Traditional Chinese medicine nursing knowledge pushing method and system

Abstract

The invention relates to the technical field of information retrieval, in particular to a traditional Chinese medicine nursing knowledge pushing method and system. According to the method, the physical sign data of the nursing object is collected in real time and normalized, a dynamically quantified health state feature set is constructed, and highly similar cases are screened out from historical nursing samples on the basis of the dynamically quantified health state feature set, so that an initial pharmacological keyword set which is most matched with the current physical sign is accurately extracted; further combining previous medication records and drug metabolism rules of nursing objects, calculating drug effect attenuation influence of each drug along with time lapse, performing timeliness weighting on keywords, continuously monitoring pharmacological attributes of new interventional drugs, adjusting weights of related pharmacological keywords in real time, and determining the pharmacological attributes of the new interventional drugs. And finally, sorting is performed according to the dynamically optimized keyword weight, the current most critical nursing direction is locked, and deep semantic matching and sorting are performed on the knowledge text in the direction to ensure that the pushed single nursing knowledge content is not highly related to individual signs.

Description

Technical Field

[0001] This invention relates to the field of information retrieval technology, and in particular to a method and system for pushing traditional Chinese medicine nursing knowledge. Background Technology

[0002] The field of information retrieval technology involves the organization, storage, retrieval, and management of large amounts of data resources. It utilizes computer systems to collect, classify, index, and match various information contents such as text, images, and audio.

[0003] Among them, the method of pushing traditional Chinese medicine nursing knowledge refers to the process of organizing nursing data through a pre-established traditional Chinese medicine knowledge base and providing knowledge content to users based on keyword matching or fixed topic classification in order to meet the needs of disseminating and learning traditional Chinese medicine nursing knowledge in the nursing information system.

[0004] Existing technologies primarily rely on preset keyword matching or fixed topic classifications to push knowledge. This approach has significant limitations when dealing with complex nursing scenarios because it cannot dynamically perceive the individualized real-time changes in the patient's vital signs. The triggering conditions for knowledge matching are relatively fixed, and the knowledge screening process completely ignores the temporal continuity of traditional Chinese medicine interventions and the interactions between drugs. It does not consider the natural decay of the efficacy of the used drugs over time and their continuous impact on the current health status. For example, when a patient receiving blood-activating and stasis-removing treatment shows new signs of heat syndrome, existing technologies may, due to their fixed matching logic, still push routine precautions for blood-activating and stasis-removing treatments, failing to identify the newly emerging need for heat-clearing. This results in the pushed knowledge being out of touch with the patient's most pressing health problem, not only reducing the effectiveness of the knowledge but also potentially delaying the optimal time for collaborative nursing care due to information asymmetry. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a method and system for disseminating traditional Chinese medicine nursing knowledge.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for disseminating traditional Chinese medicine nursing knowledge, comprising the following steps: S1: Collect and normalize the vital signs data of the nursing subjects, map the normalized vital signs data to a preset three-dimensional feature space, and divide the health status feature set according to the position of each vital signs data in the three-dimensional feature space. S2: Obtain similar vital sign data samples from the preset Chinese medicine nursing sample set, compare them with the vital sign data of nursing objects in the health status feature set, calculate the health status similarity index, and screen the candidate set of pharmacological keywords based on the index; S3: Obtain the attribute data of the Chinese medicine used by the nursing object at each medication time and calculate the efficacy decay coefficient, and obtain the time factor of each keyword in the pharmacological keyword candidate set, match the efficacy decay coefficient with the corresponding time factor, and output the pharmacological keyword weight set; S4: Continuously monitor the attribute data of newly introduced traditional Chinese medicines in the nursing plan, update the weights in the pharmacological keyword weight set with reference to the pharmacological keyword set, and generate a keyword weight optimization set; S5: Optimize the weight ranking of keywords in the set according to the keyword weight, extract the text vector data of the keyword with the highest weight from the set of traditional Chinese medicine nursing samples as nursing knowledge content to be pushed, and obtain the knowledge push result.

[0007] As a further embodiment of the present invention, the health status feature set includes the distribution range of vital signs, health status classification labels and cluster center parameters; the pharmacological keyword candidate set includes efficacy category labels, keyword correlation parameters and sample citation indexes; the pharmacological keyword weight set includes keyword matching weight values, efficacy decay coefficients and timeliness factor mapping relationships; the keyword weight optimization set includes updated keyword weights, drug attribute correlation parameters and sensitivity ratios; and the knowledge push results include nursing knowledge text content, keyword text vector data and push priority labels.

[0008] As a further aspect of the present invention, the step of obtaining the health status feature set specifically includes: S111: Collect the body temperature, heart rate and blood pressure data of the nursing subjects to form a raw vital signs data set. Calculate the mean and standard deviation of each data item in the raw vital signs data set, and perform standard deviation normalization to obtain the normalized vital signs values. S112: Call the normalized values ​​of vital signs, map the normalized body temperature data, heart rate data and blood pressure data together to form vital sign coordinate points in the three-dimensional feature space, calculate the Euclidean distance between the vital sign coordinate points, and establish a set of Euclidean distance values. S113: Perform clustering based on the density of the Euclidean distance value set in the three-dimensional feature space, identify the densely distributed areas of coordinate points as independent clusters, calculate the geometric center of all coordinate points in each cluster as the cluster center, use the cluster center as a reference, and define the range boundary of the health status according to the preset distance threshold to generate a health status feature set.

[0009] As a further aspect of the present invention, the step of obtaining the candidate set of pharmacological keywords specifically includes: S211: Call the vital sign data of the nursing objects in the health status feature set, and obtain similar vital sign data samples from the preset Chinese medicine nursing sample set. Pair and compare the vital sign data of the nursing objects with the vital sign data of each sample, and calculate the health status similarity index of each sample. S212: Set a similarity screening threshold, compare the health status similarity index of each sample with the similarity screening threshold one by one, screen out samples with index values ​​higher than the similarity screening threshold, and establish a high similarity sample set. S213: Extract the pharmacological keywords corresponding to each sample in the high similarity sample set to form a candidate set of pharmacological keywords.

[0010] As a further aspect of the present invention, the step of obtaining the pharmacological keyword weight set specifically includes: S311: Obtain the attribute data of the Chinese medicine used for each medication time of the nursing subject. The attribute data of the Chinese medicine includes the drug's duration of action parameter, metabolic rate parameter, and duration of action parameter. Calculate the time difference based on the medication time node data and the drug's duration of action parameter. Combine the time difference, metabolic rate parameter, and duration of action parameter to calculate the drug efficacy decay coefficient. S312: Obtain the time-effect factor of each keyword in the candidate set of pharmacological keywords, and match the time-effect factor of each keyword with the corresponding drug efficacy decay coefficient according to the correspondence between keywords and drug administration time, and establish the decay information corresponding to the keyword. S313: For each matching item in the decay information corresponding to the keyword, based on the time factor, apply the matching drug efficacy decay coefficient to perform an exponential decay operation, calculate a weight value for each keyword, and integrate the weight values ​​of all keywords to output a set of pharmacological keyword weights.

