A method for analyzing drug sales data

Abstract

The application relates to the technical field of sales support, and discloses a medicine sales data analysis method, which comprises the following steps: constructing a cognitive map model; performing clinical path complexity grading; applying a clinical path complexity grading algorithm to analyze clinical decision process data of a medical institution, and calculating a clinical path complexity index; implementing cognitive-clinical path matching; adopting a cognitive-clinical path matching algorithm to process sales representative cognitive state data and clinical path complexity data, and calculating a matching degree score; performing hierarchical progressive information presentation; based on a hierarchical progressive information presentation algorithm, processing the matching degree score and the clinical path data, dividing information into different priority levels, and outputting an optimized information presentation sequence; and providing personalized clinical knowledge navigation; the application significantly improves the accuracy and effectiveness of professional information transmission by intelligently adapting the cognitive ability of a sales representative to the complexity of a clinical path.

Description

Technical Field

[0001] This invention relates to the field of sales support technology, and more specifically, to a method for analyzing pharmaceutical sales data. Background Technology

[0002] Drug sales data analysis is a key capability for pharmaceutical companies to enhance their market competitiveness. Drug sales representatives need to process and convey a large amount of complex clinical pathway information when communicating with medical professionals.

[0003] Conventional technical solutions typically treat sales representative training and customer analysis as two separate processes, failing to establish a dynamic matching mechanism between sales representatives' cognitive state and complex medical expertise. Traditional sales support systems often cannot effectively resolve this contradiction, resulting in inefficient transmission of key medical information and impacting sales performance and customer experience.

[0004] In today's era of information overload in the medical field, pharmaceutical sales representatives face a dual challenge: on the one hand, they need to master and convey increasingly complex medical expertise; on the other hand, they need to adapt to the clinical pathways and procurement decision-making processes of different medical institutions. Summary of the Invention

[0005] This invention provides a method for analyzing drug sales data, which solves the technical problem that conventional solutions in related technologies usually treat sales representative training and customer analysis as two separate processes, failing to establish a dynamic matching mechanism between the sales representative's cognitive state and complex medical expertise.

[0006] This invention provides a method for analyzing drug sales data, comprising the following steps:

[0007] A cognitive graph model is constructed, and a cognitive graph construction algorithm is used to process medical knowledge data and sales representative cognitive feature data to generate a cognitive graph model that includes a knowledge node layer, a relationship network layer, a cognitive mapping layer, and a correlation calculation module.

[0008] Perform clinical pathway complexity classification, apply the clinical pathway complexity classification algorithm to analyze the clinical decision-making process data of medical institutions, and calculate the clinical pathway complexity index.

[0009] Implement cognitive-clinical pathway matching, and use the cognitive-clinical pathway matching algorithm to process sales representative cognitive status data and clinical pathway complexity data, and calculate matching score;

[0010] The system performs hierarchical and progressive information presentation, processes matching scores and clinical pathway data based on a hierarchical and progressive information presentation algorithm, divides information into different priority levels, and outputs an optimized information presentation sequence.

[0011] It provides personalized clinical knowledge navigation, using a personalized clinical knowledge navigation model to analyze sales representatives' cognitive characteristics and historical interaction data, and generates personalized clinical pathway knowledge navigation solutions.

[0012] The cognitive graph construction algorithm integrates human cognitive models with medical knowledge graphs to achieve a two-way mapping between knowledge and cognition.

[0013] Furthermore, the cognitive graph model includes a knowledge node layer, a relationship network layer, a cognitive mapping layer, and a relevance calculation module, wherein:

[0014] The knowledge node layer contains the basic units of medical expertise, with each node having a unique identifier and a set of attributes;

[0015] The relational network layer defines the semantic associations between nodes, including various medical professional relation types such as treatment relations, indication relations, and contraindication relations;

[0016] The cognitive mapping layer establishes the correspondence between knowledge nodes and cognitive load, and assigns cognitive complexity weights to each node;

[0017] The relevance calculation module calculates the relevance strength between different knowledge points based on the semantic distance and co-occurrence frequency between nodes.

[0018] Furthermore, in the step of grading the complexity of the clinical pathway, the clinical pathway complexity index is calculated based on three dimensions: the relevance of medical concepts, the depth of professional terminology, and the novelty of knowledge. Before the calculation, the indices of these three dimensions are normalized and uniformly mapped to the interval [0,1].

[0019] Furthermore, the cognitive-clinical pathway matching step includes:

[0020] Obtain cognitive status data of sales representatives, including their level of professional knowledge, cognitive load capacity, and learning style characteristics;

[0021] Extract clinical pathway complexity data, including complexity indicators and their classification results;

[0022] For each clinical pathway node, its cognitive requirement value is calculated. The cognitive requirement value is determined based on the density of professional terms and the level of conceptual abstraction of the clinical pathway node.

[0023] The cognitive resource value of the sales representative at each clinical pathway node is evaluated. The cognitive resource value is determined based on the sales representative's professional background and cognitive ability assessment results.

[0024] Calculate the cognitive satisfaction level at each clinical pathway node, representing the amount of cognitive needs that the sales representative can meet;

[0025] The cognitive satisfaction of all clinical pathway nodes is summarized and divided by the total cognitive demand value to obtain the final matching score.

[0026] Furthermore, the hierarchical and progressive information presentation steps include:

[0027] Obtain real-time cognitive load data of sales representatives, including indicators such as attention level, working memory capacity, and information processing speed;

[0028] Based on cognitive load status, clinical pathway information is divided into four levels, including the highest priority level, high priority level, medium priority level and low priority level.

