AI-based drug multi-dimensional value evaluation method and system

Abstract

The invention relates to the technical field of multi-dimensional data analysis, in particular to an AI-based drug multi-dimensional value evaluation method and system, and the method comprises the following steps: sorting and normalizing the data based on the multi-dimensional numerical data of the market performance, patient acceptability and social benefit of drugs, recognizing the weight value of a dimension parameter through numerical cross, and determining the value of the drug. And obtaining a drug value correlation parameter matrix. According to the method, accurate description of multi-dimensional data relevance is achieved by extracting parameter weights among dimensions, redundant information interference is eliminated, analysis reliability is improved, time sequence data are combined, parameter changes are dynamically tracked, distortion is reduced by means of deviation correction, trend analysis precision is improved, and external market and policy data are introduced. Interaction analysis of multi-dimensional data and external factors is enhanced, adaptability to external interference is enhanced, prediction credibility is improved, weight distribution is optimized, the parameter change trend is dynamically evaluated, data interpretation efficiency is improved, and better decision and resource configuration are supported.

Description

Technical Field

[0001] The present invention relates to the technical field of multidimensional data analysis, and in particular to an AI-based multidimensional drug value assessment method and system. Background Art

[0002] The field of multidimensional data analysis technology involves the integration, processing and analysis of data from multiple data sources or multiple dimensions to extract useful information and insights. It uses advanced algorithms and statistical methods, such as machine learning and artificial intelligence technologies, to perform exploratory data analysis, pattern recognition and predictive modeling on large-scale and complex data sets. The core of this field is to combine different aspects of data, including time series, categorical data, frequency data, etc., to form a comprehensive data view for more effective decision support. Multidimensional data analysis plays a key role in industries such as finance, retail, supply chain management, and biomedicine, providing users with the ability to gain insights from complex data.

[0003] Among them, the multidimensional drug value assessment method involves a comprehensive analysis of the overall value of drugs, aiming to support the decision-making process by evaluating different value dimensions of drugs, including the economic impact of drugs, market potential, patient satisfaction, and socioeconomic benefits. This method uses multidimensional data analysis technology to integrate and analyze complex data to help pharmaceutical companies, health policy makers, and healthcare providers understand the potential value of drugs more comprehensively. Stakeholders can make more informed choices, optimize resource allocation, and develop effective market entry strategies while ensuring the commercial success of drugs and maximizing social benefits. The application of this topic helps to improve the efficiency of drug research and development and marketing.

[0004] The existing technology has the problem of insufficient mining of the correlation between dimensions when processing multidimensional data, and it is easy to ignore the potential connection between data of different dimensions, resulting in that the results of data integration are difficult to fully reflect the overall information value. The analysis accuracy of time series data is limited, and it is impossible to dynamically track the changing trends of multidimensional parameters at different time nodes, and it is difficult to find important time differences that affect the results. The ability to integrate non-internal data such as external markets and policies is weak, resulting in the data model lacking sensitivity to external environmental interference factors, affecting the actual applicability of the analysis. For the dynamic interactive analysis of multidimensional data, the existing technology is relatively weak in the screening of key parameters and the dynamic association identification, which easily causes key value points to be omitted, affecting the comprehensiveness of the results. The lack of dynamic adjustment and hierarchical optimization of parameter weights makes the existing evaluation results lack hierarchy and adaptability, and it is difficult to meet the needs of complex decision-making scenarios, resulting in the results of multidimensional value analysis of drugs being not accurate enough, unable to fully reflect the overall value of drugs, affecting the rational allocation of corporate resources and the effectiveness of market strategies. Summary of the invention

[0005] In order to solve the problem that the existing technology is insufficient in mining the correlation between dimensions when processing multidimensional data, it is easy to ignore the potential connection between data of different dimensions, which makes it difficult to fully reflect the overall information value of the data integration results. The analysis accuracy of time series data is limited, and it is impossible to dynamically track the changing trend of multidimensional parameters at different time nodes, and it is difficult to find important time differences that affect the results. The integration ability of non-internal data such as external markets and policies is weak, resulting in the data model lacking sensitivity to external environmental interference factors, affecting the actual applicability of the analysis. For the dynamic interactive analysis of multidimensional data, the existing technology is relatively weak in the screening and dynamic association identification of key parameters, which easily causes the omission of key value points and affects the comprehensiveness of the results. The lack of dynamic adjustment and hierarchical optimization of parameter weights makes the existing evaluation results lack of hierarchy and adaptability, and it is difficult to meet the needs of complex decision-making scenarios, resulting in the results of drug multidimensional value analysis being not accurate enough, unable to fully reflect the overall value of the drug, and affecting the reasonable allocation of enterprise resources and the effectiveness of market strategies. The technical problem, the embodiment of the present invention provides a drug multidimensional value evaluation method and system based on AI. The technical solution is as follows:

[0006] On the one hand, a method for multidimensional drug value assessment based on AI is provided, the method comprising:

[0007] S1: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, the data is sorted and normalized, and the weight values ​​of dimension parameters are identified through numerical cross-identification to obtain the drug value correlation parameter matrix;

[0008] S2: Based on the drug value-related parameter matrix, extract the multidimensional parameter time series data, call the parameter change rate and the cumulative value to iteratively accumulate and calculate the offset, analyze the trend and make corrections based on the offset, and generate a time dimension parameter change trend table;

[0009] S3: Based on the time dimension parameter change trend table, market competition and policy adjustment data are introduced, external data and parameter value weights are matched, and the cumulative interference degree is analyzed by superposition. According to the interference adjustment trend, a multi-dimensional parameter dynamic superposition result is generated;

[0010] S4: Based on the dynamic superposition results of the multi-dimensional parameters, extract the differentiated dimensional interaction data, map the dimensional data through nonlinear operations and classify and extract them, screen the numerical changes and dynamic correlation values ​​of key parameters, and generate a multi-dimensional value correlation map;

[0011] S5: Based on the multi-dimensional value association map, the core parameters are normalized, the weight distribution is optimized through aggregation analysis, the hierarchical relationship and dynamic change rate of the core parameters are evaluated, the distribution relationship of differentiated time nodes is analyzed, and the dynamic evaluation results of the multi-dimensional value of the drug are obtained.

