Medical product service intelligent matching recommendation method and platform and storage medium

Abstract

The invention relates to the technical field of data processing, in particular to a medical product service intelligent matching recommendation method and platform and a storage medium, and the method comprises the steps: collecting user data and task release data; storing the task release data, analyzing a task difficulty coefficient, and analyzing the task complexity fitness of the user according to the task difficulty coefficient and the user data; constructing a user feature vector based on the user data and the task complexity fitness, and constructing a task feature vector based on the task release data and the task difficulty coefficient; constructing a task matching model based on the user feature vector and the task feature vector to obtain a demand integrating degree and a capability adaptation degree; formulating a recommendation list according to the demand integrating degree and pushing the recommendation list to the user, and formulating a candidate list according to the capability adaptability and pushing the candidate list to the task source unit; and collecting interaction data after the user receives the recommendation list, and optimizing the task matching model. According to the invention, accurate recommendation of the task data between the user and the hospital is realized.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, platform and storage medium for intelligent matching and recommendation of pharmaceutical product services. Background Art

[0002] In the digital age, task publishing platforms are constantly emerging. In digital service and collaboration scenarios, two-way matching between task publishing and users is crucial. With the development of technology, with the help of big data analysis of user skills and preferences, combined with the characteristics of task requirements, and the use of algorithm models, efficient and accurate matching can be achieved, thereby improving resource utilization and service delivery quality.

[0003] Existing medical service matching systems mostly make recommendations based on static rules or single indicators, and lack quantitative analysis of task complexity and dynamic adaptability of user skills. In addition, traditional models do not fully combine the urgency of the published tasks with the adaptability of user capabilities, resulting in a large deviation between the recommendation results and actual needs, making it difficult to meet the needs of efficient and accurate resource scheduling; existing technologies often have the following problems: the task difficulty coefficient and dynamic analysis of user complexity adaptability are not introduced, resulting in a lack of objectivity and flexibility in the recommendation results; only focusing on user needs or hospital needs, failing to achieve two-way matching; lacking dynamic optimization of the model based on user interaction data, the system cannot continuously improve the accuracy of recommendations; the user's certified skills are not structuredly associated with the necessary skills for the task, resulting in the key skill requirements not being effectively met; the weight is not dynamically adjusted according to the urgency of the task, and the recommendation priority of high-urgency tasks is insufficient. Summary of the Invention

[0004] The object of the present invention is to provide a method, platform and storage medium for intelligent matching and recommendation of pharmaceutical product services to solve at least one of the problems existing in the prior art.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for intelligent matching and recommendation of pharmaceutical product services, comprising:

[0007] Collect user data and task release data;

[0008] Store task release data and analyze task difficulty coefficients, and analyze users' task complexity adaptability based on task difficulty coefficients and user data;

[0009] Build user feature vectors based on user data and task complexity fitness, and build task feature vectors based on task release data and task difficulty coefficient;

[0010] Build a task matching model based on user feature vectors and task feature vectors to obtain demand fit and ability adaptability;

[0011] A recommendation list is created based on the degree of fit with the needs and pushed to the user; a candidate list is created based on the degree of fit with the capabilities and pushed to the task source unit;

[0012] Collect interaction data after users receive the recommendation list and optimize the task matching model.

[0013] Furthermore, the task completion status data corresponding to the task release data with the same task content is extracted as the difficulty analysis data, the task time in the difficulty analysis data is normalized, the number of difficulty analysis data with a successful task status is counted as the number of successful data, and the number of difficulty analysis data with a failed task status is counted as the number of failed data, and the task difficulty coefficient is analyzed based on the difficulty analysis data, the number of successful data and the number of failed data. The expression of the task difficulty coefficient is TD=0.6×μ(DT1)+0.4×N2 / (N1+N2), where TD represents the task difficulty coefficient, DT1 represents the normalized task time, μ() represents the average value of the data in the brackets, N1 represents the number of successful data, and N2 represents the number of failed data;

[0014] The user's task complexity adaptability is analyzed based on the task difficulty coefficient and completed tasks, and the average value of the task difficulty coefficients corresponding to the user's completed tasks is used as the task complexity adaptability.

