Search

CN-121996978-A - Method, device, equipment, medium and system for identifying decision deviation degree in real time based on interactive negative feedback

CN121996978ACN 121996978 ACN121996978 ACN 121996978ACN-121996978-A

Abstract

The application provides a method, a device, equipment, a medium and a system for identifying decision deviation degree in real time based on interactive negative feedback. The method comprises the steps of obtaining a target decision sequence, comparing a corresponding real-time decision feature flow with a preset reference decision criterion, identifying decision deviation points and associated deviation attribute features, obtaining an attribution option set, pushing the attribution option set to an external interactive entity, receiving a attribution selection signal fed back, obtaining task context features, performing dimension mapping processing on the attribution selection signal and the task context features, identifying feature deviation types, updating feature deviation nodes in a feature probability graph model, retrieving an updated feature probability graph model, extracting probability state values, performing distribution quantization, and outputting quantized deviation feature vectors. According to the application, through introducing interactive feedback and probability graph modeling, real-time objective identification of decision deviation degree and quantitative measurement of deviation intensity are realized, and the technical bottleneck that the traditional scheme is lagged in identification and the cognitive deep motivation cannot be quantized is effectively solved.

Inventors

  • Request for anonymity

Assignees

  • 北京认知涌现科技有限公司

Dates

Publication Date
20260508
Application Date
20260130

Claims (10)

  1. 1. The real-time decision deviation recognition method based on interactive negative feedback is characterized by comprising the following steps of: step 1, acquiring a target decision sequence, and comparing a corresponding real-time decision feature stream with a preset reference decision criterion to identify decision deviation points and associated deviation attribute features in the target decision sequence; Step 2, acquiring an attribution option set associated with the decision deviation point, triggering an interaction interface to push the attribution option set to an external interaction entity, and receiving attribution selection signals fed back by the external interaction entity; Step 3, acquiring task context characteristics associated with the decision deviation points, performing dimension mapping processing on the attribution selection signals and the task context characteristics to identify characteristic deviation types aiming at the decision deviation points, and updating characteristic deviation nodes in a preset characteristic probability graph model by utilizing the characteristic deviation types; And 4, calling the updated characteristic probability map model, extracting probability state values of the characteristic deviation nodes, and executing distribution quantization processing to generate quantized deviation characteristic vectors aiming at the target decision sequence.
  2. 2. The method of claim 1, wherein the reference decision criteria is obtained by performing feature modeling on expert decision logic within the field of task, comprising a sequence of standard states characterizing an ideal decision path and associated logic trigger thresholds.
  3. 3. The interactive negative feedback based decision bias real-time identification method according to claim 1, wherein the task context features include an external environmental noise component when executing the target decision sequence, a task urgency parameter, and a historical decision bias record corresponding to the target decision sequence.
  4. 4. The method for identifying the decision deviation degree in real time based on interactive negative feedback according to claim 1, wherein updating the feature deviation node in the preset feature probability map model by using the feature deviation type specifically comprises: invoking prior probability distribution of each characteristic deviation node in the characteristic probability map model; And performing posterior probability mapping on the prior probability distribution by using the characteristic deviation type as observation evidence through a Bayesian updating algorithm to generate a deviation intensity value of each characteristic deviation node which is updated as the probability state value.
  5. 5. The method for real-time recognition of decision bias based on interactive negative feedback according to claim 1, further comprising, before step 2: calculating the numerical deviation amplitude of the real-time decision feature stream relative to the reference decision criterion; And responding to the deviation amplitude of the numerical value exceeding a preset interaction triggering threshold, executing the action of triggering an interaction interface to push the attribution option set to an external interaction entity.
  6. 6. The interactive negative feedback based decision bias real-time identification method according to claim 1, wherein the attribution option set comprises a plurality of attribution tag components preset based on a cognitive feature library; the pushing the attribution option set specifically includes: matching a target number of the attribution tag components from the cognitive feature library based on the deviation attribute features associated with the decision deviation points; and performing structured encapsulation on the matched attribution label component, and sending the attribution label component to the external interaction entity through the interaction interface.
  7. 7. The utility model provides a decision deviation real-time identification device based on interactive negative feedback which characterized in that includes: the deviation recognition module is used for acquiring a target decision sequence, and comparing the corresponding real-time decision feature flow with a preset reference decision criterion to recognize decision deviation points in the target decision sequence and deviation attribute features associated with the decision deviation points; the interactive feedback module is used for acquiring an attribution option set associated with the decision deviation point, triggering an interactive interface to push the attribution option set to an external interactive entity, and receiving a feedback attribution selection signal; The model updating module is used for acquiring task context characteristics associated with the decision deviation points, performing dimension mapping on the attribution selection signals and the task context characteristics to identify characteristic deviation types aiming at the decision deviation points, and updating characteristic deviation nodes in a preset characteristic probability graph model by utilizing the characteristic deviation types; And the quantization identification module is used for retrieving the updated characteristic probability map model, extracting probability state values of each characteristic deviation node and executing distribution quantization processing to generate a quantization deviation characteristic vector aiming at the target decision sequence.
  8. 8. An electronic device comprising a memory for storing a computer program and a processor for implementing the steps of the method according to any one of claims 1-6 when the program is executed.
  9. 9. A computer-readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method according to any one of claims 1-6.
  10. 10. The utility model provides a decision deviation real-time identification system based on interactive negative feedback which characterized in that includes: the decision monitoring end is used for collecting real-time decision feature streams generated by the target decision sequences in real time; The cloud interaction engine is used for comparing the real-time decision feature flow with a reference decision criterion to identify decision deviation points and deviation attribute features associated with the decision deviation points and pushing attribution option sets to associated external interaction entities; the cognitive modeling center is used for receiving attribution selection signals fed back by the external interaction entity, and updating the feature probability map model by combining task context features so as to produce quantized deviation feature vectors; Wherein the feature probability map model is composed of a plurality of topology nodes representing the feature deviation nodes, each of the topology nodes performing a state update by the feature deviation type.