[0012] As a further scheme of the present invention, the drug value association parameter matrix includes parameter weight values, inter-dimensional correlations, and drug value parameters; the time dimension parameter change trend table includes parameter change trends, trend offsets, and time series cumulative values; the multidimensional parameter dynamic superposition results include external interference weights, trend interference intensity, and dynamically superimposed parameter values; the multidimensional value association map includes key parameter dynamic values, dimensional interaction data, and value association paths; the drug multidimensional value dynamic evaluation results include core parameter distribution, key weight dynamic changes, and hierarchical weight distribution.

[0013] As a further solution of the present invention, based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, the steps of collating and normalizing the data, and obtaining the drug value correlation parameter matrix by numerical cross-identification of dimensional parameter weight values ​​are specifically as follows:

[0014] S101: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, extract each numerical parameter and unify the format, adjust the proportion through extreme value range analysis, map the data to the corresponding interval, and generate parameter normalization results;

[0015] S102: Based on the parameter normalization result, extract the numerical data of the two dimensions, analyze the numerical ratio and the degree of cross-influence, integrate the ratio results to identify the dimension correlation, and generate a parameter weight correlation matrix according to the numerical dependency relationship;

[0016] S103: Based on the parameter weight association matrix, extract and optimize the dimension weight values ​​within the matrix, rearrange the matrix items to reduce data redundancy, perform matrix standardization processing based on the value of each item, and generate a drug value association parameter matrix.

[0017] As a further solution of the present invention, based on the drug value associated parameter matrix, extracting multidimensional parameter time series data, calling parameter change rate and cumulative value to iteratively accumulate and calculate offset, analyzing the trend and correcting it according to the offset, and generating the time dimension parameter change trend table are specifically as follows:

[0018] S201: Based on the drug value associated parameter matrix, extract the time series data of multidimensional parameters, summarize the parameter time point values ​​and change records, analyze the value change amplitude and cumulative value within the time interval, and obtain the time series cumulative change data;

[0019] S202: Based on the cumulative change data of the time series, extract the change rate of the parameter within the differentiated time interval, analyze the change offset on the time axis, compare the change amplitude to summarize the regularity of the numerical offset, and obtain parameter trend offset data;

[0020] S203: Based on the parameter trend offset data, extract the offset value to adjust the numerical trend on the time axis, integrate the parameter time series change deviation, calculate the adjusted parameter value, and generate a time dimension parameter change trend table.

[0021] As a further solution of the present invention, the adjusted parameter value is calculated according to the formula:

[0022]

[0023] Among them, Q represents the adjusted parameter value, P t represents the original parameter value at time point t, T std Represents the global trend standard value of the parameter time series, ΔT t represents the time interval difference between time point t and its adjacent time points, B t Represents the basic offset at time point t, E1 is the trend correction weight coefficient, E2 is the time interval correction weight coefficient, and E3 is the basic offset correction weight coefficient.

[0024] As a further solution of the present invention, based on the time dimension parameter change trend table, market competition and policy adjustment data are introduced, external data and parameter value weights are matched, and the cumulative interference degree is analyzed by superposition. According to the interference adjustment trend, the steps of generating a multi-dimensional parameter dynamic superposition result are specifically as follows:

[0025] S301: Based on the time dimension parameter change trend table, read the parameter value at the time point, analyze the corresponding relationship between the external data and the trend data, superimpose the external interference value to the parameter trend value, and generate an external data matching weight table;

[0026] S302: Based on the external data matching weight table, extract the matched weight data, analyze the cumulative impact of the external interference value on the parameter trend, adjust the parameter change value affected by the external influence, sort out the parameter change amplitude and cumulative trend, and generate an interference intensity adjustment trend table;

[0027] S303: Based on the interference intensity adjustment trend table, extract the dynamic change relationship of parameters in the trend data, group the time series data by weight, calculate the dynamic superposition value of the parameters, summarize the grouping numerical relationship for dynamic iterative processing, and generate a multi-dimensional parameter dynamic superposition result.

[0028] As a further solution of the present invention, the dynamic superposition value of the parameter is calculated according to the formula

[0029]

[0030] Among them, w i represents the weight factor of the i-th parameter, p i represents the time series data parameter value of the i-th parameter, Represents the weighted average of the parameters in the current group, α i represents the nonlinear adjustment index of the i-th parameter, k j represents the time weight factor in the jth time period, t j represents the value at the jth time point, t0 represents the reference time point of the time series, n represents the total number of parameters, and m represents the total number of time points.

[0031] As a further solution of the present invention, based on the dynamic superposition result of the multi-dimensional parameters, the steps of extracting differentiated dimensional interaction data, mapping the dimensional data through nonlinear operations and classifying and extracting, screening the numerical changes and dynamic correlation values ​​of key parameters, and generating a multi-dimensional value association map are specifically as follows:

[0032] S401: Based on the dynamic superposition result of the multi-dimensional parameters, the parameter values ​​are read item by item at a time point, the corresponding relationship between the external data and the trend data is analyzed, the external interference value is superimposed on the original trend of the parameter, and an interactive data arrangement table is generated;

[0033] S402: Based on the interactive data sorting table, analyzing the cumulative impact of external interference values ​​on parameter trends, adjusting parameter change values ​​affected by external factors in trend data, sorting the adjusted change amplitudes and cumulative trends, and generating a key parameter mapping set;

[0034] S403: Based on the key parameter mapping set, extract the dynamic change relationship of parameters in the adjustment trend data, group the time series data by weight, summarize the grouping numerical relationship for dynamic iterative processing, and generate a multi-dimensional value association map.

[0035] As a further solution of the present invention, the steps of normalizing the core parameters based on the multidimensional value association map, optimizing the weight distribution through aggregation analysis, evaluating the hierarchical relationship and dynamic change rate of the core parameters, and analyzing the distribution relationship of the differentiated time nodes to obtain the dynamic evaluation results of the multidimensional value of the drug are specifically as follows:

[0036] S501: Based on the multi-dimensional value association map, extract the core parameter numerical data in the map, read the core parameter time series range, map the values ​​to a unified interval in proportion, integrate the processed core parameter values, and generate a core parameter normalized numerical table;

[0037] S502: Based on the core parameter normalized value table, extract the weight values ​​corresponding to the normalized parameters, summarize the parameter change trends in the time series and reallocate the weights, sort out the weight differences, and generate a core parameter weight distribution table;

[0038] S503: Based on the core parameter weight distribution table, extract the core parameter weight value at each time node, evaluate the time series change distribution of the weight, analyze the dynamic distribution law of the core parameters, integrate the evaluated distribution law and archive it by time dimension, and generate a dynamic evaluation result of the multi-dimensional value of the drug.

