An AI intelligent interaction method and system based on digital home-based elderly care services

Abstract

The present invention discloses an AI intelligent interaction method based on digital home-based elderly care services, which obtains the home-based elderly care monitoring report of the current elderly person at home; extracts the health description vector of the current elderly person at home from the home-based elderly care monitoring report; inputs the health description vector into a preset home-based elderly care evaluation model to output the elderly care risk level of the current elderly person at home; based on the elderly care risk level of the current elderly person at home, retrieves the predefined elderly care strategies corresponding to the risk level for interaction with the user; wherein the health description vector is a flattened vector of text interaction features generated by interaction item modeling. The present invention extracts health description texts from the home-based elderly care monitoring report, conducts interaction modeling on these entity pairs, and finally generates text interaction features that can comprehensively reflect different health sign indicators, thereby facilitating the capture of the interaction effects of multi-dimensional sign indicators by the elderly care risk assessment model and providing a more personalized and accurate elderly care service experience for users.

Description

Technical Field

[0001] The present invention relates to the field of home-based elderly care services, and specifically to an AI intelligent interaction method based on digital home-based elderly care services. Background Art

[0002] In the existing home-based elderly care service technologies, the assessment of elderly care risks usually relies on the processing of traditional structured data, such as numerical health indicators like blood pressure and blood sugar. However, in actual application scenarios, home-based elderly care monitoring reports contain a large amount of unstructured text data recorded in natural language form. This text data contains rich health information, such as disease descriptions, symptom records, and doctor's suggestions. The prior art with the patent publication number CN117226866A discloses an intelligent elderly care companion robot, which provides a new home-based elderly care solution integrating smart home services, personalized companion services in terms of emotion and entertainment, and life safety guardianship monitoring services through intelligent voice interaction technology, improving the quality of life and safety of the elderly.

[0003] However, in the prior art, there are significant deficiencies in the analysis of unstructured elderly care health text data. For example:

[0004] There are often potential interaction relationships between different entities (such as "hypertension" and "obesity") in the health description text, and these relationships may significantly affect the results of elderly care risk assessment. For example, the combination of "hypertension" and "obesity" may imply a higher health risk, while the prior art usually only analyzes individual entities in isolation and fails to capture the interaction between entities, resulting in insufficient accuracy of the prediction results. Moreover, the entity descriptions in the health text may have synonyms or variants (such as "high blood sugar" and "elevated blood sugar"), and the prior art lacks unified semantic standardization processing, resulting in feature duplication or information loss. Summary of the Invention

[0005] This embodiment provides an AI intelligent interaction method and system based on digital home-based elderly care services to explore how to solve the problem of capturing the interaction between entities in home-based elderly care monitoring text.

[0006] In a first aspect, the present invention provides an AI intelligent interaction method based on digital home-based elderly care services, including:

[0007] Obtain the home-based elderly care monitoring report of the current home-based elderly;

[0008] Extract the health description vector of the current home-based elderly from the home-based elderly care monitoring report;

[0009] Input the health description vector into a preset home-based elderly care evaluation model to output the elderly care risk level of the current home-based elderly.

[0010] Retrieve the predefined elderly care strategies corresponding to the risk level based on the current elderly care risk level of the home-bound elderly and interact with the user;

[0011] Among them, the health description vector is the flattened vector of the text interaction features generated by the interaction item modeling.

[0012] In some of the embodiments, the modeling method of the home-based elderly care evaluation model includes:

[0013] S1. Obtain a home-based elderly care monitoring sample set; among them, the sample set includes: a number of home-based elderly care monitoring samples labeled with the elderly care risk level of the home-bound elderly;

[0014] S2. Perform character recognition on the home-based elderly care monitoring samples and extract their text interaction features;

[0015] S3. Flatten the text interaction features to generate a one-dimensional array of health description vectors;

[0016] S4. Match the health description vectors with the elderly care risk levels to generate home risk assessment training samples;

[0017] S5. Obtain a number of home risk assessment training samples and summarize them into a home risk assessment training set.

[0018] S6. Use a general model to receive the home risk assessment training set as input and perform iterative supervised learning until the general model is iterated into the home-based elderly care evaluation model.

[0019] In some of the embodiments, obtaining the home-based elderly care monitoring sample set includes:

[0020] S1-1. Import the electronic health report of the home-bound elderly from the electronic health record system;

[0021] S1-2. Extract a number of physical signs to be monitored from the electronic health report;

[0022] S1-3. Determine the elderly care risk level of the electronic health report according to a number of physical signs to be monitored and the predefined physical sign abnormality criteria;

[0023] The determination expression of the elderly care risk level is:

[0024] ;

[0025] Among them, RS represents the elderly care risk level score, represents the i-th physical sign to be detected, represents the risk weight of the i-th physical sign to be detected, and n represents the number of physical signs to be monitored;

[0026] S1-4. Fill in a number of physical signs to be monitored and their corresponding elderly care risk levels into a structured home elderly care monitoring template to obtain the home elderly care monitoring sample;

[0027] S1-5. Obtain a number of home elderly care monitoring samples and summarize them into the home elderly care monitoring sample set.

