Dynamic weight optimization system and method based on difference proportion and hierarchical conservation

Abstract

The invention discloses a dynamic weight optimization system and method based on difference proportion and level conservation, and the method comprises the steps: setting a to-be-scored subject in a personnel quality scoring system as a first-level index, and setting a plurality of to-be-scored items under the to-be-scored subject as second-level indexes; and according to the guidance probability of the weight of each index unit in the secondary index and the original weight ratio of each index unit in the secondary index, obtaining a difference driving adjustment coefficient of each index unit in the secondary index, and according to the difference driving adjustment coefficient, carrying out optimization processing on the original weight of each index unit in the preset secondary index in the personnel quality scoring system. And obtaining the final optimization weight of each index unit in the secondary index. According to the method, the relative importance change among the indexes can be dynamically captured, the self-adaptive weight adjustment is realized while the hierarchical constraint is maintained, the static weighting logic of the traditional method is broken through, and a new way is provided for constructing an intelligent and explainable comprehensive evaluation system.

Description

Technical Field

[0001] The present invention relates to the field of data analysis, and in particular to a dynamic weight optimization system and method based on difference ratio and hierarchy conservation. Background Art

[0002] The evaluation of graduate students' qualities shouldn't be a single-item assessment; it should be a holistic and comprehensive assessment. Graduate students' qualities aren't simply the sum of various quality indicators; rather, they are an organic whole comprised of academic proficiency, physical and mental well-being, and so on. Therefore, when establishing a hierarchical analysis, it's necessary not only to comprehensively list all qualities but also to determine the weights of each indicator based on the relative impact of each factor.

[0003] The weight allocation mechanism of the comprehensive evaluation system for graduate students has always been a core problem. Most systems use a fixed weight indicator system, which cannot dynamically adapt to the differentiated requirements for academic literacy, physical and mental literacy, and other group abilities. Static weight models such as the Analytic Hierarchy Process and the Entropy Weight Method, although they have achieved indicator quantification to a certain extent, have significant defects in dynamic adjustment and logical consistency, resulting in a disconnect between the comprehensive evaluation results of graduate students and their actual abilities. In addition, existing weight optimization technologies generally suffer from the defects of insufficient interpretability and versatility. Most systems use neural networks and ensemble learning algorithms for weight training. Although they can improve the accuracy of indicator fitting, they cannot clearly explain the internal logic of weight adjustment.

[0004] Some algorithms attempt to improve system adaptability by introducing dynamic weighting mechanisms, such as automatically updating weights through sliding window analysis of historical data or incorporating fuzzy logic to process qualitative indicators. However, these approaches still have inherent limitations: time-series-based weight updates are susceptible to data sparsity and struggle to handle newly added evaluation dimensions. While fuzzy comprehensive evaluation can accommodate subjective judgment, its membership function design relies on expert experience to rank indicator importance, and cognitive biases among experts lead to significant subjectivity in weight assignment. Existing technologies lack a comprehensive theoretical framework that simultaneously addresses the coupling issues of weight logic preservation, constraint satisfaction, and model interpretability. As a result, comprehensive evaluation systems often struggle to achieve both accuracy and credibility in practical applications. These technical bottlenecks severely hinder the evolution of graduate student evaluation systems towards intelligent and personalized approaches. Technological innovation is urgently needed to construct a weight assignment mechanism that combines mathematical rigor, dynamic adaptability, and decision-making transparency. Summary of the Invention

[0005] The purpose of the present invention is to provide a dynamic weight optimization system based on difference ratio and hierarchy conservation on the one hand, and to provide a dynamic weight optimization method based on difference ratio and hierarchy conservation on the other hand. The system and method can dynamically capture the changes in relative importance between indicators, realize adaptive adjustment of weights while maintaining hierarchical constraints, break through the static weighting logic of traditional methods, and provide a new way to build an intelligent and explainable comprehensive evaluation system.

[0006] To achieve this goal, the present invention designs a dynamic weight optimization system based on difference ratio and hierarchy conservation, which includes:

[0007] The indicator setting module is used to set the subject to be scored in the personnel quality scoring system as a first-level indicator, and multiple items to be scored under the subject to be scored as second-level indicators;

[0008] The bidirectional probability mapping module is used to obtain the guiding probability of the weight of each indicator unit in the secondary indicator based on the opinion survey data based on the secondary indicator, and to obtain the original weight ratio of each indicator unit in the secondary indicator based on the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system. The opinion survey data based on the secondary indicator is the opinion survey data obtained by subjectively scoring the personnel to be scored using the secondary indicator;

[0009] The dynamic bias control module is used to obtain the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator;

[0010] The weight adjustment and conservation processing module is used to optimize the original weight of each indicator unit in the preset secondary indicators in the personnel quality scoring system according to the difference-driven adjustment coefficient, and obtain the final optimized weight of each indicator unit in the secondary indicators.

