Data analysis method and system for drug sales

Abstract

The invention relates to the technical field of medicine sales data analysis and supply chain management, in particular to a data analysis method and system for medicine sales, and the method comprises the steps: obtaining multi-source data, constructing a multi-layer causal network through causal reasoning and quantum causal reasoning, and carrying out the multi-layer causal analysis and optimization; dynamic scene simulation and hybrid augmented reality technologies are combined to generate a market change scene and optimize a decision framework; and finally, automatic execution and real-time monitoring of the supply chain are realized through an intelligent contract and a block chain technology, and continuous optimization is performed through a feedback mechanism. According to the invention, the response speed and market adaptability of drug sales are remarkably improved, and the defects of multivariate analysis, real-time monitoring and supply chain optimization in the prior art are overcome.

Description

Technical Field

[0001] The present invention relates to the technical field of pharmaceutical sales data analysis and supply chain management, and in particular to a data analysis method and system for drug sales. Background Art

[0002] Data analysis of drug sales plays a vital role in modern pharmaceutical supply chain management. Through in-depth analysis of sales data, enterprises can more accurately predict market demand, optimize inventory management, and adjust supply chain strategies, thereby improving operational efficiency and market responsiveness. However, the existing drug sales management system (Chinese invention patent, publication number: CN117995370A, name: a pharmaceutical sales management system and its management method) has obvious deficiencies in coping with complex market changes and multi-dimensional data analysis. For example, the existing technology usually relies on traditional big data processing and statistical analysis methods, focusing mainly on the basic collation of sales data and regional demand forecasting. Although these systems can provide basic market demand forecasting and inventory management functions, they are poor in responding to market emergencies or multivariate correlation analysis due to the relatively single technical means adopted and the lack of deep causal reasoning and dynamic optimization capabilities. In addition, the feedback mechanism of the existing system is relatively lagging, making it difficult to adjust the supply chain strategy in real time, which may lead to inefficient supply chain operations and inability to fully respond to rapid market changes. Summary of the invention

[0003] In view of the many problems existing in the above-mentioned prior art, the present invention provides a data analysis method and system for drug sales. The present invention performs in-depth data analysis and dynamic optimization through the integration of causal reasoning, quantum causal reasoning, multi-layer causal network construction, dynamic scene simulation, hybrid augmented reality, smart contracts and blockchain technology. Through the coordinated application of these technologies, the present invention can monitor and adjust supply chain operations in real time, effectively improving the response speed and market adaptability of drug sales.

[0004] A data analysis method for drug sales, comprising the following steps:

[0005] Obtain sales data, market data, social media data, climate data, and public health data from multiple data sources, clean and standardize them, and generate the first integrated data through dynamic weight assignment;

[0006] Based on the first integrated data, causal reasoning and quantum causal reasoning are performed using distributed computing nodes to build a multi-layer causal network, which is optimized through an adaptive weight dynamic adjustment mechanism to generate the final causal network data;

[0007] Based on the final causal network data, a variety of heuristic algorithms and meta-heuristic algorithms are used for collaborative optimization, and the final adaptive optimization data is generated through multi-level parallel feedback loops and adaptive feedback loops;

[0008] Based on the final adaptive optimization data, dynamic scenario simulation is carried out, and market change scenarios are generated by combining mixed augmented reality technology. The decision framework is optimized and adjusted through the cross-scenario decision enhancement mechanism to generate an optimized decision framework;

[0009] Based on the optimized decision-making framework, smart contracts are generated and deployed, and the supply chain is automatically executed and monitored in real time through the blockchain platform. The supply chain operations are continuously optimized and adjusted through the feedback mechanism.

[0010] Preferably, the data cleaning and standardization step includes performing the following processing on the sales data, market data, social media data, climate data, and public health data:

[0011] Noise removal, through frequency domain filtering and time domain smoothing algorithms, eliminates random noise and interference signals in the data;

[0012] Outlier processing, using anomaly detection algorithms based on statistical analysis to identify and correct outliers in the data;

[0013] Missing data filling uses interpolation methods or machine learning prediction models to reasonably fill in the missing parts of the data.

[0014] Preferably, the causal inference step identifies and quantifies the causal relationship between data by constructing a causal relationship matrix, where the element C of the causal relationship matrix is ij Indicates data D i For data D j The causal influence strength, the causal relationship strength C ij Obtained through the following calculation formula:

[0015]

[0016] Among them, C ij Indicates data D i For data D j The causal influence strength of W k is the weight coefficient of causal relationship, which is dynamically adjusted based on the correlation between historical data and current data; f k (D i , D j ) represents a function used to calculate causal relationships, which uses different mathematical models according to different data types and causal chain characteristics.

[0017] Preferably, the quantum causal reasoning step uses a quantum computing algorithm to process high-dimensional causal relationships, wherein the ultra-parallel computing capability of quantum bits is used to quickly process multi-variable causal chains, and the calculation results of quantum causal reasoning optimize the causal relationship matrix through quantum state superposition and interference phenomena, thereby generating quantum computing result data.

[0018] Preferably, the multi-layer causal network construction step includes the following contents:

[0019] The final causal network data is divided into multiple causal levels. Each causal level is stratified based on the strength of the causal relationship. The strongest causal relationship is divided into the top causal network, and the weaker causal relationships are divided into the middle and bottom causal networks in turn.

[0020] In each layer of the causal network, the network structure is optimized through the correlation of the causal chain, so that each layer of the causal network can independently reflect the influence of a specific causal level, and finally generate multi-layer causal network data to describe the causal relationship structure at different levels in drug sales data.

[0021] Preferably, the adaptive weight dynamic adjustment mechanism optimizes the causal network through the following process:

[0022] Based on real-time market feedback data, the weights of each causal chain in the multi-layer causal network are dynamically adjusted. The weight adjustment is based on the changing trends, emergencies and abnormal fluctuations in the feedback data.

[0023] By adaptively adjusting the model, the weight of each causal chain is recalculated to generate the optimized final causal network data.

[0024] Preferably, the collaborative optimization step includes the following:

[0025] Generate preliminary optimization data through the collaborative operation of genetic algorithm, ant colony optimization algorithm and particle swarm optimization algorithm. Genetic algorithm is used to explore the global optimal solution, ant colony optimization algorithm is used to find the local optimal solution of path dependence, and particle swarm optimization algorithm searches for the global optimal solution in multi-dimensional space.

[0026] The preliminary optimization data are integrated and optimized through multi-level parallel feedback loops. The optimization data generated by different algorithms are cross-validated and weighted through the feedback loop to generate the final adaptive optimization data.

