A clinical specialist job recruitment decision-making method and system based on big data

Abstract

The present application relates to the technical field of big data processing, and in particular to a clinical specialist job recruitment decision method and system based on big data, which collects the unit characteristics of potential recruitment objects and inputs them into a development trajectory detection model to obtain associated development characteristic data; based on the unit characteristics of potential recruitment objects, search for associated reference recruitment object units in a recruitment resource library to obtain associated reference recruitment object unit-related resume data; input the unit characteristics of potential recruitment objects and the reference recruitment object unit-related resume data into a semantic segmentation model to obtain multiple semantic segmentation resume data, each semantic segmentation resume data being associated with a department or business personnel in the potential recruitment object unit; based on the reference recruitment object unit-related resume data, each semantic segmentation resume data is scored in multiple aspects to determine the score of the potential recruitment object unit. The present invention can predict the future recruitment of hospitals based on the potential object situation in multiple dimensions, thereby improving the efficiency of recruitment and the accuracy of matching results.

Description

Technical Field

[0001] The present application relates to the field of big data processing technology, and in particular to a clinical specialist job recruitment decision-making method and system based on big data. Background Art

[0002] The rapid growth of online recruitment has brought more choices to job seekers, but also increased the cost of job seekers searching for the positions they need. For this reason, all the well-known online recruitment websites have introduced job recommendation systems to help job seekers decide which positions to apply for.

[0003] Among the existing technologies, the traditional hospital recruitment method is that professionals and hospitals recruit through the talent market, which has low recruitment efficiency and low possibility of successful matching. The existing professional talent recruitment matching system has limited data processing capabilities. Although it is better than the traditional recruitment method, it cannot meet people's needs for matching between professionals and hospitals. Summary of the invention

[0004] To achieve the above objectives, this application provides the following technical solutions:

[0005] According to the first aspect of the present invention, the present invention claims protection for a clinical specialist job recruitment decision method based on big data, comprising:

[0006] Collecting unit characteristics of potential recruitment targets, wherein the unit characteristics of potential recruitment targets include resume data of the unit of potential recruitment targets, and the unit of potential recruitment targets includes multiple potential recruitment targets;

[0007] Inputting the unit characteristics of the potential recruitment object into a development trajectory detection model to obtain development characteristic data associated with the unit of the potential recruitment object;

[0008] Based on the development characteristic data, the recruitment resource library is searched for the reference recruitment target unit associated with the reference recruitment target unit, and the reference recruitment target unit associated resume data is obtained, wherein the reference recruitment target unit associated resume data includes at least one of the following: the reference recruitment target unit resume data associated with the reference recruitment target unit, the department resume data associated with each department of the reference recruitment target unit, and the business personnel resume data of each business personnel of the business department in each department of the reference recruitment target unit;

[0009] Input the characteristics of the potential recruitment unit and the resume data related to the reference recruitment unit into a semantic segmentation model to obtain a plurality of semantic segmentation resume data, each of which is associated with a department or a business person in one of the potential recruitment units;

[0010] Based on the relevant resume data of the reference recruitment unit, each of the semantically segmented resume data is scored in terms of years of employment, number of seconded positions, and public opinion;

[0011] Based on the score of each semantically segmented resume data, the score of the potential recruitment target unit is determined.

[0012] Furthermore, the step of inputting the unit characteristics of the potential recruitment object into the development trajectory detection model to obtain the development characteristic data associated with the unit of the potential recruitment object includes:

[0013] Inputting the unit characteristics of the potential recruitment object into the resume data initial screening unit to obtain semantic multimedia resume data associated with the potential recruitment object unit;

[0014] Inputting the semantic multimedia resume data into a development trajectory detection model to obtain development feature data associated with the potential recruitment unit;

[0015] The reference recruitment object unit resume data is multimedia resume data.

[0016] Furthermore, the semantic segmentation model includes a department segmentation unit, and the unit characteristics of the potential recruitment object and the relevant resume data of the reference recruitment object unit are input into the semantic segmentation model to obtain a plurality of semantic segmentation resume data, including:

[0017] Merging the unit characteristics of the potential recruitment object and the resume data of the reference recruitment object with the department resume data associated with each department of the reference recruitment object unit in sequence to obtain a plurality of multi-data source resume data;

[0018] The resume data from the plurality of multi-data sources are input into the department cutting unit to obtain the functional department resume data associated with each department in the potential recruitment unit.

[0019] Furthermore, the semantic segmentation model further includes a business personnel segmentation unit, which inputs the unit characteristics of the potential recruitment object and the relevant resume data of the reference recruitment object unit into the semantic segmentation model to obtain a plurality of semantic segmentation resume data, including:

[0020] Merging the business department resume data associated with the first business department in the functional department resume data with the resume data of a plurality of business personnel associated with the first business department into multi-data source resume data;

[0021] The resume data from multiple data sources is input into the business personnel segmentation unit to obtain the resume data of functional business personnel associated with each business personnel in the first business department.