[0015] Furthermore, different certification skills and required skills are extracted as a skill set, and each skill in the skill set is numbered to obtain a skill number, so that each skill corresponds to a unique skill number. The skill numbers are sorted in ascending order to obtain a skill sequence, and a multivariate vector is constructed as a skill vector. The number of elements in the skill vector is the same as the number of skills in the skill sequence and corresponds one to one.

[0016] Furthermore, a user feature vector is constructed based on the user's task completion, response speed, task complexity adaptability, user score, and skill vector. The user's response speed is calculated and normalized to obtain a response speed parameter. A four-element vector is constructed, in which each element corresponds to the user's task completion, response speed parameter, task complexity adaptability, and user score. The skill vector is filled in based on the user's certified skills, and the filled skill vector is used as the user skill vector. The four-element vector and the filled skill vector are concatenated to obtain the user feature vector.

[0017] A task feature vector is constructed based on the degree of urgency, task difficulty coefficient, required skills and skill vector. A ternary vector is constructed, in which each element corresponds to the degree of urgency, task difficulty coefficient and preset matching parameters. The skill vector is filled in according to the required skills, and the filled skill vector is used as the task skill vector. The ternary vector and the filled skill vector are spliced ​​together to obtain the task feature vector.

[0018] Further, skill matching parameters are analyzed based on the skill vectors in the user feature vector and the task feature vector;

[0019] The demand fit is analyzed based on the user feature vector, task feature vector and skill matching parameters. The expression of the demand fit is: Where MAB represents demand fit, w1 represents the first fit weight, w2 represents the second fit weight, w3 represents the third fit weight, w4 represents the fourth fit weight, CS represents the skill matching parameter, F represents the task complexity adaptability, E represents the urgency, T represents the response speed parameter, and MS represents the preset matching parameter;

[0020] The ability adaptability is analyzed based on the user feature vector, task feature vector, and skill matching parameters. The expression of the ability adaptability is: Where MBA represents ability fit, α1 represents the first fit weight, α2 represents the second fit weight, α3 represents the third fit weight, α4 represents the fourth fit weight, CR represents task completion, and UR represents user rating.

[0021] Furthermore, the analysis process of capability adaptability is updated according to the degree of urgency and the task difficulty coefficient. When the degree of urgency is greater than or equal to the first update threshold, the first adaptability weight is increased and the third adaptability weight is decreased to update the analysis process of capability adaptability; when the task difficulty coefficient is greater than or equal to the second update threshold, the first adaptability weight and the second adaptability weight are decreased and the third adaptability weight is increased to update the analysis process of capability adaptability.

[0022] Furthermore, a recommendation list is prepared based on the demand fit, and the tasks corresponding to the demand fit greater than or equal to the fit threshold are extracted as candidate tasks. These tasks are then sorted in descending order of demand fit, and the top Y candidate tasks are taken as the recommendation list, where Y represents the number of recommendations.

[0023] A recommendation list is made based on the ability adaptability. Users whose task completion is greater than or equal to the completion threshold and whose user rating is greater than or equal to the rating threshold are extracted as candidate users. They are then sorted in descending order according to their ability adaptability, and the top Y candidate users are taken as the candidate list.

[0024] Furthermore, the construction process of the task feature vector is optimized based on the interaction data, recommendation list and candidate list to optimize the task matching model, the number of users who accept the recommendation in the interaction data corresponding to the same task content is extracted as the number of accepted users, the number of users who ignore the recommendation in the interaction data corresponding to the same task content is extracted as the number of ignored users, and the number of users in the candidate list corresponding to the same task content whose interaction data accepts the recommendation is extracted as the number of valid users. The recommendation matching parameters are analyzed based on the number of accepted users, the number of ignored users and the number of valid users, and when the recommendation matching parameter is less than 1, the value of the preset matching parameter is updated to be equal to the recommendation matching parameter.

[0025] In another aspect, the present invention further provides a pharmaceutical product service intelligent matching and recommendation platform, comprising:

[0026] Data collection module, collects user data and task release data;

[0027] Parameter analysis module, which stores task release data and analyzes task difficulty coefficients, and analyzes the user's task complexity adaptability based on the task difficulty coefficients and user data;

[0028] Vector construction module, which constructs user feature vectors based on user data and task complexity fitness, and constructs task feature vectors based on task release data and task difficulty coefficient;

[0029] Model building module, which builds a task matching model based on user feature vectors and task feature vectors to obtain demand fit and capability adaptation;

[0030] The recommendation analysis module creates a recommendation list based on the degree of fit between requirements and pushes it to users, and creates a candidate list based on the degree of fit between capabilities and pushes it to the task source unit;

[0031] The feedback optimization module collects interaction data after users receive the recommendation list and optimizes the task matching model.