[0027] Preferably, the multi-level parallel feedback loop optimizes each link of the supply chain by:

[0028] Each feedback loop operates independently to optimize different parts of the supply chain;

[0029] Between multi-level feedback loops, the consistency of feedback data in each link is maintained through a key data sharing mechanism, and the optimization strategy of each link is adjusted accordingly based on the feedback results.

[0030] Preferably, the dynamic scenario simulation step includes generating market change scenario data by:

[0031] Generate preliminary market change scenario data using Monte Carlo simulation, in which possible market changes are predicted through simulation and statistical analysis of a large number of random samples;

[0032] Combined with mixed augmented reality technology, market changes are further simulated in virtual scenes, and optimized market change scenario data is generated on this basis, ultimately forming a cross-scenario optimization decision-making framework that is used to guide decision-making execution in actual market environments.

[0033] A system for implementing the data analysis method for drug sales, comprising:

[0034] a data collection module, configured to obtain sales data, market data, social media data, climate data, and public health data from multiple data sources, clean and standardize the sales data, market data, social media data, climate data, and public health data, and generate first integrated data through dynamic weight allocation;

[0035] A causal reasoning module, connected to the data acquisition module, is used to perform causal reasoning and quantum causal reasoning based on the first integrated data using distributed computing nodes, to construct a multi-layer causal network, to optimize the multi-layer causal network through an adaptive weight dynamic adjustment mechanism, and to generate final causal network data;

[0036] A collaborative optimization module, connected to the causal reasoning module, is used to perform collaborative optimization based on the final causal network data using a variety of heuristic algorithms and meta-heuristic algorithms, and generate final adaptive optimization data through multi-level parallel feedback loops and adaptive feedback loops;

[0037] The dynamic scenario simulation module is connected to the collaborative optimization module and is used to perform dynamic scenario simulation based on the final adaptive optimization data, generate market change scenarios in combination with mixed augmented reality technology, and optimize and adjust the decision framework through the cross-scenario decision enhancement mechanism to generate an optimized decision framework;

[0038] The smart contract execution module is connected to the dynamic scenario simulation module and is used to generate and deploy smart contracts based on the optimized decision-making framework, realize automated execution and real-time monitoring of the supply chain through the blockchain platform, and continuously optimize and adjust supply chain operations through a feedback mechanism.

[0039] Compared with the prior art, the advantages and beneficial effects of the present invention are:

[0040] The present invention achieves in-depth analysis and dynamic optimization of drug sales data through the integration and innovation of multiple advanced technical means;

[0041] The present invention uses causal reasoning and quantum causal reasoning technology to construct a multi-layer causal network that can reflect the multi-dimensional causal relationship in drug sales, overcoming the shortcomings of existing systems in multivariate association analysis;

[0042] The present invention uses dynamic scene simulation and mixed augmented reality technology to achieve highly visualized prediction of market changes and enhance the flexibility and accuracy of decision-making;

[0043] The present invention introduces smart contracts and blockchain technology to achieve automated execution and real-time monitoring of the supply chain, and continuously optimizes it through a feedback mechanism, fundamentally solving the defects of existing technologies in terms of real-time and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic diagram of the process of the present invention;

[0045] Figure 2 A flowchart for constructing and optimizing a causal network in the present invention;

[0046] Figure 3 It is a flow chart of dynamic scene simulation and decision optimization in the present invention;

[0047] Figure 4 This is a flow chart of supply chain management and smart contract execution in the present invention;

[0048] Figure 5 This is a system structure block diagram of the present invention. DETAILED DESCRIPTION

[0049] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present disclosure. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is obvious that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0050] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0051] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0052] like Figure 1 As shown, a data analysis method for drug sales includes the following steps:

[0053] Obtain sales data, market data, social media data, climate data, and public health data from multiple data sources, clean and standardize them, and generate the first integrated data through dynamic weight assignment;

[0054] Preferably, the data cleaning and standardization step includes performing the following processing on the sales data, market data, social media data, climate data, and public health data:

[0055] Noise removal, through frequency domain filtering and time domain smoothing algorithms, eliminates random noise and interference signals in the data;

[0056] Outlier processing, using anomaly detection algorithms based on statistical analysis to identify and correct outliers in the data;

[0057] Missing data filling uses interpolation methods or machine learning prediction models to reasonably fill in the missing parts of the data.

[0058] Data cleaning and standardization are key steps in the preprocessing stage. Their goal is to eliminate noise and outliers in the original data and to complete missing data to improve the consistency and integrity of the data. In actual operation, the system uses frequency domain filtering and time domain smoothing algorithms for noise removal. The principle of frequency domain filtering is to convert time domain signals into frequency domain signals through Fourier transform, and to purify data by filtering out noise components of specific frequencies. This process is particularly suitable for removing periodic or high-frequency interference signals. The time domain smoothing algorithm reduces random fluctuations in the data in the time domain through methods such as moving average or exponential smoothing to ensure the stability of the data. The effect of noise removal is to make the data closer to the actual situation and reduce errors in analysis.

[0059] Outlier processing is achieved through anomaly detection algorithms based on statistical analysis. These algorithms identify abnormal data points that deviate significantly from the normal distribution by calculating the statistical characteristics of the data distribution, such as the mean, standard deviation, etc. Outliers may be caused by data collection errors, external interference, or extreme situations, so they must be corrected or eliminated before analysis. In specific implementations, common statistical methods such as box plots and Z-score analysis can be used to automatically detect and process these outliers. The effect of this step is to reduce the interference of abnormal data on the overall analysis results and improve the accuracy of data analysis.

[0060] For the filling of missing data, the present invention adopts interpolation method or machine learning prediction model. Interpolation method infers the reasonable value of missing data points through known data points, and commonly used methods include linear interpolation and polynomial interpolation. This method is suitable for situations where there are few missing data and adjacent data have strong correlation. For situations where there are more missing data or the relationship between data is complex, a prediction model based on machine learning can be used, such as the K nearest neighbor algorithm (KNN) or a time series prediction model based on historical data. These models make reasonable predictions and completions for missing values ​​by learning the inherent laws of data. The effect of missing data filling is to maintain the integrity of the data set and avoid analysis errors caused by incomplete data.

[0061] Finally, through the above processing steps, the system will standardize the processed data to eliminate the inconsistencies between different data sources due to different measurement units or value ranges. Standardization is usually performed by scaling the data or using the Z-score standardization method. The effect of this step is to map all data to a unified scale for subsequent fusion processing and analysis.