[0011] Furthermore, the method for obtaining the guidance probability of the weight of each indicator unit in the secondary indicator based on the opinion survey data based on the secondary indicators includes: the scorer sorts the importance of each indicator unit in the secondary indicator to obtain a score vector S = [n1, n1-1, ..., 1], wherein the first place in the importance sorting result scores n1, the second place in the importance sorting result scores n1-1, and so on, the last place in the importance sorting result gets 1 point, the score vector S is quantified to obtain a ranking score S1, the total scoring weight S2 of the scorer is obtained according to the score vector S, and the guidance probability of each indicator unit in the secondary indicator is obtained according to the ranking score S1 and the total scoring weight S2.

[0012] Furthermore, the method of quantizing the score vector S to obtain the ranking score S1 includes:

[0013]

[0014] Among them, n1 is the number of indicator units in the secondary indicator, k represents the ranking position of a single secondary indicator unit, and v k represents the number of scorers who ranked a single secondary indicator unit at position k;

[0015] The method of obtaining the total scoring weight S2 of multiple scorers according to the score vector S includes:

[0016]

[0017] Among them, n2 is the number of scorers selected, and n1 is the number of indicator units in the secondary indicator.

[0018] Furthermore, the method for obtaining the weighted guidance probability of each secondary indicator according to the ranking score S1 and the total scoring weight S2 includes:

[0019]

[0020] Among them, n1 is the number of indicator units in the secondary indicator, n2 is the number of scorers drawn, S2 is the total scoring power of n2 individuals, k represents the ranking position of a single secondary indicator unit, v k It represents the number of scorers who ranked a single secondary indicator unit at position k, and P1 is the guiding probability of the weight of each indicator unit in the secondary indicator.

[0021] Furthermore, the method for obtaining the original weight ratio of each indicator unit in the secondary indicator according to the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system includes:

[0022]

[0023]

[0024] Among them, W2 is the quantified result of the weight of a single secondary indicator unit in the original setting of the personnel quality scoring system; W1 is the quantified result of the first-level indicator in the original setting of the personnel quality scoring system; P2 is the original weight ratio of each indicator unit in the second-level indicator, It is the quantitative result of the first secondary indicator unit in the original setting of the personnel quality scoring system. is the quantitative result of the second secondary indicator unit in the original setting of the personnel quality scoring system, and so on. This is the quantitative result of the third secondary indicator unit in the original setting of the personnel quality scoring system. It is the quantitative result of the n1th secondary indicator unit in the original setting of the personnel quality scoring system.

[0025] Furthermore, the method of obtaining the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator includes:

[0026] W3=P1-P2

[0027] |W3|≤min{0.5P2,0.3(1-P2)},

[0028] Among them, W3 is the difference-driven adjustment coefficient of each indicator unit in the secondary indicator, P1 is the guiding probability of the weight of each indicator unit in the secondary indicator, P2 is the original weight ratio of each indicator unit in the secondary indicator, and |W3| is the bias constraint condition of W3, which ensures that the adjustment range of the weight of each indicator unit in the secondary indicator does not exceed 50% of the original weight ratio P2, avoiding the high-weight items of each indicator unit in the secondary indicator from losing their dominant position due to a sharp downward adjustment. At the same time, the upward adjustment space of the low-weight items of each indicator unit in the secondary indicator is limited to 0.3(1-P2) to prevent them from excessively crowding out the weights of other items.

[0029] Furthermore, the process of optimizing the original weights of each indicator unit in the secondary indicators preset in the personnel quality scoring system according to the difference-driven adjustment coefficient specifically includes performing a preliminary adjustment on the original weights of each indicator unit in the secondary indicators preset in the personnel quality scoring system, and scaling the original weights of each indicator unit in the secondary indicators after the preliminary adjustment;

[0030] The specific process of making preliminary adjustments to the original weights of each indicator unit in the preset secondary indicators in the personnel quality scoring system is as follows:

[0031] W4'=W2(1+W3)=W2+W2*W3,

[0032] Among them, W2 is the result of quantifying the weights of all secondary indicator weight units in the original setting of the personnel quality scoring system; W3 is the difference-driven adjustment coefficient of each indicator unit in the secondary indicator; W4' is the weight of each indicator unit in the secondary indicator after preliminary adjustment.

[0033] Furthermore, the specific process of scaling the original weights of each indicator unit in the secondary indicators after preliminary adjustment is as follows:

[0034]

[0035] Among them, W1 is the quantitative result of the first-level indicator in the original setting of the personnel quality scoring system, W4' is the weight of each indicator unit in the secondary indicator after preliminary adjustment, W4 is the final optimized weight of each indicator unit in the secondary indicator, and ∑W4' is the sum of the weights of each indicator unit in the secondary indicator after preliminary adjustment.