[0032] On the other hand, the present invention also provides a storage medium, characterized in that it stores instructions that, when executed on a computer, enable the computer to execute any of the above-described methods for intelligent matching and recommendation of pharmaceutical product services.

[0033] The beneficial effects of the present invention are as follows: through multidimensional data modeling and dynamic optimization mechanism, efficient two-way matching between users and published tasks is achieved, combining task difficulty, user adaptability and skill matching to ensure that recommendation results are highly consistent with needs, and based on real-time optimization model of interactive data, adapting to task urgency and changes in user capabilities, shortening decision-making time and improving resource allocation efficiency through threshold screening and priority sorting. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 This is a flow chart of the intelligent matching and recommendation method for pharmaceutical product services in this embodiment.

[0036] Figure 2 Flowchart of the method for constructing a feature vector in this embodiment.

[0037] Figure 3 Flowchart of the method for constructing the task matching model in this embodiment.

[0038] Figure 4 This is a structural diagram of the medical product service intelligent matching and recommendation platform of this embodiment. DETAILED DESCRIPTION

[0039] In order to more clearly illustrate the present invention, the present invention is further described below in conjunction with preferred embodiments and accompanying drawings. Similar components in the accompanying drawings are represented by the same reference numerals. It should be understood by those skilled in the art that the following detailed description is illustrative rather than restrictive and should not be used to limit the scope of protection of the present invention.

[0040] It should be noted that, although the terms "first," "second," and "third" may be used to describe the embodiments of the present application, the description should not be limited to these terms. These terms are merely used to distinguish the descriptions. For example, without departing from the scope of the embodiments of the present application, "first" may also be referred to as "second," and similarly, "second" may also be referred to as "first."

[0041] See also Figure 1 As shown, this is the intelligent matching and recommendation method for pharmaceutical product services of this embodiment, including:

[0042] Step S1, collect user data and task release data, the user data includes task completion, response speed, certification skills, completed tasks and user ratings, the task completion is the ratio of the number of tasks completed by the user to the number of tasks accepted by the user, the response speed is the average time from task release to user acceptance, and its unit is hour, the certification skills are the skills and skill levels that the user has passed the certification, including but not limited to emergency care, medical record entry, etc., the skill levels include elementary, intermediate and advanced, the user rating is the average of the hospital's rating of the user's completion quality after the user completes the task, and the user rating range is [1,5], the task release data is the task data released by the hospital, including any The task content, required skills, urgency and task completion status data are as follows: the task release data is the data obtained after the task source unit releases the task on the platform, which can be task data released by multiple hospitals. In this embodiment, only the released task data are analyzed. The required skills for the task match the certified skills. The urgency is the urgency of the task set by the hospital, which is divided into levels 1-5. Level 1 represents a routine task and level 5 represents an urgent task. The task completion status data includes task time and task status. The task status includes success and failure. The user data and task release data are collected by importing them from the database of the medical product service intelligent matching recommendation platform.

[0043] It should be noted that the object information (including but not limited to object device information, object personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the object or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant national and regional laws, regulations and standards.

[0044] Specifically, in step S1 of this embodiment, user data and task release data are collected to ensure the comprehensiveness and real-time nature of the data source, provide basic support for subsequent analysis, and avoid recommendation deviations caused by data missing.

[0045] Please continue reading Figure 1 As shown, the pharmaceutical product service intelligent matching recommendation method further includes:

[0046] Step S2: store the task release data and analyze the task difficulty coefficient, and analyze the user's task complexity adaptability based on the task difficulty coefficient and user data.

[0047] Specifically, in step S2 of this embodiment, the task completion data corresponding to the task release data with the same task content is extracted as the difficulty analysis data, and the task time in the difficulty analysis data is normalized. The normalization formula used is: X = (xx min ) / (xmax -x min ), where X represents the data after normalization, x represents the data before normalization, and x min Represents the minimum value of the data before normalization, x max It represents the maximum value in the data before normalization, counts the number of difficulty analysis data with successful task status as the number of successful data, counts the number of difficulty analysis data with failed task status as the number of failed data, and analyzes the task difficulty coefficient based on the difficulty analysis data, the number of successful data and the number of failed data. The expression of the task difficulty coefficient is TD=0.6×μ(DT1)+0.4×N2 / (N1+N2), where D represents the task difficulty coefficient, DT1 represents the normalized task time, μ() represents the average value of the data in the brackets, N1 represents the number of successful data, and N2 represents the number of failed data.