[0039] On the other hand, an electric vehicle state monitoring system is provided, the electric vehicle state monitoring system is used to execute the above electric vehicle state monitoring method, the system comprises:

[0040] The parameter distribution calculation module extracts the ratio of sales volume to usage frequency, analyzes the ratio of sales volume to adjustment intensity, performs cross calculations, and obtains a weight distribution table of multidimensional parameters of drugs based on the multidimensional numerical data of drug market performance, patient acceptance, and social benefits.

[0041] The trend offset correction module extracts the sales volume, usage frequency and adjustment intensity cumulative values ​​in the time series based on the drug multi-dimensional parameter weight distribution table, and obtains the parameter dynamic trend correction result by accumulating the change rate within the differentiated time points;

[0042] The external data fusion module extracts market competition pricing fluctuation data and policy incentive data based on the parameter dynamic trend correction results, compares pricing fluctuations with sales volume change rates, analyzes the interference intensity of policy incentives on adjustment efforts, and obtains external factor fusion impact data;

[0043] The interactive feature extraction module extracts the interactive values ​​of sales volume changes, policy adjustments and competition impact based on the external factor fusion impact data, performs dynamic correlation feature extraction and feature grouping comparison, and obtains multi-dimensional dynamic interactive feature results;

[0044] The dynamic value assessment module extracts the dynamic change rate and distribution data of the core parameters based on the multi-dimensional dynamic interaction characteristic results, performs core dynamic change value analysis and time node parameter hierarchical assessment, and obtains the dynamic assessment results of the multi-dimensional value of the drug.

[0045] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0046] By collating and normalizing the multi-dimensional numerical data of drugs and extracting the parameter weight values ​​between different dimensions, the inherent correlation between multi-dimensional data can be accurately described. This cross-analysis between values ​​ensures that the integration of data in different dimensions has an objective basis, which helps to eliminate the interference of redundant information and improve the reliability of the analysis results. Combined with time series data, the trend of parameter changes can be dynamically tracked, and the distortion caused by single time node analysis can be reduced by trend deviation correction, thereby improving the accuracy of trend analysis. By introducing external market and policy data, the interactive analysis ability of multi-dimensional data and external factors is strengthened. The quantitative analysis of external interference factors further enhances the realistic adaptability of data dynamic modeling and improves the credibility of prediction. Through nonlinear mapping and classification of interactive data of differentiated dimensions, key parameters and their dynamic associations are screened to provide higher-level visualization support for multi-dimensional value analysis. Through hierarchical analysis and weight optimization of key parameters, their dynamic change trends and distribution relationships are accurately evaluated to form scientific and rigorous drug value evaluation results, which not only improves the precision of data processing, but also makes the interpretation of complex multi-dimensional data more efficient and intuitive, providing stakeholders with stronger data support capabilities and helping to optimize decision-making efficiency and resource allocation effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the workflow of the present invention;

[0048] Figure 2 This is a detailed flow chart of S1 of the present invention;

[0049] Figure 3 This is a detailed flow chart of S2 of the present invention;

[0050] Figure 4 This is a detailed flow chart of S3 of the present invention;

[0051] Figure 5 This is a detailed flow chart of S4 of the present invention;

[0052] Figure 6 This is a detailed flow chart of S5 of the present invention;

[0053] Figure 7 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0054] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0055] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0056] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0057] In the embodiments of the present invention, sometimes a subscript such as W1 may be written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0058] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0059] See also Figure 1 The embodiment of the present invention provides a method for evaluating the multidimensional value of drugs based on AI. The processing flow of the method may include the following steps:

[0060] S1: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, the data is sorted and normalized. By crossing the numerical values ​​in the original data, the parameter weight values ​​between the dimensions are identified to obtain the drug value association parameter matrix;

[0061] S2: Based on the drug value-related parameter matrix, extract the time series data of multidimensional related parameters, call the parameter change rate and cumulative value within the time interval, obtain the parameter trend offset through numerical iteration accumulation, analyze the parameter change trend at different time points according to the offset, perform trend deviation correction, and generate a time dimension parameter change trend table;

[0062] S3: Based on the time dimension parameter change trend table, external data of market competition and policy adjustment are introduced, the matching weights of external data and parameter values ​​in the trend table are extracted, superposition analysis between parameters is performed, the interference degree of external factors on parameter values ​​in the trend table is accumulated, and the parameter change trend is adjusted according to the interference intensity to generate multi-dimensional parameter dynamic superposition results;

[0063] S4: Based on the dynamic superposition results of multi-dimensional parameters, the interactive data of differentiated dimensions in the dynamic superposition table are extracted, the dimensional data are mapped through non-linear operations, and classified and extracted, the numerical changes and dynamic correlation values ​​of key parameters are screened, and a multi-dimensional value correlation map is generated;

[0064] S5: Based on the multidimensional value association map, the core parameters in the association map are normalized, the weight distribution of key parameters is optimized through aggregation analysis, the dynamic change rate of weights is analyzed and the core values ​​are layered, and the distribution relationship of the core parameters at differentiated time nodes is evaluated in turn to obtain the dynamic evaluation results of the multidimensional value of drugs.

[0065] The drug value-related parameter matrix includes parameter weight values, correlations between dimensions, and drug value parameters. The time dimension parameter change trend table includes parameter change trends, trend offsets, and time series cumulative values. The multidimensional parameter dynamic superposition results include external interference weights, trend interference intensity, and dynamically superimposed parameter values. The multidimensional value association map includes key parameter dynamic values, dimension interaction data, and value association paths. The dynamic evaluation results of drug multidimensional value include core parameter distribution, dynamic changes in key weights, and hierarchical weight distribution.

[0066] Specifically, if Figure 2 As shown in the figure, based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, the steps of sorting and normalizing the data, and obtaining the drug value correlation parameter matrix through numerical cross-identification of dimensional parameter weight values ​​are as follows:

[0067] S101: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, extract each numerical parameter and unify the format, adjust the proportion through extreme value range analysis, map the data to the corresponding interval, and generate parameter normalization results;

[0068] Analyze the actual value range of the data in different dimensions item by item, extract each numerical parameter in turn, convert the extracted parameters into a standard format according to the actual distribution characteristics, use the extreme range analysis method, define the minimum and maximum range of each parameter, and set the linear mapping factor according to the proportional adjustment rule to map the original data to a unified numerical range, such as standardizing the data to the [0, 1] interval to ensure that all parameters are compared and analyzed on the same scale, and generate parameter normalization results through the mapped results for further processing. Each step in the normalization process maintains data accuracy and proportional consistency to avoid weight distortion due to range adjustment, ensure that the three dimensions of market performance, patient acceptance, and social benefits can be compared under the same reference framework, and lay the foundation for subsequent matrix generation.