[0028] In some embodiments, perform character recognition on the home elderly care monitoring sample, and extract its text interaction features, including:

[0029] S2-1. Extract the set of health description entities of the home elderly care monitoring sample;

[0030] S2-2. Perform interaction item modeling on any two health description entities in the set of health description entities to generate the text interaction features.

[0031] In some embodiments, extracting the set of health description entities of the home elderly care monitoring sample includes:

[0032] S2-1-1. Select the health description text from the home elderly care monitoring sample;

[0033] S2-1-2. Import the health description text into the SC toolkit for entity recognition, and filter out the original triple entity set of the health description text;

[0034] S2-1-3. Semantically standardize all the original triple entities in the original triple entity set to generate standard triple entities, and define them as the health description entities of the home elderly care monitoring sample.

[0035] In some embodiments, selecting the health description text from the home elderly care monitoring sample includes:

[0036] S2-1-2-1. Map the health description text to a multi-dimensional vector space through word embedding to generate word vectors containing context features;

[0037] S2-1-2-2. Analyze each word of the word vectors containing context features using the sequence annotation module of the SC toolkit to identify entity names;

[0038] S2-1-2-3. Assign categories to the entity names based on the preset label classification rules of the SC toolkit to generate entity categories;

[0039] S2-1-2-4. Record the start position and end position of the entity category in the health description text, and define the enclosed area of the start position and end position as the entity boundary;

[0040] S2-1-2-5. Screen out the entity name, entity category, and entity boundary, and combine and define them as the original triple entities of the health description text; wherein, the triple of the original triple entities is: entity name, entity category, and entity boundary.

[0041] S2-1-2-6. Screen out all the original triple entities from the health description text, and combine and define them as the set of the original triple entities.

[0042] In some embodiments, the defining steps of the health description entities of the home-based elderly care monitoring samples include:

[0043] S2-1-3-1. Calculate the word vector similarity between any two entity names in the set of original triple entities.

[0044] The calculation expression of the word vector similarity between any two original triple entities is:

[0045] ;

[0046] where CS represents the word vector similarity, respectively represent the word vectors of any two original triple entities, represents the dot product of the word vectors of any two original triple entities, and respectively represent the modulus length of the word vector and the modulus length of the word vector;

[0047] S2-1-3-2. Determine whether it exceeds the first similarity threshold according to the word vector similarity between any two original triple entities.

[0048] S2-1-3-3. If the word vector similarity exceeds the first similarity threshold, define the entity name that exceeds the first similarity threshold as a synonymous variant name.

[0049] S2-1-3-4. Replace the synonymous variant name with the standard expression name, so that any two original triple entities with word vector similarity exceeding the first similarity threshold are transformed into the same standard triple entity.

[0050] In some embodiments, for any two health description entities in the set of health description entities, performing interaction term modeling to generate the text interaction features includes:

[0051] S2-2-1. Associate the word vector similarity between any two original triple entities with the set of health description entities;

[0052] S2-2-2. Determine whether the word vector similarity between any two health description entities exceeds the second similarity threshold;

[0053] S2-2-3. If the word vector similarity exceeds the second similarity threshold, screen out the pairs of health description entities with significant semantic relevance in the set of health description entities;

[0054] S2-2-4. Perform interaction term modeling on the pairs of health description entities to generate the text interaction features.

[0055] In some embodiments, a general model is used to receive the home risk assessment training set as input and perform iterative supervised learning until the general model is iterated into the home care assessment model, including:

[0056] S6-1. Extract the first batch of home risk assessment training samples in the home risk assessment training set as input features;

[0057] S6-2. Perform forward propagation on the first batch of home risk assessment training samples and output the predicted values of the first batch of old-age risk levels;

[0058] S6-3. Calculate the prediction error between the predicted values of the old-age risk levels in the first batch and the actual old-age risk levels;

[0059] S6-4. If the prediction error does not reach the minimum error, update the model parameters of the general model until the minimum error is reached;

[0060] S6-5. If the prediction error reaches the minimum error, regard the model parameters of the current batch as the model parameters of the home care assessment model;

[0061] S6-6. Export the model parameters of the home care assessment model and construct the home care assessment model based on this.

[0062] Compared with the prior art, an AI intelligent interaction method based on digital home care services of the present invention extracts health description texts from home care monitoring reports, then calculates the word vector similarity between health description entities, screens out pairs of health description entities with significant semantic relevance, and performs interaction modeling on these entity pairs, finally generating text interaction features that can comprehensively reflect different health sign indicators, so as to facilitate the home care risk assessment model to capture the interaction effects of multi-dimensional sign indicators, output a more accurate home care risk level, and provide a more personalized and accurate home care service experience for users.