[0029] Calculate the importance score for each piece of information based on factors such as its clinical decision-making impact, timeliness and novelty, and relevance to current sales targets;

[0030] Prioritize information at each level by importance score to ensure that key information is presented first;

[0031] Design a time-series pattern for information presentation, control information density and presentation rate, and avoid presenting too much information in a short period of time.

[0032] Furthermore, in the personalized clinical knowledge navigation step, the personalized clinical knowledge navigation model includes:

[0033] The cognitive characteristic modeling module is responsible for building a cognitive characteristic model for sales representatives, including a knowledge structure analysis unit, a learning style identification unit, and a cognitive preference extraction unit.

[0034] The path knowledge optimization module adjusts the knowledge presentation method according to the cognitive characteristic model, and includes a sequence planning unit, a deep control unit, and an association reinforcement unit.

[0035] The continuous learning and adaptation module optimizes navigation performance through feedback and includes an interactive data collection unit, a model update unit, and an accuracy evaluation unit.

[0036] Furthermore, the cognitive characteristic modeling module also includes:

[0037] The knowledge structure analysis unit calculates the degree of matching between the user's knowledge structure and the standard medical knowledge system through a knowledge graph similarity function;

[0038] The learning style recognition unit classifies the learning style of sales representatives using a learning style classification function. This function calculates the probability of the learning style based on the sales representatives' historical learning behavior data and a Bayesian classification model.

[0039] The cognitive preference extraction unit calculates the sales representative's preference for cognitive features using a cognitive preference scoring function based on the sales representative's interaction behavior and behavior weights.

[0040] Furthermore, the path knowledge optimization module also includes:

[0041] The sequence planning unit calculates the optimal presentation sequence for the knowledge point set of the sales representative through a sequence optimization function, which is based on the transfer suitability between knowledge points and the learning suitability of knowledge points for the sales representative.

[0042] The depth control unit calculates the optimal presentation depth of knowledge points for sales representatives using a depth adjustment function, which is based on the maximum presentable depth of knowledge points, the cognitive complexity of knowledge points, and the sales representative's cognitive ability score.

[0043] The association enhancement unit calculates the association display weight between knowledge points for sales representatives through an association weight function. This function is based on the basic relevance strength between knowledge points and the gap in the sales representative's understanding of the relationship between knowledge points.

[0044] Furthermore, the continuous learning adaptation module also includes:

[0045] The interactive data collection unit is responsible for capturing all user interactions with the system, including clicks, browsing duration, search keywords, etc.

[0046] The model update unit calculates the updated values ​​of the model parameters using a parameter update function, which is based on the current model parameters, newly collected interaction data, and the learning rate.

[0047] The accuracy assessment unit calculates the effectiveness score of the navigation scheme for the sales representative's most recent interactions at a specific time using an accuracy assessment function based on a weighting function of time difference and a difference measure function between the predicted and actual results.

[0048] This invention provides a pharmaceutical sales data analysis system for executing the aforementioned pharmaceutical sales data analysis method, comprising:

[0049] The cognitive graph construction module is used to process medical knowledge data and sales representative cognitive feature data using cognitive graph construction algorithms to generate a cognitive graph model that includes a knowledge node layer, a relationship network layer, a cognitive mapping layer, and a correlation calculation module.

[0050] The clinical pathway complexity classification module is used to analyze clinical decision-making process data of medical institutions using the clinical pathway complexity classification algorithm and calculate clinical pathway complexity index.

[0051] The cognitive-clinical pathway matching module is used to process sales representative cognitive status data and clinical pathway complexity data using a cognitive-clinical pathway matching algorithm, and calculate the matching score.

[0052] The hierarchical progressive information presentation module is used to process matching scores and clinical pathway data based on the hierarchical progressive information presentation algorithm, divide the information into different priority levels, and output an optimized information presentation sequence.

[0053] The personalized clinical knowledge navigation module is used to analyze the sales representatives' cognitive characteristics and historical interaction data using a personalized clinical knowledge navigation model to generate personalized clinical pathway knowledge navigation solutions.

[0054] The cognitive graph construction algorithm integrates human cognitive models with medical knowledge graphs to achieve a two-way mapping between knowledge and cognition.

[0055] The beneficial effects of this invention are as follows: by intelligently adapting to the cognitive abilities of sales representatives and the complexity of clinical pathways, it significantly improves the accuracy and effectiveness of professional information delivery, and solves the industry pain point of low efficiency in the delivery of high-value medical professional information during the sales process. Attached Figure Description

[0056] Figure 1 This is a flowchart of the drug sales data analysis method of the present invention;

[0057] Figure 2 This is a bar chart comparing the efficiency of sales representative information transmission in this invention, showing the comparison of the amount of effective information transmitted to doctors by sales representatives with different backgrounds within a standard time before and after the implementation of this method.

[0058] Figure 3 This is a bar chart showing the evaluation of doctors' information delivery by sales representatives before and after the implementation of this method. It compares the doctors' ratings of sales representatives in five dimensions: information accuracy, information completeness, clarity of expression, ability to answer professional questions, and overall evaluation.

[0059] Figure 4 This is a stacked bar chart of information presentation strategies for sales representatives with different cognitive matching levels according to the present invention, which shows the information presentation strategies adopted for sales representative groups with different cognitive matching levels and the proportion of information at each level.