[0022] Furthermore, based on the relevant resume data of the reference recruitment unit, each of the semantic segmentation resume data is scored in terms of years of employment, number of seconded positions, and public opinion, including:

[0023] For one of the semantic cutting history data, performing bottom line condition identification on the semantic cutting history data to obtain a bottom line condition boundary associated with the semantic cutting history data;

[0024] Determining the position of qualified candidates based on the bottom line condition boundary;

[0025] Expanding the semantic cutting history data based on the qualified candidate position to obtain expanded first cutting history data;

[0026] Compare the academic achievements and clinical practice of each of the first cutting history data with the department history data or the business personnel history data that are sequentially associated, align the one with the smaller difference in academic achievements or clinical practice, and perform standard normalization on the other to obtain aligned candidate cutting history data;

[0027] Based on the relevant resume data of the reference recruitment unit, the candidate cutting resume data is scored in terms of years of experience, number of seconded positions and public opinion.

[0028] Furthermore, the candidate cutting resume data is scored in terms of years of employment, number of seconded positions, and public opinion, including:

[0029] Performing linear regression based on the bottom line condition boundary of the candidate cutting resume data to determine the average capability level associated with the department or business personnel associated with the candidate cutting resume data;

[0030] Marking the candidate cutting history data by stages to obtain stage cutting history data;

[0031] The stage cutting history data is divided into at least one segment along the key event, and the importance value of each segment is calculated in turn based on the distance transformation algorithm;

[0032] If the importance value fails to be obtained, it is merged with the next segment to obtain the importance value after fusion;

[0033] The obtained history data segment closest to the average value of all importance values ​​is used as the key history data segment of the candidate cutting history data;

[0034] The candidate cutting resume data are scored based on the key resume data segments in terms of the number of seconded positions and / or the public opinion.

[0035] Furthermore, in terms of years of experience, the candidate cutting experience data is scored, including:

[0036] The score of the candidate cutting experience data in terms of years of experience is determined based on the following formula:

[0037] ;

[0038] in, represents the score of the candidate cutting experience data in terms of years of experience, Indicates the years of experience of the candidate cutting record data, Indicates the years of employment of the department history data associated with the candidate cutting history data, Indicates the minimum years of experience score. is a hyperparameter used to transform the acceptance of the ratings in terms of years of experience, and max() and min() are functions used to select the maximum and minimum values ​​contained in their respective brackets, respectively.

[0039] Further, in terms of the number of seconded positions and / or the public opinion, the candidate cutting resume data is scored based on the key resume data segments, including:

[0040] Based on the least squares method, the key resume data segment is subjected to linear regression to obtain the number of first seconded positions;

[0041] Based on the first number of seconded positions, the average ability level is updated to obtain an updated number of seconded positions;

[0042] The score of the candidate cutting resume data in terms of the number of seconded positions is determined based on the following formula:

[0043] ;

[0044] in, represents the score of the candidate cutting resume data in terms of the number of seconded positions, The difference between the updated number of seconded positions associated with the candidate cutting history data and the number of seconded positions associated with the department history data or the business personnel history data, Indicates the preset acceptance threshold for the number of seconded positions. Indicates the score of the minimum number of seconded positions;

[0045] and / or,

[0046] Determine a first public opinion and a second public opinion, wherein the first public opinion is a public opinion evaluation composed of three key historical data segments in the first key historical data segment group, and the second public opinion is a public opinion evaluation composed of three key historical data segments in the second key historical data segment group, wherein the first key historical data segment group includes a plurality of the two key historical data segments with the highest and lowest vertical positions in the key historical data segments, and the second key historical data segment group includes the other three key historical data segments in the key historical data segments excluding the first key historical data segment group;

[0047] Determining a first weight of the first public opinion and a second weight of the second public opinion based on a ratio of distances between the first key historical data segment group and the key historical data segment in the second key historical data segment group;

[0048] The score of the candidate cutting history data in terms of public opinion is determined based on the following formula:

[0049] ;

[0050] in, represents the score of the candidate cutting history data in terms of public opinion, represents the first weight, represents the second weight, Indicates the xth public opinion, represents the public opinion acceptance threshold of the xth public opinion association, represents the difference in the number of seconded positions between the xth public opinion associated with the candidate cutting record data and the xth public opinion associated with the department record data or the business personnel record data associated with the candidate cutting record data, Shows the guaranteed public opinion score of the xth public opinion association.

[0051] Further, based on the score of each of the semantic segmentation resume data, the score of the potential recruitment unit is determined, including:

[0052] For a piece of the semantic segmentation resume data, based on the weights associated with the years of service, the number of seconded positions and the public opinion, and the scores of the semantic segmentation resume data in terms of years of service, the number of seconded positions and the public opinion, determine the segmentation resume data score associated with the semantic segmentation resume data;

[0053] Determine the score associated with the potential recruitment unit based on the weight associated with each of the semantic segmentation resume data and the segmentation resume data score associated with each of the semantic segmentation resume data;

[0054] The weight associated with the semantic segmentation history data is positively correlated with the business degree associated with the department or business personnel associated with the semantic segmentation history data.

[0055] According to the second aspect of the present invention, the present invention seeks protection for a clinical specialist job recruitment decision system based on big data, the system comprising a processor and a non-volatile memory storing computer instructions, and when the computer instructions are executed by the processor, the system executes the clinical specialist job recruitment decision method based on big data.

