Talent competency assessment method and system

Abstract

The invention provides a talent competency assessment method and system. The talent competency assessment method comprises the steps that 1, related data of talents are collected, the collected data are preprocessed, a knowledge base for talent competency assessment is constructed, knowledge in the knowledge base is subjected to structured processing, and multi-granularity knowledge used for assessment is formed; 2, constructing an original scoring matrix; 3, constructing an ideal solution according to the original scoring matrix; 4, constructing a utility value matrix; 5, establishing a regret-appreciation matrix; step 6, establishing a perception utility value matrix; 7, determining an attribute weight through an entropy weight method; and step 8, calculating a comprehensive evaluation risk value, sorting the objects through the comprehensive evaluation risk value, and the larger the comprehensive evaluation risk value is, the stronger the competency is. According to the method, the competency of the talent is evaluated in a data and knowledge cooperative driving mode, the problems of single data and insufficient knowledge application in a traditional evaluation method are solved, and comprehensive and accurate evaluation of the competency of the talent is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of capability assessment, and in particular, to a talent competency assessment method and system, and in particular, to a "Chinese + vocational skills" talent competency assessment method and system. Background Art

[0002] At present, the traditional "Chinese + professional skills" talent competency assessment method mainly relies on a single data source, such as educational background, work experience, etc., and lacks the application of knowledge such as industry standards and professional skill requirements, resulting in incomplete and inaccurate assessment results.

[0003] Patent application document CN106021274A discloses a talent capability evaluation model system and method based on big data, including: talent big data information collection module, talent big data distributed management module, talent big data analysis module, talent capability precise recommendation module and talent capability multi-dimensional visualization display module. However, this patent cannot completely solve the existing technical problems, nor can it meet the needs of the present invention. Summary of the invention

[0004] In view of the defects in the prior art, the object of the present invention is to provide a talent competency assessment method and system.

[0005] The talent competency assessment method provided by the present invention comprises:

[0006] Step 1: Collect relevant data of talents and pre-process the collected data, including data cleaning, denoising and normalization operations, build a knowledge base for talent competency assessment, including industry standards, professional skill requirements and language proficiency standards, and structure the knowledge in the knowledge base to form multi-granular knowledge for assessment;

[0007] Step 2: Construct the original score matrix Q = (q ri ) n×m , r=1,2,…,m;i=1,2,…,n, where m is the number of experts, n is the number of attributes, q ri is the rating value;

[0008] Step 3: Based on the original score matrix Q, construct the ideal solution I = [u1,u2,…,u n ], where u i is the ideal value of the i-th evaluation indicator. The higher the risk level, the larger the score value. The maximum value is taken as the ideal value of each evaluation indicator.

[0009] Step 4: Construct the utility value matrix F;

[0010] Step 5: Establish the regret-delight matrix C;

[0011] Step 6: Establish the perceived utility value matrix G;

[0012] Step 7: Determine attribute weights by entropy weight method;

[0013] Step 8: Calculate the comprehensive evaluation risk value and rank each object according to the comprehensive evaluation risk value. The larger the comprehensive evaluation risk value, the stronger the competence.

[0014] Specifically, the step 1 includes:

[0015] Optimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), U is the domain of the research object, V is the score of the research object in terms of language or skills, and R r , r=1,2,...,m is the fuzzy approximate equivalence relationship from U to V. Then the optimistic multi-granularity intuitionistic fuzzy lower approximation of R with respect to the θ operator is and the upper approximation The expression is:

[0016]

[0017] in, and are the membership and non-membership values ​​of object x in the optimistic case with respect to the fuzzy approximate relation R; r is a sub-relation in the relation R; m is the number of sub-relation in the relation R; represents the optimistic approximation result of Z about the relation R; x is an object in the domain U; ​​μ represents the membership degree; v represents the non-membership degree;

[0018] Description of the theta operator:

[0019] Among them, θ 221 is called the θ operator, θ 365 is called θ * Operator,

[0020] Then we have:

[0021]

[0022] Among them, y is an element in the domain U; ​​μ Z (y) represents the membership of element y with respect to evaluation value Z; v Z (y) represents the non-membership degree of element y with respect to evaluation value Z; Represents objects x and y with respect to the relation R r The membership degree of ; ∧ and ∨ represent conjunction and disjunction respectively;

[0023] for For object x under m relations, the membership degree of x is combined with the θ operator to calculate IFS(U) is the set of all intuitionistic fuzzy sets in the domain U. Its optimistic decision rules based on the θ operator include:

[0024] Accept decision rule: If and Then x∈POS O (Z);

[0025] Delayed decision rule: If and Dissatisfied and Then x∈BND O (Z);

[0026] Rejection decision rule: If and Then x∈NEG O (Z);