[0060] Figure 5 This invention presents a bar chart showing the changes in the cognitive state of sales representatives and the system response effect, illustrating the system's response effect to different changes in the cognitive state of sales representatives, including evaluations of the system response effect under four states: decreased attention, knowledge comprehension difficulties, information overload, and shift in interest.

[0061] Figure 6This is a line graph showing the continuous learning effect of the personalized knowledge navigation model of the present invention, illustrating the changing trends of four indicators: model accuracy, sales representative satisfaction, knowledge mastery improvement, and sales conversion rate in the initial, intermediate, and mature stages.

[0062] Figure 7 This is a mind map of the cognitive graph model structure and component relationships proposed in the patent application. It shows the overall structure of the cognitive graph model and the relationships between its components, including four main components and their sub-components: the knowledge node layer, the relationship network layer, the cognitive mapping layer, and the relevance calculation module. Detailed Implementation

[0063] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, some features described in the examples may be combined in other examples.

[0064] Example 1

[0065] This embodiment provides a method for analyzing drug sales data, which includes the following steps:

[0066] Step 1: Construct a cognitive graph model

[0067] In this step, a cognitive graph construction algorithm is used to process medical knowledge data and sales representative cognitive characteristic data to generate a cognitive graph model. Specifically, the cognitive graph model can be represented as follows:

[0068]

[0069] in, This represents a set of concept nodes, where each node represents a medical professional knowledge point. This represents a set of relation edges, indicating the relationships between different knowledge points; This represents a cognitive complexity mapping function, used to quantify the cognitive difficulty of each knowledge point; This represents a set of association rules used to describe the association rules between different knowledge points.

[0070] Furthermore, The cognitive complexity mapping function is defined as follows: This function maps knowledge nodes to The complexity value of the interval is calculated based on a weighted combination of factors such as the abstraction level of nodes, the density of technical terms, the number of relationships between concepts, and the steepness of the learning curve. During the calculation process, each factor is normalized, and indicators of different dimensions are uniformly converted to the [0,1] interval to ensure the comparability of each factor; The association rule set contains a series of rules Each rule describes a specific relationship between knowledge points, including the strength of the relationship. This function will store the knowledge points. and Relationship mapping to The relevance strength value of the interval is calculated based on a comprehensive evaluation of factors such as knowledge point co-occurrence frequency, semantic similarity, clinical application relevance, and citation relationships. Similarly, these factors are normalized to ensure that indicators with different characteristics can be weighted in the calculation.

[0071] Furthermore, the cognitive mapping model includes the following components:

[0072] Knowledge Node Layer: Contains the basic units of medical expertise; each node has a unique identifier and a set of attributes.

[0073] Relationship network layer: Defines the semantic relationships between nodes, including various medical professional relationship types such as treatment relationship, indication relationship, and contraindication relationship;

[0074] Cognitive mapping layer: Establishes the correspondence between knowledge nodes and cognitive load, and assigns cognitive complexity weights to each node;

[0075] Relevance calculation module: Calculates the relevance strength between different knowledge points based on the semantic distance and co-occurrence frequency between nodes.

[0076] It should be understood that the core of this step lies in combining traditional knowledge graphs with cognitive science models, which not only represents the structure of knowledge itself, but also takes into account the characteristics and limitations of human cognition, thereby achieving a two-way mapping between knowledge and cognition.

[0077] Step 2: Implement clinical pathway complexity grading

[0078] In this step, a clinical pathway complexity grading algorithm is applied to analyze the clinical decision-making process data of medical institutions to calculate the clinical pathway complexity index. Specifically, the clinical pathway complexity can be calculated using the following formula:

[0079]

[0080] in, It is a path The medical concept association function is used to calculate the association density between medical concepts in a clinical pathway. It is defined as the ratio of the number of interrelated medical concept pairs in a clinical pathway to the total number of possible concept pairs. It is a path The terminology depth function is used to assess the complexity of terminology in clinical pathways. Its calculation is based on the terminology level classification and hierarchical structure. It is a path The knowledge novelty function measures the proportion of emerging medical knowledge contained in a clinical pathway. It is calculated based on the difference between the publication time of the knowledge point and the current time, as well as the frequency of citations in medical literature.

[0081] in, Indicates clinical pathway Complexity metrics; Representing a path The degree of relevance to medical concepts; Representing a path Depth of technical terminology; Representing a path The novelty of knowledge; , and is a weighting coefficient used to adjust the contribution ratio of different factors to complexity.

[0082] Before calculating the clinical pathway complexity index, the input data needs to be preprocessed:

[0083] right , and The three indicators are normalized and mapped to the [0,1] interval to eliminate the influence of dimensional differences on the calculation results;

[0084] For the technical term depth function The calculation requires encoding the categorized data such as the professional level of the terminology (e.g., primary, intermediate, advanced, expert level) and converting it into numerical data (e.g., 1, 2, 3, 4) in order to perform mathematical operations.

[0085] For the knowledge novelty function The time data needs to be standardized to convert the absolute time difference into a relative time value so that it can be weighted and calculated with other indicators.

[0086] It should be noted that this step quantifies the complexity of clinical pathways from multiple dimensions, providing basic data support for subsequent cognitive-clinical pathway matching.