[0063] In a second aspect, the present invention provides an AI intelligent interaction system based on digital home-based elderly care services, which executes the interaction method described in the first aspect. The system includes:

[0064] An elderly care monitoring report acquisition module, configured to acquire an elderly care monitoring report of the current home-based elderly;

[0065] A health description vector extraction module, configured to extract a health description vector of the current home-based elderly from the elderly care monitoring report;

[0066] An elderly care risk level output module, configured to input the health description vector into a preset home-based elderly care evaluation model and output the elderly care risk level of the current home-based elderly.

[0067] An elderly care strategy retrieval module, configured to retrieve a predefined elderly care strategy corresponding to the risk level based on the elderly care risk level of the current home-based elderly and interact with the user;

[0068] Wherein, the health description vector is a flattened vector of text interaction features generated by interaction item modeling.

[0069] Compared with the prior art, the beneficial effects of an AI intelligent interaction system based on digital home-based elderly care services of the present invention are the same as those of the above-mentioned AI intelligent interaction method based on digital home-based elderly care services, so they will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] Figure 1 is a schematic flow chart of an AI intelligent interaction method based on digital home-based elderly care services of the present invention;

[0071] Figure 2 is a schematic modeling flow chart of the elderly care evaluation model described in the present invention;

[0072] Figure 3 is a schematic transformation flow chart of the standard triple entity described in the present invention;

[0073] Figure 4 is a structural block diagram of an AI intelligent interaction system based on digital home-based elderly care services of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0074] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0075] Embodiment 1: Refer toFigures 1 to 3 , the present invention provides an AI intelligent interaction method based on digital home-based elderly care services, including the following steps:

[0076] Obtain the home-based elderly care monitoring report of the current home-based elderly;

[0077] Extract the health description vector of the current home-based elderly from the home-based elderly care monitoring report;

[0078] Input the health description vector into a preset home-based elderly care evaluation model to output the elderly care risk level of the current home-based elderly.

[0079] Based on the elderly care risk level of the current home-based elderly, retrieve the predefined elderly care strategies corresponding to the risk level for interaction with the user;

[0080] Among them, the health description vector is a flattened vector of text interaction features generated by interaction item modeling.

[0081] In this embodiment, by extracting the health description vector from the home-based elderly care monitoring report and combining the interaction item modeling and flattening operations of text interaction features, the digital expression of the health information of home-based elderly is realized. By inputting the generated health description vector into a preset home-based elderly care evaluation model, the elderly care risk level of home-based elderly can be quickly calculated and output.

[0082] Among them, the predefined elderly care strategy refers to a targeted intervention or management plan formulated in advance according to the elderly care risk level and the specific health status of the home-based elderly. The core of these strategies is to provide suitable health guidance, medical support, and life care for the elderly with different risk levels to ensure that the elderly can obtain appropriate elderly care services in the home environment.

[0083] In this embodiment:

[0084] Low risk: For the elderly with a lower elderly care risk level, the predefined elderly care strategies may include suggestions for a healthy lifestyle (such as diet adjustment, moderate exercise) and regular monitoring of physical conditions to maintain a good health state.

[0085] Medium risk: For the elderly with a medium elderly care risk level, the strategies may include more frequent health monitoring, a more detailed health management plan (such as monitoring changes in indicators such as blood sugar and blood pressure), and appropriate community medical interventions, such as regular follow-up visits and medication management.

[0086] High risk: For the elderly with a higher elderly care risk level, the predefined elderly care strategies may involve emergency intervention measures, such as arranging special examinations, personalized treatment plans, strengthening care support (such as regular visits by home caregivers), and even suggesting transfer to a professional elderly care institution for continuous care.

[0087] In this embodiment, the modeling method of the home-based elderly care evaluation model includes:

[0088] Step S1, obtaining a home-based elderly care monitoring sample set; wherein the sample set includes: a number of home-based elderly care monitoring samples marked with home-based elderly care risk levels;

[0089] Step S2, performing text recognition on the home-based elderly care monitoring sample to extract its text interaction features;

[0090] Step S3, flattening the text interaction features to generate a health description vector of a one-dimensional array;

[0091] Specifically, text interaction features may be multi-dimensional features contained in the monitoring data of an elderly person living at home, such as blood pressure (high or low), blood sugar, weight, heart rate, sleep quality and other indicators; feature flattening refers to flattening the text interaction features carrying multi-dimensional health monitoring information into a one-dimensional array. The flattened health description vector can be directly passed into the home-based elderly care evaluation model as an input feature.

[0092] Step S4: Match the health description vector and the pension risk level to generate a home risk assessment training sample;

[0093] Step S5: obtain a number of home risk assessment training samples and aggregate them into a home risk assessment training set.