[0027] Among them, POS O (Z) represents the positive domain of the evaluation value Z under the optimistic state; BND O (Z) represents the boundary region of the evaluation value Z under the optimistic state; NEG O (Z) represents the negative domain of the evaluation value Z under the optimistic state; α and β are evaluation threshold parameters; if the optimistic approximation value of object x under each evaluation dimension is greater than the threshold α, then the language or skill competence is judged to be acceptable; if the optimistic approximation value is less than the threshold α, then its competence is judged to be unacceptable; if the optimistic approximation value is between α and β, then its competence is judged to be uncertain, that is, it cannot be directly judged based on the existing information, and additional information is needed to make a judgment;

[0028] Pessimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), R r is the fuzzy approximate equivalence relationship from U to V. R on the Pessimistic Multi-granularity Intuitionistic Fuzzy Lower Approximation of θ Operator and the upper approximation as follows:

[0029]

[0030] Among them, P represents pessimism or a pessimistic state; and The membership and non-membership values ​​of object x with respect to the fuzzy approximate relation R under pessimistic conditions are as follows:

[0031]

[0032] for y is an element in Z, and the pessimistic decision rule based on the θ operator includes:

[0033] Accept decision rule: If Then x∈POS p (Z);

[0034] Delayed decision rule: If and Dissatisfied and Then x∈BND p (Z);

[0035] Rejection decision rule: If and Then x∈NEG p (Z).

[0036] Specifically, the power function is selected as the utility function of sorting, and the utility matrix expression is:

[0037]

[0038] in, It is a parameter that controls the magnitude of the change;

[0039] The regret-delight matrix C is expressed as:

[0040]

[0041] Among them, ψ is the regret aversion coefficient;

[0042] The expression of the perceived utility value matrix G is:

[0043]

[0044] Where e is a constant.

[0045] Specifically, the step 7 includes:

[0046] For a given set of n attributes A1, A2, ..., A n , where A i ={a1,a2,…,a n}, the standardized values ​​of each attribute data are B1, B2,…, B n , then:

[0047]

[0048] Calculate the information entropy of each attribute as E i :

[0049]

[0050] in, If p ri =0, then define

[0051] Calculate the weight W of each attribute through information entropy i for:

[0052]

[0053] Specifically, the step 8 includes:

[0054] According to the attribute weights, the perceived utility functions of the objects under each attribute are weighted and superimposed, and the comprehensive evaluation risk value S of the jth object is j The expression is:

[0055]

[0056] Among them, W i is the weight of attribute i, g ji The perceived utility value of object j regarding attribute i.

[0057] The talent competency assessment system provided by the present invention comprises:

[0058] Module M1: Collect relevant data on talents, pre-process the collected data, including data cleaning, denoising and normalization operations, build a knowledge base for talent competency assessment, including industry standards, professional skill requirements and language proficiency standards, and structure the knowledge in the knowledge base to form multi-granular knowledge for assessment;

[0059] Module M2: Construct the original scoring matrix Q = (q ri ) n×m , r=1,2,…,m;i=1,2,…,n, where m is the number of experts, n is the number of attributes, q ri is the rating value;

[0060] Module M3: Based on the original score matrix Q, construct the ideal solution I = [u1,u2,…,u n ], where u i is the ideal value of the i-th evaluation indicator. The higher the risk level, the larger the score value. The maximum value is taken as the ideal value of each evaluation indicator.

[0061] Module M4: Construct utility value matrix F;

[0062] Module M5: Establishing the regret-delight matrix C;

[0063] Module M6: Establishing the perceived utility value matrix G;

[0064] Module M7: Determine attribute weights by entropy weight method;

[0065] Module M8: Calculate the comprehensive evaluation risk value and rank each object according to the comprehensive evaluation risk value. The larger the comprehensive evaluation risk value, the stronger the competence.

[0066] Specifically, the module M1 includes:

[0067] Optimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), U is the domain of the research object, V is the score of the research object in terms of language or skills, and R r , r=1,2,...,m is the fuzzy approximate equivalence relationship from U to V. Then the optimistic multi-granularity intuitionistic fuzzy lower approximation of R with respect to the θ operator is and the upper approximation The expression is:

[0068]

[0069] in, and are the membership and non-membership values ​​of object x in the optimistic case with respect to the fuzzy approximate relation R; r is a sub-relation in the relation R; m is the number of sub-relation in the relation R; represents the optimistic approximation result of Z about the relation R; x is an object in the domain U; ​​μ represents the membership degree; v represents the non-membership degree;

[0070] Description of the theta operator:

[0071] Among them, θ 221 is called the θ operator, θ 365 is called θ * Operator,

[0072] Then we have:

[0073]

[0074]