[0048] Specifically, in step S2 of this embodiment, the user's task complexity adaptability is analyzed based on the task difficulty coefficient and completed tasks, and the average value of the task difficulty coefficients corresponding to the user's completed tasks is used as the task complexity adaptability.

[0049] Specifically, in step S2 of this embodiment, the objectivity of the recommendation is enhanced and subjective judgment errors are avoided by quantifying the task difficulty and user ability.

[0050] Please continue reading Figure 1 As shown, the pharmaceutical product service intelligent matching recommendation method further includes:

[0051] Step S3: construct a user feature vector based on user data and task complexity fitness, and construct a task feature vector based on task release data and task difficulty coefficient.

[0052] See also Figure 2 As shown in FIG, it is a method for constructing a feature vector, including:

[0053] Step S31: construct a skill vector based on the certified skills and the required skills.

[0054] Specifically, in step S31 described in this embodiment, different certification skills and required skills are extracted as a skill set, and each skill in the skill set is numbered to obtain a skill number, so that each skill corresponds to a unique skill number, and the skill numbers are sorted in ascending order to obtain a skill sequence, and a multivariate vector is constructed as a skill vector. The number of elements in the skill vector is the same as the number of skills in the skill sequence and corresponds one to one.

[0055] Specifically, the skill vector described in this embodiment is a multi-element vector, and the skill sequence is {first aid nursing, medical record entry, venous blood collection}, then the skill vector is a ternary vector, denoted as S, S = [s1, s2, s3], s1, s2, s3 respectively correspond to the order of first aid nursing, medical record entry, and venous blood collection in the skill sequence.

[0056] Please continue reading Figure 2 As shown, the method for constructing the feature vector further includes:

[0057] Step S32: constructing a user feature vector based on the skill vector, user data and task complexity fitness.

[0058] Specifically, in step S32 of this embodiment, a user feature vector is constructed based on the user's task completion, response speed, task complexity adaptability, user score, and skill vector. The user's response speed is obtained and normalized to obtain a response speed parameter. A four-element vector is constructed, and each element in the four-element vector corresponds to the user's task completion, response speed parameter, task complexity adaptability, and user score. The skill vector is filled in according to the user's certified skills, and the filled skill vector is used as the user skill vector. The user skill vector is set to AS, AS = [as1, as2, ..., as n ], set the value of the element in the skill vector that does not correspond to the user's certified skills to 0, and set the value of the element in the skill vector that corresponds to the user's certified skills according to the skill level. If the skill level is elementary, it is set to 1; if the skill level is intermediate, it is set to 2; if the skill level is advanced, it is set to 3; concatenate the quaternion vector with the filled skill vector to obtain the user feature vector, and set the user feature vector to A, A = [CR, T, F, UR, AS], where CR represents task completion, T represents response speed parameter, F represents task complexity adaptability, and UR represents user score, [as1, as2, ..., as n ] represents the value corresponding to each element of the user skill vector, and the value of the nth element in the user skill vector is set to as n .

[0059] Please continue reading Figure 2 As shown, the method for constructing the feature vector further includes:

[0060] Step S33: constructing a task feature vector based on the skill vector, task release data, and task difficulty coefficient.

[0061] Specifically, in step S33 of this embodiment, a task feature vector is constructed based on the urgency, task difficulty coefficient, required skills, and skill vector, and a ternary vector is constructed. Each element in the ternary vector corresponds to the urgency, task difficulty coefficient, and preset matching parameters. The skill vector is filled in based on the required skills, and the filled skill vector is used as the task skill vector. The task skill vector is set to BS, BS = [bs1, bs2, ..., bs n ], set the value of the element corresponding to the required skill in the skill vector to 1, set the value of the element corresponding to the non-required skill in the skill vector to 0, concatenate the ternary vector with the padded skill vector to obtain the task feature vector, and set the task feature vector to B, B = [E, TD, MS, BS], where E represents the urgency, TD represents the task difficulty coefficient, and MS represents the preset matching parameters, [bs1, bs2, ..., bs n ] represents the value corresponding to each element in the task skill vector, and the value of the nth element in the task skill vector is set to bs n .