[0036] Furthermore, a dynamic weight optimization method based on difference ratio and hierarchy conservation based on the dynamic weight optimization system includes:

[0037] The subject to be scored in the personnel quality scoring system is set as the first-level indicator, and the multiple items to be scored under the subject to be scored are set as the second-level indicators;

[0038] The guiding probability of the weight of each indicator unit in the secondary indicator is obtained based on the opinion survey data based on the secondary indicator, and the original weight ratio of each indicator unit in the secondary indicator is obtained based on the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system. The opinion survey data based on the secondary indicator is the opinion survey data obtained by subjectively scoring the personnel to be scored using the secondary indicator;

[0039] Obtaining the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator;

[0040] According to the difference-driven adjustment coefficient, the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system is optimized to obtain the final optimized weight of each indicator unit in the secondary indicator.

[0041] The beneficial effects of the present invention are as follows: the traditional methods used in the existing comprehensive evaluation system for graduate students (such as the subjective weighting method and the entropy weight method) are easy to destroy the relative importance ranking between the original weights when adjusting the weights. The weight adjustment logic is disconnected from the original logic, resulting in inconsistent evaluation logic. Moreover, it is difficult to ensure the consistency of the sum of weights between levels after the weight adjustment (such as the sum of the weights of each option of the lower-level indicator must still be equal to the initial weight value of the upper-level indicator after adjustment and optimization), and repeated corrections are required. At the same time, some existing algorithms are mostly black box models, which cannot intuitively explain the adjustment logic, and are difficult to adapt to the parameter requirements in different scenarios, and lack interpretability and versatility.

[0042] The present invention proposes an adaptive weight optimization algorithm based on difference ratio. By quantitatively investigating and studying the difference ratio between the guidance probability and the original weight, the algorithm realizes the dynamic adjustment of the weight. During the adjustment process, the relative stability of the original weight ranking is ensured, and the hierarchical weights and constraints are strictly satisfied. A transparent and verifiable mathematical formula is provided, which is suitable for multi-scenario analysis of ranking-type multiple-choice questions. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a structural schematic diagram of the present invention;

[0044] Figure 2 This is the algorithm flow chart and application example diagram of the present invention. DETAILED DESCRIPTION

[0045] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0046] like Figure 1 As shown in FIG, a dynamic weight optimization system based on difference ratio and hierarchy conservation includes:

[0047] The indicator setting module is used to set the subject to be scored in the personnel quality scoring system as a first-level indicator, and multiple items to be scored under the subject to be scored as second-level indicators;

[0048] The bidirectional probability mapping module is used to obtain the guiding probability of the weight of each indicator unit in the secondary indicator based on the opinion survey data based on the secondary indicator, and to obtain the original weight ratio of each indicator unit in the secondary indicator based on the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system. The opinion survey data based on the secondary indicator is the opinion survey data obtained by subjectively scoring the personnel to be scored using the secondary indicator;

[0049] The dynamic bias control module is used to obtain the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator;

[0050] The weight adjustment and conservation processing module is used to optimize the original weight of each indicator unit in the preset secondary indicators in the personnel quality scoring system according to the difference-driven adjustment coefficient, and obtain the final optimized weight of each indicator unit in the secondary indicators.

[0051] The personnel quality scoring system includes a system for school tutors to score graduate students. The subjects to be scored (first-level indicators) can be set to one or more according to actual conditions.

[0052] Figure 2 To implement the algorithm flow chart and application example diagram of the dynamic weight optimization algorithm based on difference ratio and hierarchical conservation, the first-level weight W1 and the weight set of the second-level indicator unit to which the first-level weight belongs are input into the system. In the bidirectional probability mapping stage, the quantitative ranking score S1 of the weight of a single secondary indicator unit is calculated, and the total empowerment S2 is calculated. The guidance probability P1 of a single secondary indicator unit is calculated by the quantitative ranking score S1 of the secondary indicator and the total empowerment S2. At the same time, the original weight proportion P2 of a single secondary indicator unit is calculated according to the original weight of the secondary indicator unit preset in the scoring system. In the dynamic bias control stage, the difference-driven adjustment coefficient W3 of a single secondary indicator unit is calculated according to the guidance probability P1 of a single secondary indicator unit and the original weight proportion P2 of a single secondary indicator unit. W3=P1-P2, and the bias constraint condition of the difference-driven adjustment coefficient W3 is calculated. In the weight adjustment and conservation processing stage, the weight W4'=W2(1+W3)=W2+W2*W3 after preliminary optimization of each secondary indicator unit is calculated, and the weight of the i-th secondary indicator after preliminary optimization is put into Repeat the above steps to obtain the preliminary optimization weights of each secondary indicator unit to which the first-level weight belongs, and scale the preliminary optimization weights of each secondary indicator unit to obtain the final optimization weights of each secondary indicator unit. Finally, the final optimized weights of each secondary indicator unit to which the first-level weight belongs can be obtained

