Search

CN-122023016-A - Enterprise digital asset comprehensive management method and system

CN122023016ACN 122023016 ACN122023016 ACN 122023016ACN-122023016-A

Abstract

The invention discloses a method and a system for comprehensively managing enterprise digital assets, which relate to the technical field of computers, and the method aligns risk indexes and associated factors uniformly in time in a reporting period and forms multi-element time sequence data input, so that subsequent analysis is unfolded on the same caliber and the same time axis, and interpretation deviation and repeated arrangement cost caused by cross-system caliber difference are reduced; in addition, resampling stable selection is introduced to form a stable edge set, so that causal graph generation is converted into structure screening with consistent frequency by single solution, the probability of accidental edges entering an interpretation link is reduced, and the stability and consistency of cross-batch and cross-scene analysis results are improved.

Inventors

  • CAI CHAO
  • ZHANG LIZHI

Assignees

  • 深圳轻松住信息科技有限公司

Dates

Publication Date
20260512
Application Date
20260126

Claims (10)

  1. 1. An enterprise digital asset integrated management method, comprising: Step S1, when a request for generating a risk analysis statement for a target asset combination is received, acquiring and time-aligning multi-element time series data of the target asset combination in a designated reporting period, wherein the multi-element time series data at least comprises a risk index sequence and a plurality of association factor sequences corresponding to the risk index sequence; Step S2, executing causal discovery processing on a computer based on the multivariate time series data to obtain a directed causal graph representing the association factors to the risk indexes, and determining at least one directed path from the directed causal graph as a key causal structure; the nodes of the directed causal graph respectively correspond to variables of the risk index sequence and the association factor sequence; S3, extracting the change characteristics of the risk index sequence in the appointed report period, and encoding the change characteristics and the key cause and effect structure together as conditional input; And S4, inputting the conditions into a trained risk narration generating model, and generating and outputting a risk analysis narration text corresponding to the appointed report period.
  2. 2. The method of claim 1, wherein the time alignment includes frequency unifying, missing value handling, and outlier handling of the risk indicator sequence and the association factor sequence, and appending a source system identification and a timestamp identification for each data point.
  3. 3. The method of claim 1, wherein the causal discovery process comprises constructing a continuous optimization problem comprising sparsity constraints and loop-free constraints to learn an adjacency matrix representing causal relationships of variables, and solving to obtain the directed causal graph; The continuous optimization problem expresses the variable as a combination of a current value and a plurality of hysteresis values when the hysteresis variable is expanded so as to allow the hysteresis effect to be expressed by a span directed edge, and the loop-free constraint is only applied to a causal structure corresponding to the same time slice, so that the directed causal graph output by the same time slice meets the directed loop-free interpretation requirement.
  4. 4. A method of integrated management of enterprise digital assets as claimed in claim 3, wherein a forbidden constraint is introduced into the continuous optimization problem, the forbidden constraint being generated by a financial domain knowledge pattern for prohibiting occurrence of a causal edge defined as not allowed by a knowledge pattern in the directed causal graph.
  5. 5. The method of claim 1, wherein determining a critical causal structure comprises extracting directed paths from the directed causal graph that terminate at risk indicator nodes, calculating a causal strength score for each directed path, and selecting a predetermined number of directed paths with highest causal strength scores as the critical causal structure.
  6. 6. The method of claim 1, wherein the change feature comprises identifying significant change points of the risk indicator sequence in the designated reporting period by using a change point detection algorithm, and outputting statistics of subsequences before and after each change point.
  7. 7. The method for integrated management of enterprise digital assets according to claim 4, wherein said risk narrative generation model is an autoregressive language model based on a Transformer framework, said model is supervised and trained by pairing historical risk report paragraphs with corresponding standardized risk data templates and key causal structures to form training samples, so that the generated text is constrained by said conditional inputs; And after solving the continuous optimization problem, carrying out resampling stability selection processing on the causal edges, counting the causal edge occurrence frequency obtained by resampling for a plurality of times, screening a stability edge set, and generating a directed causal graph by using the stability edge set or taking the stability edge set as an input foundation for extracting a subsequent key causal structure.
  8. 8. The method of claim 1, further comprising the step of tracing by associating and storing a statement in the risk analysis statement with a path segment in the critical causal structure supporting the statement and a corresponding source data time window identifier to form a queriable evidence chain.
  9. 9. A computer-readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1-8.
  10. 10. An enterprise digital asset integrated management system, based on the method of any one of claims 1-8, comprising: the data interface module is used for acquiring and time-aligning multi-element time sequence data of the target asset combination in a designated reporting period; a cause and effect analysis module to perform a cause and effect discovery process based on the multivariate time series data to generate a directed cause and effect graph and to determine a key cause and effect structure; The feature extraction module is used for extracting the change features of the risk index sequence; The statement generating module is used for encoding the key cause and effect structure and the change characteristic into conditional input and inputting the conditional input and the conditional input into the risk statement generating model to output a risk analysis statement text; And the traceability storage module is used for establishing an evidence chain link between the statement in the risk analysis narration text and the path segment and source data time window mark.