[0062] Specifically, in this embodiment, the preset matching parameter is set to 1. In this embodiment, no specific limitation is imposed on the value of the preset matching parameter, and those skilled in the art can freely set it. The setting of the preset matching parameter should satisfy the range [0.8, 1].

[0063] Specifically, in step S3 of this embodiment, the multidimensional data is structured to improve the interpretability and matching efficiency of the model.

[0064] Please continue reading Figure 1 As shown, the pharmaceutical product service intelligent matching recommendation method further includes:

[0065] Step S4: construct a task matching model based on the user feature vector and the task feature vector to obtain the demand fit and ability adaptability.

[0066] See also Figure 3 As shown in FIG, a method for constructing a task matching model includes:

[0067] Step S41: Analyze skill matching parameters based on the user feature vector and the task feature vector.

[0068] Specifically, in step S41 of this embodiment, the skill matching parameter is analyzed based on the skill vector in the user feature vector and the task feature vector. The expression of the skill matching parameter is: Where CS represents the skill matching parameter and N represents the number of elements in the skill vector.

[0069] Specifically, in this embodiment, if the required skills are not set in the task release data, the skill matching parameter is set to 1.

[0070] Please continue reading Figure 3 As shown, the method for constructing the task matching model further includes:

[0071] Step S42: Analyze the demand compatibility based on the user feature vector, the task feature vector, and the skill matching parameters.

[0072] Specifically, in step S42 of this embodiment, the demand fit is analyzed based on the user feature vector, the task feature vector, and the skill matching parameter. The expression of the demand fit is: In the formula, MAB represents demand fit, w1 represents the first fit weight, w2 represents the second fit weight, w3 represents the third fit weight, w4 represents the fourth fit weight, w1+w2+w3+w4=1, and ε represents the smoothing factor.

[0073] Specifically, in this embodiment, the first fit weight is set to 0.4, the second fit weight is set to 0.3, the third fit weight is set to 0.2, the fourth fit weight is set to 0.1, and the smoothing factor is set to 0.1. In this embodiment, there is no specific limitation on the values ​​of each fit weight and smoothing factor, and those skilled in the art can set them freely. The setting of the smoothing factor is to prevent the denominator from being 0, and the setting of the smoothing factor should satisfy the range [0.1, 0.2].

[0074] Please continue reading Figure 3 As shown, the method for constructing the task matching model further includes:

[0075] Step S43: analyzing the ability adaptability based on the user feature vector, the task feature vector, and the skill matching parameters.

[0076] Specifically, in step S43 of this embodiment, the ability adaptability is analyzed based on the user feature vector, the task feature vector, and the skill matching parameter. The expression of the ability adaptability is: Where MBA represents ability adaptability, α1 represents the first adaptability weight, α2 represents the second adaptability weight, α3 represents the third adaptability weight, α4 represents the fourth adaptability weight, and α1+α2+α3+α4=1.

[0077] Specifically, in this embodiment, the first fitness weight is set to 0.3, the second fitness weight is set to 0.3, the third fitness weight is set to 0.3, and the fourth fitness weight is set to 0.1. This embodiment does not specifically limit the setting of each fitness weight, and those skilled in the art can freely set it.

[0078] Please continue reading Figure 3As shown, the method for constructing the task matching model further includes:

[0079] Step S44, updating the analysis process of capability adaptability based on the urgency and task difficulty coefficient.

[0080] Specifically, in step S44 described in this embodiment, the analysis process of updating the capability adaptability is carried out according to the urgency and the task difficulty coefficient. If the urgency is greater than or equal to the first update threshold, the first adaptability weight is increased and the third adaptability weight is reduced to update the capability adaptability analysis process. The increase in the first adaptability weight is equal to the decrease in the third adaptability weight, which is L1. If the urgency is less than the first update threshold, the analysis process of the capability adaptability is not updated. If the task difficulty coefficient is greater than or equal to the second update threshold, the first adaptability weight and the second adaptability weight are reduced and the third adaptability weight is increased to update the capability adaptability analysis process. The decrease in the first adaptability weight and the second adaptability weight is equal and the sum of the decrease in the first adaptability weight and the second adaptability weight is equal to the increase in the third adaptability weight, which is L2. If the task difficulty coefficient is less than the second update threshold, L1 represents the first change parameter, 0.1≤L1≤0.2, and L2 represents the second change parameter, 0.05≤L1≤0.1.