[0056] The present application relates to the technical field of big data processing, and in particular to a clinical specialist job recruitment decision method and system based on big data, which collects the unit characteristics of potential recruitment objects and inputs them into a development trajectory detection model to obtain associated development characteristic data; based on the unit characteristics of potential recruitment objects, search for associated reference recruitment object units in a recruitment resource library to obtain associated reference recruitment object unit-related resume data; input the unit characteristics of potential recruitment objects and the reference recruitment object unit-related resume data into a semantic segmentation model to obtain multiple semantic segmentation resume data, each semantic segmentation resume data being associated with a department or business personnel in the potential recruitment object unit; based on the reference recruitment object unit-related resume data, each semantic segmentation resume data is scored in multiple aspects to determine the score of the potential recruitment object unit. The present invention can predict the future recruitment of hospitals based on the potential object situation in multiple dimensions, thereby improving the efficiency of recruitment and the accuracy of matching results. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A flowchart of the steps of a clinical specialist job recruitment decision-making method based on big data as claimed in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0059] The terms "first", "second" and "third" in this application are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" can explicitly or implicitly include multiple features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of this application are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a stage, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these stages, methods, products or devices.

[0060] Reference to "embodiments" herein means that the specific features, structures, or characteristics described in conjunction with the embodiments may be included in multiple embodiments of the present application. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0061] Please see attached Figure 1 , Figure 1 A schematic diagram of the steps of a clinical specialist job recruitment decision method based on big data provided in an embodiment of the present application, which method can be implemented by a computer device as the main body, and the steps of the clinical specialist job recruitment decision method based on big data include:

[0062] Step S101: Collect characteristics of a potential recruitment target unit, wherein the characteristics of the potential recruitment target unit include resume data of the potential recruitment target unit, and the potential recruitment target unit includes multiple potential recruitment targets.

[0063] The characteristics of the potential recruitment target unit may be, for example, scanned resume data of the potential recruitment target unit, or resume data obtained by photographing the potential recruitment target unit with a camera.

[0064] Step S102: input the unit characteristics of the potential recruitment object into the development trajectory detection model to obtain the development characteristic data associated with the unit of the potential recruitment object.

[0065] The development trajectory detection model may be a pre-trained model, which may be used to identify units in the history data.

[0066] Step S103: searching for the associated reference recruitment target units in the recruitment resource library based on the development characteristic data, and obtaining the relevant resume data of the reference recruitment target units associated with the reference recruitment target units.

[0067] Among them, the resume data related to the reference recruitment target unit includes at least one of the following: the reference recruitment target unit resume data associated with the reference recruitment target unit, the department resume data associated with each department of the reference recruitment target unit, and the business personnel resume data of each business personnel in the business department of each department of the reference recruitment target unit.

[0068] It is understandable that the recruitment resource library may pre-store relevant resume data of reference recruitment object units associated with various units. The closer the potential recruitment object is to the reference recruitment object unit, the higher the score of the potential recruitment object will be.

[0069] The resume data related to the reference recruitment target unit may include the resume data of the reference recruitment target unit as a whole, the resume data of each department obtained by dividing the reference recruitment target unit, and the resume data of each business personnel obtained by dividing the business department.

[0070] Step S104: input the semantic resume data and the resume data related to the reference recruitment object unit into a semantic segmentation model to obtain a plurality of semantic segmentation resume data.

[0071] Each of the semantic segmentation resume data is associated with a department or a business person in one of the potential recruitment units.

[0072] Among them, the semantic segmentation model can be a pre-trained artificial intelligence model based on deep learning, which can segment the semantic resume data based on the relevant resume data of the reference recruitment unit to obtain the resume data of the departments of each unit in the semantic resume data and / or the resume data of each business personnel in the business department.

[0073] Step S105: Based on the relevant resume data of the reference recruitment unit, score each of the semantically segmented resume data in terms of years of employment, number of seconded positions, and public opinion.

[0074] It is understandable that the score associated with the semantic segmentation resume data is positively correlated with the similarity between the semantic segmentation resume data and the resume data related to the reference recruitment unit. The score range can be 0-100, and those skilled in the art can change the range based on actual needs, for example, it can be 0-1, and the embodiments of this disclosure do not limit its specific value.

[0075] Step S106: Determine the score of the potential recruitment unit based on the score of each semantic segmentation resume data.

[0076] In some embodiments, different weight coefficients may be set for each unit of a potential recruitment target or each department of a potential recruitment target, and the total score of the unit of a potential recruitment target may be the weighted sum of the scores of each department. Alternatively, different weight coefficients may also be set for each business person in a business department, and the score of the business department is the weighted sum of the scores of each business person.

[0077] In the disclosed embodiment, the development trajectory of the collected potential recruitment target unit characteristics is first detected to obtain corresponding development characteristic data, and then the corresponding reference recruitment target unit-related resume data is searched in the recruitment resource library, and then the potential recruitment target unit in the potential recruitment target unit characteristics is segmented based on the reference recruitment target unit-related resume data to obtain semantically segmented resume data associated with each department or business personnel, and further based on the semantically segmented resume data and the reference recruitment target unit-related resume data, scores are performed in multiple aspects, and finally the score of the potential recruitment target unit is obtained by combining the score of each semantically segmented resume data. Such a design can accurately segment each department or business personnel in the potential recruitment target unit based on the reference recruitment target unit resume data, and obtain a more accurate potential recruitment target unit score by analyzing each department or business personnel, thereby realizing automated functional scoring and ensuring reliability, and effectively improving the efficiency and accuracy of potential recruitment target unit scoring.