[0062] After all data cleaning and standardization are completed, the system will dynamically assign weights to data from different data sources. The principle of dynamic weight assignment is to assign a weight coefficient to each type of data based on the quality, timeliness and relevance of each type of data to reflect its importance in the integrated data. The system can use an adaptive algorithm to adjust the weights in real time to ensure the optimality of the data fusion results. The effect of dynamic weight assignment is to generate a first integrated data that comprehensively considers the characteristics of multi-source data, retains the key information of each data source, and achieves data balance through weight adjustment, laying the foundation for subsequent causal reasoning and multi-layer causal network construction.

[0063] Based on the first integrated data, causal reasoning and quantum causal reasoning are performed using distributed computing nodes to build a multi-layer causal network, which is optimized through an adaptive weight dynamic adjustment mechanism to generate the final causal network data;

[0064] Causal reasoning is a key step in identifying and quantifying the causal relationship between different variables. In the present invention, causal reasoning is performed in parallel through distributed computing nodes to cope with the large-scale and high-dimensional characteristics of drug sales data. The design of distributed computing nodes allows the system to process data simultaneously on multiple computing nodes, thereby accelerating the identification and calculation process of causal relationships. Specifically, each node is responsible for calculating the causal relationship between a certain part of the variables in the data set, which can significantly improve the computing efficiency, and through the integration of the results, a global causal relationship matrix is ​​obtained. Each element of this matrix represents the causal strength between two variables, and the causal strength is calculated by a specific algorithm, which is dynamically adjusted according to the correlation between historical data and current data to ensure the accuracy of the causal relationship.

[0065] On the basis of causal reasoning, the present invention further introduces quantum causal reasoning technology. Quantum causal reasoning uses the parallel processing capability of quantum computing and is particularly suitable for the analysis of high-dimensional data sets and complex causal chains. In the process of quantum causal reasoning, the ultra-parallel computing capability of quantum bits enables the system to process multiple causal paths at the same time, and optimize the calculation results of causal relationships through the superposition and interference phenomena of quantum states. Specifically, quantum computing can explore multiple possible causal structures in a very short time and select the optimal causal path, thereby generating a more accurate causal relationship matrix. This matrix combines the results of traditional causal reasoning and forms a more accurate causal relationship model by re-evaluating and optimizing the causal strength.

[0066] After obtaining the preliminary causal relationship model, the system begins to build a multi-layer causal network. The construction of the multi-layer causal network is based on the hierarchical processing of causal relationship strength. Specifically, the system constructs the variable pairs with higher causal relationship strength as the top-level causal network, and the variable pairs with weaker causal relationship strength as the middle and bottom-level causal networks. Each layer of the network reflects the influence of different levels of causal relationships on drug sales data. This hierarchical processing enables the system to more carefully characterize the complex causal structure and clarify the mechanism of action of different levels of causal relationships.

[0067] After the multi-layer causal network is constructed, the system further optimizes the network through an adaptive weight dynamic adjustment mechanism. The principle of the adaptive weight dynamic adjustment mechanism is to dynamically adjust the weights of each causal chain in the multi-layer causal network according to real-time market feedback and data changes. Specifically, the system will continue to monitor changes in the market environment, such as fluctuations in drug demand, strategic adjustments by competitors, etc., and update the weights of each causal chain in the causal network in real time according to these changes. In this way, the system can maintain the optimal structure of the causal network under different market environments and ensure the dynamic accuracy of the causal relationship. Finally, the multi-layer causal network after weight adjustment is integrated into the final causal network data, which contains the optimized causal relationship model, and can provide a high-quality data foundation for subsequent collaborative optimization and decision support.

[0068] In summary, the present invention successfully generates the final causal network data through distributed causal reasoning, quantum causal reasoning, multi-layer causal network construction and adaptive weight dynamic adjustment. It not only accurately depicts the complex causal relationship in drug sales data, but also adapts to changes in the market environment through dynamic adjustment, providing reliable support for subsequent data optimization and decision-making processes. The effect of this process is that through in-depth analysis and optimization, the system can better understand and predict the key factors in drug sales, thereby improving the intelligence level of the entire supply chain management.

[0069] Preferably, the causal inference step identifies and quantifies the causal relationship between data by constructing a causal relationship matrix, where the element C of the causal relationship matrix is ij Indicates data D i For data D j The causal influence strength, the causal relationship strength C ij Obtained through the following calculation formula:

[0070]

[0071] Among them, C ij Indicates data D i For data D j The causal influence strength of k is the weight coefficient of causal relationship, which is dynamically adjusted based on the correlation between historical data and current data; f k (D i , D j ) represents a function used to calculate causal relationships, which uses different mathematical models according to different data types and causal chain characteristics.

[0072] Weight coefficient W kIt will make adaptive adjustments as the market environment changes. For example, if the impact of climate data on drug sales increases in a certain period, the system will automatically increase the weight of the climate-related causal path to ensure that the causal relationship matrix can reflect the actual market changes in real time. Function f k (D i , D j ) is the calculation core of causal relationships. This function uses different mathematical models for calculation according to different data types and the characteristics of the causal chain. Common mathematical models include linear regression models, nonlinear models, time series models, etc. These models are suitable for different types of causal relationships. For example, for simple linear relationships, a linear regression model may be used, while for complex time series data, a time series model is used to capture long-term trends and seasonal fluctuations.

[0073] In practical applications, the causal relationship matrix can help reveal the key driving factors in drug sales. For example, through analysis, it is found that the sharp fluctuations in market data within a certain period of time significantly affect the sales trend of the drug. At this time, the corresponding element C in the causal relationship matrix ij It will show the strong influence of market data on sales data. This information can help decision makers adjust market strategies in time to cope with changes. In addition, if the system finds that the weight of social media data W k Increase, and the causal relationship strength between it and sales data is C ij If the sales volume of drugs increases, it can be judged that the influence of social media on drug sales has increased, which provides a strong basis for the adjustment of marketing strategies.

[0074] Through this method, the present invention can construct a multi-dimensional, multi-level causal network, systematically reveal the complex relationships in drug sales data, and provide a solid foundation for further optimization and decision-making. The resulting causal matrix can not only clearly show the influence of various factors, but also maintain sensitivity and adaptability to changes in the market environment through dynamic adjustment. The effect of this causal reasoning is that it enables companies to more accurately grasp market dynamics, optimize sales strategies, and improve overall operating efficiency in a data-driven manner.