Description

Method, device, equipment, medium and system for identifying decision deviation degree in real time based on interactive negative feedback Technical Field The application relates to the technical field of artificial intelligence and cognitive computing, in particular to a method, a device, equipment, a medium and a system for identifying decision deviation degree in real time based on interactive negative feedback. Background With the rapid development of financial science and technology and personalized education systems, intelligent decision support systems are increasingly widely applied in the fields of personal investment financial management, professional skill training and the like. The system needs to identify the intrinsic cognitive rules and potential bias by analyzing the decision-making behavior of the user, thereby providing personalized guiding suggestions. In the existing cognitive analysis scheme, a statistical analysis method based on passive behavior tracks such as click flow and residence time is generally adopted. The method comprises the steps of firstly collecting operation logs of users on an interactive interface, extracting simple behavior indexes, then classifying the indexes by using a preset classification algorithm, defining static preference labels of the users, and finally matching fixed prompt information according to the labels. However, this passive analysis scheme has significant technical drawbacks. Meanwhile, the scheme lacks a real-time feedback correction mechanism, the cognitive model cannot be dynamically adjusted according to the context of the current task, so that the recognition of the decision deviation degree of a user is lagged and has low objectivity, and the accurate deviation correction requirement in a complex decision scene cannot be met. Disclosure of Invention In order to solve the technical problems, the application provides a method, a device, equipment, a medium and a system for identifying decision deviation degree in real time based on interactive negative feedback so as to at least alleviate the technical problems. A decision deviation degree real-time identification method based on interactive negative feedback comprises the steps of 1, obtaining a target decision sequence, comparing a corresponding real-time decision feature stream with a preset reference decision criterion to identify a decision deviation point in the target decision sequence and a deviation attribute feature associated with the decision deviation point, 2, obtaining an attribution option set associated with the decision deviation point, triggering an interactive interface to push the attribution option set to an external interactive entity, receiving attribution selection signals fed back by the external interactive entity, 3, obtaining task context features associated with the decision deviation point, performing dimension mapping processing on the attribution selection signals and the task context features to identify a feature deviation type for the decision deviation point, updating feature deviation nodes in a preset feature probability graph model by utilizing the feature deviation type, and 4, extracting probability state values of the feature deviation nodes after updating, and performing distribution quantization processing to produce a deviation feature vector for the target decision sequence. Optionally, the reference decision criterion is obtained by performing feature modeling on expert decision logic within the field of task, comprising a sequence of standard states characterizing an ideal decision path and associated logic trigger thresholds. Optionally, the task context feature includes an external ambient noise component when the target decision sequence is executed, a task urgency parameter, and a historical decision bias record corresponding to the target decision sequence. Optionally, updating the feature deviation nodes in the preset feature probability map model by using the feature deviation type specifically includes calling prior probability distribution of each feature deviation node in the feature probability map model, and performing posterior probability mapping on the prior probability distribution by using the feature deviation type as observation evidence through a Bayesian updating algorithm to generate a deviation intensity value of each feature deviation node updated as the probability state value. Optionally, before the step 2, the method further comprises the steps of calculating a value deviation amplitude of the real-time decision feature stream relative to the reference decision criterion, and executing the action of triggering an interaction interface to push the attribution option set to an external interaction entity in response to the value deviation amplitude exceeding a preset interaction triggering threshold. Optionally, the attribution option set includes a plurality of attribution tag components preset based on a cognitive featu