[0078] In some embodiments, inputting the unit characteristics of the potential recruitment object into a development trajectory detection model to obtain the development characteristic data associated with the unit of the potential recruitment object includes:

[0079] Input the unit characteristics of the potential recruitment object into the resume data initial screening unit to obtain semantic multimedia resume data associated with the potential recruitment object unit; input the semantic multimedia resume data into the development trajectory detection model to obtain development characteristic data associated with the potential recruitment object unit; wherein the reference recruitment object unit resume data is multimedia resume data.

[0080] By adopting the above scheme, by integrating the unit characteristics of potential recruitment candidates into multimedia, the amount of calculation can be effectively reduced while ensuring the reliability of the scoring.

[0081] In some embodiments, the semantic segmentation model includes a department segmentation unit, and the unit characteristics of the potential recruitment object and the reference recruitment object unit-related resume data are input into the semantic segmentation model to obtain a plurality of semantic segmentation resume data, including:

[0082] The characteristics of the potential recruitment object unit and the relevant resume data of the reference recruitment object unit are sequentially integrated with the department resume data associated with each department of the reference recruitment object unit to obtain multiple multi-data source resume data; the multiple multi-data source resume data are input into the department cutting unit to obtain the functional department resume data associated with each department in the potential recruitment object unit.

[0083] Optionally, the semantic multimedia resume data and the resume data of the reference recruitment object unit may be sequentially merged with the department resume data associated with each department of the reference recruitment object unit to obtain multiple multi-data source resume data.

[0084] Among them, in the stage of inputting multiple multi-data source resume data into the department cutting unit and obtaining multiple functional department resume data, data source A in the multi-data source resume data can be department resume data, data source B can be semantic multimedia resume data, and data source C can be reference recruitment object unit resume data. Among them, the potential recruitment object "A City Third Medical University" has 10 departments, the number of multi-data source resume data can also be 10, and the number of functional department resume data obtained by the department cutting unit can be 10, which are sequentially associated with the various departments of "A City Third Medical University" in the potential recruitment object unit characteristics.

[0085] In some embodiments, the semantic segmentation model further includes a business personnel segmentation unit, wherein the semantic multimedia resume data and the resume data of the reference recruitment unit are input into the semantic segmentation model to obtain a plurality of semantic segmentation resume data, including:

[0086] Merging the business department resume data associated with the first business department in the functional department resume data with the resume data of a plurality of business personnel associated with the first business department into multi-data source resume data;

[0087] The resume data from multiple data sources is input into the business personnel segmentation unit to obtain the resume data of functional business personnel associated with each business personnel in the first business department.

[0088] Among them, in the multi-data source resume data input into the business personnel cutting unit in the stage of inputting the resume data of multi-data sources into the business personnel cutting unit and obtaining the resume data of functional business personnel associated with each business personnel, data source 5 can be the business department resume data of a functional business department, data source 6 can be the department resume data associated with the business department in the reference recruitment object unit, and data sources 1-4 can be the four business personnel of the business department in turn. After the resume data of multi-data sources is input into the business personnel cutting unit, the functional business personnel resume data of each business personnel associated with the functional department in the business department resume data can be obtained.

[0089] In some embodiments, each of the semantic segmentation resume data is scored based on years of employment, number of seconded positions, and public opinion, including:

[0090] Perform bottom-line condition identification on each of the semantic cutting resume data to obtain bottom-line condition boundaries associated with each of the semantic cutting resume data; determine qualified candidate positions based on the bottom-line condition boundaries; expand each of the semantic cutting resume data based on the qualified candidate positions to obtain expanded first cutting resume data; compare each of the first cutting resume data with the department resume data or the business personnel resume data that are sequentially associated with the data in terms of academic achievements and clinical practice, align the one with the smaller difference in academic achievements or clinical practice, and perform standard normalization on the other to obtain aligned candidate cutting resume data; score the candidate cutting resume data based on multiple aspects such as years of experience, number of seconded positions, and public opinion.

[0091] By adopting the above scheme, before scoring each semantic segmented resume data, each semantic segmented resume data is first aligned, which can further improve the confidence of the unit score of the potential recruitment object.

[0092] In some embodiments, the bottom line condition can be set as a doctoral degree in a medical-related major.

[0093] In some embodiments, the candidate cutting resume data is scored based on years of experience, number of seconded positions, and public opinion, including:

[0094] For one of the candidate cutting resume data, linear regression is performed based on the bottom line condition boundary of the candidate cutting resume data to determine the average ability level associated with the department or business personnel associated with the candidate cutting resume data; the candidate cutting resume data is stage-marked to obtain stage cutting resume data; the stage cutting resume data is divided into at least one segment along the key event, and the importance value of each segment is calculated in turn based on the distance transformation algorithm; if the importance value fails to be calculated, it is fused with the next segment to calculate the fused importance value; the resume data segment that is closest to the average value of all importance values ​​is used as the key resume data segment of the candidate cutting resume data; based on the key resume data segment, the candidate cutting resume data is scored in terms of the number of seconded positions and / or the public opinion.