[0081] Specifically, in this embodiment, the first update threshold is set to 4, and the second update threshold is set to 0.8. In this embodiment, there is no specific limitation on the values ​​of the update threshold and the change parameter, and those skilled in the art can set them freely. The setting of the second update threshold should satisfy: second update threshold = first update threshold × 0.2.

[0082] Specifically, in step S4 of this embodiment, by constructing a task matching model and calculating the demand fit and capability adaptability, and combining linear weights with a dynamic parameter adjustment mechanism, accurate personalized recommendations are achieved to meet the two-way needs of users and hospitals.

[0083] Please continue reading Figure 1 As shown, the pharmaceutical product service intelligent matching recommendation method further includes:

[0084] Step S5: A recommendation list is prepared based on the degree of fit with the requirements and pushed to the user; a candidate list is prepared based on the degree of fit with the capabilities and pushed to the task source unit.

[0085] Specifically, in step S5 described in this embodiment, a recommendation list is formulated based on the demand fit, and task contents corresponding to demand fit greater than or equal to the fit threshold are extracted as candidate tasks, and sorted in descending order of demand fit, and the first Y candidate tasks are taken as the recommendation list, where Y represents the recommendation quantity parameter.

[0086] Specifically, in this embodiment, the recommended quantity parameter is set to 5, and the fit threshold is set to 0.7. In this embodiment, there is no specific limitation on the values ​​of the recommended quantity parameter and the fit threshold, and those skilled in the art can set them freely. The setting of the recommended quantity parameter should satisfy [3,5], and the setting of the fit threshold should satisfy [0.6,0.9].

[0087] Specifically, in step S5 described in this embodiment, a recommendation list is formulated based on the ability adaptability, and users whose task completion is greater than or equal to the completion threshold and whose user scores are greater than or equal to the score threshold are extracted as candidate users. The users are sorted in descending order according to the ability adaptability, and the first Y candidate users are taken as the candidate list.

[0088] Specifically, in this embodiment, the completion threshold is set to 0.8, and the score threshold is set to 4. In this embodiment, there is no specific limitation on the setting of the completion threshold and the score threshold, and those skilled in the art can set them freely. The setting of the completion threshold should satisfy [0.8, 0.9], and the setting of the score threshold should satisfy [3.5, 4].

[0089] Specifically, in step S5 of this embodiment, threshold screening and sorting optimization are performed to ensure the priority and practicality of the recommendation results, thereby improving the decision-making efficiency of users and hospitals.

[0090] Please continue reading Figure 1 As shown, the pharmaceutical product service intelligent matching recommendation method further includes:

[0091] Step S6: Collect interaction data after the user receives the recommendation list and optimize the task matching model. The interaction data includes accepting recommendations and ignoring recommendations.

[0092] Specifically, in step S6 described in this embodiment, the construction process of the task feature vector is optimized based on the interaction data, the recommendation list and the candidate list to optimize the task matching model, and the number of users who accept the recommendation in the interaction data corresponding to the same task content is extracted as the number of accepted users, the number of users who ignore the recommendation in the interaction data corresponding to the same task content is extracted as the number of ignored users, and the number of users in the candidate list corresponding to the same task content whose interaction data accepts the recommendation is extracted as the number of valid users. The recommendation matching parameter is analyzed based on the number of accepted users, the number of ignored users and the number of valid users. The expression of the recommendation matching parameter is: Q = (n1+n3) / (n1+n2), where Q represents the recommendation matching parameter, n1 represents the number of accepted users, n2 represents the number of ignored users, and n3 represents the number of valid users. The preset matching parameter is updated based on the recommended matching parameter. If the recommended matching parameter is greater than or equal to 1, the preset matching parameter is not updated. Otherwise, the value of the preset matching parameter is updated to be equal to the recommended matching parameter.

[0093] Specifically, in step S6 of this embodiment, a feedback mechanism is introduced to dynamically adjust the model parameters to enhance the adaptive capability of the system and continuously improve the recommendation accuracy.