[0011] As a further aspect of the present invention, the step of obtaining the keyword weight optimization set specifically includes: S411: Continuously monitor the use of new traditional Chinese medicines in the nursing plan, collect attribute data of the new traditional Chinese medicines, and establish a dataset of new pharmacological attributes; S412: Call the new pharmacological attribute dataset and refer to the corresponding parameters of the associated drugs in the candidate set of pharmacological keywords to calculate the correlation between the pharmacological attribute data of the newly introduced drug and the parameters of the associated drugs, and generate the cross-sensitivity ratio. S413: Based on the cross-sensitivity ratio and by calling the current weight values ​​in the pharmacological keyword weight set, reassign the weights of the relevant keywords, and recombine the updated weight values ​​to generate a keyword weight optimization set.

[0012] As a further aspect of the present invention, the step of obtaining the knowledge push result specifically includes: S511: Sort the keywords in the keyword weight optimization set in descending order according to the keyword weight values, lock the keyword with the highest weight, and extract the corresponding knowledge text data from the traditional Chinese medicine nursing sample set based on the keyword with the highest weight to form the knowledge text to be processed. S512: Calculate the frequency of each word in the knowledge text to be processed and the inverse document frequency of the word in the entire Chinese medicine nursing sample set text. Based on the word frequency and inverse document frequency, obtain the keyword text vector data. S513: Calculate the comprehensive matching degree between each knowledge text and the keyword with the highest initial weight based on the keyword text vector data, compare the comprehensive matching degree of all candidate knowledge texts, select the nursing knowledge content with the highest comprehensive matching degree as the nursing knowledge content for push, and obtain the knowledge push result.

[0013] As a further aspect of the present invention, the process of calculating the comprehensive matching degree between each knowledge text and the keyword with the highest initial weight is specifically as follows: Extract the term weights directly corresponding to the keyword with the highest initial weight from the keyword text vector data, and use them as the core word matching scores; Identify all statements with the highest initial keyword weight from the keyword text vector data, and perform a weighted average calculation on the statement weights of these statements to obtain the related statement matching score; Set the weight coefficients for core keywords and related sentences; Multiply the core word matching score by the core word weight coefficient, and multiply the related sentence matching score by the related sentence weight coefficient. Sum the two products to obtain the overall matching degree.

[0014] A traditional Chinese medicine (TCM) nursing knowledge push system, the TCM nursing knowledge push system being used to execute the above-mentioned TCM nursing knowledge push method, the system comprising: The health status feature extraction module collects and normalizes the vital signs data of the nursing subjects, maps the normalized vital signs data to a preset three-dimensional feature space, and divides the health status feature set according to the position of each vital sign data in the three-dimensional feature space. The pharmacological keyword screening module obtains similar vital sign data samples from a preset set of traditional Chinese medicine nursing samples, compares them with the vital sign data of nursing subjects in the set of health status characteristics, calculates a health status similarity index, and screens a candidate set of pharmacological keywords based on the index. The pharmacological keyword weight calculation module obtains the attribute data of the Chinese medicine used by the nursing subject at each medication time and calculates the efficacy decay coefficient, and obtains the time factor of each keyword in the pharmacological keyword candidate set, matches the efficacy decay coefficient with the corresponding time factor, and outputs the pharmacological keyword weight set. The keyword weight optimization module continuously monitors the attribute data of newly introduced traditional Chinese medicines in the nursing plan, updates the weights in the pharmacological keyword weight set with reference to the pharmacological keyword set, and generates a keyword weight optimization set. The nursing knowledge content push module optimizes the weight ranking of keywords in the set according to the keyword weight optimization, extracts the text vector data of the keyword with the highest weight from the traditional Chinese medicine nursing sample set as nursing knowledge content to be pushed, and obtains the knowledge push results.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by collecting and normalizing the vital signs data of nursing subjects in real time, a dynamically quantified set of health status features is constructed. Based on this, highly similar cases are selected from historical nursing samples, thereby accurately extracting the initial set of pharmacological keywords that best match the current vital signs. Furthermore, by combining the nursing subjects' past medication records and drug metabolism patterns, the effect of drug efficacy decay over time is calculated, and the keywords are weighted according to their timeliness. At the same time, the pharmacological properties of newly introduced drugs can be continuously monitored. By quantifying the interaction correlation between new and old drugs, the weights of relevant pharmacological keywords are adjusted in real time. Finally, the keywords are sorted according to the dynamically optimized keyword weights to identify the most critical nursing direction. Deep semantic matching and sorting are then performed on the knowledge text under this direction to ensure that the pushed single nursing knowledge content is not only highly relevant to individual vital signs but also dynamically responds to changes in treatment plans. This achieves the extraction of accurate and timely personalized guidance from a massive knowledge base, greatly improving the pertinence and practical value of nursing knowledge. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the workflow of the present invention; Figure 2 This is a flowchart of step S1 of the present invention; Figure 3 This is a flowchart of step S2 of the present invention; Figure 4 This is a flowchart of step S3 of the present invention; Figure 5 This is a flowchart of step S4 of the present invention; Figure 6 This is a flowchart of step S5 of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] Please see Figure 1This invention provides a technical solution: a method for disseminating traditional Chinese medicine nursing knowledge, comprising the following steps: S1: Collect and normalize the vital signs data of the nursing subjects, map the normalized vital signs data to a preset three-dimensional feature space, and divide the health status feature set according to the position of each vital signs data in the three-dimensional feature space. S2: Obtain similar vital sign data samples from the preset Chinese medicine nursing sample set, compare them with the vital sign data of nursing objects in the health status feature set, calculate the health status similarity index, and screen the candidate set of pharmacological keywords based on the index; S3: Obtain the attribute data of the Chinese medicine used by the nursing object at each medication time and calculate the efficacy decay coefficient, and obtain the time factor of each keyword in the pharmacological keyword candidate set, match the efficacy decay coefficient with the corresponding time factor, and output the pharmacological keyword weight set; S4: Continuously monitor the attribute data of newly introduced traditional Chinese medicines in the nursing plan, update the weights in the pharmacological keyword weight set with reference to the pharmacological keyword set, and generate a keyword weight optimization set; S5: Optimize the weight ranking of keywords in the set according to the keyword weight, extract the text vector data of the keyword with the highest weight from the set of traditional Chinese medicine nursing samples as nursing knowledge content to be pushed, and obtain the knowledge push result.

[0019] The health status feature set includes the distribution range of vital signs, health status classification labels, and cluster center parameters; the pharmacological keyword candidate set includes efficacy category labels, keyword relevance parameters, and sample citation indexes; the pharmacological keyword weight set includes keyword matching weight values, efficacy decay coefficients, and timeliness factor mapping relationships; the keyword weight optimization set includes updated keyword weights, drug attribute association parameters, and sensitivity ratios; and the knowledge push results include nursing knowledge text content, keyword text vector data, and push priority labels.

