environment data of the target elderly person between the medical starting time and the current period, and evaluating an optimization value based on the medical environment data; Calculating a second probability based on the optimization value and the first probability, and using the second probability as the update probability.

[0040] In this embodiment, based on the above time difference, the first probability that the current period belongs to the rehabilitation period can be initially determined. However, this first probability is a theoretical value. The rehabilitation durations of different patients vary, and there is also a great relationship between the rehabilitation duration and their nursing environment. Therefore, the present invention further obtains the medical environment data of the target elderly person between the medical starting time and the current period, and evaluates and obtains an optimization value based on the medical environment data. This optimization value is used to characterize the quality of the medical environment in which the target elderly person is located during the rehabilitation period. When the medical environment is better, the rehabilitation speed exceeds the average rehabilitation speed, and vice versa. Finally, multiplying the optimization value by the previously obtained first probability yields the second probability, and this second probability is used as the updated probability.

[0041] As an alternative embodiment, the obtaining of the medical environment data of the target elderly person between the medical starting time and the current period, and evaluating and obtaining an optimization value based on the medical environment data includes: Obtain the medical environment data of the target elderly person between the medical starting time and the current period, where the medical environment data includes the type of medical environment and the corresponding duration of stay; Use an evaluation model based on a large model to evaluate the medical environment data to obtain the optimization value.

[0042] In this embodiment, the target elderly person may be in multiple medical environments during the rehabilitation period. For example, after surgery, they stay in the hospital for one month first, and then stay in a nursing home / home for two months. Using an evaluation model based on a large model to evaluate this medical environment data can obtain the optimization value. Obviously, the optimization value is also positively correlated with the duration of stay in a high-level medical environment (for example, the hospital is high-level, the nursing home is medium-level, and home is low-level). That is, the longer the time spent in a high-level medical environment, the larger the optimization value.

[0043] The above-mentioned large model refers to general large models such as GPT and BERT. These large models are machine learning models with huge scale and powerful functions. Usually, they far exceed conventional machine learning models in terms of training data, model parameters, and computing resources, and there is no need to locally build the model. However, it is necessary to first perform fine-tuning training on it to obtain an evaluation model that can be used for the above evaluation purpose of the present invention.

[0044] As an alternative embodiment, the file acquisition module is further configured to: After transmitting the pause push signal to the push information screening module, periodically obtain the personalized file of the target elderly person, and perform a difference comparison with the personalized file obtained last time. If there is a difference, generate a restart signal and transmit the restart signal to the push information screening module.

[0045] In this embodiment, after detecting that the personalized profile may be updated, the profile acquisition module periodically acquires the personalized profile of the target elderly person and analyzes whether it is really updated. If so, it triggers the system to restart the semantic understanding, recommended information screening and pushing operations for the personalized profile.

[0046] An embodiment of the present invention also discloses an electronic device, which is applied to the intelligent recommendation system based on the elderly person's personalized profile; it includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

[0047] An embodiment of the present invention also discloses a computer storage medium, which is applied to the intelligent recommendation system based on the elderly person's personalized profile; the computer-readable storage medium stores a computer program.

[0048] An embodiment of the present invention also discloses a computer program product, which is applied to the intelligent recommendation system based on the elderly person's personalized profile; the computer program product contains computer code that can be executed by the processor of the electronic device.

[0049] The above-mentioned computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the above. Alternatively, the computer-readable storage medium may be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0050] In order to provide interaction with the user, the systems and techniques described herein may be implemented on an electronic device having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the electronic device. Other kinds of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0051] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0052] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub - combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. The present invention belongs to the technical field of intelligent recommendation. An intelligent recommendation system based on the personalized profile of the elderly is provided, including: a profile acquisition module for retrieving the personalized profile of the target elderly person and transmitting it to the push information screening module; a push information screening module for receiving the personalized profile, using a semantic analysis plugin to perform semantic understanding on the personalized profile, extracting several pieces of interest information therefrom; screening multiple corresponding recommended information according to each piece of interest information and performing sorting processing on each piece of recommended information; an information push module for pushing each piece of recommended information to the terminal device associated with the target elderly person based on the sorting result, so that the terminal device outputs each piece of recommended information. The present invention can determine the most suitable recommended information according to the personalized profile of the elderly, with higher push accuracy, which is more conducive to bringing a richer spiritual life to the elderly.