[0094] See also Figure 4 As shown, it is the pharmaceutical product service intelligent matching recommendation platform of this embodiment, including:

[0095] Data collection module, collects user data and task release data;

[0096] Parameter analysis module, which stores task release data and analyzes task difficulty coefficients, and analyzes the user's task complexity adaptability based on the task difficulty coefficients and user data;

[0097] Vector construction module, which constructs user feature vectors based on user data and task complexity fitness, and constructs task feature vectors based on task release data and task difficulty coefficient;

[0098] Model building module, which builds a task matching model based on user feature vectors and task feature vectors to obtain demand fit and capability adaptation;

[0099] The recommendation analysis module creates a recommendation list based on the degree of fit between requirements and pushes it to users, and creates a candidate list based on the degree of fit between capabilities and pushes it to the task source unit;

[0100] The feedback optimization module collects interaction data after users receive the recommendation list and optimizes the task matching model.

[0101] An embodiment of the present application also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enables the computer to execute the pharmaceutical product service intelligent matching recommendation method as described in the above method embodiment.

[0102] Those skilled in the art will appreciate that all or some of the steps and systems in the method disclosed above can be implemented as software, firmware, hardware, and appropriate combinations thereof. Some physical components or all physical components can be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, and the computer-readable medium can include computer storage media (or non-transitory media) and communication media (or temporary media). As known to those skilled in the art, the term computer storage media is included in any method or technology for storing information (such as a computer-readable program, a data structure, a program module, or other data) and is volatile and non-volatile, removable, and non-removable. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory, or other memory technology, CD-ROM, digital versatile disks (DVD), or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage, or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, as is well known to those skilled in the art, communication media typically embodies computer-readable programs, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0103] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in this field, other different forms of changes or modifications can be made based on the above description. It is impossible to list all the implementation methods here. All obvious changes or modifications derived from the technical solution of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for intelligent matching and recommendation of pharmaceutical product services, characterized in that: include: Collect user data and task release data; Store task release data and analyze task difficulty coefficients, and analyze users' task complexity adaptability based on task difficulty coefficients and user data; Build user feature vectors based on user data and task complexity fitness, and build task feature vectors based on task release data and task difficulty coefficient; Build a task matching model based on user feature vectors and task feature vectors to obtain demand fit and ability adaptability; A recommendation list is created based on the degree of fit with the needs and pushed to the user; a candidate list is created based on the degree of fit with the capabilities and pushed to the task source unit; Collect interaction data after users receive recommendation lists and optimize the task matching model.

2. The intelligent matching and recommendation method for pharmaceutical product services according to claim 1, characterized in that: The task completion status data corresponding to the task release data with the same task content are extracted as the difficulty analysis data, the task time in the difficulty analysis data is normalized, the number of difficulty analysis data with a successful task status is counted as the number of successful data, and the number of difficulty analysis data with a failed task status is counted as the number of failed data. The task difficulty coefficient is analyzed based on the difficulty analysis data, the number of successful data, and the number of failed data. The expression of the task difficulty coefficient is TD=0.6×μ(DT1)+0.4×N2 / (N1+N2), where TD represents the task difficulty coefficient, DT1 represents the normalized task time, μ() represents the average value of the data in the brackets, N1 represents the number of successful data, and N2 represents the number of failed data. The user's task complexity adaptability is analyzed based on the task difficulty coefficient and completed tasks, and the average value of the task difficulty coefficients corresponding to the user's completed tasks is taken as the task complexity adaptability.

3. The intelligent matching and recommendation method for pharmaceutical product services according to claim 2, characterized in that: Extract different certification skills and required skills as a skill set, and number each skill in the skill set to obtain a skill number, so that each skill corresponds to a unique skill number. Sort the skill numbers in ascending order to obtain a skill sequence, and construct a multivariate vector as the skill vector. The number of elements in the skill vector is the same as the number of skills in the skill sequence and corresponds one to one.