[0053] The above technical solution primarily addresses the survey and statistical problem of "ranking the importance of n1 options (secondary indicator units) in a question (first-level indicator) by selecting n2 individuals (sample number) to achieve effective data statistics and analysis." The n2 individuals (sample number) are the raters using the personnel quality rating system. The bidirectional probability mapping phase is the first stage of the algorithm. Essentially, it establishes two independent probability spaces: the guiding probability P1 and the original weight percentage P2 for any option in the question, mapping subjective survey data to objective, pre-set weights. The guiding probability P1 is achieved through ranking score quantization and total score normalization. Ranking score quantization maps the rater's ranking position for any option (i.e., any second-level indicator unit) in the question into a linear score, resulting in a score vector S. The ranking weights are then quantified using the score vector S = [n1, n1-1, ..., 1]. The higher the ranking position in the score vector S = [n1, n1-1, ..., 1], the higher the score, and the higher the importance of the option (i.e., any second-level indicator unit) in the question corresponding to that ranking position.

[0054] A method for obtaining the guidance probability of the weight of each indicator unit in the secondary indicator based on the opinion survey data of the secondary indicator includes: the scorer sorts the importance of each indicator unit in the secondary indicator to obtain a score vector S = [n1, n1-1, ..., 1], wherein the first place in the importance sorting result scores n1, the second place in the importance sorting result scores n1-1, and so on, the last place in the importance sorting result gets 1 point, the score vector S is quantified to obtain a ranking score S1, the total scoring weight S2 of the scorer is obtained according to the score vector S, and the guidance probability of each indicator unit in the secondary indicator is obtained according to the ranking score S1 and the total scoring weight S2.

[0055] The method of quantizing the score vector S to obtain the ranking score S1 includes:

[0056] Among them, n1 is the number of indicator units in the secondary indicator, k represents the ranking position of a single secondary indicator unit, and v k represents the number of scorers who ranked a single secondary indicator unit at position k;

[0057] The method of obtaining the total scoring weight S2 of multiple scorers according to the score vector S includes:

[0058]

[0059] Where n2 is the number of selected raters, and n1 is the number of indicator units in the secondary indicator. The process of obtaining the total scoring weight S2 of multiple raters based on the score vector S is the process of total score normalization. Total score normalization involves further calculating the total scoring weight S2 of n2 raters (samples) under the condition that the selected raters have a scoring weight of (n1+n1-1+,...,+1) for a given question. Then, through normalization, the quantitative ranking score S1 of each option (secondary indicator unit) is further converted into a guidance probability P1, eliminating the impact of sample size differences.

[0060] The method for obtaining the weighted guidance probability of each secondary indicator according to the ranking score S1 and the total scoring weight S2 includes:

[0061]

[0062] Among them, n1 is the number of indicator units in the secondary indicator, n2 is the number of scorers drawn, S2 is the total scoring power of n2 individuals, k represents the ranking position of a single secondary indicator unit, v k It represents the number of scorers who ranked a single secondary indicator unit at position k, and P1 is the guiding probability of the weight of each indicator unit in the secondary indicator.

[0063] In the above technical solution, the weight optimization algorithm aims to investigate and analyze the rationality of the weight ratio of each option (secondary indicator unit) relative to the total weight (first-level indicator) in the setting of multiple options (secondary indicator units) of a certain question (first-level indicator). Therefore, it is also necessary to obtain the original weight ratio P2. The method for obtaining the original weight ratio of each indicator unit in the second-level indicator based on the original weight of each indicator unit in the second-level indicator preset in the personnel quality scoring system includes:

[0064]

[0065] Among them, W2 is the quantified result of the weight of a single secondary indicator unit in the original setting of the personnel quality scoring system; W1 is the quantified result of the first-level indicator in the original setting of the personnel quality scoring system; P2 is the original weight ratio of each indicator unit in the second-level indicator, It is the quantitative result of the first secondary indicator unit in the original setting of the personnel quality scoring system. This is the quantitative result of the second secondary indicator unit in the original setting of the personnel quality scoring system. is the quantitative result of the third secondary indicator unit in the original setting of the personnel quality scoring system, and so on. This is the quantified result of the n1th secondary indicator unit in the original setup of the personnel quality scoring system. The original weight P2 is intended to be a benchmark anchor, providing a reference frame for algorithm adjustments and ensuring that weight adjustments do not deviate from the original design framework.