[0075] Among them, y is an element in the domain U; ​​μ z (y) represents the membership of element y with respect to evaluation value Z; v z (y) represents the non-membership degree of element y with respect to evaluation value Z; Represents objects x and y with respect to the relation R r The membership degree of ; ∧ and ∨ represent conjunction and disjunction respectively;

[0076] for For object x under m relations, the membership degree of x is combined with the θ operator to calculate IFS(U) is the set of all intuitionistic fuzzy sets in the domain U. Its optimistic decision rules based on the θ operator include:

[0077] Accept decision rule: If and Then x∈POS O (Z);

[0078] Delayed decision rule: If and Dissatisfied and Then x∈BND O (Z);

[0079] Rejection decision rule: If and Then x∈NEG O (Z);

[0080] Among them, POS O (Z) represents the positive domain of the evaluation value Z under the optimistic state; BND O (Z) represents the boundary region of the evaluation value Z under the optimistic state; NEG O (Z) represents the negative domain of the evaluation value Z under the optimistic state; α and β are evaluation threshold parameters; if the optimistic approximation value of object x under each evaluation dimension is greater than the threshold α, then the language or skill competence is judged to be acceptable; if the optimistic approximation value is less than the threshold α, then its competence is judged to be unacceptable; if the optimistic approximation value is between α and β, then its competence is judged to be uncertain, that is, it cannot be directly judged based on the existing information, and additional information is needed to make a judgment;

[0081] Pessimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), R r is the fuzzy approximate equivalence relationship from U to V. R on the Pessimistic Multi-granularity Intuitionistic Fuzzy Lower Approximation of θ Operator and the upper approximation as follows:

[0082]

[0083] Among them, P represents pessimism or a pessimistic state; and The membership and non-membership values ​​of object x with respect to the fuzzy approximate relation R under pessimistic conditions are as follows:

[0084]

[0085] for y is an element in Z, and the pessimistic decision rule based on the θ operator includes:

[0086] Accept decision rule: If and Then x∈POS p (Z);

[0087] Delayed decision rule: If and Dissatisfied and Then x∈BND p (Z);

[0088] Rejection decision rule: If and Then x∈NEG p (Z).

[0089] Specifically, the power function is selected as the utility function of sorting, and the utility matrix expression is:

[0090]

[0091] in, It is a parameter that controls the magnitude of the change;

[0092] The regret-delight matrix C is expressed as:

[0093]

[0094] Among them, ψ is the regret aversion coefficient;

[0095] The expression of the perceived utility value matrix G is:

[0096]

[0097] Where e is a constant.

[0098] Specifically, the module M7 includes:

[0099] For a given n attributes A1, A2, ..., A n , where A i ={a1,a2,...,a n}, the standardized values ​​of each attribute data are B1, B2,…, B n , then:

[0100]

[0101] Calculate the information entropy of each attribute as E i :

[0102]

[0103] in, If p ri =0, then define

[0104] Calculate the weight W of each attribute through information entropy i for:

[0105]

[0106] Specifically, the module M8 includes:

[0107] According to the attribute weights, the perceived utility functions of the objects under each attribute are weighted and superimposed, and the comprehensive evaluation risk value S of the jth object is j The expression is:

[0108]

[0109] Among them, W i is the weight of attribute i, g ji The perceived utility value of object j regarding attribute i.

[0110] Compared with the prior art, the present invention has the following beneficial effects:

[0111] The present invention provides a talent competency assessment method, which solves the problems of single data and insufficient knowledge application in traditional assessment methods by adopting a data and knowledge collaboratively driven approach, achieves a comprehensive and accurate assessment of talent competency, and improves the scientificity and practicality of the assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0112] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0113] Figure 1 A flow chart of the talent competency assessment method;

[0114] Figure 2 Schematic diagram for scoring multi-granularity knowledge;

[0115] Figure 3 is the perceived utility value of employees under different attributes. DETAILED DESCRIPTION

[0116] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0117] Example 1

[0118] like Figure 1 The present invention provides a "Chinese + occupation" talent competency assessment method, comprising:

[0119] Step 1: Preprocessing and multi-granularity knowledge mining;

[0120] Collect relevant data of "Chinese + professional" talents, including but not limited to educational background, professional skills, work experience, language ability, etc. Pre-process the collected data, including data cleaning, denoising, normalization and other operations to ensure data quality and consistency. Then, build a knowledge base for the competency assessment of "Chinese + professional" talents, including industry standards, professional skill requirements, language ability standards, etc. Structuralize the knowledge in the knowledge base to form multi-granular knowledge that can be used for assessment.