[0075] Preferably, the quantum causal reasoning step uses a quantum computing algorithm to process high-dimensional causal relationships, wherein the ultra-parallel computing capability of quantum bits is used to quickly process multi-variable causal chains, and the calculation results of quantum causal reasoning optimize the causal relationship matrix through quantum state superposition and interference phenomena, thereby generating quantum computing result data.

[0076] Quantum computing has shown significant advantages in processing multi-variable causal chains due to its unique parallel computing capabilities and efficient ability to solve complex problems. In the traditional computing framework, multi-variable causal chains usually lead to a sharp increase in computational complexity due to high dimensions and a large number of variables, which makes a comprehensive analysis of causal relationships difficult and time-consuming. Quantum computing, through the super-parallel computing capabilities of quantum bits, can process multiple causal paths simultaneously, thereby greatly improving the computational speed and accuracy of causal reasoning.

[0077] In quantum causal reasoning, quantum state superposition and interference phenomena are the core principles. Quantum state superposition allows the system to process multiple possible causal paths at the same time, while quantum interference is used to optimize the selection of these paths. Specifically, quantum bits in quantum computing can exist in multiple states (superposition states) at the same time, which enables the system to evaluate the possibility of multiple causal chains in one round of calculations. When these chains undergo quantum calculations, the quantum interference effect will cause the chain paths that conform to the causal relationship to be strengthened, while the paths that do not conform or are interfered by noise will be weakened or offset, thereby optimizing the calculation results of the causal relationship matrix. The final quantum computing result data contains the optimal causal chain structure, and these structures have been verified and optimized by quantum computing to ensure high accuracy and reliability.

[0078] For example, in the analysis of drug sales data, suppose it is necessary to evaluate the combined effects of multiple market factors (such as price fluctuations, publicity efforts, and consumer sentiment) on drug sales. In traditional computing methods, it is necessary to evaluate the mutual influence of these factors and their cumulative effects one by one, which is almost impossible in the context of high-dimensional data sets. Through quantum causal reasoning, the system can evaluate all possible combinations of factors at once and quickly derive possible causal paths through quantum state superposition. The quantum interference effect further optimizes these paths and ultimately outputs the most likely causal relationship model, indicating which market factors have the greatest impact on sales in a specific time period. This provides a basis for the rapid adjustment of marketing strategies, enabling companies to take effective countermeasures in a short period of time to avoid sales declines caused by market fluctuations.

[0079] like Figure 2 As shown, preferably, the multi-layer causal network construction step includes the following contents:

[0080] The final causal network data is divided into multiple causal levels. Each causal level is stratified based on the strength of the causal relationship. The strongest causal relationship is divided into the top causal network, and the weaker causal relationships are divided into the middle and bottom causal networks in turn.

[0081] In each layer of the causal network, the network structure is optimized through the correlation of the causal chain, so that each layer of the causal network can independently reflect the influence of a specific causal level, and finally generate multi-layer causal network data to describe the causal relationship structure at different levels in drug sales data.

[0082] The strength of a causal relationship is a key indicator used to quantify the degree of causal influence between two variables. In the process of constructing a multi-layer causal network, the system will divide it into different causal levels according to the strength of the causal relationship between each pair of data variables. Specifically, the causal relationships with the highest strength are classified into the top-level causal network. These relationships are usually the factors that have the most direct and significant impact on drug sales, such as the direct impact of pricing strategies on sales. Weaker causal relationships are divided into middle-level and bottom-level causal networks in turn. These weaker relationships may involve indirect or long-term impacts, such as the impact of seasonal changes on consumer sentiment, which in turn indirectly affects the market demand for drugs. Through this hierarchical processing, the system can clearly distinguish and understand the different causal levels in drug sales data, ensuring that the causal relationships between each level are fully and independently analyzed.

[0083] In each layer of the causal network, the system further optimizes the network structure through the correlation of the causal chain. The correlation of the causal chain refers to the mutual relationship and degree of dependence between different variables in a causal path. In order to optimize the network structure, the system will identify and strengthen those variable nodes that play a key role in the causal chain, and weaken or remove chains with low correlation or redundant. This optimization ensures that each layer of the causal network can independently and accurately reflect the influence of causal relationships at a specific level, avoiding interference caused by redundant or weakly correlated factors. For example, in the top-level causal network, the system may find that drug pricing, promotional activities, and sales channels are highly correlated. The chains of these variables will be strengthened during the optimization process and become the core basis for decision-making. In the bottom-level causal network, seasonal factors may indirectly affect sales through multiple mediating variables. These chains will be retained, but their priority in the network will be lower.

[0084] Through the above stratification and optimization process, the system generates multi-layer causal network data. This dataset describes the causal relationship structure at different levels in drug sales data and provides a multi-dimensional causal model. The main effect of this model is that it can help decision makers understand the complex causal mechanism of drug sales as a whole and make targeted strategic adjustments from different levels. For example, the direct relationship in the top-level causal network may need to be optimized through price adjustment or promotion strategies, while the indirect relationship in the middle and bottom causal networks may need to be gradually influenced through long-term brand building or market education.

[0085] Preferably, the adaptive weight dynamic adjustment mechanism optimizes the causal network through the following process:

[0086] Based on real-time market feedback data, the weights of each causal chain in the multi-layer causal network are dynamically adjusted. The weight adjustment is based on the changing trends, emergencies and abnormal fluctuations in the feedback data.

[0087] By adaptively adjusting the model, the weight of each causal chain is recalculated to generate the optimized final causal network data.

[0088] Real-time market feedback data refers to the latest information continuously obtained from the market, including but not limited to changes in sales data, fluctuations in market demand, strategic adjustments by competitors, and changes in consumer behavior. In drug sales data analysis, these feedback data are crucial for timely understanding of market dynamics. For example, when the sales volume of a certain drug rises or falls sharply in a short period of time, the system will capture this change through market feedback data and adjust the weights of related variables in the causal network accordingly.

[0089] The working principle of the adaptive weight dynamic adjustment mechanism is to calculate and update the weights of each causal chain in the multi-layer causal network in real time through an adaptive adjustment model. Specifically, the system will re-evaluate the influence of each causal chain based on the changing trends in the feedback data (such as an increase or decrease in sales), emergencies (such as policy changes or public health events), and abnormal fluctuations (such as sudden price wars or supply disruptions in the market). In this process, the weight adjustment is dynamic, that is, as time goes by and the market environment changes, the system can continuously update the weights of the causal chain to ensure that the causal network always reflects the current market reality. For example, if the impact of climate data on drug sales in a certain period increases significantly, the system will automatically increase the weights of climate-related causal chains, so that the causal network can better reflect the market characteristics of this period.