[0066] In the above technical solution, the dynamic bias control stage is the control stage of the algorithm. Its core function is to use the guidance probability P1 and original weight ratio P2 of a single option (secondary indicator unit) obtained in the previous stage, and further quantify the difference to obtain the difference-driven adjustment coefficient W3, which serves as the direct basis for weight bias to achieve data-driven dynamic adjustment. The method of obtaining the difference-driven adjustment coefficient of each indicator unit in the secondary indicator based on the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator includes:

[0067] W3=P1-P2

[0068] |W3|≤min{0.5P2,0.3(1-P2)},

[0069] Where W3 is the difference-driven adjustment coefficient for each indicator unit in the secondary indicator, P1 is the guiding probability of the weight of each indicator unit in the secondary indicator, P2 is the original weight proportion of each indicator unit in the secondary indicator, and |W3| is the bias constraint on W3. This ensures that the adjustment range of the weight of each indicator unit in the secondary indicator does not exceed 50% of the original weight proportion P2, preventing the high-weight items in each indicator unit from losing their dominance due to a significant downward adjustment. At the same time, the upward adjustment space of low-weight items in each indicator unit in the secondary indicator is limited to 0.3(1-P2) to prevent them from excessively crowding out the weight of other items. If W3 is a positive value, it indicates that the original weight should be adjusted upward based on the questionnaire results, so W4' increases; conversely, if W3 is a negative value, it indicates that the original weight should be adjusted downward based on the questionnaire results, so W4' decreases. This stage can clearly express the "degree of deviation of user opinions (sample survey opinions) from the original design" through simple difference calculations. It can also automatically generate difference coefficients based on actual data, eliminating the need for manual adjustment rules, and is intuitively interpretable and dynamically adaptable.

[0070] In the above technical solution, the weight adjustment and conservation processing stage is the core implementation and final calibration stage of the algorithm. Its core function is to integrate the sample survey opinions into the original weight data of each option based on the difference-driven adjustment coefficient W3 obtained in the previous stage. Specifically, it is through a two-stage adjustment strategy (preliminary adjustment and scaling processing) that the overall conservation of the hierarchical weight system is strictly guaranteed while achieving dynamic optimization of weights. The process of optimizing the original weights of each indicator unit in the secondary indicators preset in the personnel quality scoring system according to the difference-driven adjustment coefficient specifically includes preliminary adjustments to the original weights of each indicator unit in the secondary indicators preset in the personnel quality scoring system, and scaling the original weights of each indicator unit in the secondary indicators after the preliminary adjustment;

[0071] The specific process of making preliminary adjustments to the original weights of each indicator unit in the preset secondary indicators in the personnel quality scoring system is as follows:

[0072] W4'=W2(1+W3)=W2+W2*W3,

[0073] Among them, W2 is the result of quantifying the weights of all secondary indicator weight units in the original setting of the personnel quality scoring system; W3 is the difference-driven adjustment coefficient of each indicator unit in the secondary indicator; W4' is the weight of each indicator unit in the secondary indicator after preliminary adjustment.

[0074] During this phase, the weights of each option (secondary indicator unit) are freely adjusted based on the variance coefficient W3, maximizing the incorporation of user feedback. Scaling is achieved through the addition of 1+W3, avoiding the lack of sensitivity of low-weighted items within each option (secondary indicator unit) due to absolute difference adjustments, thus achieving multiplicative adjustment. Furthermore, based on dynamic bias control (|W3|), the relative order of the initially adjusted weights W4' of each option (secondary indicator unit) is adjusted to maintain a consistent order relative to the original quantified weights W2 of each option (secondary indicator unit), thus minimizing the influence of subjective opinions on objective facts.

[0075] In the above technical solution, in order to prevent and correct the problem of weight sum deviation that may be caused by the preliminary adjustment, that is, to ensure that the sum of the weights W4' of each option (secondary indicator unit) after adjustment and optimization is strictly equal to the original total weight (primary indicator) W1, a scaling technology is required. The specific process of scaling the original weights of each indicator unit in the secondary indicator after the preliminary adjustment is as follows:

[0076] Among them, W1 is the quantitative result of the first-level indicator in the original setting of the personnel quality scoring system, W4' is the weight of each indicator unit in the secondary indicator after preliminary adjustment, W4 is the final optimized weight of each indicator unit in the secondary indicator, and ∑W4' is the sum of the weights of each indicator unit in the secondary indicator after preliminary adjustment. The final optimized weight W4 of the final single option (secondary indicator unit) is obtained by first calculating the scaling factor = original total weight W1 / total weight of each option (secondary indicator unit) after optimization ∑W4', and then multiplying it by the preliminary adjusted weight W4' of the single option (secondary indicator unit). Through the scaling operation, the strict mathematical consistency of the multi-level weight system is ensured, avoiding "weight leakage" or "weight redundancy".