[0094] Step S6: Use the general model to receive the home risk assessment training set as input, and perform iterative supervised learning until the general model is iterated into the home-based elderly care assessment model.

[0095] In this embodiment, the general model can be any supervised learning model with AI intelligent learning capability, which is used to learn rules from the home risk assessment training set, and finally form a home-based elderly care assessment model that can predict the elderly care risk level, such as a random forest or a neural network.

[0096] This embodiment obtains data from a home-based elderly care monitoring sample set marked with elderly care risk levels, extracts text interaction features in the monitoring data through text recognition, and further converts these multi-dimensional health data into health description vectors in a one-dimensional array format through feature flattening operations. This ensures that the data can be easily passed as input to the home-based elderly care assessment model for learning. Next, the health description vector is matched with the elderly care risk level to generate a home-based risk assessment training sample set. Using the training samples, the general model is iteratively trained through supervised learning, and finally forms a home-based elderly care assessment model that can predict the elderly care risk level of the elderly at home.

[0097] Furthermore, the step S1 specifically includes:

[0098] S1-1. Import the electronic health reports of the elderly living at home from the electronic health record system;

[0099] The electronic health record (EHR, Electronic Health Record) is a general term widely used in various medical and health management systems. Many hospitals, community health service centers, and health management institutions use EHR systems to record patients' health information. Exemplarily, it includes: the national health record system or the local community health management system. Among them, the electronic health report can be the physical examination report, health follow-up record, chronic disease management report, or personal health assessment result recorded by the elderly living at home in this file.

[0100] Exemplarily, it includes the physical examination results of the elderly living at home regularly or irregularly, such as indicators like blood glucose, blood pressure, electrocardiogram, etc.; the records of regular follow-ups of the elderly living at home for chronic disease management or data tracking after health intervention; the health score results generated through questionnaires and doctor evaluations, etc.

[0101] S1-2. Extract a number of physical signs to be monitored from the electronic health report;

[0102] The physical signs to be monitored can be blood pressure, blood glucose, heart rate, body weight, etc., which represent the current health status of the elderly living at home.

[0103] S1-3. Determine the elderly care risk level of the electronic health report according to a number of physical signs to be monitored and the predefined physical sign abnormality criteria;

[0104] Among them, the predefined physical sign abnormality criteria refer to setting abnormality criteria for each physical sign indicator. For example:

[0105] Blood pressure: If the systolic blood pressure is greater than 180 mmHg or the diastolic blood pressure is greater than 110 mmHg, it is regarded as hypertension.

[0106] Blood glucose: If the fasting blood glucose is greater than 126 mg / dL, it is regarded as hyperglycemia.

[0107] Body weight: If the body weight is too light or too heavy, exceeding a certain BMI threshold (such as BMI < 18.5 or BMI > 30), it is judged as abnormal. By converting the abnormality of each physical sign indicator into a numerical value (usually binary classification: normal or abnormal), and then summarizing the abnormal scores of all physical sign indicators. Finally, a comprehensive risk score is calculated to determine the elderly care risk level.

[0108] The determination expression of the elderly care risk level is:

[0109] ;

[0110] Among them, RS represents the elderly care risk level score, represents the i-th physical sign index to be detected, represents the risk weight of the i-th physical sign index to be detected, and n represents the number of physical sign indexes to be monitored.

[0111] For the pension risk level score, it can be divided into different pension risk levels:

[0112] When is less than 3, it is determined as a low risk;

[0113] When RS is greater than or equal to 3 and less than 6, it is determined as a medium risk;

[0114] When RS is greater than or equal to 6, it is determined as a high risk;

[0115] Of course, the actual pension situation can be further refined based on the score, such as no risk, extremely low risk, medium-high risk, etc.

[0116] S1-4. Fill a number of physical sign indexes to be monitored and the corresponding pension risk levels into a structured home pension monitoring template to obtain the home pension monitoring sample;

[0117] S1-5. Obtain a number of home pension monitoring samples and summarize them into the home pension monitoring sample set.

[0118] In this embodiment, by importing the electronic health reports of home-based elderly people, detailed records of multiple health indicators such as blood glucose, blood pressure, weight, and heart rate are provided. Then, a number of physical sign indexes to be monitored are extracted from the electronic health reports to reflect the health status of home-based elderly people. According to the predefined physical sign abnormality criteria (such as blood pressure higher than a specific value, blood glucose exceeding the standard, etc.), the health abnormality of each physical sign is determined, and the pension risk level is calculated according to the abnormality. Finally, the extracted number of physical sign indexes to be monitored and the risk levels are filled into a structured home pension monitoring template to form a home pension monitoring sample and summarized into a home pension monitoring sample set.

[0119] Further, step S2 specifically includes:

[0120] S2-1. Extract the set of health description entities of the home pension monitoring sample;

[0121] S2-2. Perform interaction item modeling on any two health description entities in the set of health description entities to generate the text interaction features.