[0069] S102: based on the parameter normalization results, extract the numerical data of the two dimensions, analyze the numerical ratio and the degree of cross-influence, integrate the ratio results to identify the dimension correlation, and generate a parameter weight correlation matrix based on the numerical dependency relationship;

[0070] Through ratio calculation and cross-analysis, the degree of mutual influence between data of each dimension is evaluated, the numerical ratio between two dimensions is calculated by normalizing data, the key ratio interval is screened, and the strength of the correlation between dimensions is determined based on the distribution trend of the ratio, and a cross-impact assessment is performed. By statistically analyzing the numerical fluctuation interval corresponding to each ratio and its collateral impact on the parameters, the ratio results are summarized to identify the degree of correlation between the two dimensions. Combined with the correlation results, a parameter weight correlation matrix is ​​generated, in which the matrix entries identify the degree of dependence of the two dimensions according to the results of the ratio calculation, and an identification value is assigned to each matrix entry. The establishment of the dependency relationship is based on the clear relationship between the numerical values, ensuring the accuracy and interpretability of the correlation matrix, and providing a basis for the subsequent generation of parameter weights.

[0071] S103: Based on the parameter weight association matrix, extract and optimize the dimension weight values ​​in the matrix, rearrange the matrix items to reduce data redundancy, perform matrix standardization processing based on the values ​​of each item, and generate a drug value association parameter matrix;

[0072] Select the dimension weight values ​​in the matrix items, improve the calculation efficiency of the matrix through screening and optimization steps, use the arrangement rules of weight values ​​in the matrix to identify repeated or redundant items, merge similar or highly correlated items to reduce data redundancy, redistribute the weights and rearrange the order of matrix items according to the optimized items to ensure the balanced distribution of weight values ​​in the matrix, combine with the normalization processing rules to further standardize the values ​​in each item, and use the scaling rules to enable all parameter items in the matrix to reflect relative value under the same reference standard to generate a drug value-related parameter matrix. During the standardization process of the matrix, ensure that the calculation results of each item are balanced between operability and usability, and provide a complete and efficient calculation framework for subsequent multidimensional value evaluation.

[0073] Specifically, if Figure 3 As shown in the figure, based on the drug value-related parameter matrix, the multi-dimensional parameter time series data is extracted, the parameter change rate and the cumulative value are called to iteratively accumulate and calculate the offset, the trend is analyzed and corrected according to the offset, and the steps of generating the time dimension parameter change trend table are as follows:

[0074] S201: based on the drug value associated parameter matrix, extract the time series data of multi-dimensional parameters, summarize the parameter time point values ​​and change records, analyze the value change amplitude and cumulative value within the time interval, and obtain the time series cumulative change data;

[0075] For each dimensional data in the parameter matrix, its original value is recorded at the specified time point, and its change value and change direction between adjacent time points are gradually calculated. By accumulating the change data at consecutive time points, the cumulative change value of the time series is formed. The drug parameters are set to include stability, solubility and absorption rate. The numerical changes at time points T1, T2 and T3 are recorded respectively. The change values ​​of adjacent time points of each item are accumulated in sequence, and the degree of influence of the time interval is marked in the change record. At the same time, the change amplitude is classified and identified. By statistically analyzing the change data within the time interval, a result table including the cumulative change value, the proportion of the change amplitude and the trend of the total change amount is generated, so as to further evaluate the correlation and dynamic change characteristics of the multidimensional parameters and obtain the cumulative change data of the time series.

[0076] S202: based on the cumulative change data of the time series, extract the change rate of the parameter within the differentiated time interval, analyze the change offset on the time axis, compare the change amplitude to summarize the regularity of the numerical offset, and obtain the parameter trend offset data;

[0077] For each parameter dimension, the time axis is divided into three intervals: short-term, medium-term and long-term. The average offset value and range value of the changes in each interval are calculated. At the same time, combined with the cumulative change data in the previous step, the time distribution characteristics of the change rate are analyzed. Taking the solubility data from time points T1 to T4 as an example, the change rate and fluctuation range in the short term from T1 to T2 are calculated, and compared with the change rate in the medium term (T2 to T3) and long term (T3 to T4) to generate a set of change offsets. By comparing the change ranges at different time intervals, the change trends and offset regularities of multidimensional parameters are summarized, and the focus is on analyzing whether there are periodic offsets or cumulative change imbalances on the time axis to obtain parameter trend offset data.

[0078] S203: Based on the parameter trend offset data, extract the offset value to adjust the numerical trend on the time axis, integrate the parameter time series change deviation, calculate the adjusted parameter value, and generate a time dimension parameter change trend table;

[0079] Calculate the adjusted parameter value according to the formula:

[0080]

[0081] Among them, Q represents the adjusted parameter value, P t represents the original parameter value at time point t, T std Represents the global trend standard value of the parameter time series, ΔT t represents the time interval difference between time point t and its adjacent time points, B tRepresents the basic offset at time point t, E1 is the trend correction weight coefficient, E2 is the time interval correction weight coefficient, and E3 is the basic offset correction weight coefficient;

[0082] Detailed explanation of the formula and calculation process:

[0083] This formula is used to calculate the adjusted parameter value Q at time point t, and the result is used to correct the parameter deviation in the time series and generate a normalized parameter time series change trend table;

[0084] P t : The original parameter value at time point t, obtained by monitoring and collecting time series data. For example, when the time point t=5 is set, the monitored original parameter value is 120;

[0085] T std : The global trend standard value of the parameter time series is calculated by trend analysis of the entire time series. For example, after analysis, the global trend standard value is determined to be 100;

[0086] ΔT t : The time interval difference between time point t and its adjacent time points, obtained by calculating the time difference between adjacent time points. For example, if the interval between time points is 1 day, then ΔT t =1;

[0087] B t : The basic offset at time point t, obtained by calculating the difference between the time point and the global average trend value. For example, if the basic offset at time point t=5 is 15, then B t =15;

[0088] E1: Trend correction weight coefficient, used to adjust the influence of the global trend standard value on the offset correction. The setting basis is the fitting effect of historical data. The value is between 0 and 1. The setting is 0.8, which means that the global trend standard value has a greater influence on the correction;

[0089] E2: Time interval correction weight coefficient, used to adjust the influence of time span on parameter adjustment value. The setting basis is the influence of time interval on data fluctuation. The value is between 0 and 1. The setting is 0.5, which means that the time interval has a medium influence on the correction.