[0020] Please see Figure 2 The specific steps for obtaining the health status feature set are as follows: S111: Collect the body temperature, heart rate and blood pressure data of the nursing subjects to form a raw vital signs data set. Calculate the mean and standard deviation of each data item in the raw vital signs data set, and perform standard deviation normalization to obtain the normalized vital signs values. By using a smart vital sign monitoring device worn on the wrist of the patient, data was collected continuously for 2 hours at a frequency of once every 10 minutes to obtain a raw set of vital sign data containing 12 sets of data. The body temperature data sequence is {36.7, 36.8, 36.7, 36.9, 36.8, 36.8, 37.0, 36.9, 36.8, 36.7, 36.9, 36.8}, in degrees Celsius; the heart rate data sequence is {72, 74, 73, 75, 74, 74, 76, 75, 74, 73, 75, 74}, in beats per minute; the blood pressure data sequence is {121, 123, 122, 124, 123, 123, 125, 124, 123, 122, 124, 123}, and the diastolic blood pressure data sequence is {81, 82, 81, 83, 82, 82, 84, 83, 82, 81, 83, 82}, in millimeters of mercury. For the four sets of raw vital signs data collected, their arithmetic means were calculated. The mean of the body temperature data was calculated by summing the 12 temperature values ​​in the set and dividing by 12, i.e. Celsius. The mean of heart rate data is calculated as follows: Bpm. The mean of systolic blood pressure data is calculated as follows: mmHg. The mean of diastolic blood pressure data was calculated as follows: The standard deviation of each data point is calculated based on the mean of the data points. The standard deviation for body temperature is calculated as follows: the difference between each measured temperature value and the mean temperature of 36.82 is squared, the sum of these 12 squared differences is divided by 12, and the square root is taken, resulting in a standard deviation of 0.09 degrees Celsius. The standard deviation for heart rate is calculated similarly to body temperature, resulting in a standard deviation of 1.00 beats / minute. The standard deviation for systolic blood pressure and diastolic blood pressure is also 1.00 mmHg. Next, standard deviation normalization is performed, converting each data point in the original vital signs dataset into a normalized value. The normalization process involves taking an original data point, subtracting the mean of the corresponding data type, and then dividing the difference by the standard deviation of the corresponding data type. For example, for the body temperature data of 36.7 degrees Celsius at the first data point, the normalized value is... For the heart rate data of 72 beats / minute at the first collection point, its normalized value is: For the systolic blood pressure data of 121 mmHg at the first collection point, its normalized value is: For the diastolic blood pressure data of 81 mmHg at the first collection point, its normalized value is: Following this method, the same calculation was performed on each of the four vital signs data at all 12 time points, ultimately obtaining a set of normalized vital sign values ​​consisting of 48 values.

[0021] S112: Call the normalized values ​​of vital signs, map the normalized body temperature data, heart rate data and blood pressure data together to form vital sign coordinate points in the three-dimensional feature space, calculate the Euclidean distance between vital sign coordinate points, and establish a set of Euclidean distance values. The normalized vital signs data set is retrieved, and normalized body temperature, normalized heart rate, and normalized systolic blood pressure data are selected as the three dimensions and mapped to a three-dimensional feature space. Diastolic blood pressure data is not used for spatial mapping in this step, but its normalized value is retained. In this three-dimensional feature space, the X-axis represents normalized body temperature, the Y-axis represents normalized heart rate, and the Z-axis represents normalized systolic blood pressure. The three normalized values ​​at each acquisition time point constitute a vital signs coordinate point. Taking the aforementioned example data, the data at the first acquisition time point is mapped to coordinate point P1, with coordinates (-1.33, -2.00, -2.00). The normalized body temperature value at the second acquisition time point is... Normalized heart rate value Normalized systolic blood pressure value Therefore, the coordinates of point P2 are (-0.22, 0.00, 0.00). Similarly, all the vital sign data from the 12 acquisition time points are mapped to 12 coordinate points in the three-dimensional feature space, designated P1 to P12. After mapping all coordinate points, the Euclidean distance between each pair of these points is calculated. The Euclidean distance is calculated as follows: for any two coordinate points P1(x1, y1, z1) and P2(x2, y2, z2), the distance between them is equal to... squared plus squared plus Square the product, and finally take the square root of the sum. Specifically, calculate the Euclidean distance between P1 and P2, substitute the coordinate values, and the distance is... Following this method, continue calculating the distances between all point pairs from P1 to P3, P1 to P4, and so on up to P11 to P12. Since there are 66 point pairs in total for 12 points, this will generate 66 Euclidean distance values. Integrate these 66 calculated Euclidean distance values ​​to establish an unordered set of Euclidean distance values.

[0022] S113: Perform clustering based on the density of the Euclidean distance value set in the three-dimensional feature space, identify the dense area of ​​coordinate points as an independent cluster, calculate the geometric center of all coordinate points in each cluster as the cluster center, and use the cluster center as the reference to define the range boundary of the health status according to the preset distance threshold to generate a health status feature set. Clustering is performed based on the set of Euclidean distance values ​​and the specific location distribution of 12 coordinate points in the 3D feature space. A point... The condition for being identified as a core point is that it falls within a preset neighborhood radius. The number of neighboring points within the area reaches the minimum point threshold. .point Neighbor set The definition is as follows: ,in, This refers to the coordinates of a specific vital sign that is currently being examined. It refers to any point in the set of vital sign coordinate points; It refers to the set of all 12 mapped feature coordinate points generated by S112; It is a guide and points The Euclidean distance between them is calculated as described in S112; (epsilon) is the neighborhood radius, determined by performing S111 and S112 processes on a baseline database of individuals with stable health, calculating the Euclidean distances between all coordinate points, and multiplying the median of these distances by a specific coefficient (e.g., 0.8), which is set to 1.5 here. It is a dimensionless value, calculated from the Euclidean distances of the normalized data. Dimensions are consistent. Point The criteria for determining whether something is a core point are: ,in, Neighbor set The number of points contained therein, its dimension is the number of points; This is the minimum threshold for the number of points. It is set based on clustering experiments conducted on the aforementioned benchmark database. By testing different values, the value that most effectively aggregates core health status points while best excluding occasional physiological fluctuations was selected; here, it is set to 4. Its unit of measurement is also the number of points, similar to... Maintain consistency.

[0023] Taking checking whether coordinate point P2 (coordinates (-0.22, 0.00, 0.00)) is a core point as an example, the complete calculation process is as follows: 1. Parameter setting: Check point : Neighborhood radius of P2(-0.22, 0.00, 0.00) 1.5 Minimum number of points threshold 42. Calculate the Euclidean distance between P2 and other points. Suppose that in addition to P2, there are four other points with coordinates P3(-0.22, 1.00, 1.00), P5(0.89, 0.00, 0.00), P7(2.00, 1.00, 1.00), and P8(0.89, -1.00, -1.00).