[0095] Specifically, we can first use the least squares method to perform the first linear regression based on the bottom line condition boundary, collect the department and average ability level, and mark the department resume data according to the number of seconded positions; secondly, divide the department into several segments along the key events and calculate the importance value of each segment in turn based on the distance transformation algorithm. If the calculation fails, it will be merged with the next segment to continue to calculate the importance value; the variance of all the importance values ​​calculated will be used as the key resume data segment of the department or business personnel.

[0096] Among them, the specific implementation of linear regression based on the bottom line condition boundary of the resume data can be found in the implementation of the related technology, and the embodiments of the present disclosure will not be described in detail.

[0097] In some embodiments, the candidate cutting experience data is scored in terms of years of experience, including:

[0098] The score of the candidate cutting experience data in terms of years of experience is determined based on the following formula:

[0099] ;

[0100] in, represents the score of the candidate cutting experience data in terms of years of experience, Indicates the years of experience of the candidate cutting record data, Indicates the years of employment of the department history data associated with the candidate cutting history data, Indicates the minimum years of experience score. is a hyperparameter used to transform the acceptance of the ratings in terms of years of experience, and max() and min() are functions used to select the maximum and minimum values ​​contained in their respective brackets, respectively.

[0101] In the above embodiment, through the above publicity, the score of the department of the potential recruitment unit can be accurately determined, that is, the similarity of the years of employment between the department of the potential recruitment unit and the department of the associated reference recruitment unit, and the device can be used to as well as The value of can effectively ensure the robustness of the score.

[0102] In some embodiments, in terms of the number of seconded positions, the candidate cutting resume data is scored based on the key resume data segments, including:

[0103] Based on the least squares method, the key resume data segment is linearly regressed to obtain the first number of seconded positions; based on the first number of seconded positions, the average ability level is updated to obtain the updated number of seconded positions; based on the following formula, the score of the candidate cutting resume data in terms of the number of seconded positions is determined:

[0104] ;

[0105] in, represents the score of the candidate cutting resume data in terms of the number of seconded positions, The difference between the updated number of seconded positions associated with the candidate cutting history data and the number of seconded positions associated with the department history data or the business personnel history data, Indicates the preset acceptance threshold for the number of seconded positions. Indicates the score of the minimum number of seconded positions;

[0106] By adopting the above scheme, the key resume data segment can be subjected to a second linear regression, and the number of seconded positions in the first linear regression can be updated according to the obtained number of seconded positions, thereby further ensuring the accuracy of the scoring of the number of seconded positions.

[0107] In some embodiments, in terms of the public opinion, the candidate cutting resume data is scored based on the key resume data segment, including:

[0108] Determine a first public opinion and a second public opinion, wherein the first public opinion is a public opinion evaluation composed of three key historical data segments in the first key historical data segment group, and the second public opinion is a public opinion evaluation composed of three key historical data segments in the second key historical data segment group, wherein the first key historical data segment group includes a plurality of the two key historical data segments with the highest and lowest vertical positions in the key historical data segments, and the second key historical data segment group includes the other three key historical data segments in the key historical data segments excluding the first key historical data segment group;

[0109] Determining a first weight of the first public opinion and a second weight of the second public opinion based on a ratio of distances between the first key historical data segment group and the key historical data segment in the second key historical data segment group;

[0110] The score of the candidate cutting history data in terms of public opinion is determined based on the following formula:

[0111] ;

[0112] in, represents the score of the candidate cutting history data in terms of public opinion, represents the first weight, represents the second weight, Indicates the xth public opinion, represents the public opinion acceptance threshold of the xth public opinion association, represents the difference in the number of seconded positions between the xth public opinion associated with the candidate cutting record data and the xth public opinion associated with the department record data or the business personnel record data associated with the candidate cutting record data, Represents the guaranteed public opinion score for the xth public opinion association.

[0113] Optionally, the stage of determining the first public opinion and the second public opinion may specifically include: (1) obtaining the lowest point and the highest point in the vertical position of the key history data segment; (2) taking the above two points as the center, traversing the left and right key points in both directions, and collecting the minimum public opinion evaluation formed by the three points; (3) taking the smaller value of the lowest point public opinion evaluation and the highest point public opinion evaluation as the first public opinion of the functional department (i.e., the first public opinion); (4) repeating operations 1-3 with the remaining key points as the area to obtain the second public opinion (i.e., the second public opinion).

[0114] Among them, the weight between two public opinions is calculated according to the ratio of the three-point distances of their public opinion evaluation, that is, the greater the distance, the greater the weight.

[0115] In some embodiments, determining the score of the potential recruitment unit based on the score of each semantic segmentation resume data includes:

[0116] For one of the semantically cut resume data, the score of the cut resume data associated with the semantically cut resume data is determined based on the weights associated with the years of service, the number of seconded positions and the public opinion in turn, and the scores of the semantically cut resume data in terms of years of service, the number of seconded positions and the public opinion; the score associated with the potential recruitment unit is determined based on the weight associated with each of the semantically cut resume data and the score of the cut resume data associated with each of the semantically cut resume data; wherein the weight associated with the semantically cut resume data is positively correlated with the business degree associated with the department or business personnel associated with the semantically cut resume data.