[0090] E3: Basic offset correction weight coefficient, used to adjust the contribution of the original offset value to the final correction result. The setting basis is the influence of the basic offset on the data accuracy. The value is between 0 and 1. The setting is 0.3, which means that the basic offset has a small influence on the correction.

[0091] Substitute the parameters into the formula for calculation:

[0092] Calculate P t -E1·T std :P t -E1·T std =120-0.8×100=120-80=40;

[0093] Calculate the absolute value |P t -E1·T std |:|40|=40;

[0094] calculate

[0095] calculate

[0096] Calculate the square root of the denominator

[0097] Calculate the adjusted parameter value Q:

[0098] The result Q≈17.88 shows that after the original parameter value at time point t=5 is corrected, the adjusted parameter value is about 17.88. This result is used to generate a normalized parameter time series change trend table to reflect more accurate trend information.

[0099] Specifically, if Figure 4 As shown in the figure, based on the time dimension parameter change trend table, market competition and policy adjustment data are introduced, external data and parameter value weights are matched, and the cumulative interference degree is analyzed by superposition. According to the interference adjustment trend, the steps to generate the dynamic superposition results of multi-dimensional parameters are as follows:

[0100] S301: Based on the time dimension parameter change trend table, read the parameter value at the time point, analyze the corresponding relationship between the external data and the trend data, superimpose the external interference value to the parameter trend value, and generate an external data matching weight table;

[0101] A comprehensive analysis is conducted in combination with external interference factors at different time points. The parameter change values ​​at the time points recorded in the trend table are read item by item and a time series change model is established. The trend change value at each time point is compared with the actual value. External interference data, such as market policy changes, raw material price fluctuations and other external variables, are introduced to quantify the degree of influence of interference data on trend values, and they are mapped to parameter trend values ​​in an overlay manner. During the overlay process, interference values ​​are extracted at each time point, and the interference amplitude is adjusted through linear mapping rules to achieve a unified dimension with the trend data. In order to further clarify the impact of external factors, the overlaid values ​​are classified and processed. After the external interference data is classified according to the source of interference, they are cross-matched with the values ​​in the time trend table one by one. Finally, an external data matching weight table is generated to identify the specific weights of each external interference factor on the parameter trend changes at different time points, laying the foundation for subsequent interference adjustments.

[0102] S302: Based on the external data matching weight table, extract the matched weight data, analyze the cumulative impact of the external interference value on the parameter trend, adjust the parameter change value affected by the external influence, sort out the parameter change amplitude and cumulative trend, and generate an interference intensity adjustment trend table;

[0103] The matching weight data of each time point is extracted, and the cumulative impact of external interference values ​​on time trend parameters is analyzed by calculation. The matching weight data is calculated one by one with the parameter change values ​​of the corresponding time points in the trend table, and the trend part affected by external interference and the original parameter change part are separated. The total impact of external factors on parameter changes in the entire time series is calculated by accumulating interference values, and the parameter values ​​affected by interference in the time trend are adjusted accordingly. First, the amplitude of the interference impact at a single time point is corrected to ensure that the adjusted trend change meets the smoothness requirements of the overall time series. At the same time, the cumulative effect of the interference value is normalized to avoid excessive concentration of external influences in a certain time period. The adjusted parameter change amplitudes are sorted out, combined with the cumulative trend calculation, and an interference intensity adjustment trend table is generated to clearly show the overall interference intensity and change direction of external factors on the time dimension trend, providing a basis for the next step of dynamic superposition processing.

[0104] S303: Based on the interference intensity adjustment trend table, extract the dynamic change relationship of parameters in the trend data, group the time series data by weight, calculate the dynamic superposition value of the parameters, summarize the grouping numerical relationship, perform dynamic iterative processing, and generate a multi-dimensional parameter dynamic superposition result;

[0105] Calculate the dynamic superposition value of the parameter according to the formula

[0106]

[0107] Among them, w irepresents the weight factor of the i-th parameter, p i represents the time series data parameter value of the i-th parameter, Represents the weighted average of the parameters in the current group, α i represents the nonlinear adjustment index of the i-th parameter, k j represents the time weight factor in the jth time period, t j represents the value at the jth time point, t0 represents the reference time point of the time series, n represents the total number of parameters, and m represents the total number of time points;

[0108] Detailed explanation of the formula and the process of formula calculation and derivation:

[0109] The formula is used to calculate the dynamic superposition value of multi-dimensional parameters in time series data. The obtained results are used to describe the weighted dynamic change trend of different parameters in the time series and their combined impact with time changes;

[0110] w i is the weight factor of the ith parameter, reflecting the importance of the parameter to the overall dynamic superposition calculation. The weight is set by analyzing the sensitivity of parameter changes and the correlation with the parameters. For example, through monitoring, the sensitivity scores of parameters p1, p2, and p3 are 0.8, 0.6, and 0.9, respectively. Then their normalized weights are w1 = 0.36, w2 = 0.27, and w3 = 0.37, respectively.

[0111] p i is the parameter value in the i-th time series, representing the actual collected data in the time series. For example, the measured values ​​of parameters p1, p2, and p3 at time points t1, t2, and t3 are 25, 30, and 28, respectively;

[0112] is the weighted average of the parameters in the current group, which is used to smooth the central trend of dynamic changes in the group. calculate;

[0113] α i is the nonlinear adjustment index of the ith parameter, which adjusts the contribution of the parameter in the calculation of the dynamic superposition value. The nonlinear index is determined by analyzing the fluctuation amplitude of the historical trend. For example, if the variance values ​​of the collected historical data are 1.2, 0.8, and 1.5, respectively, then the adjustment indexes are α1=1.1, α2=1.05, and α3=1.15 respectively;

[0114] k j is the time weight factor of the jth time period, which is determined by the quantification of the importance of the time period. For example, the weight factors of time periods t1, t2, and t3 are 0.5, 0.3, and 0.2, respectively, reflecting the relative contribution to the overall time period;

[0115] tj is the value at the jth time point, obtained through actual time series acquisition, for example, time point t1=1, t2=2, t3=3, and the time unit is day;

[0116] t0 is the reference time point of the time series, which is set by selecting a reference point, for example, the reference time point t0 = 0;

[0117] Calculate the weighted dynamic change term of the numerator part;