[0087] Step 3: Implement cognitive-clinical pathway matching

[0088] In this step, a cognitive-clinical pathway matching algorithm is used to process the sales representative's cognitive state data and clinical pathway complexity data to generate a matching score. Specifically, the matching score can be calculated using the following formula:

[0089]

[0090] in, It is a minimum value function used to calculate cognitive demand values. and cognitive resource value The smaller of the two values ​​represents the sales representative's position at the clinical pathway node. The system returns the amount of cognitive needs that can actually be met. When the cognitive resource value is less than the cognitive need value, the system returns the cognitive resource value, indicating that the sales representative can only partially meet the cognitive needs of this clinical pathway node; when the cognitive resource value is greater than or equal to the cognitive need value, the system returns the cognitive need value, indicating that the sales representative fully meets the cognitive needs of this clinical pathway node.

[0091] in, Sales Representative With clinical pathways Match score; Representing clinical pathway nodes The cognitive demand value; This indicates that the sales representative is at a clinical pathway node. Cognitive resource value; This indicates that the smaller of the two values ​​is taken, representing the sales representative's position at the clinical pathway node. The cognitive needs that can be met.

[0092] Before performing cognitive-clinical pathway matching calculations, the input data needs to be preprocessed:

[0093] Cognitive demand value and cognitive resource value Normalization is performed to map them to the same numerical range (such as the [0,10] interval) to ensure that they are compared under the same units;

[0094] The categorical features (such as learning style, professional background type, etc.) in the cognitive state data of sales representatives are encoded and converted into quantifiable numerical features.

[0095] Standardize time-related characteristics (such as the sales representative's years of work experience and knowledge update frequency) to eliminate the impact of time unit differences on calculations.

[0096] Furthermore, the cognitive-clinical pathway matching algorithm includes the following steps:

[0097] Step 3.1: Obtain sales representatives' cognitive status data, including their professional knowledge level, cognitive load capacity, and learning style characteristics;

[0098] Step 3.2: Extract clinical pathway complexity data, including the complexity index and its classification results calculated in Step 2 above;

[0099] Step 3.3, for each clinical pathway node Calculate its cognitive needs value The cognitive needs value is determined based on the density of professional terms and the degree of conceptual abstraction of clinical pathway nodes;

[0100] Step 3.4, evaluate the sales representative at each clinical pathway node. Cognitive resource value Cognitive resource value is determined based on the sales representative's professional background and cognitive ability assessment results;

[0101] Step 3.5: Calculate the cognitive satisfaction level at each clinical pathway node. This indicates the amount of cognitive needs that a sales representative can meet;

[0102] Step 3.6: Summarize the cognitive satisfaction of all clinical pathway nodes and divide by the total cognitive needs value to obtain the final matching score. ;

[0103] Step 3.7: Set a threshold based on the matching score. When the score is below the low threshold, trigger the information simplification strategy; when the score is in the middle range, adopt the standard information presentation strategy; when the score is above the high threshold, enable the information enrichment strategy.

[0104] Furthermore, this implementation method can dynamically adjust the information presentation strategy based on the matching score. When the matching score is low, the system will automatically simplify the information presentation to reduce cognitive load; when the matching score is at a medium level, the system will provide standardized professional information; when the matching score is high, the system will provide richer professional information to meet the cognitive needs of sales representatives.

[0105] Step 4: Implement hierarchical and progressive information presentation

[0106] In this step, a hierarchical progressive information presentation algorithm is used to process matching scores and clinical pathway data, outputting an optimized information presentation sequence. Specifically, the hierarchical progressive information presentation algorithm includes the following steps:

[0107] Step 4.1: Obtain real-time cognitive load data of sales representatives, including indicators such as attention level, working memory capacity, and information processing speed;

[0108] Step 4.2: Based on cognitive load status, the clinical pathway information is divided into four levels:

[0109] Level 1 (highest priority): Contains key information such as the basic definition of the disease, main indications, and core medication regimens;

[0110] Tier 2 (High Priority): Contains important information such as treatment mechanism, key clinical evidence, and common adverse reactions;

[0111] Tier 3 (Medium Priority): Contains advanced information such as alternative treatments, drug interactions, and medication use in special populations;

[0112] Level 4 (low priority): Contains professional information such as the latest research progress, rare adverse reactions, and academic controversies;

[0113] Step 4.3, calculate the importance score for each piece of information, based on the following factors:

[0114] The impact of information on clinical decision-making;

[0115] The timeliness and novelty of information;

[0116] The relevance of information to current sales targets;

[0117] Specifically, the information importance score calculation function Defined as:

[0118]

[0119] Information items are assigned to different presentation levels based on their importance scores. (Level assignment function) Defined as:

[0120]

[0121] in, Information Importance score; Information The clinical decision influence function is used to assess the extent to which the information affects medical decisions. Its calculation is based on the recommendation level and evidence level of the information in clinical guidelines. Information The timeliness and novelty function is used to evaluate the time value of information. Its calculation is based on the difference between the information's publication time and the current time, as well as its citation frequency in the latest medical literature. Information The relevance function to the current sales target is used to evaluate the matching degree between information and sales tasks. Its calculation is based on the semantic similarity between the information content and the characteristics of the product being sold. , and is a weighting coefficient used to adjust the contribution ratio of different factors to the information importance score.