[0122] In this embodiment, by extracting the set of health description entities in the home pension monitoring sample and performing interaction item modeling based on the relationships between the health description entities, the text interaction features are generated. Effectively capture the interactions between health description entities and provide more comprehensive feature information for subsequent pension risk assessment.

[0123] Further, the specific steps of step S2-1 further include:

[0124] S2-1-1. Select the health description text from the home care monitoring samples;

[0125] S2-1-2. Import the health description text into the SC toolkit for entity recognition, and screen out the original triple entity set of the health description text;

[0126] In this embodiment, the SC toolkit refers to the SciSpaCy toolkit, which is a toolkit based on natural language processing (NLP) technology and is specifically used for tasks such as information extraction, entity recognition, and semantic standardization in medical texts. It is usually used to extract key entities from health-related texts, such as disease names, symptoms, signs, treatment methods, etc. Entity recognition by the SC toolkit means using the toolkit to analyze the health description text, identify the medical entity information contained therein, and organize these entities into an original triple entity set.

[0127] S2-1-3. Semantically standardize all the original triple entities in the original triple entity set to generate standard triple entities, and define them as the health description entities of the home care monitoring samples.

[0128] In this embodiment, by selecting the health description text from the home care monitoring samples and using the SC toolkit for entity recognition of the text, an original triple entity set reflecting health information is extracted. The SC toolkit efficiently extracts medical-related entity information from the text through natural language processing technology and organizes it in a structured manner. Then, semantic standardization is performed on these original triple entities to ensure that entities in different expression forms have a unified standard representation, generating a standard triple entity set, significantly improving the normativity of health description entities, and laying a high-quality data foundation for subsequent text interaction feature modeling.

[0129] In step S2-1-2, the screening steps include:

[0130] S2-1-2-1. Map the health description text to a multi-dimensional vector space through word embedding to generate word vectors containing context features;

[0131] Among them, word embedding is a technique in natural language processing (NLP). By mapping words to a continuous, low-dimensional vector space, it can capture the semantic similarity and context information of words. This mapping makes the vectors of semantically similar words (such as "doctor" and "nurse") in the same vector space relatively close.

[0132] S2-1-2-2. Analyze each word of the word vectors containing context features using the sequence annotation module of the SC toolkit to identify entity names;

[0133] S2-1-2-3. Based on the predefined tag classification rules of the SC toolkit, assign categories to the entity names to generate entity categories;

[0134] S2-1-2-4. Record the start position and end position of the entity category in the health description text, and define the enclosed area of the start position and end position as the entity boundary;

[0135] S2-1-2-5. Filter out the entity names, entity categories, and entity boundaries, and combine them to define the original triple entities of the health description text; among them, the triple of the original triple entity is: entity name, entity category, and entity boundary.

[0136] Exemplarily, the triple is:

[0137] Entity name: Hypertension

[0138] Entity category: Disease

[0139] Entity location: The position where hypertension appears in the text (such as "from the 5th to the 8th character").

[0140] S2-1-2-6. Filter out all the original triple entities from the health description text, and combine them to define the set of the original triple entities.

[0141] In this embodiment, by mapping the health description text to a multi-dimensional vector space to generate word vectors containing context features, the accurate expression of the semantic information of the words in the text is ensured. Using the sequence annotation module of the SC toolkit, analyze the text content word by word, identify the entity names and assign categories according to the predefined rules, and then locate the position of the entity in the text to construct a set of original triple entities containing entity names, categories, and boundaries. This step can extract structured key information from the health description text, such as disease names and their positions, and generate a comprehensive and standardized set of original triple entities.

[0142] In step S2-1-3, the definition steps of the health description entity include:

[0143] S2-1-3-1. Calculate the word vector similarity between any two original triple entities according to any two entity names in the set of original triple entities;

[0144] The calculation expression of the word vector similarity between any two original triple entities is:

[0145] ;

[0146] Among them, CS represents the word vector similarity, respectively represent the word vectors of any two original triple entities, represents the dot product of the word vectors of any two original triple entities, and respectively represent the modulus length of the word vector and the modulus length of the word vector

[0147] In this implementation reward, the cosine similarity is used for the word vector similarity, and the result of the cosine similarity ranges from -1 to 1. If the similarity is 1, it means that the two word vectors are exactly the same. If the similarity is 0, it means that the two word vectors are orthogonal and have no similarity. If the similarity is -1, it means that the two word vectors are exactly opposite. In the field of natural language processing, using the cosine similarity to calculate the word vector similarity between different entities (such as disease names, symptoms, drugs, etc.) can measure the semantic proximity between the two, so as to judge whether they belong to the same category or whether there is a certain semantic relationship.