[0117] Among them, different weight coefficients can be set for the years of experience, number of seconded positions and public opinion of each department type. The department score of a simple department or business personnel is the weighted sum of the years of experience score, the number of seconded positions score and the public opinion score.

[0118] In some embodiments, the clinical specialist job recruitment decision-making method based on big data may specifically include the following steps:

[0119] S1, transferring the functional resume data to the resume data preliminary screening unit.

[0120] S2, in the resume data initial screening unit, the functional resume data is firstly obtained through multimedia to obtain the unit multimedia resume data; then it is input into the unit classifier for development trajectory detection; the identified unit will search the recruitment resource library for the prior information of the reference recruitment target unit, including the multimedia resume data of the reference recruitment target unit, the multimedia resume data of each department of the reference recruitment target unit and the multimedia resume data of each business personnel in the business department.

[0121] S2.1, the business personnel mentioned in S2 refers to the business department decomposed into multiple sections according to the trend, such as horizontal fold, which can be decomposed into two business personnel, horizontal and fold.

[0122] S3, merge the multimedia resume data of the function, the multimedia resume data of the reference recruitment unit and the multimedia resume data of the first department of the reference recruitment unit in terms of data source to form a multi-data source resume data; the operation of the remaining departments is consistent with the first department, forming a total of N multi-data source resume data equal to the number of unit departments, forming a batch, and inputting them into the pre-trained department cutting unit. The department cutting unit encodes and decodes the input resume data and outputs the multimedia resume data of each department of the function.

[0123] S4, if there is a business department in the functional department, further process the business department. The multimedia resume data of the functional business department, the multimedia resume data of the business department of the reference recruitment unit, and the multimedia resume data of each business personnel of the business department of the reference recruitment unit are integrated in terms of data source to form a multi-data source resume data. The pre-trained business personnel cutting unit receives the above multi-data source resume data, and after encoding and decoding the resume data, outputs the multimedia resume data of each business personnel of the functional business department. After this step, the business department can be disassembled into multiple business personnel, and the business personnel can be regarded as equivalent to a simple department.

[0124] S5, the cut departments and business personnel enter the department alignment unit to eliminate the interference of the overall size of the unit on the departmental assessment. First, the multimedia resume data of the functional departments and business personnel are identified for the bottom line conditions to determine the bottom line condition boundaries, and then the qualified candidate positions are calculated, and the functional departments and business personnel are expanded; the academic achievements and practical clinical practice of the expanded resume data are compared with the academic achievements and practical clinical practice of the resume data of the related departments of the reference recruitment unit, and the one with the smaller difference (academic achievements or practical clinical practice) is taken for alignment, and the other is scaled according to the original academic achievements and practical clinical practice ratio of the functional department resume data, and the aligned multimedia resume data of the functional departments and business personnel is obtained.

[0125] S6, for the multimedia resume data of a single department or business personnel, firstly, the least squares method is used to perform the first linear regression based on the bottom line condition value, the department and average ability level are collected, and the department resume data is marked according to the number of seconded positions; secondly, the department is divided into several segments along the key events and the importance value of each segment is calculated in turn based on the distance transformation algorithm. If the calculation fails, it is merged with the next segment to continue to calculate the importance value; the variance of all the importance values ​​calculated is used as the key resume data segment of the department or business personnel for the calculation of the number of seconded positions and public opinion of the unit.

[0126] S7. The evaluation of functional departments or business personnel is divided into three aspects: years of experience, number of seconded positions and public opinion.

[0127] S7.1, the length of service of a functional department is defined as the length of service of a department in the horizontal direction when the department is placed horizontally in S6, recorded as , refer to the recruitment unit department's years of employment , the calculation formula for the functional department's years of experience is as follows:

[0128] ;

[0129] in, is the hyperparameter of the transformation score acceptance, To guarantee the bottom score.

[0130] S7.2, perform a second linear regression based on the least squares method on the department key resume data segments obtained in S6, and update the number of seconded positions obtained in the first linear regression according to the number of seconded positions obtained. The number of seconded positions in the functional department is defined as the updated number of seconded positions, and the calculation formula for the number of seconded positions in the functional department is as follows:

[0131] ;

[0132] in, It is the difference between the number of seconded positions in the functional department and the number of seconded positions in the reference recruitment unit department. The acceptance threshold for the number of seconded positions. To guarantee the bottom score.

[0133] S7.3, the method of collecting public opinions of functional departments is as follows: (1) Obtain the lowest and highest points in the vertical position of the department's key historical data segment; (2) Take the above two points as the center, traverse the left and right key points in turn, and collect the minimum public opinion evaluation formed by the three points; (3) Take the smaller value of the lowest point public opinion evaluation and the highest point public opinion evaluation as the first public opinion of the functional department; (4) Repeat operations 1-3 with the remaining key points as the area to obtain the second public opinion. (5) The weight between the two public opinions is calculated according to the ratio of the distance between the three points of the public opinion evaluation, that is, the farther the distance, the greater the weight. The formula for calculating a single public opinion score is as follows, and the total score of the department's public opinion is the weighted sum of the two public opinion scores:

[0134] ;

[0135] in, It is the difference between the public opinion of the functional department and the public opinion of the reference recruitment unit department. The acceptance threshold for the number of seconded positions. To guarantee the bottom score.