4. The intelligent matching and recommendation method for pharmaceutical product services according to claim 3, characterized in that: A user feature vector is constructed based on the user's task completion, response speed, task complexity adaptability, user rating, and skill vector. The user's response speed is calculated and normalized to obtain the response speed parameter. A four-element vector is constructed, in which each element corresponds to the user's task completion, response speed parameter, task complexity adaptability, and user rating. The skill vector is filled in based on the user's certified skills, and the filled skill vector is used as the user skill vector. The four-element vector and the filled skill vector are concatenated to obtain the user feature vector. A task feature vector is constructed based on the degree of urgency, task difficulty coefficient, required skills and skill vector. A ternary vector is constructed, in which each element corresponds to the degree of urgency, task difficulty coefficient and preset matching parameters. The skill vector is filled in according to the required skills, and the filled skill vector is used as the task skill vector. The ternary vector and the filled skill vector are spliced ​​together to obtain the task feature vector.

5. The intelligent matching and recommendation method for pharmaceutical product services according to claim 4, characterized in that: Analyze skill matching parameters based on skill vectors in user feature vectors and task feature vectors; The demand fit is analyzed based on the user feature vector, task feature vector and skill matching parameters. The expression of the demand fit is: Where MAB represents demand fit, w1 represents the first fit weight, w2 represents the second fit weight, w3 represents the third fit weight, w4 represents the fourth fit weight, CS represents the skill matching parameter, F represents the task complexity adaptability, E represents the urgency, T represents the response speed parameter, and MS represents the preset matching parameter; The ability adaptability is analyzed based on the user feature vector, task feature vector, and skill matching parameters. The expression of the ability adaptability is: Where MBA represents ability fit, α1 represents the first fit weight, α2 represents the second fit weight, α3 represents the third fit weight, α4 represents the fourth fit weight, CR represents task completion, and UR represents user rating.

6. The intelligent matching and recommendation method for pharmaceutical product services according to claim 5, characterized in that: An analysis process for updating the capability adaptability based on the urgency and the task difficulty coefficient, and when the urgency is greater than or equal to a first update threshold, increasing the first adaptability weight and decreasing the third adaptability weight to update the capability adaptability analysis process; When the task difficulty coefficient is greater than or equal to the second update threshold, the first fitness weight and the second fitness weight are reduced and the third fitness weight is increased to update the ability fitness analysis process.

7. The intelligent matching and recommendation method for pharmaceutical product services according to claim 6, characterized in that: Create a recommendation list based on the demand fit, extract tasks with a demand fit greater than or equal to the fit threshold as candidate tasks, and sort them in descending order of demand fit. Take the top Y candidate tasks as the recommendation list, where Y represents the number of recommendations. A recommendation list is made based on the ability adaptability. Users whose task completion is greater than or equal to the completion threshold and whose user rating is greater than or equal to the rating threshold are extracted as candidate users. They are then sorted in descending order according to their ability adaptability, and the top Y candidate users are taken as the candidate list.

8. The intelligent matching and recommendation method for pharmaceutical product services according to claim 7, characterized in that: The construction process of the task feature vector is optimized based on the interaction data, recommendation list and candidate list to optimize the task matching model. The number of users who accept the recommendation in the interaction data corresponding to the same task content is extracted as the number of accepted users, the number of users who ignore the recommendation in the interaction data corresponding to the same task content is extracted as the number of ignored users, and the number of users in the candidate list corresponding to the same task content is extracted as the number of valid users who accept the recommendation. The recommendation matching parameters are analyzed based on the number of accepted users, the number of ignored users and the number of valid users. When the recommendation matching parameter is less than 1, the value of the preset matching parameter is updated to be equal to the recommendation matching parameter.

9. A pharmaceutical product service intelligent matching and recommendation platform, characterized by: include: Data collection module, collects user data and task release data; Parameter analysis module, which stores task release data and analyzes task difficulty coefficients, and analyzes the user's task complexity adaptability based on the task difficulty coefficients and user data; Vector construction module, which constructs user feature vectors based on user data and task complexity fitness, and constructs task feature vectors based on task release data and task difficulty coefficient; Model building module, which builds a task matching model based on user feature vectors and task feature vectors to obtain demand fit and capability adaptation; The recommendation analysis module creates a recommendation list based on the degree of fit between requirements and pushes it to users, and creates a candidate list based on the degree of fit between capabilities and pushes it to the task source unit; The feedback optimization module collects interaction data after users receive the recommendation list and optimizes the task matching model.

10. A storage medium, characterized in that: Instructions are stored which, when executed on a computer, cause the computer to execute the method for intelligent matching and recommendation of pharmaceutical product services as described in any one of claims 1 to 8.

Images