[0148] S2-1-3-2. Determine whether it exceeds the first similarity threshold according to the word vector similarity between any two original triple entities;

[0149] S2-1-3-3. If the word vector similarity exceeds the first similarity threshold, define the entity name that exceeds the first similarity threshold as a synonymous variant name;

[0150] S2-1-3-4. Replace the synonymous variant name with the standard expression name, so that any two original triple entities with word vector similarity exceeding the first similarity threshold are transformed into the same standard triple entity.

[0151] Among them, the standard expression name is the standard expression name defined based on a unified standard thesaurus or industry norms in the field of home-based elderly care services.

[0152] This embodiment effectively evaluates the semantic proximity of entity names by calculating the word vector similarity between original triple entities. The semantic similarity between two entity word vectors is accurately measured by the cosine similarity, and based on the first similarity threshold, the entity name that exceeds the threshold is determined as a synonymous variant name. By replacing the synonymous variant name with the standard expression name, semantic unification and standardization are achieved. This embodiment ensures that in the field of home-based elderly care services, entities with different expression methods can be standardized, providing high-quality semantic input for health description modeling.

[0153] Further, the specific steps of S2-2 further include:

[0154] S2-2-1. Associate the word vector similarity between any two original triple entities to the set of health description entities;

[0155] S2-2-2. Determine whether the word vector similarity between any two health description entities exceeds the second similarity threshold;

[0156] S2-2-3. If the word vector similarity exceeds the second similarity threshold, screen out the pairs of health description entities with significant semantic relevance in the set of health description entities;

[0157] Further, the significant semantic relevance means that there is a strong semantic connection between two health description entities, which is measured by the word vector similarity. If the word vector similarity of two health description entities is high enough (exceeding the preset similarity threshold), it is considered that they have significant semantic relevance. This relevance means that these two entities express similar concepts or have related health information in a certain context and may jointly affect the elderly care risk.

[0158] In this embodiment, by calculating the word vector similarity between health description entities, their significant semantic relationship is determined. When the similarity of two health description entities exceeds the second similarity threshold, it can be considered that they are significantly related semantically, and then these entity pairs are used as interaction terms for modeling.

[0159] S2-2-4. Perform interaction term modeling on the pairs of health description entities to generate the text interaction features.

[0160] In this embodiment, by calculating the word vector similarity between health description entities, pairs of entities with significant semantic relevance are screened out. The significant semantic relevance is measured by the word vector similarity exceeding the second similarity threshold, reflecting the strong semantic association between entities, which usually means having potential common impacts on health information or risk factors. By performing interaction term modeling on the screened significantly related entity pairs, more multi-dimensional text interaction features are generated. This embodiment can not only capture the complex relationships between health description entities but also provide richer features for elderly care risk assessment.

[0161] In this embodiment, the training steps of the home elderly care evaluation model include:

[0162] S6-1. Extract the first batch of home risk assessment training samples in the home risk assessment training set as input features;

[0163] S6-2. Perform forward propagation on the home risk assessment training samples of the first batch, and output the predicted values of the old-age care risk levels for the first batch.

[0164] S6-3. Calculate the prediction error between the predicted values of the old-age care risk levels and the actual old-age care risk levels in the first batch.

[0165] S6-4. If the prediction error does not reach the minimum error, update the model parameters of the general model until the minimum error is reached.

[0166] S6-5. If the prediction error reaches the minimum error, regard the model parameters of the current batch as the model parameters of the home-based old-age care assessment model.

[0167] S6-6. Export the model parameters of the home-based old-age care assessment model, and construct the home-based old-age care assessment model based on these parameters.

[0168] In this embodiment, through an iterative supervised learning process, the parameters of the general model are gradually optimized, and finally a home-based old-age care assessment model is generated. Specifically, the system extracts training samples from the home risk assessment training set, performs forward propagation to predict the old-age care risk level, and continuously adjusts the model parameters by calculating the error between the predicted value and the actual value. Through this iterative optimization process, the prediction accuracy of the model is gradually improved until the minimum error standard is reached, ensuring that the finally generated home-based old-age care assessment model has high accuracy and reliability.

[0169] Embodiment 2: The embodiment of the present invention also provides an AI intelligent interaction system based on digital home-based old-age care services. This system is used to implement the above method embodiment, and the parts that have been described will not be repeated here. The following terms such as "module", "unit", "sub-unit", etc. can be a combination of software and / or hardware that can achieve a predetermined function. Although the system described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0170] As Figure 4 shown, Figure 4 is the structural block diagram of an AI intelligent interaction system based on digital home-based old-age care services according to the present invention. This system includes:

[0171] An old-age care monitoring report acquisition module, which is used to acquire the home-based old-age care monitoring report of the current home-based elderly.

[0172] A health description vector extraction module, which is used to extract the health description vector of the current home-based elderly from the home-based old-age care monitoring report.