[0077] In the above technical solution, after obtaining the final optimized weights of each option (secondary indicator unit) in the scoring subject, the personnel quality scoring system can be put into use. The scoring can be calculated based on the specific scores of the scored personnel by the scorers to ensure the scientificity and fairness of the evaluation work. The calculation method of the basic score includes:

[0078] Weight distribution: First-level indicator weight: Assume that there are 5 first-level evaluation indicators, Y1, Y2, Y3, Y4, and Y5, and their weights are set to W1, W2, W3, W4, and W5 respectively.

[0079] Secondary indicator weight: For the i-th primary indicator, set its secondary indicator weight to W i1 ,W i2 ,…,W ij , under each first-level evaluation indicator, the sum of the weights of all second-level evaluation indicators is 1.

[0080] Evaluation point weight: The weight of each evaluation point is set to V ijk , where i is the first-level indicator, j is the second-level indicator, k is the k-th evaluation point under the j-th second-level indicator, and the sum of the weights of each evaluation point is 1.

[0081] Methods for calculating the basic score of the rated person based on the final optimized weights of each option (secondary indicator unit) and the score of the rater include:

[0082] Calculate the secondary indicator score: Score ij =∑S ijk ×V ijk ,

[0083] Among them, Score ij is the score of the jth secondary indicator under the ith first-level indicator, S ijk is the score of the kth evaluation point under the jth secondary indicator under the i-th primary indicator, V ijk The score weight of the kth evaluation point under the jth secondary indicator under the i-th primary indicator;

[0084] Calculate the first-level indicator score:

[0085] Among them, Score i is the score of the i-th first-level indicator, Score ij is the score of the jth secondary indicator under the ith first-level indicator, W ij The final optimized weight of the jth secondary indicator under the i-th primary indicator;

[0086] Calculate the base score:

[0087] Among them, Score is the basic score of the person being rated, Score i is the score of the i-th first-level indicator, W i is the score weight of the i-th first-level indicator.

[0088] In the above technical solution, in addition to the subjective score of the scorer, the scoring formula can be used to calculate the additional score of the scored person based on his or her own honors or other additional points. The additional score of the scored person is calculated based on the additional score items of the scored person, and the sum of the basic score and the additional score is used as the final score of the scored person. The method includes:

[0089] Assignment formula:

[0090] X is the maximum score of the additional score, m is the minimum score of the additional score; r is the participant ranking; N is the total number of evaluators, and AdditionalScore is the additional score;

[0091] FinalScore=Score+AdditionalScore,

[0092] Among them, FinalScore is the final score of the rated person, Score is the basic score of the rated person, and AdditionalScore is the additional score of the rated person.

[0093] Example 2

[0094] A preferred embodiment:

[0095] Question: The options for the graduate student assessment subject "Qualities" include: "1. Morality, 2. Physical Strength, 3. Teamwork, 4. Honor, 5. Practice, 6. Skills". Please write down the order you think is appropriate according to the importance of each option. Among them, the subject "Qualities" is a first-level indicator, and the six options are six second-level indicator units. This question has 6 options in total, and the number of respondents is 27. That is, n1=6, n2=27, a total of 6 options. The option ranked first in importance corresponds to the highest score vector of 6, and so on. The option ranked sixth in importance corresponds to the lowest score vector of 1. The score vector S=(6, 5, 4, 3, 2, 1).

[0096] (1) Calculation of the guidance probability P1: By statistically analyzing the questionnaire results, taking the option "morality" as an example, 18 people ranked morality first (score 18*6), 6 people ranked morality second (score 6*5), 2 people ranked morality third (score 2*4), and 1 person ranked morality fourth (score 1*3). Thus, S2 = (6+5+4+3+2+1)*27 = 567, S1 = 18*6+6*5+2*4+1*3 = 149, and P1 = S1 / S2 = 149 / 567 = 26.27%. Similarly, the guidance probability P1 of the remaining options can be obtained: morality 26.27%, physical strength 17.64%, teamwork 18.52%, honor 10.41%, skills 12.70%, and practice 14.46%.