[0122] Before calculating the importance score of information, the input data needs to be preprocessed:

[0123] right , and The output values ​​of the three functions are normalized, mapping them uniformly to the [0,1] interval to eliminate dimensional differences. This is especially important for... Time-sensitivity functions require converting time data (such as days, months, and years) into uniform relative time values;

[0124] Encode the categorical data such as recommendation levels and evidence levels in clinical guidelines, and convert text-based classifications (such as A, B, and C level recommendations) into numerical data (such as 3, 2, and 1).

[0125] Vectorization is performed on the text data in semantic similarity calculation, converting unstructured text information into computable numerical vectors.

[0126] Step 4.4: Prioritize the information at each level according to its importance score to ensure that key information is presented first;

[0127] Step 4.5: Design the timing pattern of information presentation, control the information density and presentation rate, and avoid presenting too much information in a short period of time;

[0128] Step 4.6: Collect interaction feedback data from sales representatives, including dwell time, click behavior, and query patterns;

[0129] Step 4.7: Based on the interactive feedback data, dynamically adjust the depth and breadth of information presentation to achieve a personalized information progression strategy.

[0130] It should be understood that the core of this step lies in dynamically adjusting the information presentation strategy based on the sales representative's cognitive state, avoiding cognitive overload, and improving information transmission efficiency.

[0131] Step 5: Provide personalized clinical knowledge navigation

[0132] In this step, a personalized clinical knowledge navigation model is used to analyze the sales representatives' cognitive characteristics data and historical interaction data to generate a personalized clinical pathway knowledge navigation plan.

[0133] Furthermore, the personalized clinical knowledge navigation model includes the following components:

[0134] Cognitive Characteristic Modeling Module: Responsible for building cognitive characteristic models for sales representatives, and includes the following sub-components:

[0135] Knowledge Structure Analysis Unit: This unit assesses the sales representative's existing medical knowledge system and structure through a knowledge graph similarity function. The formula for calculating the degree of matching between a user's knowledge structure and the standard medical knowledge system is as follows:

[0136]

[0137] in, Sales Representative The knowledge set, Represents a set of standard medical knowledge. This represents the size of the intersection of two sets. This represents the size of the union of two sets. Before calculation, the knowledge sets need to be standardized to uniformly encode the same knowledge points in different forms, ensuring the accuracy of set operations.

[0138] The learning style recognition unit identifies the information acquisition and processing preferences of sales representatives. This unit learns a style classification function. For sales representatives The learning styles can be categorized using the following formula:

[0139]

[0140] in, This represents a predefined set of learning style categories (such as visual, auditory, reading, and practical). Sales Representative Historical learning behavior data, Represents given behavioral data Under these conditions, sales representatives belong to the learning style. The learning style probability is calculated using a Bayesian classification model. Before applying the Bayesian classification model, the learning style categories (classification data) need to be encoded, and the behavioral data needs to be feature extracted and normalized.

[0141] Cognitive Preference Extraction Unit: This unit analyzes the cognitive habits exhibited by sales representatives during information interaction. It utilizes a cognitive preference rating function. Calculate sales representative Cognitive characteristics The degree of preference is expressed by the following formula:

[0142]

[0143] in, Indicates the sales representative's first An interactive behavior, It is an indicator function, when the behavior Reflecting cognitive characteristics The value is 1 if it is true, and 0 otherwise. It is behavior The weight of the behavior is related to its time decay and importance. Before calculation, the interaction behavior data needs to be preprocessed, including behavior classification coding and time data standardization, to ensure the comparability of behavior data at different time points;

[0144] Path Knowledge Optimization Module: Adjusts the knowledge presentation method based on the cognitive characteristic model, and includes the following sub-components:

[0145] Sequence Planning Unit: Determines the optimal presentation order of clinical pathway knowledge points. This unit uses a sequence optimization function. Calculation for sales representatives knowledge point collection The optimal presentation sequence is given by the following formula:

[0146]

[0147] in, Represents a set of knowledge points All possible permutations, This represents one of the arrangements. Indicates the first in the permutation One knowledge point, Representing knowledge points and The suitability of the transfer between them Representing knowledge points For sales representatives Learning fitness. Before calculation, the transfer fitness and learning fitness data need to be normalized and mapped to the [0,1] interval to ensure that the product operation is performed under the same units;

[0148] Depth Control Unit: This unit adjusts the depth and complexity of knowledge points based on cognitive ability through a depth adjustment function. Calculation knowledge points For sales representatives The optimal rendering depth is determined by the following formula:

[0149]

[0150] in, Representing knowledge points The maximum renderable depth, Representing knowledge points Cognitive complexity represents the amount of cognitive resources required to understand a particular knowledge point. Sales Representative Cognitive ability score, Indicates to The rounding up operation. Before calculation, the cognitive complexity needs to be considered. and cognitive ability score Standardize the data to ensure that both are calculated within the same numerical range.

[0151] The Reinforcement Unit highlights the connections between knowledge points to promote overall understanding. This unit utilizes a correlation weighting function. Calculation for sales representatives knowledge points and The correlation weights between them are shown by the following formula:

[0152]

[0153] in, Representing knowledge points and The strength of the basic correlation between them Sales Representative Knowledge points and The gap in understanding the relationship between them This is an adjustment coefficient used to control the strength of personalized adjustments. Before calculation, the strength of the underlying correlation needs to be determined. and the gap in understanding Normalization is performed to ensure that both are calculated within the same numerical range;

[0154] Continuous Learning and Adaptation Module: This module continuously optimizes the navigation experience based on feedback and includes the following sub-components:

[0155] Interactive Data Collection Unit: Records data on the interaction between sales representatives and the system. This unit is responsible for capturing all user interactions with the system, including clicks, browsing duration, search keywords, etc.