[0090] The adaptive adjustment model is the core tool for weight adjustment. It identifies the main factors affecting causal relationships through deep learning or statistical analysis of feedback data, and recalculates the weights of the causal chain accordingly. The adaptive adjustment model can use a variety of algorithms, such as gradient descent, genetic algorithm, or adjustment method based on Bayesian theory, which can accurately capture small changes in causal relationships while ensuring computational efficiency. After generating new weights, the system will apply them to the causal network to generate optimized final causal network data. These data not only contain the latest market information, but also ensure the accuracy and real-time nature of causal relationships through dynamic adjustment.

[0091] For example, suppose that in a certain market area, a climate-related public health event (such as a flu epidemic) suddenly breaks out, causing a surge in demand for a certain antiviral drug. The system finds through real-time market feedback data that the causal relationship between climate factors and drug sales has become extremely significant in a short period of time. The adaptive weight dynamic adjustment mechanism will immediately respond to this change, and through the adaptive adjustment model, it will greatly increase the weight of the climate-related causal chain, so that the optimized causal network can accurately reflect the current market changes. This adjustment enables companies to quickly realize the significant impact of climate change on drug sales and adopt corresponding market strategies, such as increasing inventory or strengthening marketing, to cope with sudden demand.

[0092] Based on the final causal network data, a variety of heuristic algorithms and meta-heuristic algorithms are used for collaborative optimization, and the final adaptive optimization data is generated through multi-level parallel feedback loops and adaptive feedback loops;

[0093] Heuristic algorithms and meta-heuristic algorithms are commonly used methods for solving complex optimization problems. In the present invention, heuristic algorithms such as genetic algorithms, ant colony optimization algorithms and particle swarm optimization algorithms are used to search for optimal solutions in multidimensional data space. The advantage of these algorithms is that they can find solutions close to the global optimal solution in a relatively short time, and are particularly suitable for processing high-dimensional, multi-variable data sets. For example, genetic algorithms iteratively update solutions by simulating the process of natural selection to find the best combination of sales data; ant colony optimization algorithms optimize supply chain logistics problems by simulating the path of ants looking for food; and particle swarm optimization algorithms search for optimal solutions for drug pricing strategies in multidimensional space. Meta-heuristic algorithms, such as simulated annealing and taboo search, further optimize the search path of the heuristic algorithm to prevent falling into the local optimal solution, thereby improving the effect of the entire optimization process.

[0094] In the analysis of drug sales data, the collaborative work of these algorithms is particularly important. Through collaborative optimization, the system can integrate the advantages of different algorithms, comprehensively consider the complex relationships between multiple factors, and generate more accurate optimization data. Specifically, the system first starts different heuristic algorithms based on the final causal network data to optimize multiple variables related to drug sales, including but not limited to pricing strategies, promotional activities, inventory management, and marketing strategies. Each algorithm will independently generate preliminary optimization data to reflect the potential effects of different strategy combinations. Then, the system integrates and re-optimizes these preliminary optimization data through meta-heuristic algorithms to ensure that the optimization results are not only optimal in all aspects, but also globally consistent and optimal.

[0095] The introduction of multi-level parallel feedback loops and adaptive feedback loops is to further enhance the reliability and adaptability of optimization results. In the multi-level parallel feedback loop, the system will run independently for different market feedback loops, such as the optimization loop for inventory management, the optimization loop for logistics routes, and the optimization loop for marketing strategies. In each loop, the system adjusts the optimization strategy through real-time market feedback data, and inputs the adjusted results back into the optimization algorithm for further optimization. This parallel multi-level feedback mechanism ensures that different market strategies can maintain dynamic optimization under different market conditions, avoiding the possible inadaptability of a single strategy in different market environments.

[0096] The adaptive feedback loop is to ensure that the optimization results can remain long-term in a changing market environment. Through the adaptive feedback loop, the system can adjust the parameters of the optimization strategy in real time to respond to unexpected events or long-term trend changes in the market. For example, if drug sales suddenly surge in a certain market area, the system will capture this change through the feedback loop and immediately adjust the optimization parameters of the supply chain and market strategy, such as increasing inventory or increasing marketing efforts, to ensure that the supply chain and market strategy can quickly respond to market changes and maximize sales revenue.

[0097] Preferably, the collaborative optimization step includes the following:

[0098] Generate preliminary optimization data through the collaborative operation of genetic algorithm, ant colony optimization algorithm and particle swarm optimization algorithm. Genetic algorithm is used to explore the global optimal solution, ant colony optimization algorithm is used to find the local optimal solution of path dependence, and particle swarm optimization algorithm searches for the global optimal solution in multi-dimensional space.

[0099] The preliminary optimization data are integrated and optimized through multi-level parallel feedback loops. The optimization data generated by different algorithms are cross-validated and weighted through the feedback loop to generate the final adaptive optimization data.

[0100] Genetic algorithm is an optimization method based on natural selection and genetic mechanism, which is particularly suitable for solving global optimization problems. In drug sales data analysis, the main task of genetic algorithm is to explore the global optimal solution, that is, to find a strategy combination that can maximize sales benefits in the multi-dimensional sales data space. The algorithm simulates the process of biological evolution, including selection, crossover and mutation, and gradually iterates the optimization solution. For example, in the optimization of drug pricing strategy, genetic algorithm can evaluate the combination of different pricing strategies and finally find an optimal pricing plan that can both meet market demand and maximize corporate profits.

[0101] The ant colony optimization algorithm is good at solving path-dependent local optimization problems. It finds the best path or solution by simulating the process of ants foraging. In the present invention, the ant colony optimization algorithm is used to optimize the supply chain path or logistics route to ensure that the drug can reach the market in the shortest time and at the lowest cost. By simulating the pheromone traces left by ants on the path, the algorithm gradually strengthens the selection of the optimal path and ultimately achieves effective optimization of the logistics path. For example, when optimizing the drug delivery path, the ant colony optimization algorithm can find the shortest delivery path through continuous trial and optimization, thereby reducing delivery time and cost.

[0102] Particle swarm optimization is an optimization algorithm based on swarm intelligence, which is particularly suitable for searching for the global optimal solution in multidimensional space. The algorithm gradually approaches the optimal solution by simulating the movement of particle swarms in the search space and mutual learning. In drug sales data analysis, particle swarm optimization can be used to optimize the combination of multiple variables, such as the comprehensive optimization of marketing strategies, inventory management, and production scheduling. The particle swarm moves in multidimensional space, evaluates the effect of each combination, and finally converges to a global optimal solution. For example, in marketing strategy optimization, the particle swarm optimization algorithm can find an optimal marketing plan by continuously adjusting the combination of marketing strategies, which can attract target customers and effectively control marketing costs.