[0121] like Figure 2 ,Multi-granularity knowledge acquisition includes:

[0122] Optimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), where U is the domain of the research object, V is the score of the research object in terms of language or skills, and R r (r=1,2,...,m) is the fuzzy approximate equivalence relationship from U to V. (Z is any element in V), then the optimistic multi-granularity intuitionistic fuzzy lower approximation of R with respect to the θ operator is and the upper approximation as follows:

[0123]

[0124] in, and are the membership and non-membership values ​​of object x in the optimistic case with respect to the fuzzy approximate relation R; r is a sub-relation in relation R; m is the number of sub-relationships in relation R; V is the score of the research object in terms of language or skills, Z is an element in the set V, is a whole, representing the optimistic approximation result of Z about the relation R; x is an object in the domain U; ​​μ represents the membership degree; v represents the non-membership degree;

[0125] Description of the θ operator: θ: [0,1]×[0,1]→[0,1], but:

[0126]

[0127] Among them, θ 221 is called the θ operator, θ 365 is called θ * Operator.

[0128] in:

[0129]

[0130] Among them, y is an element in the domain U; ​​μ Z (y) represents the membership of element y with respect to evaluation value Z; v Z (y) represents the non-membership degree of element y with respect to evaluation value Z; Represents objects x and y with respect to the relation R r The membership degree of ; ∧ and ∨ represent conjunction and disjunction respectively. Take the example of m as an example, that is, the membership degree of object x is combined with the θ operator for calculation under m relations. (IFS(U): the set of all intuitionistic fuzzy sets in the domain U), the optimistic three-branch decision rule based on the θ operator is as follows:

[0131] (1) Accept the decision rule (denoted as: P O ):

[0132] If satisfied and Then x∈POS O (Z);

[0133] (2) Delayed decision rule (denoted as: B O ):

[0134] If satisfied and Dissatisfied and Then x∈BND O (Z);

[0135] (3) Rejection decision rule (denoted as: N O ):

[0136] If satisfied and Then x∈NEG O (Z).

[0137] Among them, POS O (Z) represents the positive domain of the evaluation value Z under the optimistic state; BNDO (Z) represents the boundary region of the evaluation value Z under the optimistic state; NEG O (Z) represents the negative domain of the evaluation value Z under an optimistic state; α and β are evaluation threshold parameters. According to the evaluation rules under the above three optimistic situations, if the optimistic approximate value of the object x under each evaluation dimension is greater than the threshold α, then the language or skill competence is judged to be acceptable; if its value is less than the threshold α, then its competence is judged to be unacceptable; if its value is between α and β, then its competence is judged to be uncertain, that is, it cannot be directly judged based on the existing information, and additional information is needed before making a judgment. The pessimistic situation is similar to the optimistic situation, and its calculation rules are relatively stricter than those of the optimistic situation, as follows:

[0138] Pessimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), R r (r=1,2,...,m) is the fuzzy approximate equivalence relationship from U to V. Then the pessimistic multi-granularity intuitionistic fuzzy lower approximation of R with respect to the θ operator is and the upper approximation as follows:

[0139]

[0140] Among them, P represents pessimism or a pessimistic state; and The membership and non-membership values ​​of object x with respect to the fuzzy approximate relation R under pessimistic conditions are as follows:

[0141]

[0142]

[0143] (For any intuitionistic fuzzy set Z in the domain U), y is an element in Z, and the pessimistic three-branch decision rule based on the θ operator is as follows:

[0144] (1) Accept the decision rule (denoted as: P P ): If satisfied and Then x∈POS p (Z);

[0145] (2) Delayed decision rule (denoted as: B P ): If satisfied and Dissatisfied and Then x∈BND p (Z);

[0146] (3) Rejection decision rule (denoted as: N P ): If satisfied and Then x∈NEG p (Z).

[0147] Step 2: Construct an evaluation matrix;

[0148] Construct the original score matrix Q = (q ri ) n×m (r=1,2,…,m;i=1,2,…,n), where m is the number of experts, n is the number of attributes, q ri is the rating value.

[0149] Step 3: Build an ideal solution;

[0150] According to the original score matrix Q, construct the ideal solution I = [u1,u2,…,u n ], where u i (i=1,2,…,n) is the ideal value of the i-th evaluation indicator. It can be seen from Table 1 that the higher the risk level, the greater the score value. In order to reduce the decision maker’s regret, the maximum value is taken as the ideal value of each evaluation indicator.

[0151] Table 1 Risk level classification and scoring values

[0152]

[0153]

[0154] Step 4: Construct the utility value matrix;

[0155] The power function is selected as the utility function of the sorting method, and the parameter Take 0.9 and construct the utility matrix as follows:

[0156]

[0157] Where F is the utility matrix, The value of the object in F. According to existing research, When the value is 0.9, it is relatively stable. When constructing the utility value matrix, the power function is selected as the utility function of the sorting method, where It is a parameter that controls the magnitude of the change.

[0158] Step 5: Establish the regret-delight matrix C;

[0159]

[0160] Among them, C is the regret-delight matrix, ψ is the regret avoidance coefficient, and ψ is 0.0133.

