Search

CN-121980176-A - Multi-dimensional evaluation reliability aggregation calculation method integrating Kendell harmonic coefficients

CN121980176ACN 121980176 ACN121980176 ACN 121980176ACN-121980176-A

Abstract

The invention relates to the technical field of evaluation reliability calculation, in particular to a multi-dimensional evaluation reliability aggregation calculation method integrating Kendel and harmonic coefficients, which comprises the following steps of collecting multi-source evaluation information and arranging the multi-source evaluation information into an evaluation sequence in sequence to identify evaluation behavior characteristics, constructing an independent time window according to the total number of evaluation objects, calculating the Kendel and harmonic coefficients to quantify the evaluation consistency, constructing a three-dimensional interaction tensor based on the evaluation behavior characteristics and the Kendel coefficients, identifying initial reliability degree through tensor decomposition, calculating evaluation sequencing dispersion degree, constructing reliability decrementing relation, correcting and resetting the initial reliability degree to generate an evaluation reliability index, and establishing a reliability verification closed-loop mechanism according to the index to generate a reliability evaluation report; the invention realizes a comprehensive and accurate multi-dimensional evaluation reliability aggregation calculation method.

Inventors

  • ZENG BIN
  • YANG JINXIN

Assignees

  • 科睿特软件集团股份有限公司

Dates

Publication Date
20260505
Application Date
20260115

Claims (9)