[0118] According to the formula, calculate the numerator part: First, calculate the weighted average Calculate the dynamic change of each parameter:

[0119] For p1:

[0120] For p2:

[0121] For p3:

[0122] Sum of the numerator parts: Calculate the time weighted square sum of the denominator;

[0123] According to the formula, calculate the denominator: For t1: k1·(t1-t0) 2 =0.5·(1-0) 2 =0.5·1=0.5;

[0124] For t2: k2·(t2-t0) 2 =0.3·(2-0) 2 =0.3·4=1.2;

[0125] For t3: k3·(t3-t0) 2 =0.2·(3-0) 2 =0.2·9=1.8;

[0126] The sum of the denominators: Calculate the dynamic superposition value D and substitute the results of the numerator and denominator into the formula:

[0127] The results show that the dynamic superposition value of multidimensional parameters in time series data is 1.05. This result describes the overall change trend of time series parameters under the dynamic weight grouping conditions, and comprehensively considers the nonlinear dynamic relationship between time and parameters. This value can be further used to predict the dynamic evolution trend of time series or optimize subsequent adjustment strategies.

[0128] Specifically, if Figure 5 As shown in the figure, based on the dynamic superposition results of multi-dimensional parameters, the differentiated dimensional interaction data is extracted, the dimensional data is mapped and classified through non-linear operations, the numerical changes and dynamic correlation values ​​of key parameters are screened, and the steps of generating a multi-dimensional value association map are as follows:

[0129] S401: Based on the dynamic superposition result of multi-dimensional parameters, the parameter values ​​are read item by item at each time point, the corresponding relationship between external data and trend data is analyzed, the external interference value is superimposed on the original trend of the parameter, and an interactive data arrangement table is generated;

[0130] It is also combined with external interference data at different time points for analysis. The time series parameters in the superposition results are mapped and matched with the external interference data one by one to confirm the extent of the impact of external interference factors on the change of trend parameters. The parameters read include core dimensions such as sales volume changes in market performance and long-term value growth ratio in social benefits, and external interference is introduced one by one, such as the impact of market environment changes and policy fluctuations on drug value. When superimposing external interference values, linear rules are used to decompose the interference data, and the interference factors are divided into direct interference and indirect interference according to the influence weights. The parameter trend value is adjusted according to the corresponding numerical change ratio. After the interference data is superimposed, the original trend data is compared with the adjusted trend data, and the parameter change results and interference matching data at each time point are sorted out. Finally, an interactive data sorting table is generated to clarify the correspondence between external data and dynamic parameter trends, providing basic support for subsequent analysis.

[0131] S402: Based on the interactive data sorting table, analyzing the cumulative impact of external interference values ​​on parameter trends, adjusting parameter change values ​​affected by external factors in trend data, sorting out the adjusted change ranges and cumulative trends, and generating a key parameter mapping set;

[0132] By calculating the relationship between the parameter change amplitude and the cumulative interference value at each time point, the numerical value affected by the interference in the trend data is gradually adjusted. During the adjustment process, the interference data is decomposed into two types: time point influence and interval influence, and each influence is quantitatively analyzed. For the time point influence, it is directly superimposed on the trend parameter value of the corresponding time point, and the range is restricted to ensure smooth data changes; for the interval influence, it is evenly distributed to multiple time points according to the distribution characteristics of the interference source to ensure that the cumulative interference can reflect the gradually increasing trend changes. The adjusted parameter change amplitude is reclassified to clarify which data segments are the core impact areas and which are the secondary fluctuation areas. All adjusted change amplitudes and cumulative trend data are organized into a set of key parameter mapping sets, which clarifies the mapping relationship between interference influence and parameter change, providing an intuitive basis for dynamic change analysis.

[0133] S403: Based on the key parameter mapping set, extract the dynamic change relationship of parameters in the adjustment trend data, group the time series data by weight, summarize the group value relationship for dynamic iterative processing, and generate a multi-dimensional value association map;

[0134] Time series are grouped according to the weights of different parameters to further clarify the relationship between parameters and time dimension. Time series are divided into high-weight group, medium-weight group and low-weight group based on the weight values ​​in the mapping set, and their dominant roles in dynamic changes are analyzed respectively. For each group of data, the parameter change relationship within the group is recursively calculated for multiple rounds through dynamic iteration method to extract the core trend dynamic value, and form a new dynamic superposition relationship based on the mutual influence between parameters. When summarizing the group values, the dynamic change data of different groups are merged, and the interaction analysis between high-weight parameters and low-weight parameters in the same time period is integrated into the overall dynamic trend to generate a multi-dimensional value association map. This map clearly shows the dynamic evolution process of multi-dimensional parameters under the influence of time and external interference, and provides panoramic numerical support for the comprehensive evaluation of multi-dimensional value.

[0135] Specifically, if Figure 6 As shown in the figure, the steps of normalizing the core parameters based on the multidimensional value association map, optimizing the weight distribution through aggregation analysis, evaluating the hierarchical relationship and dynamic change rate of the core parameters, and analyzing the distribution relationship of differentiated time nodes to obtain the dynamic evaluation results of the multidimensional value of drugs are as follows:

[0136] S501: Based on the multi-dimensional value association map, extract the core parameter numerical data in the map, read the core parameter time series range, map the values ​​to a unified interval in proportion, integrate the processed core parameter values, and generate a core parameter normalized numerical table;

[0137] By reading the range values ​​of the core parameters in the time series one by one, the core parameter data at each time point is accurately extracted. For the extracted core parameters, they are proportionally mapped to a unified standardized interval according to the actual change amplitude and fluctuation characteristics of their time series. The mapping process relies on the difference range between the maximum and minimum values ​​to convert the core parameter values ​​proportionally to the preset interval range, while ensuring that the dynamic changes after the numerical mapping are consistent with the original trend. After completing the numerical mapping, the core parameter values ​​at different time points are classified and integrated. During the processing, the redundant parts between the data are eliminated, and the interval values ​​with large fluctuations are smoothly adjusted to ensure that the normalized results of the core parameter changes have good stability and continuity. A core parameter normalized numerical table is generated to clarify the mapping values ​​of the core parameters at each time point in the standardized interval, providing a unified basic data support for further parameter analysis and dynamic evaluation.