[0161] Step 6: Establish the perceived utility value matrix G;

[0162]

[0163] Among them, G is the utility value matrix, which is composed of the regret-delight matrix C and the utility matrix F, and e is a constant.

[0164] Step 7: Determine attribute weights by entropy weight method;

[0165] For a given set of n attributes (A1, A2, ..., A n ), where A i ={a1,a2,…,a n}, the standardized values ​​of each attribute data are B1, B2,…, B n ,So:

[0166]

[0167] Calculate the information entropy of each attribute as E i (i=1,2,…,n):

[0168]

[0169] in, If p ri =0, then define

[0170] Calculate the weight W of each attribute through information entropy i (i=1,2,…,n) is:

[0171]

[0172] Step 8: Calculate the comprehensive evaluation risk value;

[0173] According to the attribute weights, the perceived utility functions of the objects under each attribute are weighted and superimposed, and the comprehensive evaluation risk value S of the jth object is j as follows:

[0174]

[0175] Among them, W i is the weight of attribute i, g ji is the perceived utility value of object j regarding attribute i, S j is the comprehensive evaluation risk value of object j.

[0176] Step 9: Finally pass S j Sort the objects, S j The larger the value, the greater the competence.

[0177] Example 2

[0178] A multinational logistics company needs to evaluate the "Chinese + professional skills" competency of the talents it recruits in order to select one "Outstanding Employee of the Year". It invited three experts to form a jury and built a decision-making information system (U, A, R r ,Z), where (r=1,2,3). The jury examines 9 employees U={x1,x2,…,x9} in the following five aspects: A={y1,y2,y3,y4,y5}, representing: Chinese language ability (y1), professional skill level (y2), cross-cultural communication ability (y3), teamwork ability (y4) and professional ethics (y5). The intuitive fuzzy set Z is used to represent the leader's favor for the employee, and the threshold is (0.65,0.30). Where:

[0179] Z={<x1,0.50,0.39> ,<x2,0.54,0.36> ,<x3,0.55,0.35> ,

[0180] <x4,0.51,0.37> ,<x5,0.52,0.37> ,<x6,0.52,0.38> ,

[0181] <x7,0.48,0.43> ,<x8,0.44,0.49> ,<x9,0.44,0.45>}

[0182] The evaluation values ​​given by the experts are shown in Table 1.

[0183] Table 1 Evaluation values ​​of expert R1 on 9 employees

[0184]

[0185]

[0186] Table 2 Expert R2 evaluation values ​​of 9 employees

[0187] <![CDATA[y1]]> <![CDATA[y2]]> <![CDATA[y3]]> <![CDATA[y4]]> <![CDATA[y5]]> <![CDATA[x1]]> (0.57,0.32) (0.52,0.37) (0.48,0.41) (0.49,0.40) (0.44,0.45) <![CDATA[x2]]> (0.53,0.36) (0.45,0.44) (0.55,0.34) (0.53,0.36) (0.59,0.30) <![CDATA[x3]]> (0.64,0.25) (0.57,0.32) (0.56,0.33) (0.63,0.27) (0.58,0.31) <![CDATA[x4]]> (0.49,0.40) (0.56,0.33) (0.44,0.46) (0.46,0.43) (0.47,0.42) <![CDATA[x5]]> (0.42,0.47) (0.63,0.26) (0.45,0.44) (0.59,0.30) (0.48,0.41) <![CDATA[x6]]> (0.60,0.29) (0.62,0.27) (0.41,0.48) (0.48,0.41) (0.49,0.40) <![CDATA[x7]]> (0.47,0.42) (0.46,0.43) (0.43,0.46) (0.49,0.40) (0.45,0.44) <![CDATA[x8]]> (0.54,0.36) (0.48,0.41) (0.42,0.47) (0.41,0.48) (0.48,0.42) <![CDATA[x9]]> (0.52,0.37) (0.44,0.45) (0.43,0.46) (0.48,0.41) (0.52,0.37)