[0136] S8, set different weight coefficients for each department type's years of service, number of seconded positions, and public opinion. The department score of a simple department or business personnel is the weighted sum of the years of service score, number of seconded positions score, and public opinion score. Set different weight coefficients for each business personnel in the business department, and the score of the business department is the weighted sum of the scores of each business personnel. Set different weight coefficients for each department of the unit, and the total department score of the unit is the weighted sum of the scores of each department.

[0137] S9, in the model training stage, the pre-labeled multimedia resume data of functional departments are used as the real labels, and the multimedia resume data of the reference recruitment object unit, the multimedia resume data of the reference recruitment object unit department and the functional multimedia resume data are fused in terms of data source to form multi-data source resume data as the input of the model. After the encoding and decoding stages of the model, a department probability map of the same size as the input resume data is output. The loss of the current iteration of the model is calculated through the probability map and used to transform the model parameters.

[0138] S10, the training phase of the business personnel segmentation unit described in S4 is similar to that described in S9, the difference being that the business personnel segmentation unit uses the annotated multimedia resume data of functional business personnel as the real label, and the input resume data is modified to the multimedia resume data of the reference recruitment unit department, the multimedia resume data of each business personnel of the reference recruitment unit, and the multimedia resume data of the functional department are integrated in terms of data source to form seven data source resume data. The number of business personnel in a business department shall not exceed five, and if it is less than five, blank resume data shall be used to fill the data source.

[0139] According to the second embodiment of the present invention, the present invention requests protection for a clinical specialist job recruitment decision system based on big data, the system comprising a processor and a non-volatile memory storing computer instructions, when the computer instructions are executed by the processor, the system executes the clinical specialist job recruitment decision method based on big data.

[0140] In the above embodiment, one or more of the above steps can be used to segment functions into departments and business personnel levels, and to conduct multi-faceted evaluations to obtain scores for functional departments. Not only can each functional department be accurately disassembled in order of departments, but business department issues can also be decomposed into simple departments through business personnel segmentation units for detailed analysis. This can effectively ensure the accuracy and reliability of automated unit scoring of potential recruits.

[0141] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0142] In addition, each functional unit in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of software functional units. The above is only an implementation method of the present application, and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the specification and drawings of this application, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present application.

[0143] The specific implementation methods of the invention are described in detail above, but they are only examples, and the present application is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the present application, and therefore, the equalization, modification, and improvement made without departing from the spirit and principle of the present application should be included in the scope of the present application.

Claims

1. A clinical specialist job recruitment decision-making method based on big data, characterized in that: include: Collecting unit characteristics of potential recruitment targets, wherein the unit characteristics of potential recruitment targets include resume data of the unit of potential recruitment targets, and the unit of potential recruitment targets includes multiple potential recruitment targets; Inputting the unit characteristics of the potential recruitment object into a development trajectory detection model to obtain development characteristic data associated with the unit of the potential recruitment object; Based on the development characteristic data, the recruitment resource library is searched for the reference recruitment target unit associated with the reference recruitment target unit, and the reference recruitment target unit associated resume data is obtained, wherein the reference recruitment target unit associated resume data includes at least one of the following: the reference recruitment target unit resume data associated with the reference recruitment target unit, the department resume data associated with each department of the reference recruitment target unit, and the business personnel resume data of each business personnel of the business department in each department of the reference recruitment target unit; Input the characteristics of the potential recruitment unit and the resume data related to the reference recruitment unit into a semantic segmentation model to obtain a plurality of semantic segmentation resume data, each of which is associated with a department or a business person in one of the potential recruitment units; Based on the relevant resume data of the reference recruitment unit, each of the semantically segmented resume data is scored in terms of years of employment, number of seconded positions, and public opinion; Determine the score of the potential recruitment unit based on the score of each semantic segmentation resume data; Based on the relevant resume data of the reference recruitment unit, each of the semantic segmentation resume data is scored in terms of years of employment, number of seconded positions, and public opinion, including: For one of the semantic cutting history data, performing bottom line condition identification on the semantic cutting history data to obtain a bottom line condition boundary associated with the semantic cutting history data; Determining the position of qualified candidates based on the bottom line condition boundary; Expanding the semantic cutting history data based on the qualified candidate position to obtain expanded first cutting history data; Compare the academic achievements and clinical practice of each of the first cutting history data with the department history data or the business personnel history data that are sequentially associated, align the one with the smaller difference in academic achievements or clinical practice, and perform standard normalization on the other to obtain aligned candidate cutting history data; Based on the relevant resume data of the reference recruitment unit, the candidate's resume data is scored in terms of years of employment, number of seconded positions, and public opinion; The candidate's cutting resume data is scored based on years of experience, number of seconded positions, and public opinion, including: Performing linear regression based on the bottom line condition boundary of the candidate cutting resume data to determine the average capability level associated with the department or business personnel associated with the candidate cutting resume data; Marking the candidate cutting history data by stages to obtain stage cutting history data; The stage cutting history data is divided into at least one segment along the key event, and the importance value of each segment is calculated in turn based on the distance transformation algorithm; If the importance value fails to be obtained, it is merged with the next segment to obtain the importance value after fusion; The obtained history data segment closest to the average value of all importance values ​​is used as the key history data segment of the candidate cutting history data; Scoring the candidate cutting resume data based on the key resume data segments in terms of the number of seconded positions and / or the public opinion; In terms of years of experience, the candidate's cutting experience data is scored, including: The score of the candidate cutting experience data in terms of years of experience is determined based on the following formula: ； in, represents the score of the candidate cutting experience data in terms of years of experience, Indicates the years of experience of the candidate cutting record data, Indicates the years of employment of the department history data associated with the candidate cutting history data, Indicates the minimum years of experience score. is a hyperparameter used to transform the acceptance of the ratings in terms of years of experience, max() and min() are functions used to select the maximum and minimum values ​​of the values ​​contained in their respective brackets; In terms of the number of seconded positions and / or the public opinion, the candidate cutting resume data is scored based on the key resume data segments, including: Based on the least squares method, the key resume data segment is subjected to linear regression to obtain the number of first seconded positions; Based on the first number of seconded positions, the average ability level is updated to obtain an updated number of seconded positions; The score of the candidate cutting resume data in terms of the number of seconded positions is determined based on the following formula: ； in, represents the score of the candidate cutting resume data in terms of the number of seconded positions, The difference between the updated number of seconded positions associated with the candidate cutting history data and the number of seconded positions associated with the department history data or the business personnel history data, Indicates the preset acceptance threshold for the number of seconded positions. Indicates the score of the minimum number of seconded positions; and / or, Determine a first public opinion and a second public opinion, wherein the first public opinion is a public opinion evaluation composed of three key historical data segments in the first key historical data segment group, and the second public opinion is a public opinion evaluation composed of three key historical data segments in the second key historical data segment group, wherein the first key historical data segment group includes a plurality of the two key historical data segments with the highest and lowest vertical positions in the key historical data segments, and the second key historical data segment group includes the other three key historical data segments in the key historical data segments excluding the first key historical data segment group; Determining a first weight of the first public opinion and a second weight of the second public opinion based on a ratio of distances between the first key historical data segment group and the key historical data segment in the second key historical data segment group; The score of the candidate cutting history data in terms of public opinion is determined based on the following formula: ； in, represents the score of the candidate cutting history data in terms of public opinion, represents the first weight, represents the second weight, Indicates the xth public opinion, represents the public opinion acceptance threshold of the xth public opinion association, represents the difference in the number of seconded positions between the xth public opinion associated with the candidate cutting record data and the xth public opinion associated with the department record data or the business personnel record data associated with the candidate cutting record data, Shows the guaranteed public opinion score of the xth public opinion association.