[0156] Model update unit: This unit dynamically adjusts the parameters of the cognitive characteristic model based on new data. This is achieved through a parameter update function. The updated values ​​of the model parameters are calculated using the following formula:

[0157]

[0158] in, Indicates the current model parameters. This represents newly collected interaction data. Indicates parameters In data loss function on, The model represents the new data. loss function on Relative to parameters The gradient is used to calculate the partial derivatives of the loss function with respect to each parameter, indicating the direction and magnitude of parameter adjustment. The learning rate controls the step size for parameter updates. Before updating the parameters, the newly collected interaction data needs to be processed. Preprocessing includes:

[0159] Standardize time series data to eliminate the impact of periodic fluctuations in data from different time periods;

[0160] Encode categorized interaction behaviors (such as click type, search type, etc.) and convert them into quantifiable numerical features;

[0161] Normalize continuous data (such as browsing time, scrolling distance, etc.) to unify indicators with different dimensions to the same numerical range;

[0162] Accuracy Evaluation Unit: Periodically evaluates the effectiveness of the navigation scheme and provides optimization suggestions. This unit uses an accuracy evaluation function. Computational navigation solutions for sales representatives In time The recent The effectiveness score of each interaction, among which , Indicates the first The interaction, in which For interactive indexing, the range is

[0163] , indicating from a point in time arrive Continuous interaction, The model represents the first The prediction results of this interaction Indicates the first The actual result of this interaction This is a function that measures the difference between predicted and actual results, calculating the distance or error between them. A smaller value indicates a more accurate prediction. This represents a weighting function based on time difference, where This is the current assessment point in time. It is the first The time of each interaction. This represents the time difference; the smaller the time difference (i.e., the closer the interaction time), the greater the weight. Before the evaluation calculation, the interaction behavior data needs to be preprocessed, including:

[0164] Standardize different types of interactive behaviors to make their support scores comparable;

[0165] For time period The behavioral data within the period is time-weighted and normalized to ensure that recent behavior has a greater impact on the evaluation results.

[0166] It should be noted that this step, through a personalized knowledge navigation solution, helps sales representatives more efficiently master and convey complex medical expertise, thereby improving sales performance and customer experience.

[0167] The following is an example of an application of the present invention, such as... Figure 2-7 As shown, the implementation process is as follows:

[0168] First, a cognitive graph model containing knowledge related to anti-tumor drug A was constructed. The graph contains the following main set of nodes:

[0169] Basic medical knowledge: lung cancer classification, gene mutation types, cell signaling pathways, etc.;

[0170] Drug characteristics: mechanism of action, pharmacokinetics, pharmacodynamics, etc.;

[0171] Clinical data: clinical trial design, efficacy indicators, safety data, etc.;

[0172] Medication guidance: dosing regimen, adverse reaction management, drug interactions, etc.;

[0173] Through expert evaluation and literature analysis, a cognitive complexity mapping function was applied to each knowledge node. Quantification is performed to obtain the cognitive complexity value of the nodes. For example, the cognitive complexity of "EGFR gene mutation type in non-small cell lung cancer" is 0.75 (high complexity), while the cognitive complexity of "common adverse drug reactions" is 0.35 (medium complexity).

[0174] Based on the sales process of drug A, five clinical pathways were designed:

[0175] Basic Disease Understanding Pathway: Introducing the classification, epidemiology, and gene mutation characteristics of lung cancer;

[0176] Drug mechanism pathways: elucidating drug molecular targets and mechanisms of action;

[0177] Clinical evidence pathway: showcasing key clinical trial data and real-world research findings;

[0178] Medication management pathway: Explain the dosing regimen, adverse reaction management, and medication use in special populations;

[0179] Medical insurance policy pathway: Introduction to drug reimbursement policies and patient assistance programs;

[0180] The complexity index of each pathway is calculated using a clinical pathway complexity grading algorithm. The calculation formula is:

[0181]

[0182] in, It is a path The medical concept association function is used to calculate the association density between medical concepts in a clinical pathway. It is defined as the ratio of the number of interrelated medical concept pairs in a clinical pathway to the total number of possible concept pairs. It is a path The terminology depth function is used to assess the complexity of terminology in clinical pathways. Its calculation is based on the terminology level classification and hierarchical structure. It is a path The knowledge novelty function measures the proportion of emerging medical knowledge contained in a clinical pathway. It is calculated based on the difference between the publication time of the knowledge point and the current time, as well as the frequency of citations in medical literature.

[0183] in, Indicates clinical pathway Complexity metrics; Representing a path The degree of relevance to medical concepts; Representing a path Depth of technical terminology; Representing a path The novelty of knowledge; , and is a weighting coefficient used to adjust the contribution ratio of different factors to complexity.

[0184] A cognitive ability assessment was conducted on 30 sales representatives to obtain their cognitive resource scores. The assessment includes tests of medical knowledge, information processing skills, and learning styles.

[0185] A cognitive-clinical pathway matching algorithm was applied to calculate the matching score between each sales representative and each clinical pathway. The matching score is calculated using the following formula:

[0186]

[0187] in, It is a minimum value function used to calculate cognitive demand values. and cognitive resource value The smaller of the two values ​​represents the sales representative's position at the clinical pathway node. The system returns the amount of cognitive needs that can actually be met. When the cognitive resource value is less than the cognitive need value, the system returns the cognitive resource value, indicating that the sales representative can only partially meet the cognitive needs of this clinical pathway node; when the cognitive resource value is greater than or equal to the cognitive need value, the system returns the cognitive need value, indicating that the sales representative fully meets the cognitive needs of this clinical pathway node.