[0097] (2) Calculate the original probability P2. In the preset scheme, the weight of the subject "Qualities" is 15, including 2 points for morality, 4 points for physical strength, 1 point for teamwork, 2 points for honor, 2 points for skills, and 4 points for practice. The calculation results of P2 are as follows:

[0098] Morality: 2 / 15 = 13.33%

[0099] Physical strength: 4 / 15=26.67%

[0100] Team: 1 / 15 = 6.67%

[0101] Honor: 2 / 15 = 13.33%

[0102] Skill: 2 / 15 = 13.33%

[0103] Practice: 4 / 15 = 26.67%

[0104] (3) Calculate the adjustment coefficient W3 and the adjusted weight W4, still taking "morality" as an example

[0105] Adjustment coefficient W3 ("morality") = 26.27% - 13.33% = 12.94% = 0.1294; if W3 is a positive value, it means that the original weight should be adjusted upward based on the results of the questionnaire, so W4' increases; conversely, if W3 is a negative value, it means that the original weight should be adjusted downward based on the results of the questionnaire, so W4' decreases.

[0106] Adjusted weight W4'("morality") = 2 + (0.1294*2) = 2.26

[0107] And so on, we can get the options W4':

[0108] Scaling is performed on {'Character': 2.26, 'Physical Strength': 3.64, 'Teamwork': 1.12, 'Honor': 1.94, 'Skills': 1.99, 'Practice': 3.51}: Summing all options W4' yields an optimized total weight of 14.46, which is not equal to the "Quality" weight W1 = 15. Further scaling is required. The process is as follows, still using "Character" as an example.

[0109] Corrected W4 ("morality") = W4' * (W1 / 14.46) = 2.26 * (15 / 14.46) = 2.34

[0110] And so on to get the final weight of each option:

[0111] {'Character': 2.34, 'Physical Strength': 3.78, 'Teamwork': 1.16, 'Honor': 2.01, 'Skills': 2.06, 'Practice': 3.64}

[0112] Example 3

[0113] A dynamic weight optimization method based on the dynamic weight optimization system and based on difference ratio and hierarchy conservation includes:

[0114] Step 1: Set the subject to be scored in the personnel quality scoring system as the first-level indicator, and set the multiple items to be scored under the subject to be scored as the second-level indicators;

[0115] Step 2: Obtain the guiding probability of the weight of each indicator unit in the secondary indicator based on the opinion survey data based on the secondary indicator, and obtain the original weight ratio of each indicator unit in the secondary indicator based on the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system. The opinion survey data based on the secondary indicator is the opinion survey data obtained by subjectively scoring the personnel to be scored using the secondary indicator;

[0116] Step 3: Obtain the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator;

[0117] Step 4: Optimize the original weights of each indicator unit in the secondary indicators preset in the personnel quality scoring system according to the difference-driven adjustment coefficient to obtain the final optimized weights of each indicator unit in the secondary indicators.

[0118] Example 4

[0119] A computer program product includes a computer program, which implements the steps of the method described in Example 2 when executed by a processor.

[0120] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

[0121] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0122] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1A system that specifies the functions of a box or boxes.

[0123] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction system that is implemented in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0124] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0125] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A dynamic weight optimization system based on difference ratio and hierarchy conservation, characterized in that: It includes: The indicator setting module is used to set the subject to be scored in the personnel quality scoring system as a first-level indicator, and multiple items to be scored under the subject to be scored as second-level indicators; The bidirectional probability mapping module is used to obtain the guiding probability of the weight of each indicator unit in the secondary indicator based on the opinion survey data based on the secondary indicator, and to obtain the original weight ratio of each indicator unit in the secondary indicator based on the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system. The opinion survey data based on the secondary indicator is the opinion survey data obtained by subjectively scoring the personnel to be scored using the secondary indicator; The dynamic bias control module is used to obtain the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator; The weight adjustment and conservation processing module is used to optimize the original weight of each indicator unit in the preset secondary indicators in the personnel quality scoring system according to the difference-driven adjustment coefficient, and obtain the final optimized weight of each indicator unit in the secondary indicators.

2. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 1, characterized in that: A method for obtaining the guidance probability of the weight of each indicator unit in the secondary indicator based on opinion survey data based on the secondary indicators includes: the scorer sorts the importance of each indicator unit in the secondary indicator to obtain a score vector S = [n1, n1-1,..., 1], wherein the first place in the importance sorting result scores n1, the second place in the importance sorting result scores n1-1, and so on, the last place in the importance sorting result gets 1 point, quantizing the score vector S to obtain a ranking score S1, obtaining the total scoring weight S2 of the scorer according to the score vector S, and obtaining the guidance probability of each indicator unit in the secondary indicator according to the ranking score S1 and the total scoring weight S2.

3. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 2, characterized in that: The method of quantizing the score vector S to obtain the ranking score S1 includes: Among them, n1 is the number of indicator units in the secondary indicator, k represents the ranking position of a single secondary indicator unit, and v k represents the number of scorers who ranked a single secondary indicator unit at position k; The method of obtaining the total scoring weight S2 of multiple scorers according to the score vector S includes: Among them, n2 is the number of scorers selected, and n1 is the number of indicator units in the secondary indicator.

4. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 2 or 3, characterized in that: The method for obtaining the weighted guidance probability of each secondary indicator according to the ranking score S1 and the total scoring weight S2 includes: Among them, n1 is the number of indicator units in the secondary indicator, n2 is the number of scorers drawn, S2 is the total scoring power of n2 individuals, k represents the ranking position of a single secondary indicator unit, v k It represents the number of scorers who ranked a single secondary indicator unit at position k, and P1 is the guiding probability of the weight of each indicator unit in the secondary indicator.

5. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 1, characterized in that: Methods for obtaining the original weight proportions of each indicator unit in the secondary indicators according to the original weights of each indicator unit in the secondary indicators preset in the personnel quality scoring system include: Among them, W2 is the quantified result of the weight of a single secondary indicator unit in the original setting of the personnel quality scoring system; W1 is the quantified result of the first-level indicator in the original setting of the personnel quality scoring system; P2 is the original weight ratio of each indicator unit in the second-level indicator, It is the quantitative result of the first secondary indicator unit in the original setting of the personnel quality scoring system. This is the quantitative result of the second secondary indicator unit in the original setting of the personnel quality scoring system. is the quantitative result of the third secondary indicator unit in the original setting of the personnel quality scoring system, and so on. It is the quantitative result of the n1th secondary indicator unit in the original setting of the personnel quality scoring system.

6. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 1, characterized in that: The method for obtaining the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator includes: W3=P1-P2 |W3|≤min{0.5P2,0.3(1-P2)}, Among them, W3 is the difference-driven adjustment coefficient of each indicator unit in the secondary indicator, P1 is the guiding probability of the weight of each indicator unit in the secondary indicator, P2 is the original weight ratio of each indicator unit in the secondary indicator, and |W3| is the bias constraint condition of W3, which ensures that the adjustment range of the weight of each indicator unit in the secondary indicator does not exceed 50% of the original weight ratio P2, avoiding the high-weight items of each indicator unit in the secondary indicator from losing their dominant position due to a sharp downward adjustment. At the same time, the upward adjustment space of the low-weight items of each indicator unit in the secondary indicator is limited to 0.3(1-P2) to prevent them from excessively crowding out the weights of other items.

7. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 6, characterized in that: The process of optimizing the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system according to the difference-driven adjustment coefficient specifically includes preliminarily adjusting the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system, and scaling the original weight of each indicator unit in the secondary indicator after the preliminarily adjustment; The specific process of making preliminary adjustments to the original weights of each indicator unit in the preset secondary indicators in the personnel quality scoring system is as follows: W4'=W2(1+W3)=W2+W2*W3, Among them, W2 is the result of quantifying the weights of all secondary indicator weight units in the original setting of the personnel quality scoring system; W3 is the difference-driven adjustment coefficient of each indicator unit in the secondary indicator; W4' is the weight of each indicator unit in the secondary indicator after preliminary adjustment.

8. The dynamic weight optimization system based on difference ratio and hierarchy conservation according to claim 7, characterized in that: The specific process of scaling the original weights of each indicator unit in the secondary indicators after preliminary adjustment is as follows: Among them, W1 is the quantitative result of the first-level indicator in the original setting of the personnel quality scoring system, W4' is the weight of each indicator unit in the secondary indicator after preliminary adjustment, W4 is the final optimized weight of each indicator unit in the secondary indicator, and ∑W4' is the sum of the weights of each indicator unit in the secondary indicator after preliminary adjustment.

9. A dynamic weight optimization method based on difference ratio and hierarchy conservation based on the dynamic weight optimization system according to any one of claims 1 to 8, characterized in that: It includes: The subject to be scored in the personnel quality scoring system is set as the first-level indicator, and the multiple items to be scored under the subject to be scored are set as the second-level indicators; The guiding probability of the weight of each indicator unit in the secondary indicator is obtained based on the opinion survey data based on the secondary indicator, and the original weight ratio of each indicator unit in the secondary indicator is obtained based on the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system. The opinion survey data based on the secondary indicator is the opinion survey data obtained by subjectively scoring the personnel to be scored using the secondary indicator; Obtaining the difference driving adjustment coefficient of each indicator unit in the secondary indicator according to the guidance probability of the weight of each indicator unit in the secondary indicator and the original weight ratio of each indicator unit in the secondary indicator; According to the difference-driven adjustment coefficient, the original weight of each indicator unit in the secondary indicator preset in the personnel quality scoring system is optimized to obtain the final optimized weight of each indicator unit in the secondary indicator.

10. A computer program product comprising a computer program / instruction, which implements the steps of the method according to claim 9 when the computer program / instruction is executed by a processor.

Images