[0024] ; ; ; ; Building a Neighbor Set :because Therefore, P3 is within the neighborhood of P2. Because Therefore, P5 is within the neighborhood of P2. Because Therefore, P7 is not within the neighborhood of P2. Because Therefore, P8 is not within the neighborhood of P2. Assume that, after calculation, there are two other points, P9 and P10, that are also within the 1.5 neighborhood radius of P2. A point itself is also its own neighbor; therefore, the neighborhood set of P2... It must contain at least {P2, P3, P5, P9, P10}. 4. Determine if it is a core point: The number of points in the neighbor set is... Compare the number of neighbors with the minimum number of points threshold: .because The condition is met, so coordinate point P2 is identified as a core point.

[0025] Subsequently, P2 and all its neighboring points {P2, P3, P5, P9, P10} are grouped together to form an initial cluster. Then, an unchecked point (such as P3) is selected from this cluster, and the above process is repeated, incorporating new points in its neighborhood into the current cluster until the cluster no longer expands. If the number of points in the neighborhood of the starting point is less than 4, it is temporarily marked as a noise point. After identifying a cluster, a new starting point is selected from the remaining undivided points, and the entire process is repeated. Assume the calculated cluster center coordinates are (0.15, 0.50, 0.55). Then, based on a preset distance threshold of 2.5, the boundary of the healthy status range is defined using the cluster center as the reference. The distance threshold is set based on: analyzing data from two different groups (one group of confirmed healthy individuals, and the other group of individuals with clear records of abnormal physical signs), comparing the maximum offset distance of the physical sign coordinates of the healthy group and the offset distance distribution of the coordinates of the abnormal group, and selecting a value that maximizes the distinction between these two groups as the threshold. Finally, the cluster center coordinates (0.15, 0.50, 0.55) and the distance threshold of 2.5 are defined together as the health status feature set of the care subject.

[0026] Please see Figure 3 The specific steps for obtaining the candidate set of pharmacological keywords are as follows: S211: Call the vital sign data of the nursing subjects in the health status feature set, and obtain similar vital sign data samples from the preset Chinese medicine nursing sample set. Pair and compare the vital sign data of the nursing subjects with the vital sign data of each sample, and calculate the health status similarity index of each sample. The original vital signs data of the nursing subjects were retrieved, specifically the mean values ​​of body temperature 36.82 degrees Celsius, heart rate 74.00 beats / minute, systolic blood pressure 123.00 mmHg, and diastolic blood pressure 82.00 mmHg. Simultaneously, similar vital sign data samples were obtained from a traditional Chinese medicine (TCM) nursing sample set. This TCM nursing sample set is a database constructed by collecting and structuring effective nursing cases from previous years. Each case (sample) contains one piece of structured data, which includes the patient's mean vital signs at a specific treatment stage, the TCM intervention plan used at that time, and the corresponding pharmacological keywords. For example, three samples were extracted from the sample set: Sample 1, vital signs data {body temperature 37.1 degrees Celsius, heart rate 78 bpm, systolic blood pressure 128 bpm, diastolic blood pressure 85 bpm}; Sample 2, vital signs data {body temperature 36.9 degrees Celsius, heart rate 75 bpm, systolic blood pressure 124 bpm, diastolic blood pressure 83 bpm}; Sample 3, vital signs data {body temperature 36.5 degrees Celsius, heart rate 68 bpm, systolic blood pressure 115 bpm, diastolic blood pressure 75 bpm}. The vital sign data of the current nursing subject are paired and compared with the vital sign data of each sample in the sample set to calculate the health status similarity index for each sample. First, the four vital sign data of the nursing subject and each sample are normalized to the same standard deviation as in S111. The mean and standard deviation used are global baseline values ​​obtained from a prior statistical analysis of all data in the traditional Chinese medicine nursing sample set. Assume the global mean is {body temperature 37.0, heart rate 75, systolic blood pressure 125, diastolic blood pressure 84}, and the global standard deviation is {body temperature 0.5, heart rate 5, systolic blood pressure 8, diastolic blood pressure 6}. The normalized vital sign vector of the nursing subject is calculated as follows: Body temperature... Heart rate systolic blood pressure diastolic blood pressure The normalized vital sign vector for sample two is calculated as: body temperature. Heart rate systolic blood pressure diastolic blood pressure Then, the Euclidean distance between the normalized vital sign vector of the nursing subject and the normalized vital sign vector of each sample is calculated. For example, the distance with sample two is... Finally, the calculated Euclidean distance is passed through the inverse function. Converted to a similarity index. The similarity index between the health status of the nursing subject and sample two is... Following this method, the similarity index of health status between each nursing subject and all samples in the sample set is calculated one by one.

[0027] S212: Set a similarity screening threshold, compare the health status similarity index of each sample with the similarity screening threshold one by one, screen out samples with index values ​​higher than the similarity screening threshold, and establish a high similarity sample set. A similarity screening threshold is set based on the following: A retrospective analysis of a validated database of effective traditional Chinese medicine nursing cases is conducted to calculate the similarity index distribution between successful cases and subsequent similar matching cases. This threshold is set at a level that ensures a high probability of consistency between the pharmacological keywords corresponding to the selected samples and the subsequent improvement direction of the patients; here, it is set to 0.75. The calculated health status similarity index of each sample is compared numerically with the similarity screening threshold of 0.75. For sample one, the calculated similarity index is assumed to be 0.65; for sample two, the similarity index is 0.781; and for sample three, the similarity index is 0.45. The comparison process is as follows: The index of sample one, 0.65, is compared with 0.75, because... Therefore, it was not selected. Comparing the index of sample two, 0.781 and 0.75, because... Therefore, sample two was selected. Comparing the index of sample three (0.45 and 0.75), because... Therefore, they were not selected. This screening process was repeated for all samples in the traditional Chinese medicine nursing sample set. All samples with a health status similarity index value higher than 0.75 were selected and integrated to establish a high similarity sample set.