2. The intelligent recommendation system based on the personalized profile of the elderly according to claim 1, wherein: Using a semantic analysis plugin to perform semantic understanding on the personalized profile, and extracting several pieces of interest information therefrom, including: Using a semantic analysis plugin to perform semantic understanding on the personalized profile, extracting several pieces of first interest information therefrom, and determining the preference probability of each piece of first interest information, and sorting each piece of first interest information from largest to smallest according to the preference probability; Identifying the first quantity of the first interest information that does not match each historical interest information in the historical interest information set of the target elderly person, and calculating the ratio of the first quantity to the total number of each piece of historical interest information in the historical interest information set; Determining a second quantity according to the ratio and the preset total number of recommended information, and screening out the second quantity of the first interest information located at the head of the sorted queue; And screening out a third quantity of the historical interest information from the historical interest information set, where the historical interest information is different from the screened first interest information; wherein, the third quantity is equal to the preset total number of recommended information minus the second quantity.

3. The intelligent recommendation system based on the personalized profile of the elderly according to claim 2, wherein: The semantic analysis plugin includes a semantic analysis model constructed based on Transformer and a multi-domain knowledge graph; wherein, the multi-domain knowledge graph is used to be input into the semantic analysis model together with the personalized profile to assist it in performing semantic understanding.

4. The intelligent recommendation system based on the personalized profile of the elderly according to claim 3, characterized in that: The profile acquisition module is further configured to obtain personalized profile update information related to the target elderly person, and determine an update probability according to the personalized profile update information, including: Obtaining personalized profile update information related to the target elderly person, where the personalized profile update information at least includes the medical and health information of the target elderly person; Determining a medical start time according to the medical and health information, predicting the probability that the current time period belongs to a rehabilitation time period according to the time difference between the medical start time and the current time period, and using this probability as the update probability.

5. An intelligent recommendation system based on the personalized profile of the elderly according to claim 4, characterized in that: Predicting the probability that the current time period belongs to a rehabilitation time period according to the time difference between the medical start time and the current time period, and using this probability as the update probability, including: Predicting a first probability that the current time period belongs to a rehabilitation time period according to the time difference between the medical start time and the current time period; Obtain the medical environment data of the target elderly person between the medical starting time and the current period, and evaluate the optimization value according to the medical environment data; Calculate the second probability according to the optimization value and the first probability, and use the second probability as the updated probability.

6. The intelligent recommendation system based on the personalized profile of the elderly according to claim 5, characterized in that: Obtain the medical environment data of the target elderly person between the medical starting time and the current period, and evaluate the optimization value according to the medical environment data, including: Obtain the medical environment data of the target elderly person between the medical starting time and the current period, where the medical environment data includes the medical environment type and the corresponding duration; Use an evaluation model based on a large model to evaluate the medical environment data to obtain the optimization value.

7. An intelligent recommendation system based on the personalized profile of the elderly according to claim 6, characterized in that: The file acquisition module is further configured to: After transmitting the pause push signal to the push information screening module, periodically obtain the personalized file of the target elderly person, and compare it with the personalized file obtained last time. If there are differences, generate a restart signal and transmit the restart signal to the push information screening module.

8. An electronic device, characterized in that: Applied to the intelligent recommendation system based on the elderly person's personalized file as described in any one of claims 1-7; includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor.

9. A computer storage medium, characterized in that: Applied to the intelligent recommendation system based on the elderly person's personalized file as described in any one of claims 1-7; the computer-readable storage medium stores a computer program.

10. A computer program product, characterized in that: Applied to the intelligent recommendation system based on the elderly person's personalized file as described in any one of claims 1-7; the computer program product contains computer code that can be executed by the processor of an electronic device.

Images