[0173] An old-age care risk level output module, which is used to input the health description vector into a preset home-based old-age care assessment model and output the old-age care risk level of the current home-based elderly.

[0174] The elderly care strategy retrieval module is used to retrieve the predefined elderly care strategy corresponding to the risk level and interact with the user based on the elderly care risk level of the current home-based elderly

[0175] Wherein, the health description vector is a flattened vector of text interaction features generated by interaction item modeling

[0176] In the above system, the home-based elderly care monitoring report of the current home-based elderly is obtained through the elderly care monitoring report acquisition module; the health description vector of the current home-based elderly is extracted through the health description vector extraction module; the elderly care risk level of the current home-based elderly is obtained through the elderly care risk level output module; the predefined elderly care strategy is obtained through the elderly care strategy retrieval module; the problem of how to capture the entity interaction effect in the home-based elderly care monitoring text is solved

[0177] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable systems. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wire (such as infrared, wireless, microwave, etc.)

[0178] The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more collections of available media. The available media can be magnetic media (for example, floppy disks, hard disks, magnetic tapes), optical media (for example, DVD ) or semiconductor media. The semiconductor media can be a solid-state drive

[0179] In several embodiments provided in the present application, it should be understood that the disclosed systems, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the units is only a logical function division of an underwater topographic change analysis system and method for waterways. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of the system or unit can be in electrical, mechanical, or other forms.

[0180] As described above, the above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application.

Claims

1. An AI intelligent interaction method based on digital home-based elderly care services, characterized in that: include: Obtain the current home-based elderly care monitoring report; Extracting the health description vector of the elderly living at home from the home-based elderly care monitoring report; Input the health description vector into a preset home-based elderly care assessment model to output the elderly care risk level of the current elderly at home; Based on the current elderly care risk level of the elderly living at home, the predefined elderly care strategy corresponding to the risk level is retrieved to interact with the user; Wherein, the health description vector is a flattened vector of text interaction features generated by interaction item modeling; The modeling method of the home-based elderly care evaluation model includes: S1. Obtain a home-based elderly care monitoring sample set; wherein the sample set includes: a number of home-based elderly care monitoring samples marked with home-based elderly care risk levels; S2. Performing text recognition on the home-based elderly care monitoring samples to extract text interaction features; S3, flattening the text interaction features to generate a health description vector of a one-dimensional array; S4, matching the health description vector with the elderly care risk level to generate home risk assessment training samples; S5. Obtain a number of home risk assessment training samples and aggregate them into a home risk assessment training set; S6. Using the general model to receive the home risk assessment training set as input, and performing iterative supervised learning until the general model is iterated into the home-based elderly care assessment model; The S2 includes: S2-1, extracting the health description entity set of the home-based elderly care monitoring sample; S2-2, performing interaction item modeling on any two health description entities in the health description entity set to generate the text interaction feature; The S2-2 includes: S2-2-1, associate the word vector similarity between any two original triple entities to the health description entity set; S2-2-2, judging whether the word vector similarity between any two health description entities exceeds a second similarity threshold; S2-2-3. If the word vector similarity exceeds the second similarity threshold, select health description entity pairs with significant semantic relevance in the health description entity set, and the entity pairs have a common impact on the pension risk; S2-2-4. Model the interaction items of the health description entity pairs and generate the text interaction features to capture the interaction between the health description entities.

2. According to claim 1, an AI intelligent interaction method based on digital home-based elderly care services is characterized in that: Obtain home-based elderly care monitoring sample set, including: S1-1. Import the electronic health reports of elderly people living at home from the electronic health record system; S1-2, extracting a number of physical signs to be monitored from the electronic health report; S1-3, determine the pension risk level of the electronic health report based on a number of monitored physical signs and indicators and predefined physical sign abnormality standards; The determination expression of the pension risk level is: ； Among them, RS represents the pension risk level score, represents the i-th physical sign index to be detected, represents the risk weight of the i-th physical sign indicator to be detected, and n represents the number of physical sign indicators to be monitored; S1-4, filling a number of physical signs to be monitored and the corresponding elderly care risk levels into a structured home-based elderly care monitoring template to obtain the home-based elderly care monitoring sample; S1-5. Obtain a number of home-based elderly care monitoring samples, and aggregate them into the home-based elderly care monitoring sample set.

3. According to claim 1, an AI intelligent interaction method based on digital home-based elderly care services is characterized in that: Extract the health description entity set of the home-based elderly care monitoring sample, including: S2-1-1. Select health description text from the home-based elderly care monitoring sample; S2-1-2, importing the health description text into the SC toolkit for entity recognition, and filtering out the original triple entity set of the health description text; S2-1-3. Semantically standardize the original triple entities in the original triple entity set to generate standard triple entities, and define them as health description entities of the home-based elderly care monitoring sample.