[0028] S213: Extract the pharmacological keywords corresponding to each sample in the set of highly similar samples to form a candidate set of pharmacological keywords; From the set of highly similar samples, the pharmacological keywords corresponding to each sample are extracted. These pharmacological keywords are manually assigned by TCM experts based on the TCM intervention plan and its main efficacy for each sample during the construction of the TCM nursing sample set. Examples include "clearing heat and detoxifying," "activating blood circulation and removing blood stasis," and "soothing the liver and regulating qi." The extraction process involves iterating through each sample in the highly similar sample set, accessing the pharmacological keyword field stored in its data structure, and copying the text content of that field. Assume that in the highly similar sample set, besides sample two, two other samples, sample four and sample five, are selected. The pharmacological keywords corresponding to sample two are "soothing the liver and regulating qi" and "strengthening the spleen and stomach." The pharmacological keywords corresponding to sample four are "soothing the liver and regulating qi" and "calming the mind and soothing the nerves." The pharmacological keywords corresponding to sample five are "nourishing yin and reducing fire" and "strengthening the spleen and soothing the stomach." All the pharmacological keywords corresponding to these three samples are then combined to obtain a list containing "soothing the liver and regulating qi," "strengthening the spleen and soothing the stomach," "soothing the liver and regulating qi," "calming the mind and soothing the nerves," "nourishing yin and reducing fire," and "strengthening the spleen and soothing the stomach." The keywords in the list are deduplicated, with each keyword retained only once. The final set of pharmacological keyword candidates is {"soothe the liver and regulate qi", "strengthen the spleen and stomach", "calm the mind and soothe the nerves", "nourish yin and reduce fire"}.

[0029] Please see Figure 4The specific steps for obtaining the pharmacological keyword weight set are as follows: S311: Obtain the attribute data of the Chinese medicine used for each medication time of the nursing subject. The attribute data of the Chinese medicine includes the drug's duration of action parameter, metabolic rate parameter, and duration of action parameter. Calculate the time difference based on the medication time node data and the drug's duration of action parameter. Combine the time difference, metabolic rate parameter, and duration of action parameter to calculate the drug efficacy decay coefficient. Data on the Chinese herbal medicines used by the patient at each medication time point over a past period and their attributes were obtained. Assume the patient had two medication records in the past three days: the first medication was taken 72 hours ago (medicine A); the second medication was taken 24 hours ago (medicine B). The attribute data for the Chinese herbal medicines were retrieved from a Chinese herbal medicine attribute database. The attribute data for drug A are: duration of action parameter of 48 hours, metabolic rate parameter of 0.03, and duration of action parameter of 24 hours. The attribute data for drug B are: duration of action parameter of 36 hours, metabolic rate parameter of 0.04, and duration of action parameter of 18 hours. These parameters are based on pharmacokinetic research literature and clinical trial data. The time difference between the current time point and each medication time point was calculated. For drug A, the time difference is 72 hours. For drug B, the time difference is 24 hours. Next, combining the time difference, metabolic rate parameter, and duration of action parameter, the efficacy decay coefficient of each drug at the current time point was calculated. The calculation process is as follows: First, determine whether the time difference exceeds the duration of drug effect parameter. If it does, the efficacy attenuation coefficient is set to 0. For drug A, the time difference of 72 hours is greater than its duration of drug effect of 48 hours; therefore, the efficacy attenuation coefficient for drug A is 0. For drug B, its time difference of 24 hours does not exceed its duration of drug effect of 36 hours, and the efficacy attenuation coefficient is calculated as follows: After substituting the values, it becomes Calculated .

[0030] S312: Obtain the time-effect factor of each keyword in the candidate set of pharmacological keywords, and match the time-effect factor of each keyword with the corresponding drug efficacy decay coefficient based on the correspondence between keywords and drug administration time, and establish the decay information corresponding to the keyword. The timeliness factor of each keyword in the candidate set of pharmacological keywords is obtained. The timeliness factor characterizes the closeness of the association between the keyword and recent medication use. It is set based on analyzing the temporal correlation between keywords and medication behavior in historical case data, and quantified according to the frequency of occurrence of the keyword in a high-similarity sample set. The higher the frequency of the keyword, the larger its basic timeliness factor. Assuming that based on frequency, the basic timeliness factor for "soothing the liver and regulating qi" is 0.9, for "strengthening the spleen and stomach" it is 0.8, for "clearing the heart and calming the mind" it is 0.6, and for "nourishing yin and reducing fire" it is 0.5. Based on the correspondence between keywords and medication time, the timeliness factor of each keyword is matched with its corresponding efficacy decay coefficient. This correspondence comes from a traditional Chinese medicine attribute database, which records the pharmacological keywords associated with each type of traditional Chinese medicine. Assuming drug A is associated with "nourishing yin and reducing fire," and drug B is associated with "soothing the liver and regulating qi" and "strengthening the spleen and stomach," since the efficacy decay coefficient of drug A is 0, the decay coefficient matched for its associated keyword "nourishing yin and reducing fire" is also 0. The efficacy attenuation coefficient of drug B is 0.9468, therefore, its associated keywords "soothing the liver and regulating qi" and "strengthening the spleen and stomach" both match an attenuation coefficient of 0.9468. For the keyword "calming the mind and soothing the nerves," which exists in the candidate set but is not directly related to recent medication use, a default minimum attenuation coefficient of 0.1 is assigned. This coefficient is set based on an experimentally determined minimum value that preserves the basic information of the dominant keyword without interfering with it. Thus, the attenuation information corresponding to the keywords is established: "soothing the liver and regulating qi" corresponds to 0.9468, "strengthening the spleen and stomach" corresponds to 0.9468, "calming the mind and soothing the nerves" corresponds to 0.1, and "nourishing yin and reducing fire" corresponds to 0.

[0031] S313: For each set of matching items in the decay information corresponding to the keyword, based on the time factor, apply the matching drug efficacy decay coefficient to perform an exponential decay operation, calculate a weight value for each keyword, and integrate the weight values ​​of all keywords to output a set of pharmacological keyword weights. For each matching item in the keyword decay information, an operation is performed to calculate the weight value of each keyword. The operation involves multiplying the keyword's base time-effectiveness factor by the matched drug efficacy decay coefficient. For the keyword "soothe the liver and regulate qi," its weight value is calculated as follows: The weight value for the keyword "strengthening the spleen and stomach" is calculated as follows: The weight value for the keyword "calming the mind and soothing the nerves" is calculated as follows: The weight value for the keyword "nourishing yin and reducing fire" is calculated as follows: The calculated weight values ​​of all keywords are integrated to output a structured set of pharmacological keyword weights, which is as follows: {"Soothe the liver and regulate qi": 0.852, "Strengthen the spleen and stomach": 0.757, "Calm the mind and soothe the nerves": 0.06, "Nourish yin and reduce fire": 0}.

[0032] Please see Figure 5 The specific steps for obtaining the keyword weight optimization set are as follows: S411: Continuously monitor the use of new traditional Chinese medicines in the nursing plan, collect attribute data of the new traditional Chinese medicines, and establish a dataset of new pharmacological attributes; During the implementation of the nursing plan, the use of newly introduced traditional Chinese medicine (TCM) is continuously monitored through the medical order entry interface. Assume that after generating the pharmacological keyword weight set, a new TCM herb, denoted as drug C, is introduced into the nursing plan. Once the introduction of the new drug is detected, the attribute data of drug C is immediately collected. By accessing the TCM attribute database, the pharmacological attribute information of drug C is obtained, such as: {Drug Name: "Drug C", Main Efficacy Category: "Qi-regulating", Property: "Warm", Meridian Tropism: "Liver, Spleen", Main Chemical Component Group: "Volatile Oils"}. These collected attribute data are then integrated to establish a separate new pharmacological attribute dataset.