[0188] in, Sales Representative With clinical pathways Match score; Representing clinical pathway nodes The cognitive demand value; This indicates that the sales representative is at a clinical pathway node. Cognitive resource value; This indicates that the smaller of the two values ​​is taken, representing the sales representative's position at the clinical pathway node. The cognitive needs that can be met.

[0189] Based on matching scores, sales representatives were grouped, and differentiated information presentation strategies were designed for each group:

[0190] High-match group (match ≥ 0.85): Information enrichment strategy is enabled to provide comprehensive and in-depth professional information, including all information from layer 1 to layer 4, especially the latest research progress and analysis of complex mechanisms;

[0191] Medium matching degree group (0.65≤match degree<0.85): Adopts standard information presentation strategy, provides core professional information from layer 1 to layer 3, and maintains a moderate level of complexity;

[0192] Low matching degree group (match degree <0.65): Triggers information simplification strategy, mainly provides basic professional information of the first and second layers, and simplifies complex concepts by using graphical and analogy methods;

[0193] Apply an importance score calculation function to each piece of information. Rate it:

[0194]

[0195] in, Information Importance score; Information The clinical decision influence function is used to assess the extent to which the information affects medical decisions. Its calculation is based on the recommendation level and evidence level of the information in clinical guidelines. Information The timeliness and novelty function is used to evaluate the time value of information. Its calculation is based on the difference between the information's publication time and the current time, as well as its citation frequency in the latest medical literature. Information The relevance function to the current sales target is used to evaluate the matching degree between information and sales tasks. Its calculation is based on the semantic similarity between the information content and the characteristics of the product being sold. , and is a weighting coefficient used to adjust the contribution ratio of different factors to the information importance score.

[0196] A personalized clinical knowledge navigation model is constructed based on the cognitive characteristics and interaction data of sales representatives. This model includes:

[0197] Cognitive characteristic modeling module:

[0198] The knowledge structure of sales representatives is assessed through a knowledge structure analysis unit, and calculations are performed. ;

[0199] The learning style of the sales representative is determined by the learning style recognition unit, and the calculation is performed. ;

[0200] The cognitive habits of sales representatives are analyzed using a cognitive preference extraction unit, and calculations are performed. ;

[0201] Path knowledge optimization module:

[0202] The optimal presentation order of knowledge points is determined by sequence planning units;

[0203] Adjust the depth and complexity of knowledge points through the depth control unit;

[0204] The connections between knowledge points are highlighted by reinforcing related units;

[0205] Continuous learning and adaptation module:

[0206] The learning behavior of sales representatives is recorded through interactive data collection units;

[0207] The navigation model parameters are dynamically adjusted through the model update unit.

[0208] The effectiveness of the navigation scheme is evaluated using an accuracy evaluation unit.

[0209] It is understood that data preprocessing methods known to those skilled in the art include data cleaning, data transformation, and data reduction. Data transformation includes type conversion and normalization and standardization. Although the dimensions and types of data were omitted in the description of the preceding embodiments, data preprocessing is a technical knowledge known to those skilled in the art and a prerequisite step in data processing. Therefore, the previously described well-known data preprocessing steps were not described independently.

[0210] The embodiments of the present invention have been described above. However, the embodiments are not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make more equivalent embodiments under the guidance of the present embodiments, and all of them are within the protection scope of the present embodiments.

Claims

1. A method for analyzing pharmaceutical sales data, characterized in that, Includes the following steps: A cognitive graph model is constructed, and a cognitive graph construction algorithm is used to process medical knowledge data and sales representative cognitive feature data to generate a cognitive graph model that includes a knowledge node layer, a relationship network layer, a cognitive mapping layer, and a correlation calculation module. The clinical pathway complexity is graded, and the clinical pathway complexity grading algorithm is applied to analyze the clinical decision-making process data of medical institutions to calculate the clinical pathway complexity index. In the clinical pathway complexity grading step, the clinical pathway complexity index is calculated based on three dimensions: medical concept relevance, professional terminology depth, and knowledge novelty. Before the calculation, the indices of these three dimensions are normalized and uniformly mapped to the [0,1] interval. Implement cognitive-clinical pathway matching, and use the cognitive-clinical pathway matching algorithm to process sales representative cognitive status data and clinical pathway complexity data, and calculate matching score; The cognitive-clinical pathway matching steps include: Acquire data on the cognitive status of sales representatives, including their level of professional knowledge, cognitive load capacity, and learning style characteristics; Extract clinical pathway complexity data, including complexity indicators and their classification results; For each clinical pathway node, its cognitive demand value is calculated. The cognitive demand value is determined based on the density of professional terms and the level of conceptual abstraction of the clinical pathway node. The cognitive resource value of the sales representative at each clinical pathway node is evaluated. The cognitive resource value is determined based on the sales representative's professional background and cognitive ability assessment results. Calculate the cognitive satisfaction level at each clinical pathway node, representing the amount of cognitive needs that the sales representative can meet; The cognitive satisfaction of all clinical pathway nodes is summarized and divided by the total cognitive demand value to obtain the final matching score. The system performs hierarchical and progressive information presentation, processes matching scores and clinical pathway data based on a hierarchical and progressive information presentation algorithm, divides information into different priority levels, and outputs an optimized information presentation sequence. Personalized clinical knowledge navigation is provided by analyzing sales representatives' cognitive characteristics and historical interaction data using a personalized clinical knowledge navigation model to generate personalized clinical pathway knowledge navigation solutions. The personalized clinical knowledge navigation model includes the following steps: The cognitive characteristic modeling module is responsible for building a cognitive characteristic model for sales representatives, including a knowledge structure analysis unit, a learning style identification unit, and a cognitive preference extraction unit. The path knowledge optimization module adjusts the knowledge presentation method according to the cognitive characteristic model, and includes a sequence planning unit, a deep control unit, and an association reinforcement unit. The continuous learning and adaptation module continuously optimizes the navigation effect through feedback, and includes an interactive data collection unit, a model update unit, and an accuracy evaluation unit; the continuous learning and adaptation module includes: The interaction data collection unit is responsible for capturing all user interactions with the system, including clicks, browsing duration, and search keywords; The model update unit calculates the updated values ​​of the model parameters using a parameter update function, which is based on the current model parameters, newly collected interaction data, and the learning rate. The accuracy assessment unit calculates the effectiveness score of the navigation scheme for the sales representative's most recent interactions at a specific time using an accuracy assessment function. This function is based on a weighting function of time difference and a difference measurement function between the predicted and actual results. The cognitive graph construction algorithm integrates human cognitive models with medical knowledge graphs to achieve a two-way mapping between knowledge and cognition.