Description

Enterprise digital asset comprehensive management method and system Technical Field The invention relates to the technical field of computers, in particular to a comprehensive management method and system for enterprise digital assets. Background In the scene facing financial supervision reporting and internal risk management, the existing enterprise digital asset (such as digital representation of assets, liabilities and related risk exposure) comprehensive management system can generally collect and integrate multi-source data such as transactions, valuations, market data, credit and quota, risk metering results and the like, and form a risk signboard/instrument panel through a visual analysis component so as to support risk monitoring and report output. The prior proposal generally combines a risk metering model with a chart display for presenting the trend, the distribution characteristics and part of statistical association relation of the indexes in order to meet the requirements. However, in the aspect of generating descriptive contents such as management layer analysis and conclusion in a monitoring report, the conventional scheme depends on a fixed template or regularized interpretation of a single index, so that it is difficult to establish traceable attribution links between risk changes and driving factors (such as market factor changes, bin holding structure adjustment, parameter/model changes, business events and external impacts) of cross-system and cross-caliber, and form consistent and recheckable coherent descriptions, therefore, in the actual compiling process, risk management personnel still need to manually interpret multi-source data based on experience, connect events in series with text organizations, and when multiple scenes, rapid iterative analysis and repeated reporting are required to be carried out on the same asset combination, the problems of inconsistent analysis frame, insufficient reusability, long compiling period and the like easily occur. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. The invention provides a comprehensive management method and system for enterprise digital assets, which solve the problems that multisource risk data is complete in visualization, but supervision and description attribution still depends on manual work and is difficult to iterate. In order to solve the technical problems, the invention provides the following technical scheme: In a first aspect, an embodiment of the present invention provides a method for integrated management of digitized assets of an enterprise, including: Step S1, when a request for generating a risk analysis statement for a target asset combination is received, acquiring and time-aligning multi-element time series data of the target asset combination in a designated reporting period, wherein the multi-element time series data at least comprises a risk index sequence and a plurality of association factor sequences corresponding to the risk index sequence; Step S2, executing causal discovery processing on a computer based on the multivariate time series data to obtain a directed causal graph representing the association factors to the risk indexes, and determining at least one directed path from the directed causal graph as a key causal structure; the nodes of the directed causal graph respectively correspond to variables of the risk index sequence and the association factor sequence; S3, extracting the change characteristics of the risk index sequence in the appointed report period, and encoding the change characteristics and the key cause and effect structure together as conditional input; And S4, inputting the conditions into a trained risk narration generating model, and generating and outputting a risk analysis narration text corresponding to the appointed report period. The method for comprehensively managing the enterprise digital assets is characterized in that time alignment comprises the steps of carrying out frequency unification, missing value processing and outlier processing on the risk index sequence and the association factor sequence, and adding a source system identifier and a timestamp identifier to each data point. The causal discovery process comprises constructing a continuous optimization problem containing sparsity constraint and loop-free constraint to learn an adjacency matrix representing variable causal relation, and obtaining the directed causal graph by solving; The continuous optimization problem expresses the variable as a combination of a current value and a plurality of hysteresis values when the hysteresis variable is expanded so as to allow the hysteresis effect to be expressed by a span directed edge, and the loop-free constraint is only applied to a causal structure corresponding to the same time slice, so that the directed causal graph output by the same time slice meets the directed loop-free interpret