[0033] S412: Call the new pharmacological attribute dataset and, with reference to the corresponding parameters of the associated drugs in the candidate set of pharmacological keywords, calculate the correlation between the pharmacological attribute data of the newly introduced drug and the parameters of the associated drugs, and generate the cross-sensitivity ratio. The pharmacological attribute dataset of drug C is accessed, and the correlation between the newly introduced drug C and the already associated drug B is calculated by referring to the corresponding parameters of existing related drugs in the candidate set of pharmacological keywords. The correlation is obtained by weighted summation after quantitative comparison of multiple attributes, and the specific calculation method is as follows: ,in, The degree of association between the newly introduced drug C and the already associated drug B; Drug C represents the newly introduced drug; This represents existing drug B; The number of pharmacological properties included in the comparison is 4 in this example; It is an index of pharmacological properties, from 1 to n; It is the first The weights of each attribute are set based on expert scoring and historical data regression analysis. Attributes that better reflect the essence of the drug (such as "main efficacy category") are given greater weights. Here, the weight of the main efficacy category is set to 0.4, and the weights of the other three items are each 0.2. Drugs C and B are in the first The quantitative similarity score on each attribute is determined according to preset pharmacological rules; and These are the first two drugs, C and B. Each attribute value; It is the first The contribution factors of each attribute are determined based on statistical analysis of data on synergistic or antagonistic drug interactions. They are used to amplify or reduce the influence of a particular attribute in a specific scenario. If it is known that a certain attribute contributes more to synergistic effects, then its contribution factor value will be larger. Here, they are set to 1.1, 1.0, 1.0, and 1.0. All parameters... , and All values ​​are dimensionless, ensuring the final result is accurate. Dimensional consistency.

[0034] Taking the calculation of the correlation between drug C and drug B as an example, the complete calculation process is as follows: 1. Parameter setting: number of comparison attributes 4-attribute weights : (Main efficacy categories) (Medicinal properties), (Meridian Tropism) (Major Chemical Component Group) Attribute Contribution Factor : , , , Properties of drug C: =“Qi-regulating category” = "warm" =“Liver and Spleen” =Attributes of drug B, which is a "volatile oil" product: =“Qi-regulating category” = "flat", =“Liver, Stomach” =“Flavones” 2. Calculate the quantitative similarity score for each attribute. According to preset pharmacological rules (e.g., same attribute value = 1.0, similar attribute value = 0.7, partially overlapping attribute value = 0.5, completely different attribute value = 0): The efficacy category "Regulating Qi" is the same as "Regulating Qi," with a score of 1.0. The medicinal properties of "warm" and "neutral" were defined as similar, with a score of 0.7. The meridian tropism of "Liver and Spleen" overlaps with that of "Liver and Stomach" by a single "Liver," which is defined as partial overlap and scored as 0.5. The chemical composition groups "volatile oils" and "flavonoids" are completely different, with a score of 0.3. Substitute them into the formula to calculate the correlation. i=1 (Main Function Category): i=2 (medicinal properties): i=3 (meridian tropism): i=4 (Main chemical component group): This correlation score is used to generate the cross-sensitivity ratio. The cross-sensitivity ratio is calculated by multiplying the correlation score by a preset sensitivity modifier. This modifier is preset based on the synergistic or antagonistic effects of the drugs. Assuming drug C and drug B have a synergistic effect, the modifier is set to 0.2. The cross-sensitivity ratio equals... .

[0035] S413: Based on the cross-sensitivity ratio and by calling the current weight values ​​in the pharmacological keyword weight set, reassign the weights of the relevant keywords, recombine the updated weight values, and generate a keyword weight optimization set. Based on the calculated cross-sensitivity ratio of 0.136, and by accessing the current weight values ​​of each keyword in the pharmacological keyword weight set, the weights of the relevant keywords are reassigned. Relevant keywords refer to those jointly influenced by the newly introduced drug C and the associated existing drug B. Drug B is associated with "soothing the liver and regulating qi" and "strengthening the spleen and harmonizing the stomach," therefore the weights of these two keywords will be adjusted. The reassignment calculation process is as follows: The current weight for "soothing the liver and regulating qi" is 0.852, and the updated weight is... For "strengthening the spleen and stomach," its current weight is 0.757, and the updated weight is... For other unrelated keywords "calm the mind and soothe the nerves" and "nourish yin and reduce fire", their weights remain unchanged. All updated weight values ​​are then recombined with the unupdated weight values ​​to generate a keyword weight optimization set, updated as follows: {"soothe the liver and regulate qi": 0.968, "strengthen the spleen and stomach": 0.860, "calm the mind and soothe the nerves": 0.06, "nourish yin and reduce fire": 0}.

[0036] Please see Figure 6 The specific steps for obtaining knowledge push results are as follows: S511: Optimize the keywords in the keyword weight set by sorting them in descending order, lock the keyword with the highest weight, and extract the corresponding knowledge text data from the traditional Chinese medicine nursing sample set based on the keyword with the highest weight to form the knowledge text to be processed. Optimize the weight values of each keyword in accordance with the keyword weights, and perform a descending order arrangement. The arrangement result is: "soothing the liver and regulating qi" (0.968), "invigorating the spleen and harmonizing the stomach" (0.860), "clearing the heart and tranquilizing the mind" (0.06), "nourishing yin and reducing fire" (0). Lock the keyword "soothing the liver and regulating qi" with the highest weight value after arrangement. Based on the keyword "soothing the liver and regulating qi" with the highest weight, retrieve from the traditional Chinese medicine nursing sample set and extract the knowledge text data corresponding to all the samples marked with the pharmacological keyword "soothing the liver and regulating qi". The knowledge text data is unstructured text written by experts associated with each case in the traditional Chinese medicine nursing sample set, and the content includes detailed nursing measures, dietary suggestions, emotional regulation methods, etc. For example, the retrieval operation found 5 knowledge texts related to "soothing the liver and regulating qi", which are respectively recorded as Text 1 to Text 5. Pool these 5 knowledge texts together to form a set of knowledge texts to be processed.