2. The method for analyzing drug sales data according to claim 1, characterized in that, The cognitive graph model includes a knowledge node layer, a relationship network layer, a cognitive mapping layer, and a relevance calculation module, wherein: The knowledge node layer contains the basic units of medical expertise, with each node having a unique identifier and a set of attributes; The relational network layer defines the semantic relationships between nodes, including treatment relationships, indication relationships, and contraindication relationships; The cognitive mapping layer establishes the correspondence between knowledge nodes and cognitive load, and assigns cognitive complexity weights to each node; The relevance calculation module calculates the relevance strength between different knowledge points based on the semantic distance and co-occurrence frequency between nodes.

3. The method for analyzing drug sales data according to claim 1, characterized in that, The hierarchical and progressive information presentation steps include: Acquire real-time cognitive load data of sales representatives, including attention level, working memory capacity, and information processing speed indicators; Based on cognitive load status, clinical pathway information is divided into four levels, including the highest priority level, high priority level, medium priority level and low priority level. Calculate the importance score for each piece of information based on factors such as its clinical decision-making impact, timeliness and novelty, and relevance to current sales targets; Prioritize information at each level by importance score to ensure that key information is presented first; Design a time-series pattern for information presentation, control information density and presentation rate, and avoid presenting too much information in a short period of time.

4. The method for analyzing drug sales data according to claim 1, characterized in that, In the cognitive characteristic modeling module: The knowledge structure analysis unit calculates the degree of matching between the user's knowledge structure and the standard medical knowledge system through a knowledge graph similarity function; The learning style recognition unit classifies the learning style of sales representatives using a learning style classification function. This function calculates the probability of the learning style based on the sales representatives' historical learning behavior data and a Bayesian classification model. The cognitive preference extraction unit calculates the sales representative's preference for cognitive features using a cognitive preference scoring function based on the sales representative's interaction behavior and behavior weights.

5. The method for analyzing drug sales data according to claim 1, characterized in that, In the path knowledge optimization module: The sequence planning unit calculates the optimal presentation sequence for the knowledge point set of the sales representative through a sequence optimization function, which is based on the transfer suitability between knowledge points and the learning suitability of knowledge points for the sales representative. The depth control unit calculates the optimal presentation depth of knowledge points for sales representatives using a depth adjustment function, which is based on the maximum presentable depth of knowledge points, the cognitive complexity of knowledge points, and the sales representative's cognitive ability score. The association enhancement unit calculates the association display weight between knowledge points for sales representatives through an association weight function. This function is based on the basic relevance strength between knowledge points and the gap in the sales representative's understanding of the relationship between knowledge points.

6. A pharmaceutical sales data analysis system, characterized in that, A method for performing drug sales data analysis as described in any one of claims 1-5, comprising: The cognitive graph construction module is used to process medical knowledge data and sales representative cognitive feature data using cognitive graph construction algorithms to generate a cognitive graph model that includes a knowledge node layer, a relationship network layer, a cognitive mapping layer, and a correlation calculation module. The clinical pathway complexity classification module is used to analyze clinical decision-making process data of medical institutions using the clinical pathway complexity classification algorithm and calculate clinical pathway complexity index. The cognitive-clinical pathway matching module is used to process sales representative cognitive status data and clinical pathway complexity data using a cognitive-clinical pathway matching algorithm, and calculate the matching score. The hierarchical progressive information presentation module is used to process matching scores and clinical pathway data based on the hierarchical progressive information presentation algorithm, divide the information into different priority levels, and output an optimized information presentation sequence. The personalized clinical knowledge navigation module is used to analyze the sales representatives' cognitive characteristics and historical interaction data using a personalized clinical knowledge navigation model to generate personalized clinical pathway knowledge navigation solutions. The cognitive graph construction algorithm integrates human cognitive models with medical knowledge graphs to achieve a two-way mapping between knowledge and cognition.

Images