2. The method according to claim 1, characterized in that The step of inputting the unit characteristics of the potential recruitment object into the development trajectory detection model to obtain the development characteristic data associated with the unit of the potential recruitment object includes: Inputting the unit characteristics of the potential recruitment object into the resume data initial screening unit to obtain semantic multimedia resume data associated with the potential recruitment object unit; Inputting the semantic multimedia resume data into a development trajectory detection model to obtain development feature data associated with the potential recruitment unit; The reference recruitment object unit resume data is multimedia resume data.

3. The method according to claim 1, characterized in that The semantic segmentation model includes a department segmentation unit, and the unit characteristics of the potential recruitment object and the relevant resume data of the reference recruitment object unit are input into the semantic segmentation model to obtain a plurality of semantic segmentation resume data, including: Merging the unit characteristics of the potential recruitment object and the resume data of the reference recruitment object with the department resume data associated with each department of the reference recruitment object unit in sequence to obtain a plurality of multi-data source resume data; The resume data from the plurality of multi-data sources are input into the department cutting unit to obtain the functional department resume data associated with each department in the potential recruitment unit.

4. The method according to claim 3, characterized in that The semantic segmentation model further includes a business personnel segmentation unit, which inputs the unit characteristics of the potential recruitment object and the relevant resume data of the reference recruitment object unit into the semantic segmentation model to obtain a plurality of semantic segmentation resume data, including: Merging the business department resume data associated with the first business department in the functional department resume data with the resume data of a plurality of business personnel associated with the first business department into multi-data source resume data; The resume data from multiple data sources is input into the business personnel segmentation unit to obtain the resume data of functional business personnel associated with each business personnel in the first business department.

5. The method according to claim 1, characterized in that Determining the score of the potential recruitment unit based on the score of each semantic segmentation resume data includes: For a piece of the semantic segmentation resume data, based on the weights associated with the years of service, the number of seconded positions and the public opinion, and the scores of the semantic segmentation resume data in terms of years of service, the number of seconded positions and the public opinion, determine the segmentation resume data score associated with the semantic segmentation resume data; Determine the score associated with the potential recruitment unit based on the weight associated with each of the semantic segmentation resume data and the segmentation resume data score associated with each of the semantic segmentation resume data; The weight associated with the semantic segmentation history data is positively correlated with the business degree associated with the department or business personnel associated with the semantic segmentation history data.

6. A clinical specialist job recruitment decision system based on big data, characterized in that: The system includes a processor and a non-volatile memory storing computer instructions. When the computer instructions are executed by the processor, the system executes the clinical specialist position recruitment decision-making method based on big data as described in any one of claims 1 to 5.

Images