[0037] S512: Statistically analyze the frequency of occurrence of each word in the knowledge texts to be processed and calculate the inverse document frequency of the word in all the texts of the traditional Chinese medicine nursing sample set. Based on the word frequency and the inverse document frequency, obtain the keyword text vector data; For each text in the formed set of knowledge texts to be processed, as well as all the knowledge texts in the entire traditional Chinese medicine nursing sample set, perform the calculation of word frequency and inverse document frequency. First, perform word segmentation on each knowledge text to be processed (such as Text 1), and statistically analyze the frequency of occurrence of each word in it (TF). For example, in Text 1, the word "emotion" appears 3 times, and the total number of words in the text is 300, then the word frequency of "emotion" is . Secondly, calculate the inverse document frequency (IDF) of each word in all the texts of the traditional Chinese medicine nursing sample set. Assume that the entire sample set contains N knowledge texts, and among them, M texts contain the word "emotion". Then the calculation of the inverse document frequency of "emotion" is . Taking N = 10000 and M = 500 as an example, the inverse document frequency of "emotion" is . Based on the word frequency and the inverse document frequency, calculate the weight value (TF-IDF value) of each word. The calculation method is the multiplication of the word frequency and the inverse document frequency. In Text 1, the TF-IDF value of "emotion" is . Perform this calculation on all the words in Text 1 to obtain a set of keyword text vector data representing Text 1. Similarly, perform the same calculation on the other four knowledge texts to be processed respectively to obtain their respective keyword text vector data.

[0038] S513: Calculate the comprehensive matching degree of each knowledge text with the keyword with the highest initial weight based on the keyword text vector data, and compare the comprehensive matching degrees of all the candidate knowledge texts. Select the nursing knowledge content with the highest comprehensive matching degree as the nursing knowledge content to be pushed to obtain the knowledge push result; Among them, the process of calculating the comprehensive matching degree between each piece of knowledge text and the keyword with the highest initial weight is as follows: Extract the term weight directly corresponding to the keyword with the highest initial weight from the keyword text vector data as the core word matching score; Identify all the sentences of the keyword with the highest initial weight from the keyword text vector data, and perform a weighted average operation on the sentence weights of these sentences to obtain the associated sentence matching score; Set the core word weight coefficient and the associated sentence weight coefficient; Multiply the core word matching score by the core word weight coefficient, multiply the associated sentence matching score by the associated sentence weight coefficient, and sum the two product results to obtain the comprehensive matching degree; According to the keyword text vector data of each piece of knowledge text, calculate the comprehensive matching degree between each piece of knowledge text and the keyword with the highest initial weight "soothing the liver and regulating qi". First, extract the term weight directly corresponding to the keyword with the highest initial weight "soothing the liver and regulating qi" from the keyword text vector data of each text, that is, the TF-IDF value of the word "soothing the liver and regulating qi" itself in this text. Suppose in text one, the TF-IDF value of "soothing the liver and regulating qi" is 0.08, and this value is the core word matching score. Then, identify all the sentences containing "soothing the liver and regulating qi" in text one, and perform a weighted average operation on the weights of these sentences. The calculation method of the sentence weight is: add the TF-IDF values of all the words in the sentence. Suppose text one has two sentences containing "soothing the liver and regulating qi", and the sentence weights are 1.2 and 1.5 respectively, then the associated sentence matching score is . Next, set the core word weight coefficient and the associated sentence weight coefficient. The setting of these two coefficients is based on the prior knowledge in the field of information retrieval and is determined through tuning experiments on a dedicated test set to achieve the best knowledge text sorting effect. Here, the core word weight coefficient is determined to be 0.7, and the associated sentence weight coefficient is 0.3. Multiply the core word matching score by the core word weight coefficient, multiply the associated sentence matching score by the associated sentence weight coefficient, and then sum the two product results to obtain the comprehensive matching degree of text one. The calculation process is . According to exactly the same steps, calculate the comprehensive matching degrees of texts two, three, four, and five, and suppose they are 0.390, 0.512, 0.488, and 0.420 respectively. Compare the comprehensive matching degrees of all five candidate knowledge texts, and it is found that the comprehensive matching degree of text three, 0.512, is the highest. Therefore, select text three with the highest comprehensive matching degree as the final nursing knowledge content to be pushed, and obtain the knowledge push result.

[0039] A traditional Chinese medicine nursing knowledge push system, which is used to execute the above-mentioned traditional Chinese medicine nursing knowledge push method. The system includes: The health status feature extraction module collects and normalizes the vital signs data of the nursing subjects, maps the normalized vital signs data to a preset three-dimensional feature space, and divides the health status feature set according to the position of each vital sign data in the three-dimensional feature space. The pharmacology keyword screening module obtains similar physical sign data samples from a preset set of traditional Chinese medicine nursing samples, compares them with the physical sign data of nursing subjects in a set of health status characteristics, calculates the health status similarity index, and screens a candidate set of pharmacology keywords based on the index. The pharmacological keyword weight calculation module obtains the attribute data of the Chinese medicine used by the nursing subject at each medication time and calculates the efficacy decay coefficient, and obtains the time factor of each keyword in the pharmacological keyword candidate set. It matches the efficacy decay coefficient with the corresponding time factor and outputs the pharmacological keyword weight set. The keyword weight optimization module continuously monitors the attribute data of newly introduced traditional Chinese medicines in the nursing plan, updates the weights in the pharmacological keyword weight set with reference to the pharmacological keyword set, and generates a keyword weight optimization set. The nursing knowledge content push module optimizes the weight ranking of keywords in the set according to keyword weight, extracts the text vector data of the keywords with the highest weight from the traditional Chinese medicine nursing sample set as nursing knowledge content to be pushed, and obtains the knowledge push results.

[0040] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for disseminating traditional Chinese medicine nursing knowledge, characterized in that, Includes the following steps: S1: Collect and normalize the vital signs data of the nursing subjects, map the normalized vital signs data to a preset three-dimensional feature space, and divide the health status feature set according to the position of each vital signs data in the three-dimensional feature space. S2: Obtain similar vital sign data samples from the preset Chinese medicine nursing sample set, compare them with the vital sign data of nursing objects in the health status feature set, calculate the health status similarity index, and screen the candidate set of pharmacological keywords based on the index; S3: Obtain the attribute data of the Chinese medicine used by the nursing object at each medication time and calculate the efficacy decay coefficient, and obtain the time factor of each keyword in the pharmacological keyword candidate set, match the efficacy decay coefficient with the corresponding time factor, and output the pharmacological keyword weight set; S4: Continuously monitor the attribute data of newly introduced traditional Chinese medicines in the nursing plan, update the weights in the pharmacological keyword weight set with reference to the pharmacological keyword set, and generate a keyword weight optimization set; S5: Optimize the weight ranking of keywords in the set according to the keyword weight, extract the text vector data of the keyword with the highest weight from the set of traditional Chinese medicine nursing samples as nursing knowledge content to be pushed, and obtain the knowledge push result.

2. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 1, characterized in that, The health status feature set includes the distribution range of vital signs, health status classification labels, and cluster center parameters; the pharmacological keyword candidate set includes efficacy category labels, keyword relevance parameters, and sample citation indexes; the pharmacological keyword weight set includes keyword matching weight values, efficacy decay coefficients, and timeliness factor mapping relationships; the keyword weight optimization set includes updated keyword weights, drug attribute association parameters, and sensitivity ratios; and the knowledge push results include nursing knowledge text content, keyword text vector data, and push priority labels.

3. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 1, characterized in that, The specific steps for obtaining the health status feature set are as follows: S111: Collect the body temperature, heart rate and blood pressure data of the nursing subjects to form a raw vital signs data set. Calculate the mean and standard deviation of each data item in the raw vital signs data set, and perform standard deviation normalization to obtain the normalized vital signs values. S112: Call the normalized values ​​of vital signs, map the normalized body temperature data, heart rate data and blood pressure data together to form vital sign coordinate points in the three-dimensional feature space, calculate the Euclidean distance between the vital sign coordinate points, and establish a set of Euclidean distance values. S113: Perform clustering based on the density of the Euclidean distance value set in the three-dimensional feature space, identify the densely distributed areas of coordinate points as independent clusters, calculate the geometric center of all coordinate points in each cluster as the cluster center, use the cluster center as a reference, and define the range boundary of the health status according to the preset distance threshold to generate a health status feature set.

4. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 3, characterized in that, The specific steps for obtaining the candidate set of pharmacological keywords are as follows: S211: Call the vital sign data of the nursing objects in the health status feature set, and obtain similar vital sign data samples from the preset Chinese medicine nursing sample set. Pair and compare the vital sign data of the nursing objects with the vital sign data of each sample, and calculate the health status similarity index of each sample. S212: Set a similarity screening threshold, compare the health status similarity index of each sample with the similarity screening threshold one by one, screen out samples with index values ​​higher than the similarity screening threshold, and establish a high similarity sample set. S213: Extract the pharmacological keywords corresponding to each sample in the high similarity sample set to form a candidate set of pharmacological keywords.

5. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 4, characterized in that, The specific steps for obtaining the pharmacological keyword weight set are as follows: S311: Obtain the attribute data of the Chinese medicine used for each medication time of the nursing subject. The attribute data of the Chinese medicine includes the drug's duration of action parameter, metabolic rate parameter, and duration of action parameter. Calculate the time difference based on the medication time node data and the drug's duration of action parameter. Combine the time difference, metabolic rate parameter, and duration of action parameter to calculate the drug efficacy decay coefficient. S312: Obtain the time-effect factor of each keyword in the candidate set of pharmacological keywords, and match the time-effect factor of each keyword with the corresponding drug efficacy decay coefficient according to the correspondence between keywords and drug administration time, and establish the decay information corresponding to the keyword. S313: For each matching item in the decay information corresponding to the keyword, based on the time factor, apply the matching drug efficacy decay coefficient to perform an exponential decay operation, calculate a weight value for each keyword, and integrate the weight values ​​of all keywords to output a set of pharmacological keyword weights.

6. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 5, characterized in that, The specific steps for obtaining the keyword weight optimization set are as follows: S411: Continuously monitor the use of new traditional Chinese medicines in the nursing plan, collect attribute data of the new traditional Chinese medicines, and establish a dataset of new pharmacological attributes; S412: Call the new pharmacological attribute dataset and refer to the corresponding parameters of the associated drugs in the candidate set of pharmacological keywords to calculate the correlation between the pharmacological attribute data of the newly introduced drug and the parameters of the associated drugs, and generate the cross-sensitivity ratio. S413: Based on the cross-sensitivity ratio and by calling the current weight values ​​in the pharmacological keyword weight set, reassign the weights of the relevant keywords, and recombine the updated weight values ​​to generate a keyword weight optimization set.

7. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 6, characterized in that, To calculate the correlation between the pharmacological property data of newly introduced drugs and the parameters of already associated drugs, the following formula is used: ； in, The degree of association between the newly introduced drug C and the already associated drug B. Drug C represents the newly introduced drug. This represents existing drug B. The number of pharmacological properties included in the comparison. It is an index of pharmacological properties. It is the first The weight of each attribute, Drugs C and B are in the first Quantitative similarity scores on each attribute and These are the first steps of the newly introduced drug C and the already associated drug B. Each attribute value; It is the first The contribution factor of each attribute.

8. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 6, characterized in that, The specific steps for obtaining the knowledge push results are as follows: S511: Sort the keywords in the keyword weight optimization set in descending order according to the keyword weight values, lock the keyword with the highest weight, and extract the corresponding knowledge text data from the traditional Chinese medicine nursing sample set based on the keyword with the highest weight to form the knowledge text to be processed. S512: Calculate the frequency of each word in the knowledge text to be processed and the inverse document frequency of the word in the entire Chinese medicine nursing sample set text. Based on the word frequency and inverse document frequency, obtain the keyword text vector data. S513: Calculate the comprehensive matching degree between each knowledge text and the keyword with the highest initial weight based on the keyword text vector data, compare the comprehensive matching degree of all candidate knowledge texts, select the nursing knowledge content with the highest comprehensive matching degree as the nursing knowledge content for push, and obtain the knowledge push result.

9. The method for disseminating traditional Chinese medicine nursing knowledge according to claim 8, characterized in that, The specific process for calculating the overall matching degree between each knowledge text and the keyword with the highest initial weight is as follows: Extract the term weights directly corresponding to the keyword with the highest initial weight from the keyword text vector data, and use them as the core word matching scores; Identify all statements with the highest initial keyword weight from the keyword text vector data, and perform a weighted average calculation on the statement weights of these statements to obtain the related statement matching score; Set the weight coefficients for core keywords and related sentences; Multiply the core word matching score by the core word weight coefficient, and multiply the related sentence matching score by the related sentence weight coefficient. Sum the two products to obtain the overall matching degree.

10. A traditional Chinese medicine nursing knowledge delivery system, characterized in that, The method for pushing traditional Chinese medicine nursing knowledge according to any one of claims 1-9, wherein the system comprises: The health status feature extraction module collects and normalizes the vital signs data of the nursing subjects, maps the normalized vital signs data to a preset three-dimensional feature space, and divides the health status feature set according to the position of each vital sign data in the three-dimensional feature space. The pharmacological keyword screening module obtains similar vital sign data samples from a preset set of traditional Chinese medicine nursing samples, compares them with the vital sign data of nursing subjects in the set of health status characteristics, calculates a health status similarity index, and screens a candidate set of pharmacological keywords based on the index. The pharmacological keyword weight calculation module obtains the attribute data of the Chinese medicine used by the nursing subject at each medication time and calculates the efficacy decay coefficient, and obtains the time factor of each keyword in the pharmacological keyword candidate set, matches the efficacy decay coefficient with the corresponding time factor, and outputs the pharmacological keyword weight set. The keyword weight optimization module continuously monitors the attribute data of newly introduced traditional Chinese medicines in the nursing plan, updates the weights in the pharmacological keyword weight set with reference to the pharmacological keyword set, and generates a keyword weight optimization set. The nursing knowledge content push module optimizes the weight ranking of keywords in the set according to the keyword weight optimization, extracts the text vector data of the keyword with the highest weight from the traditional Chinese medicine nursing sample set as nursing knowledge content to be pushed, and obtains the knowledge push results.

Images