[0103] In the coordinated operation of these three algorithms, the system first generates preliminary optimization data. Each algorithm processes different dimensions or focuses of the data and generates optimization data that can reflect the effects of different strategies. These preliminary optimization data provide the basis for subsequent integration and further optimization.

[0104] The preliminary optimization data will be integrated and optimized through multi-level parallel feedback loops. The principle of multi-level parallel feedback loops is to cross-validate and adjust the weights of the optimization data generated by different algorithms through multiple independent feedback loops. Each feedback loop runs independently for a specific optimization goal (such as price, inventory or logistics). The system continuously adjusts the parameters of the optimization strategy based on real-time market feedback data and inputs the adjusted results into the optimization algorithm again. This multi-level parallel feedback mechanism ensures the adaptability and optimality of each strategy in different market environments. Through cross-validation, the system can identify the complementarity or conflict points between various optimization schemes, and balance these factors through weight adjustment to generate more robust final adaptive optimization data.

[0105] For example, suppose that in a certain market, the demand for drugs suddenly increases, resulting in increased inventory pressure. The system detects this change through real-time market feedback data and immediately adjusts the weights of various optimization strategies through multi-level parallel feedback loops, such as increasing the priority of inventory management, and re-evaluating marketing strategies through particle swarm optimization algorithms to ensure that marketing activities are consistent with inventory supply. This dynamic adjustment mechanism enables the system to respond quickly to market changes and generate optimization data that can maximize sales revenue and best meet current market conditions.

[0106] Preferably, the multi-level parallel feedback loop optimizes each link of the supply chain in the following ways:

[0107] Each feedback loop operates independently to optimize different parts of the supply chain;

[0108] Between multi-level feedback loops, the consistency of feedback data in each link is maintained through a key data sharing mechanism, and the optimization strategy of each link is adjusted accordingly based on the feedback results.

[0109] The independent operation of each feedback loop is the basis for optimizing each link of the supply chain. The supply chain involves multiple key links, including procurement, production, inventory management, logistics distribution, and marketing. Each link has its own unique optimization goals and challenges, so it needs to be handled by an independent optimization loop. For example, in the inventory management link, the optimization goal may be to reduce inventory costs while ensuring sufficient supply; while in the logistics distribution link, the optimization goal may be to shorten delivery time and reduce transportation costs. In order to achieve these goals, the system will run a feedback loop for each link independently, using real-time data to evaluate the current operating status and optimize the strategy by adaptively adjusting the model. This independently running feedback loop ensures that each supply chain link can be optimized according to its own needs without interference from other links.

[0110] However, each link of the supply chain does not exist in isolation, and there are close connections and dependencies between them. Therefore, in order to ensure the coordinated operation of the entire supply chain, the present invention maintains the consistency of feedback data in each link through a key data sharing mechanism. The principle of the key data sharing mechanism is that between each independently running feedback loop, the system will automatically identify and share key data that have an impact on multiple links. For example, market sales data not only affects the inventory management link, but also has an important impact on the production and procurement links. Therefore, the system will share these market data in the relevant feedback loops to ensure that all relevant links can be optimized and adjusted based on the latest data. This data sharing mechanism ensures the synchronization and consistency between the feedback loops, and avoids decision-making errors caused by data islands or information asymmetry.

[0111] Based on the key data sharing mechanism, each feedback loop can refer to and coordinate with each other during operation. For example, when market demand fluctuates, the feedback loop in the market sales link will quickly capture this change and pass the change information to the feedback loop in the inventory management and production links through the key data sharing mechanism. After receiving the information, these links will immediately adjust the inventory strategy and production plan, such as increasing the frequency of inventory replenishment or adjusting the production rhythm to ensure that new market demands can be met. At the same time, the feedback loop in the logistics and distribution link will also optimize the distribution strategy based on these adjustments, such as adjusting the distribution route or speeding up the transportation speed to ensure that the product can be delivered to the market in a timely manner.

[0112] In practical applications, for example, suppose that the market demand for a certain drug suddenly surges during the flu season. Changes in sales data will be quickly captured and fed back into the feedback loop of inventory management and production. The inventory management link may find that the current inventory is insufficient to meet the surge in demand, so it increases the replenishment frequency through the feedback loop. At the same time, the production link will increase production shifts or speed up the production line based on the feedback from inventory management to ensure that inventory can be replenished in time. At the same time, the logistics link will optimize transportation routes or increase transportation batches based on these adjustments to ensure that drugs can be quickly supplied to various sales points. This series of adjustments are achieved through multi-level parallel feedback loops, and the optimization decisions of each link are based on the latest shared data, so as to achieve efficient operation of the entire supply chain.

[0113] Based on the final adaptive optimization data, dynamic scenario simulation is carried out, and market change scenarios are generated by combining mixed augmented reality technology. The decision framework is optimized and adjusted through the cross-scenario decision enhancement mechanism to generate an optimized decision framework;

[0114] like Figure 3 As shown, preferably, the dynamic scenario simulation step includes generating market change scenario data in the following manner:

[0115] Generate preliminary market change scenario data using Monte Carlo simulation, in which possible market changes are predicted through simulation and statistical analysis of a large number of random samples;

[0116] Combined with mixed augmented reality technology, market changes are further simulated in virtual scenes, and optimized market change scenario data is generated on this basis, ultimately forming a cross-scenario optimization decision-making framework that is used to guide decision-making execution in actual market environments.

[0117] The first step of dynamic scenario simulation is to generate preliminary market change scenario data using Monte Carlo simulation. Monte Carlo simulation is a statistical analysis method based on random samples, which simulates various possible change scenarios in the market by generating a large number of random samples. In drug sales data analysis, Monte Carlo simulation predicts the future trend of the market by considering the random fluctuations of multiple variables, such as price, demand, competitive environment, and external factors (such as seasonal changes and health events). Specifically, the system generates thousands or even millions of random samples, each representing a possible market scenario, and then statistically analyzes the results of these samples to obtain the possible range and probability distribution of market changes. For example, before the flu season, the system can use Monte Carlo simulation to predict the possible fluctuations in drug sales under various hypothetical conditions (such as the severity of the epidemic, market demand fluctuations, supply chain stability, etc.). This process can help decision makers identify potential risks and opportunities and provide data support for further strategy formulation.