[0188] Table 3 Expert R3's evaluation values ​​for 9 employees

[0189] <![CDATA[y1]]> <![CDATA[y2]]> <![CDATA[y3]]> <![CDATA[y4]]> <![CDATA[y5]]> <![CDATA[x1]]> (0.58,0.31) (0.53,0.36) (0.49,0.40) (0.50,0.39) (0.45,0.44) <![CDATA[x2]]> (0.52,0.38) (0.44,0.45) (0.53,0.36) (0.52,0.37) (0.57,0.32) <![CDATA[x3]]> (0.62,0.27) (0.55,0.34) (0.55,0.34) (0.61,0.28) (0.56,0.33) <![CDATA[x4]]> (0.51,0.38) (0.58,0.31) (0.45,0.44) (0.48,0.41) (0.49,0.40) <![CDATA[x5]]> (0.43,0.46) (0.64,0.25) (0.46,0.43) (0.61,0.28) (0.49,0.40) <![CDATA[x6]]> (0.58,0.31) (0.60,0.29) (0.40,0.49) (0.46,0.43) (0.47,0.42) <![CDATA[x7]]> (0.48,0.41) (0.47,0.42) (0.44,0.45) (0.49,0.40) (0.45,0.44) <![CDATA[x8]]> (0.55,0.34) (0.49,0.40) (0.43,0.46) (0.42,0.47) (0.49,0.40) <![CDATA[x9]]> (0.51,0.38) (0.43,0.46) (0.42,0.47) (0.47,0.42) (0.51,0.39)

[0190] We can further obtain the relationship matrix as follows.

[0191]

[0192] Step 1: Calculate the intuitionistic fuzzy rough sets at optimistic and pessimistic granularities as shown in Table 4.

[0193] Table 4 Multi-granularity intuitionistic fuzzy rough sets

[0194]

[0195] Step 2 divides the positive domain POS, boundary domain BND and negative domain NEG according to the decision rules and thresholds as shown in Table 5.

[0196] Table 5 Division of positive domain, boundary domain and negative domain

[0197]

[0198]

[0199] Therefore, the optimal objects in both optimistic and pessimistic states are x2 and x3, and it is impossible to judge the superiority and inferiority of the objects in the POS domain. The following is to sort the objects in the domain using the three sorting methods proposed in this chapter.

[0200] Step 3: Construct the original scoring matrix Q and construct the ideal solution based on the original scoring matrix. The higher the risk level, the greater the score value. In order to reduce the decision maker's regret, the maximum value is taken as the ideal value of each evaluation indicator, and the final ideal solution I is obtained.

[0201]

[0202] I=[7.907.969.547.919.569.055.576.976.36]

[0203] Step 4: Establish the utility value matrix F and the regret-delight matrix C.

[0204]

[0205] Step 5: Construct the perceived utility matrix G through the utility value matrix F and the regret-delight matrix C.

[0206] In order to intuitively show the relationship between the perceived utility values ​​of each employee under different attributes, the matrix G is plotted as follows: Figure 3 shown.

[0207] From the utility value matrix F and the regret-delight matrix C, we can see that under the same attribute, the smaller the utility value, the greater the degree of regret when choosing the current employee. Due to the existence of the risk aversion coefficient, the smaller the perceived utility value, the lower the comprehensive evaluation risk value of the current employee under this attribute, which in turn affects the ranking result.

[0208] Step 6 Determine the attribute weight W i , as shown in Table 6.

[0209] Table 6 Attribute weights

[0210]

[0211] Step 7 Calculate the comprehensive evaluation risk value S j , as shown in Table 7.

[0212] Table 7 Comprehensive evaluation risk value

[0213] staff <![CDATA[x1]]> <![CDATA[x2]]> <![CDATA[x3]]> <![CDATA[x4]]> <![CDATA[x5]]> <![CDATA[x6]]> <![CDATA[x7]]> <![CDATA[x8]]> <![CDATA[x9]]> Value at Risk 7.75 10.34 12.27 7.85 9.61 8.33 7.51 6.23 7.44

[0214] Step 8: According to the three sorting algorithms, we get the sorting results in optimistic and pessimistic states, as follows:

[0215] Optimistic: x3>x2>x5>x6>x1>x4>x9>x7>x8

[0216] Pessimistic: x3>x2>x5>x6>x1>x4>x7>x9>x8

[0217] Judging from the optimal selection results, in the "Chinese + Vocational Skills" talent victory assessment in this example, object x3 has the best result.

[0218] Those skilled in the art know that, in addition to implementing the system, device and its various modules provided by the present invention in a purely computer-readable program code, it is entirely possible to implement the same program in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers and embedded microcontrollers by logically programming the method steps. Therefore, the system, device and its various modules provided by the present invention can be considered as a hardware component, and the modules included therein for implementing various programs can also be considered as structures within the hardware component; the modules for implementing various functions can also be considered as both software programs for implementing the method and structures within the hardware component.