4. According to claim 3, an AI intelligent interaction method based on digital home-based elderly care services is characterized in that: Select health description text from the home-based elderly care monitoring sample, including: S2-1-2-1, mapping the health description text to a multidimensional vector space through word embedding to generate a word vector containing context features; S2-1-2-2, using the sequence labeling module of the SC toolkit to analyze the word vector containing the context features word by word to identify the entity name; S2-1-2-3. Based on the label classification rules preset by the SC toolkit, the entity name is assigned a category to generate an entity category; S2-1-2-4, recording the starting position and the ending position of the entity category in the health description text, and defining the enclosed area of ​​the starting position and the ending position as the entity boundary; S2-1-2-5, filter out the entity name, entity category and entity boundary, and define their combination as the original triple entity of the health description text; wherein the triple of the original triple entity is: entity name, entity category and entity boundary; S2-1-2-6. Filter out all original triple entities from the health description text, and define their combination as the original triple entity set.

5. According to claim 3, an AI intelligent interaction method based on digital home-based elderly care services is characterized in that: The steps for defining the health description entity of the home-based elderly care monitoring sample include: S2-1-3-1. Based on any two entity names in the original triple entity set, calculate the word vector similarity between any two original triple entities; The calculation expression of the word vector similarity between any two original triple entities is: ； Among them, CS represents the word vector similarity, and Represent the word vectors of any two original triple entities, represents the dot product of the word vectors of any two original triple entities, and Represent word vectors respectively The modulus and word vector The module length; S2-1-3-2, judging whether the word vector similarity between any two original triple entities exceeds the first similarity threshold; S2-1-3-3, if the word vector similarity exceeds the first similarity threshold, define the entity name exceeding the first similarity threshold as a synonym variant name; S2-1-3-4. Replace the synonymous variant names with standard expression names so that any two original triple entities whose word vector similarity exceeds the first similarity threshold are converted into the same standard triple entity.

6. The AI ​​intelligent interaction method based on digital home-based elderly care service according to claim 1 is characterized in that: Using a general model to receive the home risk assessment training set as input, and performing iterative supervised learning until the general model is iterated into the home-based elderly care assessment model, including: S6-1, extract the first batch of home risk assessment training samples from the home risk assessment training set as input features; S6-2, performing forward propagation on the first batch of home risk assessment training samples, and outputting the first batch of pension risk level prediction values; S6-3, calculating the prediction error between the predicted value of the pension risk level in the first batch and the actual pension risk level; S6-4, if the prediction error does not reach the minimum error, updating the model parameters of the general model until the minimum error is reached; S6-5. If the prediction error reaches the minimum error, the model parameters of the current batch are regarded as the model parameters of the home-based elderly care evaluation model; S6-6. Export the model parameters of the home-based elderly care evaluation model, and use them to construct the home-based elderly care evaluation model.

7. An AI intelligent interactive system based on digital home-based elderly care services, executing the interactive method according to any one of claims 1 to 6, characterized in that: The system comprises: The elderly care monitoring report acquisition module is used to obtain the home elderly care monitoring report of the elderly living at home; A health description vector extraction module, used to extract the health description vector of the current elderly person at home from the home-based elderly care monitoring report; The elderly care risk level output module is used to input the health description vector into a preset home-based elderly care assessment model and output the elderly care risk level of the current elderly at home; The pension strategy retrieval module is used to retrieve the predefined pension strategy corresponding to the risk level based on the pension risk level of the elderly living at home and interact with the user; Wherein, the health description vector is a flattened vector of text interaction features generated by interaction item modeling; The modeling method of the home-based elderly care evaluation model includes: S1. Obtain a home-based elderly care monitoring sample set; wherein the sample set includes: a number of home-based elderly care monitoring samples marked with home-based elderly care risk levels; S2. Performing text recognition on the home-based elderly care monitoring samples to extract text interaction features; S3, flattening the text interaction features to generate a health description vector of a one-dimensional array; S4, matching the health description vector with the elderly care risk level to generate home risk assessment training samples; S5. Obtain a number of home risk assessment training samples and aggregate them into a home risk assessment training set; S6. Using the general model to receive the home risk assessment training set as input, and performing iterative supervised learning until the general model is iterated into the home-based elderly care assessment model; The S2 includes: S2-1, extracting the health description entity set of the home-based elderly care monitoring sample; S2-2, performing interaction item modeling on any two health description entities in the health description entity set to generate the text interaction feature; The S2-2 includes: S2-2-1, associate the word vector similarity between any two original triple entities to the health description entity set; S2-2-2, judging whether the word vector similarity between any two health description entities exceeds a second similarity threshold; S2-2-3. If the word vector similarity exceeds the second similarity threshold, select health description entity pairs with significant semantic relevance in the health description entity set, and the entity pairs have a common impact on the pension risk; S2-2-4. Model the interaction items of the health description entity pairs and generate the text interaction features to capture the interaction between the health description entities.

Images