[0118] After generating preliminary market change scenario data, the system will combine mixed augmented reality (AR) technology to further simulate market changes in virtual scenes. Mixed augmented reality technology provides an immersive simulation experience by combining virtual information with the real environment. In the present invention, mixed augmented reality technology is used to build a more intuitive and dynamic market change scenario, so that decision makers can foresee the implementation effects of different market strategies in a virtual environment. For example, the system can display the impact of different pricing strategies on market demand in an augmented reality environment, or simulate the impact of a certain public health event outbreak on the drug supply chain. Through this intuitive simulation, decision makers can better understand the possible consequences of complex market changes and adjust strategies based on the simulation results. Mixed augmented reality technology not only improves the realism and interactivity of scene simulation, but also makes complex market changes and strategy adjustments easier to understand and analyze.

[0119] Based on these two steps, the system ultimately generates optimized market change scenario data and forms a cross-scenario optimization decision-making framework. The role of the cross-scenario decision enhancement mechanism is to optimize the decision framework by analyzing and comparing the strategy effects under different market change scenarios. Specifically, based on the simulation results, the system will identify the decision paths that perform best in multiple market scenarios and integrate these paths into the optimization decision framework. For example, the system may find that in most market scenarios, increasing the advertising intensity of drugs and optimizing inventory management can significantly increase market share. Therefore, the optimization decision framework will prioritize these strategy combinations and provide clear guidance for actual implementation.

[0120] Through the generated cross-scenario optimization decision framework, enterprises can respond to various changes more flexibly in the actual market environment. The effect of this framework is that it not only takes into account the current market dynamics, but also predicts possible market changes in the future and provides decision makers with a variety of response strategies. For example, suppose the system finds through dynamic scenario simulation that in a certain season in the future, due to the sudden outbreak of health events, market demand may rise sharply. At this time, the optimization decision framework will recommend that companies increase inventory in advance, increase market publicity, and optimize the supply chain to ensure a quick response. This kind of advance planning and multi-scenario response strategy can effectively reduce the risks brought by market fluctuations and improve the market adaptability and competitiveness of enterprises.

[0121] Based on the optimized decision-making framework, smart contracts are generated and deployed, and the supply chain is automatically executed and monitored in real time through the blockchain platform. The supply chain operations are continuously optimized and adjusted through the feedback mechanism.

[0122] like Figure 4 As shown in the figure, the optimization decision framework is generated based on the previous data analysis and optimization process. The framework integrates the optimal strategy combination under various market change scenarios, aiming to guide the supply chain operations of enterprises in complex market environments. A smart contract refers to a self-executing contract, in which the terms of the contract are directly written into the blockchain in the form of code. When specific conditions are met, the contract will be automatically executed. In the present invention, the smart contract is generated based on the optimization decision framework, and the specific content includes various operations in the supply chain, such as the automatic generation of purchase orders, the scheduling of logistics distribution, and the triggering conditions for inventory replenishment. For example, when market demand reaches a certain threshold, the smart contract will automatically trigger the operation of the production line or issue a new purchase order, thereby ensuring that inventory can be replenished in time to meet market demand.

[0123] The blockchain platform plays a vital role in this process. The distributed ledger technology of the blockchain can ensure the transparency and traceability of all supply chain operations. Every transaction or operation will be recorded on the blockchain, forming an immutable transaction record. This feature enables full monitoring of every link in the supply chain, avoiding misoperation or delays caused by information asymmetry or human intervention in the traditional supply chain. For example, every step of the transportation of drugs from the manufacturer to the terminal sales point will be recorded on the blockchain, and any party can view the transportation status, estimated arrival time and possible reasons for delays, thereby improving the transparency and efficiency of the supply chain.

[0124] The automated execution of the supply chain through the blockchain platform not only improves the efficiency of operations, but also reduces the risks and costs of manual intervention. Traditional supply chain management usually requires the coordination of multiple departments and personnel, but on the blockchain platform, all operations can be automatically executed through smart contracts, greatly simplifying the process. For example, when an abnormality occurs in a certain link in the supply chain (such as delayed delivery or sudden increase in demand), the system can automatically adjust according to the records on the blockchain, triggering the procurement contract of the backup supplier or rescheduling logistics, thereby ensuring the continuity and stability of the entire supply chain.

[0125] Real-time monitoring and feedback mechanisms are another key element in this step. The real-time nature of blockchain technology enables every transaction or operation in the supply chain to be monitored in real time, which provides a powerful dynamic adjustment capability for supply chain management. Through real-time monitoring, companies can keep track of the operating status of the supply chain at any time, and the system can respond immediately when problems or abnormalities are found. For example, if a batch of drugs is found to be delayed in the logistics process, the system can automatically adjust the delivery route or notify relevant departments to take emergency measures to avoid the delay from expanding the impact on the entire supply chain.

[0126] In addition, the feedback mechanism can continuously optimize the strategies and processes of supply chain management through continuous monitoring and data analysis of the results of supply chain operations. The results of each operation will be fed back into the system to form new data inputs, and the smart contract will make adaptive adjustments based on these data. For example, if a particular supplier's delivery time is often delayed, the system will automatically lower the priority of that supplier and select a more reliable supplier in future purchases. This feedback optimization mechanism ensures continuous improvement of supply chain management, allowing it to remain efficient and flexible in a changing market environment.

[0127] like Figure 5 As shown, a system for implementing the data analysis method for drug sales comprises:

[0128] a data collection module, configured to obtain sales data, market data, social media data, climate data, and public health data from multiple data sources, clean and standardize the sales data, market data, social media data, climate data, and public health data, and generate first integrated data through dynamic weight allocation;

[0129] A causal reasoning module, connected to the data acquisition module, is used to perform causal reasoning and quantum causal reasoning based on the first integrated data using distributed computing nodes, to construct a multi-layer causal network, to optimize the multi-layer causal network through an adaptive weight dynamic adjustment mechanism, and to generate final causal network data;

[0130] A collaborative optimization module, connected to the causal reasoning module, is used to perform collaborative optimization based on the final causal network data using a variety of heuristic algorithms and meta-heuristic algorithms, and generate final adaptive optimization data through multi-level parallel feedback loops and adaptive feedback loops;

[0131] The dynamic scenario simulation module is connected to the collaborative optimization module and is used to perform dynamic scenario simulation based on the final adaptive optimization data, generate market change scenarios in combination with mixed augmented reality technology, and optimize and adjust the decision framework through the cross-scenario decision enhancement mechanism to generate an optimized decision framework;

[0132] The smart contract execution module is connected to the dynamic scenario simulation module and is used to generate and deploy smart contracts based on the optimized decision-making framework, realize automated execution and real-time monitoring of the supply chain through the blockchain platform, and continuously optimize and adjust supply chain operations through a feedback mechanism.