[0219] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A talent competency assessment method, characterized in that: include: Step 1: Collect relevant data of talents and pre-process the collected data, including data cleaning, denoising and normalization operations, build a knowledge base for talent competency assessment, including industry standards, professional skill requirements and language proficiency standards, and structure the knowledge in the knowledge base to form multi-granular knowledge for assessment; Step 2: Construct the original score matrix Q = (q ri ) n×m , r=1,2,…,m;i=1,2,…,n, where m is the number of experts, n is the number of attributes, q ri is the rating value; Step 3: Based on the original score matrix Q, construct the ideal solution I = [u1,u2,…,u n ], where u i is the ideal value of the i-th evaluation indicator. The higher the risk level, the larger the score value. The maximum value is taken as the ideal value of each evaluation indicator. Step 4: Construct the utility value matrix F; Step 5: Establish the regret-delight matrix C; Step 6: Establish the perceived utility value matrix G; Step 7: Determine attribute weights by entropy weight method; Step 8: Calculate the comprehensive evaluation risk value and rank each object according to the comprehensive evaluation risk value. The larger the comprehensive evaluation risk value, the stronger the competence.

2. The talent competency assessment method according to claim 1, characterized in that: The step 1 comprises: Optimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), U is the domain of the research object, V is the score of the research object in terms of language or skills, and R r , r=1,2,…,m is the fuzzy approximate equivalence relationship from U to V. Then the optimistic multi-granularity intuitionistic fuzzy lower approximation of R with respect to the θ operator is and the upper approximation The expression is: in, and are the membership and non-membership values ​​of object x in the optimistic case with respect to the fuzzy approximate relation R; r is a sub-relation in the relation R; m is the number of sub-relationships in the relation R; represents the optimistic approximation result of Z about the relation R; x is an object in the domain U; ​​μ represents the membership degree; v represents the non-membership degree; Description of the theta operator: Among them, θ 221 is called the θ operator, θ 365 is called θ * Operator, Then we have: Among them, y is an element in the domain U; ​​μ Z (y) represents the membership of element y with respect to evaluation value Z; v Z (y) represents the non-membership degree of element y with respect to evaluation value Z; Represents objects x and y with respect to the relation R r The membership degree of ; ∧ and ∨ represent conjunction and disjunction respectively; for For object x under m relations, the membership degree of x is combined with the θ operator to calculate IFS(U) is the set of all intuitionistic fuzzy sets in the domain U. Its optimistic decision rules based on the θ operator include: Accept decision rule: If and Then x∈POS O (Z); Delayed decision rule: If and Dissatisfied and Then x∈BND O (Z); Rejection decision rule: If and Then x∈NEG O (Z); Among them, POS O (Z) represents the positive domain of the evaluation value Z under the optimistic state; BND O (Z) represents the boundary region of the evaluation value Z under the optimistic state; NEG O (Z) represents the negative domain of the evaluation value Z under the optimistic state; α and β are evaluation threshold parameters; if the optimistic approximation value of object x under each evaluation dimension is greater than the threshold α, then the language or skill competence is judged to be acceptable; if the optimistic approximation value is less than the threshold α, then its competence is judged to be unacceptable; if the optimistic approximation value is between α and β, then its competence is judged to be uncertain, that is, it cannot be directly judged based on the existing information, and additional information is needed to make a judgment; Pessimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), R r is the fuzzy approximate equivalence relationship from U to V. R on the Pessimistic Multi-granularity Intuitionistic Fuzzy Lower Approximation of θ Operator and the upper approximation as follows: Among them, P represents pessimism or a pessimistic state; and The membership and non-membership values ​​of object x with respect to the fuzzy approximate relation R under pessimistic conditions are as follows: for y is an element in Z, and the pessimistic decision rule based on the θ operator includes: Accept decision rule: If and Then x∈POS p (Z); Delayed decision rule: If and Dissatisfied and Then x∈BND p (Z); Rejection decision rule: If and Then x∈NEG p (Z).

3. The talent competency assessment method according to claim 2, characterized in that: The power function is selected as the utility function of the sorting, and the utility matrix expression is: in, It is a parameter that controls the magnitude of the change; The regret-delight matrix C is expressed as: Among them, ψ is the regret aversion coefficient; The expression of the perceived utility value matrix G is: Where e is a constant.

4. The talent competency assessment method according to claim 3, characterized in that: The step 7 comprises: For a given n attributes A1, A2, ..., A n , where A i ={a1,a2,…,a n }, the standardized values ​​of each attribute data are B1, B2,…, B n , then: Calculate the information entropy of each attribute as E i : in, If p ri =0, then define Calculate the weight W of each attribute through information entropy i for:

5. The talent competency assessment method according to claim 4, characterized in that: The step 8 comprises: According to the attribute weights, the perceived utility functions of the objects under each attribute are weighted and superimposed, and the comprehensive evaluation risk value S of the jth object is j The expression is: Among them, W i is the weight of attribute i, g ji The perceived utility value of object j regarding attribute i.