  1. 1. The multi-dimensional evaluation reliability aggregation calculation method integrating Kendell harmonic coefficients is characterized by comprising the following steps of: s1, collecting multi-source evaluation information, arranging the multi-source evaluation information into an evaluation sequence according to a sequence, and identifying evaluation behavior characteristics of the evaluation sequence; S2, constructing an independent time window according to the total number of evaluation objects of the multi-source evaluation information, and calculating the Kendell harmony coefficient of the evaluation objects in the independent time window; Step S3, constructing a three-dimensional interaction tensor based on the evaluation behavior characteristics and Kendell harmonic coefficients, and recognizing the initial credibility of the evaluation information through three-dimensional interaction tensor decomposition; s4, calculating evaluation sequence dispersion degree of the multi-source evaluation information, and constructing a reliability decreasing relation through the evaluation sequence dispersion degree; S5, carrying out influence correction and reliability index resetting on the initial reliability based on the reliability decreasing relation to generate an evaluation reliability index; and step S6, establishing a credibility verification closed-loop mechanism according to the credibility index and generating a credibility evaluation report.
  2. 2. The multi-dimensional evaluation reliability aggregation calculation method of fusion kendell harmonic coefficients according to claim 1, wherein the specific steps of step S1 are as follows: Collecting multi-source evaluation information; performing time sequence synchronization processing on the multi-source evaluation information to obtain time alignment evaluation information; The time alignment evaluation information is arranged into an evaluation sequence according to the sequence; Performing scoring difference calculation on each adjacent sequence in the evaluation sequence, and mapping the scoring difference into the local fluctuation rate of the sequence scoring; and eliminating extreme ordering frequency in the evaluation sequence through the local fluctuation rate, and identifying evaluation behavior characteristics in the evaluation sequence.
  3. 3. The multi-dimensional evaluation reliability aggregation calculation method of the fusion kendell harmonic coefficient according to claim 1, wherein the specific steps of step S2 are as follows: Collecting the identification of the evaluation objects in the multi-source evaluation information, and counting the total number of the objects based on the identification of the evaluation objects to obtain the total number of the evaluation objects; Performing time window width matching calculation according to the total number of evaluation objects, and determining a time span threshold value of an independent time window; Constructing a plurality of independent time windows for the multi-source evaluation information through the time span threshold value of the independent time windows; performing pairwise ordering consistency comparison on the evaluation sequence in each independent time window, and calculating rank correlation coefficients among ordering objects; Performing Kendell coordination coefficient conversion on rank correlation coefficients among the sequencing objects to obtain Kendell coordination factors of evaluation objects in independent time windows; and arranging the Kendell harmony factors of the evaluation objects in all independent time windows in time sequence to generate Kendell harmony coefficients.
  4. 4. The multi-dimensional evaluation reliability aggregation calculation method of fusion kendell harmonic coefficients according to claim 1, wherein the specific steps of step S3 are as follows: Extracting a behavior mode and a behavior stability parameter in the behavior feature evaluation; Calculating a ranking order mean and a ranking order variance of each examination object based on the examination ranking sequence; tensor construction is carried out by taking the behavior stability parameter, the Kendel harmonic coefficient and the ranking order variance as three independent dimensions, and a three-dimensional interaction tensor is generated; calculating the coupling degree of behavior-harmony-ordering based on the three-dimensional interaction tensor, and mapping the coupling degree into an interaction influence coefficient; and carrying out reliability mapping on the multisource evaluation information according to the interaction influence coefficient to obtain an initial reliability degree.
  5. 5. The multi-dimensional evaluation reliability aggregation calculation method of fusion kendell harmonic coefficients according to claim 1, wherein the specific steps of step S4 are as follows: Extracting an evaluation sequencing result of each evaluation source in the multi-source evaluation information; performing sorting order difference calculation on the evaluation sorting result to obtain sorting order deviation; Performing variance calculation based on the ranking order deviation, and determining the evaluation ranking dispersion degree of the multi-source evaluation information through the variance; performing association analysis by evaluating the sorting dispersity and the initial credibility to obtain dispersity-credibility association data; Carrying out gradient change rate calculation on the dispersity-credibility association data, and extracting a decreasing gradient coefficient of the credibility changing along with the dispersity; And constructing a reliability decreasing relation according to the decreasing gradient coefficient.
  6. 6. The multi-dimensional evaluation reliability aggregation calculation method of fusion kendell harmonic coefficients according to claim 5, wherein the specific steps of constructing reliability decreasing relations according to decreasing gradient coefficients are as follows: Identifying gradient change amplitude in the decreasing gradient coefficient, and dividing the decreasing gradient coefficient into sections based on the gradient change amplitude to generate a distinguishable gradient interval set; Extracting the whole gradient decreasing direction and the local reversal point by comparing the ascending and descending directions of adjacent intervals in the gradient interval set so as to identify the gradient decreasing trend; Introducing a smooth correction mechanism at a local inversion point based on the gradient decreasing trend to obtain a steady decreasing trend line; extracting the attenuation rate and the stepwise inflection point of the steadily decreasing trend line through layering and accumulating gradient changes to generate a credibility decreasing track; And dividing a decreasing boundary and a steady-state interval according to the reliability decreasing track, and generating a reliability decreasing relation.
  7. 7. The multi-dimensional evaluation reliability aggregation calculation method of fusion kendell harmonic coefficients according to claim 1, wherein the specific steps of step S5 are as follows: performing reliability corresponding pairing on the basis of the reliability decreasing relation and the initial reliability data, and establishing a reliability comparison structure; according to the credibility comparison structure, carrying out segment-by-segment difference calculation and direction sequencing on the initial credibility data according to time and group dimensions to obtain a credibility change path sequence; Performing mutation point identification on the credibility change path sequence, and respectively applying a rebound weight and an equilibrium weight on the sudden drop section and the sudden rise section to generate an equilibrium adjustment sequence; And carrying out reliability resetting calculation based on the balance adjustment sequence, and generating an evaluation reliability index.
  8. 8. The method for calculating the multi-dimensional evaluation reliability of the fusion kendel harmonic coefficient according to claim 1, wherein the specific steps of performing mutation point identification on the reliability change path sequence and applying a rebound weight and an equilibrium weight to the sudden drop segment and the sudden rise segment respectively are as follows: Performing difference calculation between adjacent nodes in the credibility change path sequence, extracting the amplitude and direction information of credibility change, and generating a path fluctuation difference value; Performing mutation point identification on the credibility change path sequence based on the path fluctuation difference value to obtain position indexes of the sudden drop point and the sudden rise point; Positioning a sudden drop point interval of the credibility change path sequence based on the position index of the sudden drop point, and performing stabilization balance treatment on the sudden drop point interval to generate a credibility return-to-stability sequence; Positioning a sudden rise point section of the credibility change path sequence based on the position index of the sudden rise point, and performing convergence adjustment on the sudden rise point section to generate a credibility balance sequence; and merging and progressively correcting the credibility stabilizing sequence and the credibility balancing sequence to generate a balancing adjustment sequence.
  9. 9. The multi-dimensional evaluation reliability aggregation calculation method of fusion kendell harmonic coefficients according to claim 1, wherein the specific steps of step S6 are as follows: performing stage division processing on the evaluation reliability index data to obtain a stage reliability sequence; comparing the deviation of the bisection stage credibility sequences, and identifying a stable region and an abnormal region through front and rear stage difference detection; establishing a recurrence control relation based on the stable region and the abnormal region, and constructing a credibility verification closed-loop mechanism through the recurrence control relation; Determining a reliability deviation record by utilizing a reliability verification closed-loop mechanism; and extracting stability, fluctuation amplitude and recovery trend of each stage in the credibility deviation record, and aggregating the stability, fluctuation amplitude and recovery trend into a credibility assessment report.