[0133] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0134] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A data analysis method for drug sales, characterized in that: The following steps are involved: Obtain sales data, market data, social media data, climate data, and public health data from multiple data sources, clean and standardize them, and generate the first integrated data through dynamic weight assignment; Based on the first integrated data, causal reasoning and quantum causal reasoning are performed using distributed computing nodes to build a multi-layer causal network, which is optimized through an adaptive weight dynamic adjustment mechanism to generate the final causal network data; Based on the final causal network data, a variety of heuristic algorithms and meta-heuristic algorithms are used for collaborative optimization, and the final adaptive optimization data is generated through multi-level parallel feedback loops and adaptive feedback loops; Based on the final adaptive optimization data, dynamic scenario simulation is carried out, and market change scenarios are generated by combining mixed augmented reality technology. The decision framework is optimized and adjusted through the cross-scenario decision enhancement mechanism to generate an optimized decision framework; Based on the optimized decision-making framework, smart contracts are generated and deployed, and the supply chain is automatically executed and monitored in real time through the blockchain platform. The supply chain operations are continuously optimized and adjusted through the feedback mechanism.

2. A data analysis method for drug sales according to claim 1, characterized in that: The data cleaning and normalization steps include the following for sales data, market data, social media data, climate data, and public health data: Noise removal, through frequency domain filtering and time domain smoothing algorithms, eliminates random noise and interference signals in the data; Outlier processing, using anomaly detection algorithms based on statistical analysis to identify and correct outliers in the data; Missing data filling uses interpolation methods or machine learning prediction models to reasonably fill in the missing parts of the data.

3. A data analysis method for drug sales according to claim 1, characterized in that: The causal inference step identifies and quantifies the causal relationships between data by constructing a causal relationship matrix, where the elements C ij Indicates data D i For data D j The causal influence strength, the causal relationship strength C ij Obtained through the following calculation formula: Among them, C ij Indicates data D i For data D j The causal influence strength of W k is the weight coefficient of causal relationship, which is dynamically adjusted based on the correlation between historical data and current data; f k (D i , D j ) represents a function used to calculate causal relationships, which uses different mathematical models according to different data types and causal chain characteristics.

4. A data analysis method for drug sales according to claim 1, characterized in that: The quantum causal reasoning step uses quantum computing algorithms to process high-dimensional causal relationships, where the ultra-parallel computing capabilities of quantum bits are used to quickly process multi-variable causal chains. The calculation results of quantum causal reasoning optimize the causal relationship matrix through quantum state superposition and interference phenomena, thereby generating quantum computing result data.

5. A data analysis method for drug sales according to claim 1, characterized in that: The steps of building a multi-layer causal network include the following: The final causal network data is divided into multiple causal levels, and each causal level is stratified based on the strength of the causal relationship. The strongest causal relationship is divided into the top causal network, and the weaker causal relationships are divided into the middle and bottom causal networks in turn; In each layer of the causal network, the network structure is optimized through the correlation of the causal chain, so that each layer of the causal network can independently reflect the influence of a specific causal level, and finally generate multi-layer causal network data to describe the causal relationship structure at different levels in drug sales data.

6. A data analysis method for drug sales according to claim 1, characterized in that: The adaptive weight dynamic adjustment mechanism optimizes the causal network through the following process: Based on real-time market feedback data, the weights of each causal chain in the multi-layer causal network are dynamically adjusted. The weight adjustment is based on the changing trends, emergencies and abnormal fluctuations in the feedback data. By adaptively adjusting the model, the weight of each causal chain is recalculated to generate the optimized final causal network data.

7. A data analysis method for drug sales according to claim 1, characterized in that: The collaborative optimization steps include the following: Generate preliminary optimization data through the collaborative operation of genetic algorithm, ant colony optimization algorithm and particle swarm optimization algorithm. Genetic algorithm is used to explore the global optimal solution, ant colony optimization algorithm is used to find the local optimal solution of path dependence, and particle swarm optimization algorithm searches for the global optimal solution in multi-dimensional space. The preliminary optimization data are integrated and optimized through multi-level parallel feedback loops. The optimization data generated by different algorithms are cross-validated and weighted through the feedback loop to generate the final adaptive optimization data.

8. A data analysis method for drug sales according to claim 1, characterized in that: Multi-level parallel feedback loops optimize each link of the supply chain by: Each feedback loop operates independently to optimize different parts of the supply chain; Between multi-level feedback loops, the consistency of feedback data in each link is maintained through a key data sharing mechanism, and the optimization strategy of each link is adjusted accordingly based on the feedback results.

9. A data analysis method for drug sales according to claim 1, characterized in that: The dynamic scenario simulation step includes generating market change scenario data in the following ways: Generate preliminary market change scenario data using Monte Carlo simulation, in which possible market changes are predicted through simulation and statistical analysis of a large number of random samples; Combined with mixed augmented reality technology, market changes are further simulated in virtual scenes, and optimized market change scenario data is generated on this basis, ultimately forming a cross-scenario optimization decision-making framework that is used to guide decision-making execution in actual market environments.

10. A system for implementing the data analysis method for drug sales according to any one of claims 1 to 9, characterized in that: include: a data collection module, configured to obtain sales data, market data, social media data, climate data, and public health data from multiple data sources, clean and standardize the sales data, market data, social media data, climate data, and public health data, and generate first integrated data through dynamic weight allocation; A causal reasoning module, connected to the data acquisition module, is used to perform causal reasoning and quantum causal reasoning based on the first integrated data using distributed computing nodes, to construct a multi-layer causal network, to optimize the multi-layer causal network through an adaptive weight dynamic adjustment mechanism, and to generate final causal network data; A collaborative optimization module, connected to the causal reasoning module, is used to perform collaborative optimization based on the final causal network data using a variety of heuristic algorithms and meta-heuristic algorithms, and generate final adaptive optimization data through multi-level parallel feedback loops and adaptive feedback loops; The dynamic scenario simulation module is connected to the collaborative optimization module and is used to perform dynamic scenario simulation based on the final adaptive optimization data, generate market change scenarios in combination with mixed augmented reality technology, and optimize and adjust the decision framework through the cross-scenario decision enhancement mechanism to generate an optimized decision framework; The smart contract execution module is connected to the dynamic scenario simulation module and is used to generate and deploy smart contracts based on the optimized decision-making framework, realize automated execution and real-time monitoring of the supply chain through the blockchain platform, and continuously optimize and adjust supply chain operations through a feedback mechanism.

Images