6. A talent competency assessment system, characterized in that: include: Module M1: Collect relevant data on talents, pre-process the collected data, including data cleaning, denoising and normalization operations, build a knowledge base for talent competency assessment, including industry standards, professional skill requirements and language proficiency standards, and structure the knowledge in the knowledge base to form multi-granular knowledge for assessment; Module M2: Construct the original scoring matrix Q = (q ri ) n×m , r=1,2,…,m;i=1,2,…,n, where m is the number of experts, n is the number of attributes, q ri is the rating value; Module M3: Based on the original score matrix Q, construct the ideal solution I = [u1,u2,…,u n ], where u i is the ideal value of the i-th evaluation indicator. The higher the risk level, the larger the score value. The maximum value is taken as the ideal value of each evaluation indicator. Module M4: Construct utility value matrix F; Module M5: Establishing the regret-delight matrix C; Module M6: Establishing the perceived utility value matrix G; Module M7: Determine attribute weights by entropy weight method; Module M8: Calculate the comprehensive evaluation risk value and rank each object according to the comprehensive evaluation risk value. The larger the comprehensive evaluation risk value, the stronger the competence.

7. The talent competency assessment system according to claim 6, characterized in that: The module M1 comprises: Optimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), U is the domain of the research object, V is the score of the research object in terms of language or skills, and R r , r=1,2,...,m is the fuzzy approximate equivalence relationship from U to V. Then the optimistic multi-granularity intuitionistic fuzzy lower approximation of R with respect to the θ operator is and the upper approximation The expression is: in, and are the membership and non-membership values ​​of object x in the optimistic case with respect to the fuzzy approximate relation R; r is a sub-relation in the relation R; m is the number of sub-relationships in the relation R; represents the optimistic approximation result of Z about the relation R; x is an object in the domain U; ​​μ represents the membership degree; v represents the non-membership degree; Description of the theta operator: Among them, θ 221 is called the θ operator, θ 365 is called θ * Operator, Then we have: Among them, y is an element in the domain U; ​​μ Z (y) represents the membership of element y with respect to evaluation value Z; v Z (y) represents the non-membership degree of element y with respect to evaluation value Z; Represents objects x and y with respect to the relation R r The membership degree of ; ∧ and ∨ represent conjunction and disjunction respectively; for For object x under m relations, the membership degree of x is combined with the θ operator to calculate IFS(U) is the set of all intuitionistic fuzzy sets in the domain U. Its optimistic decision rules based on the θ operator include: Accept decision rule: If and Then x∈POS O (Z); Delayed decision rule: If and Dissatisfied and Then x∈BND O (Z); Rejection decision rule: If and Then x∈NEG O (Z); Among them, POS O (Z) represents the positive domain of the evaluation value Z under the optimistic state; BND O (Z) represents the boundary region of the evaluation value Z under the optimistic state; NEG O (Z) represents the negative domain of the evaluation value Z under the optimistic state; α and β are evaluation threshold parameters; if the optimistic approximation value of object x under each evaluation dimension is greater than the threshold α, then the language or skill competence is judged to be acceptable; if the optimistic approximation value is less than the threshold α, then its competence is judged to be unacceptable; if the optimistic approximation value is between α and β, then its competence is judged to be uncertain, that is, it cannot be directly judged based on the existing information, and additional information is needed to make a judgment; Pessimistic case: In the intuitionistic fuzzy approximation space (U, V, R r ), R r is the fuzzy approximate equivalence relationship from U to V. R on the Pessimistic Multi-granularity Intuitionistic Fuzzy Lower Approximation of θ Operator and the upper approximation as follows: Among them, P represents pessimism or a pessimistic state; and The membership and non-membership values ​​of object x with respect to the fuzzy approximate relation R under pessimistic conditions are as follows: for y is an element in Z, and the pessimistic decision rule based on the θ operator includes: Accept decision rule: If and Then x∈POS p (Z); Delayed decision rule: If and Dissatisfied and Then x∈BND p (Z); Rejection decision rule: If and Then x∈NEG p (Z).

8. The talent competency assessment system according to claim 7, characterized in that: The power function is selected as the utility function of the sorting, and the utility matrix expression is: in, It is a parameter that controls the magnitude of the change; The regret-delight matrix C is expressed as: Among them, ψ is the regret aversion coefficient; The expression of the perceived utility value matrix G is: Where e is a constant.

9. The talent competency assessment system according to claim 8, characterized in that: The module M7 comprises: For a given set of n attributes A1, A2, ..., A n , where A i ={a1,a2,…,a n }, the standardized values ​​of each attribute data are B1, B2,…, B n , then: Calculate the information entropy of each attribute as E i : in, If p ri =0, then define Calculate the weight W of each attribute through information entropy i for:

10. The talent competency assessment system according to claim 9, characterized in that: The module M8 comprises: According to the attribute weights, the perceived utility functions of the objects under each attribute are weighted and superimposed, and the comprehensive evaluation risk value S of the jth object is j The expression is: Among them, W i is the weight of attribute i, g ji The perceived utility value of object j regarding attribute i.

Images