[0138] S502: Based on the core parameter normalized value table, extract the weight values ​​corresponding to the normalized parameters, summarize the parameter change trends in the time series and reallocate the weights, sort out the weight differences, and generate a core parameter weight distribution table;

[0139] Combined with the changing trend of the core parameters in the time series, the weight distribution is redistributed. During the redistribution process, the parameter change amplitude of each time period in the time series is first compared with the corresponding weight value, and the weight adjustment value is calculated to ensure that the time period with a high change amplitude can obtain a higher weight share. When sorting out the weight differences, the normalized weight values ​​are divided into several intervals, and the weight differences in adjacent intervals are smoothed to avoid weight imbalance due to sudden changes. After the processing is completed, the weight adjustment results of each time period are classified and summarized to form a core parameter weight distribution table. The table clearly reflects the distribution law of the weights of each core parameter in the time series at different time points, providing basic support for further analysis of dynamic distribution and overall value assessment.

[0140] S503: Based on the core parameter weight distribution table, extract the core parameter weight value at each time node, evaluate the time series change distribution of the weight, analyze the dynamic distribution law of the core parameter, integrate the evaluated distribution law and archive it by time dimension, and generate the dynamic evaluation result of the multi-dimensional value of the drug;

[0141] By continuously reading the changing distribution of weight values ​​in the time series, the dynamic distribution law of core parameters in the time dimension is comprehensively evaluated, and the weight value of each time node is correlated and compared with the previous and next nodes to identify the trend and periodic characteristics of weight changes. According to the weight distribution in the time series, the weight differences in different time periods are quantified, the key nodes of the changing law are extracted, and the dominant role and marginal influence of the core parameters in the dynamic distribution law are analyzed. After the evaluation is completed, the analyzed weight distribution law is archived according to the time dimension, and the core parameter distribution characteristics of all time nodes are integrated to form a dynamic evaluation result of the multi-dimensional value of drugs. This evaluation result clearly shows the comprehensive performance of the core parameters in time and weight distribution, providing dynamic and precise support for the multi-dimensional value evaluation of drugs.

[0142] like Figure 7 As shown, a drug multi-dimensional value assessment system based on AI includes:

[0143] The parameter distribution calculation module extracts the ratio of sales volume to usage frequency, analyzes the ratio of sales volume to adjustment intensity, performs cross calculations, and obtains a weight distribution table of multidimensional parameters of drugs based on the multidimensional numerical data of drug market performance, patient acceptance, and social benefits.

[0144] The trend offset correction module extracts the cumulative values ​​of sales volume, usage frequency and adjustment strength in the time series based on the drug multi-dimensional parameter weight distribution table, and obtains the parameter dynamic trend correction results by accumulating the change rate within the differentiated time points;

[0145] The external data fusion module extracts market competition pricing fluctuation data and policy incentive data based on the parameter dynamic trend correction results, compares pricing fluctuations with sales volume change rates, analyzes the interference intensity of policy incentives on adjustment efforts, and obtains external factor fusion impact data;

[0146] The interactive feature extraction module extracts the interactive values ​​of sales volume changes, policy adjustments and competition impact based on the integration of external factors and influence data, extracts dynamic correlation features and compares feature groups to obtain multi-dimensional dynamic interactive feature results;

[0147] The dynamic value assessment module extracts the dynamic change rate and distribution data of core parameters based on the results of multi-dimensional dynamic interaction characteristics, performs core dynamic change value analysis and hierarchical evaluation of time node parameters, and obtains the dynamic evaluation results of the multi-dimensional value of drugs.

[0148] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A multidimensional drug value assessment method based on AI, characterized in that: The following steps are involved: S1: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, the data is sorted and normalized, and the weight values ​​of dimension parameters are identified through numerical cross-identification to obtain the drug value correlation parameter matrix; S2: Based on the drug value-related parameter matrix, extract the multidimensional parameter time series data, call the parameter change rate and the cumulative value to iteratively accumulate and calculate the offset, analyze the trend and make corrections based on the offset, and generate a time dimension parameter change trend table; S3: Based on the time dimension parameter change trend table, market competition and policy adjustment data are introduced, external data and parameter value weights are matched, and the cumulative interference degree is analyzed by superposition. According to the interference adjustment trend, a multi-dimensional parameter dynamic superposition result is generated; S4: Based on the dynamic superposition results of the multi-dimensional parameters, extract the differentiated dimensional interaction data, map the dimensional data through nonlinear operations and classify and extract them, screen the numerical changes and dynamic correlation values ​​of key parameters, and generate a multi-dimensional value correlation map; S5: Based on the multi-dimensional value association map, the core parameters are normalized, the weight distribution is optimized through aggregation analysis, the hierarchical relationship and dynamic change rate of the core parameters are evaluated, the distribution relationship of differentiated time nodes is analyzed, and the dynamic evaluation results of the multi-dimensional value of the drug are obtained.

2. The AI-based multidimensional drug value assessment method according to claim 1, characterized in that: The drug value association parameter matrix includes parameter weight values, inter-dimensional correlations, and drug value parameters; the time dimension parameter change trend table includes parameter change trends, trend offsets, and time series cumulative values; the multidimensional parameter dynamic superposition results include external interference weights, trend interference intensity, and dynamically superimposed parameter values; the multidimensional value association map includes key parameter dynamic values, dimensional interaction data, and value association paths; the drug multidimensional value dynamic evaluation results include core parameter distribution, key weight dynamic changes, and hierarchical weight distribution.

3. The AI-based multidimensional drug value assessment method according to claim 1, characterized in that: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, the data is sorted and normalized, and the weight values ​​of dimension parameters are identified through numerical cross-identification to obtain the drug value correlation parameter matrix. The specific steps are as follows: S101: Based on the multi-dimensional numerical data of drug market performance, patient acceptance and social benefits, extract each numerical parameter and unify the format, adjust the proportion through extreme value range analysis, map the data to the corresponding interval, and generate parameter normalization results; S102: Based on the parameter normalization result, extract the numerical data of the two dimensions, analyze the numerical ratio and the degree of cross-influence, integrate the ratio results to identify the dimension correlation, and generate a parameter weight correlation matrix according to the numerical dependency relationship; S103: Based on the parameter weight association matrix, extract and optimize the dimension weight values ​​within the matrix, rearrange the matrix items to reduce data redundancy, perform matrix standardization processing based on the value of each item, and generate a drug value association parameter matrix.