Description

Multi-dimensional evaluation reliability aggregation calculation method integrating Kendell harmonic coefficients Technical Field The invention relates to the technical field of evaluation reliability calculation, in particular to a multi-dimensional evaluation reliability aggregation calculation method integrating Kendell harmonic coefficients. Background In the current evaluation reliability aggregation calculation method, the traditional technology depends on single-source information, so that the real situation of an evaluation object cannot be comprehensively reflected, the inaccuracy and reliability of an evaluation result are insufficient, and particularly when multi-dimensional and multi-source evaluation information is faced, the traditional method lacks an effective fusion mechanism, complexity among information is difficult to process, the one-sided property and the reliability of the evaluation result are caused, and the scientificity and the reliability of a decision are influenced. In addition, the existing auditing reliability calculation method usually focuses on the influence of time factors when processing behavior characteristics, and a dynamic auditing model cannot be effectively established, so that the method cannot adapt to a rapidly-changing environment in practical application, and particularly lacks sufficient consideration on auditing sequencing dispersion degree in the process of evaluating auditing images, so that the existing reliability evaluation result increases deviation, accurate reference relevance cannot be provided for users finally, and auditing effectiveness is influenced. In an actual application scene, the traditional method often lacks systematicness and comprehensiveness, so that an effective credibility mechanism cannot be established to verify, the application value of an audit result is weakened, the method particularly relates to the field of complex decision making, and the credibility of information is directly related to the accuracy and effect of decision making. Disclosure of Invention The invention provides a multi-dimensional evaluation reliability aggregation calculation method fused with Kendell harmonic coefficients to solve at least one technical problem. In order to achieve the above object, a multi-dimensional evaluation reliability aggregation calculation method integrating Kendell harmonic coefficients includes the following steps: s1, collecting multi-source evaluation information, arranging the multi-source evaluation information into an evaluation sequence according to a sequence, and identifying evaluation behavior characteristics of the evaluation sequence; S2, constructing an independent time window according to the total number of evaluation objects of the multi-source evaluation information, and calculating the Kendell harmony coefficient of the evaluation objects in the independent time window; Step S3, constructing a three-dimensional interaction tensor based on the evaluation behavior characteristics and Kendell harmonic coefficients, and recognizing the initial credibility of the evaluation information through three-dimensional interaction tensor decomposition; s4, calculating evaluation sequence dispersion degree of the multi-source evaluation information, and constructing a reliability decreasing relation through the evaluation sequence dispersion degree; S5, carrying out influence correction and reliability index resetting on the initial reliability based on the reliability decreasing relation to generate an evaluation reliability index; and step S6, establishing a credibility verification closed-loop mechanism according to the credibility index and generating a credibility evaluation report. According to the invention, by collecting multi-source evaluation information and identifying evaluation behavior characteristics, the comprehensiveness and diversity of data are ensured, the accuracy and the representativeness of evaluation are enhanced, the construction of an independent time window enables effective monitoring of dynamic changes of an evaluation object, the calculated Kendel harmony coefficient provides a quantification basis for subsequent analysis, the construction of a three-dimensional interaction tensor enables the relationship between different evaluation dimensions to be fully represented, the tensor decomposition technology can identify the initial reliability degree of the evaluation information, the reliability of the evaluation result is effectively improved, in the calculation of sequencing dispersion degree, the consistence and the stability of the evaluation result are revealed through analyzing the dispersion condition of the multi-source information, the constructed reliability decreasing relationship provides a scientific basis for dynamic adjustment of the reliability, the implementation of influence correction and reliability index reset is enabled to be optimized, the finally generated evaluation reliab