4. The AI-based multidimensional drug value assessment method according to claim 1, characterized in that: Based on the drug value-related parameter matrix, extract multidimensional parameter time series data, call parameter change rate and cumulative value to iteratively accumulate and calculate offset, analyze and correct the trend according to the offset, and generate the time dimension parameter change trend table in the following steps: S201: Based on the drug value associated parameter matrix, extract the time series data of multidimensional parameters, summarize the parameter time point values ​​and change records, analyze the value change amplitude and cumulative value within the time interval, and obtain the time series cumulative change data; S202: Based on the cumulative change data of the time series, extract the change rate of the parameter within the differentiated time interval, analyze the change offset on the time axis, compare the change amplitude to summarize the regularity of the numerical offset, and obtain parameter trend offset data; S203: Based on the parameter trend offset data, extract the offset value to adjust the numerical trend on the time axis, integrate the parameter time series change deviation, calculate the adjusted parameter value, and generate a time dimension parameter change trend table.

5. The AI-based multidimensional drug value assessment method according to claim 4, characterized in that: Calculate the adjusted parameter value according to the formula: Among them, Q represents the adjusted parameter value, P t represents the original parameter value at time point t, T std Represents the global trend standard value of the parameter time series, ΔT t represents the time interval difference between time point t and its adjacent time points, B t Represents the basic offset at time point t, E1 is the trend correction weight coefficient, E2 is the time interval correction weight coefficient, and E3 is the basic offset correction weight coefficient.

6. The AI-based multidimensional drug value assessment method according to claim 1, characterized in that: Based on the time dimension parameter change trend table, market competition and policy adjustment data are introduced, external data and parameter value weights are matched, and the cumulative interference degree is analyzed by superposition. According to the interference adjustment trend, the steps of generating multi-dimensional parameter dynamic superposition results are as follows: S301: Based on the time dimension parameter change trend table, read the parameter value at the time point, analyze the corresponding relationship between the external data and the trend data, superimpose the external interference value to the parameter trend value, and generate an external data matching weight table; S302: Based on the external data matching weight table, extract the matched weight data, analyze the cumulative impact of the external interference value on the parameter trend, adjust the parameter change value affected by the external influence, sort out the parameter change amplitude and cumulative trend, and generate an interference intensity adjustment trend table; S303: Based on the interference intensity adjustment trend table, extract the dynamic change relationship of parameters in the trend data, group the time series data by weight, calculate the dynamic superposition value of the parameters, summarize the grouping numerical relationship for dynamic iterative processing, and generate a multi-dimensional parameter dynamic superposition result.

7. The AI-based multidimensional drug value assessment method according to claim 6, characterized in that: Calculate the dynamic superposition value of the parameter according to the formula Among them, w i represents the weight factor of the i-th parameter, p i represents the time series data parameter value of the i-th parameter, Represents the weighted average of the parameters in the current group, α i represents the nonlinear adjustment index of the i-th parameter, k j represents the time weight factor in the jth time period, t j represents the value at the jth time point, t0 represents the reference time point of the time series, n represents the total number of parameters, and m represents the total number of time points.

8. The AI-based multidimensional drug value assessment method according to claim 1, characterized in that: Based on the dynamic superposition results of the multi-dimensional parameters, the steps of extracting differentiated dimensional interaction data, mapping dimensional data through nonlinear operations and classifying and extracting, screening key parameter value changes and dynamic correlation values, and generating a multi-dimensional value correlation map are as follows: S401: Based on the dynamic superposition result of the multi-dimensional parameters, the parameter values ​​are read item by item at a time point, the corresponding relationship between the external data and the trend data is analyzed, the external interference value is superimposed on the original trend of the parameter, and an interactive data arrangement table is generated; S402: Based on the interactive data sorting table, analyzing the cumulative impact of external interference values ​​on parameter trends, adjusting parameter change values ​​affected by external factors in trend data, sorting the adjusted change amplitudes and cumulative trends, and generating a key parameter mapping set; S403: Based on the key parameter mapping set, extract the dynamic change relationship of parameters in the adjustment trend data, group the time series data by weight, summarize the grouping numerical relationship for dynamic iterative processing, and generate a multi-dimensional value association map.

9. The AI-based multidimensional drug value assessment method according to claim 1, characterized in that: The steps of normalizing the core parameters based on the multidimensional value association map, optimizing the weight distribution through aggregation analysis, evaluating the hierarchical relationship and dynamic change rate of the core parameters, and analyzing the distribution relationship of differentiated time nodes to obtain the dynamic evaluation results of the multidimensional value of the drug are as follows: S501: Based on the multi-dimensional value association map, extract the core parameter numerical data in the map, read the core parameter time series range, map the values ​​to a unified interval in proportion, integrate the processed core parameter values, and generate a core parameter normalized numerical table; S502: Based on the core parameter normalized value table, extract the weight values ​​corresponding to the normalized parameters, summarize the parameter change trends in the time series and reallocate the weights, sort out the weight differences, and generate a core parameter weight distribution table; S503: Based on the core parameter weight distribution table, extract the core parameter weight value at each time node, evaluate the time series change distribution of the weight, analyze the dynamic distribution law of the core parameters, integrate the evaluated distribution law and archive it by time dimension, and generate a dynamic evaluation result of the multi-dimensional value of the drug.

10. An AI-based multidimensional drug value assessment system, characterized in that: According to any one of claims 1 to 9, the AI-based multidimensional drug value assessment method comprises: The parameter distribution calculation module extracts the ratio of sales volume to usage frequency, analyzes the ratio of sales volume to adjustment intensity, performs cross calculations, and obtains a weight distribution table of multidimensional parameters of drugs based on the multidimensional numerical data of drug market performance, patient acceptance, and social benefits. The trend offset correction module extracts the sales volume, usage frequency and adjustment intensity cumulative values ​​in the time series based on the drug multi-dimensional parameter weight distribution table, and obtains the parameter dynamic trend correction result by accumulating the change rate within the differentiated time points; The external data fusion module extracts market competition pricing fluctuation data and policy incentive data based on the parameter dynamic trend correction results, compares pricing fluctuations with sales volume change rates, analyzes the interference intensity of policy incentives on adjustment efforts, and obtains external factor fusion impact data; The interactive feature extraction module extracts the interactive values ​​of sales volume changes, policy adjustments and competition impact based on the external factor fusion impact data, performs dynamic correlation feature extraction and feature grouping comparison, and obtains multi-dimensional dynamic interactive feature results; The dynamic value assessment module extracts the dynamic change rate and distribution data of the core parameters based on the multi-dimensional dynamic interaction characteristic results, performs core dynamic change value analysis and time node parameter hierarchical assessment, and obtains the dynamic assessment results of the multi-dimensional value of the drug.

Images