Search

CN-121980155-A - Method, system, electronic equipment and storage medium for predicting cut-over optical cable result

CN121980155ACN 121980155 ACN121980155 ACN 121980155ACN-121980155-A

Abstract

The invention provides a method, a system, electronic equipment and a storage medium for predicting a cut-over optical cable result, which belong to the technical field of artificial intelligence and comprise the steps of extracting a plurality of basic characteristics from multi-mode data related to the cut-over optical cable, and analyzing the multi-mode data to obtain a plurality of enhancement characteristics; the method comprises the steps of carrying out feature cross combination processing on a plurality of basic features and a plurality of enhancement features to generate a plurality of associated features, integrating the plurality of basic features, the plurality of enhancement features and the plurality of associated features to generate a multi-dimensional feature set, inputting the multi-dimensional feature set into an artificial intelligent model to obtain a prediction result output by the artificial intelligent model, wherein the artificial intelligent model comprises a main prediction model, an auxiliary optimization model and a large language model. According to the invention, a multi-dimensional characteristic interaction system integrating multi-mode data is constructed, and a hybrid prediction framework based on the cooperation of an AI large model and traditional machine learning is adopted, so that high-precision cutting-connection fruit prediction is realized.

Inventors

  • Hou Lingbing
  • GONG ZHI

Assignees

  • 浪潮通信信息系统有限公司

Dates

Publication Date
20260505
Application Date
20251126

Claims (13)

  1. 1. The method for predicting the result of cutting the optical cable is characterized by comprising the following steps: Extracting a plurality of basic features from multi-mode data related to the cut-over optical cable, and analyzing the multi-mode data to obtain a plurality of enhancement features; Performing feature cross combination processing on the plurality of basic features and the plurality of enhancement features to generate a plurality of associated features; Integrating the plurality of basic features, the plurality of enhancement features and the plurality of association features to generate a multi-dimensional feature set; the multi-dimensional feature set is input into an artificial intelligent model to obtain a prediction result output by the artificial intelligent model, wherein the artificial intelligent model comprises a main prediction model, an auxiliary optimization model and a large language model, the main prediction model is obtained by fine tuning a multi-mode large model based on a multi-mode data sample, and the auxiliary optimization model adopts a random forest algorithm or a gradient lifting tree algorithm.
  2. 2. The method of claim 1, wherein the multimodal data includes historical cutover data, resource status information, network topology data, device performance data, real-time operation and maintenance data, text-like data, and time-sequence-related data.
  3. 3. The method of claim 2, wherein extracting the plurality of base features from the multi-modal data associated with the drop cable comprises: extracting the type and the core number of the optical cable from the network topology data and the resource state information; extracting a planned cut duration and a buffer duration from the historical cut data; extracting the current resource occupancy rate and the historical peak value from the resource state information; and extracting the ambient temperature and humidity from the real-time operation and maintenance data.
  4. 4. The method for predicting the outcome of a cut-over optical cable according to claim 2, wherein the analyzing the multi-modal data to obtain a plurality of enhancement features comprises: constructing an association relation between the network topology data and the fault cases in the historical cut-over data by utilizing a knowledge graph technology, and calculating the similarity between the current cut-over scheme and the historical fault cases to output a fault association coefficient as a risk association characteristic; Predicting network flow and resource load in the time sequence associated data by adopting a long-short-term memory network, and generating time sequence prediction characteristics including flow peak value offset and load super-threshold probability; and carrying out semantic analysis on the text data through a large language model, and extracting operation risk keywords and experience weight values as text semantic features.
  5. 5. The method for predicting a cut-over cable result according to claim 1, wherein the performing feature cross-combination processing on the plurality of base features and the plurality of enhancement features to generate a plurality of associated features comprises: Performing multi-order interaction operation on the plurality of basic features and the plurality of enhancement features to generate a plurality of high-order interaction features; calculating SHAP values of the high-order interaction features; and eliminating the high-order interaction features with the SHAP value lower than a preset threshold value, and determining the remaining high-order interaction features as the associated features.
  6. 6. The method for predicting the result of a cut optical cable according to claim 1, wherein the inputting the multi-dimensional feature set into an artificial intelligence model to obtain the predicted result output by the artificial intelligence model comprises: inputting the multi-dimensional feature set into the main prediction model to obtain a first result output by the main prediction model and the confidence of the first result; when the confidence coefficient is lower than a preset threshold value, inputting the multidimensional feature set into the auxiliary optimization model to obtain a second result output by the auxiliary optimization model; Performing integrated voting on the first result and the second result to generate a third result; And inputting the third result into the large language model to obtain a predicted result which is output by the large language model and contains natural language description.
  7. 7. The method of predicting the outcome of a cut-over fiber optic cable of claim 6, further comprising: and when the confidence coefficient is not lower than the preset threshold value, determining the first result as the third result.
  8. 8. The method of claim 6, wherein the integrally voting the first result with the second result to generate a third result comprises: Performing feature importance analysis on the multi-dimensional feature set by using the auxiliary optimization model, and adjusting weights of the main prediction model and the auxiliary optimization model according to analysis results; And carrying out weighted calculation on the first result and the second result based on the adjusted weight to obtain the third result.
  9. 9. The method of claim 6, wherein inputting the third result into the large language model generates the predicted result comprising a natural language description, comprising: Mapping the numerical value in the third result into a natural language text to generate a risk interpretation report, wherein the risk interpretation report comprises descriptions of characteristic factors causing risks; invoking a pre-constructed knowledge graph, and matching a risk association path according to the third result; And combining the risk association path with the risk interpretation report to generate the prediction result.
  10. 10. A cut-over cable outcome prediction system, comprising: The first processing module is used for extracting a plurality of basic features from multi-mode data related to the cut-over optical cable, and analyzing the multi-mode data to obtain a plurality of enhancement features; The second processing module is used for carrying out feature cross combination processing on the plurality of basic features and the plurality of enhancement features to generate a plurality of associated features; the third processing module is used for integrating the plurality of basic features, the plurality of enhancement features and the plurality of association features to generate a multi-dimensional feature set; The fourth processing module is used for inputting the multi-dimensional feature set into an artificial intelligent model to obtain a prediction result output by the artificial intelligent model, wherein the artificial intelligent model comprises a main prediction model, an auxiliary optimization model and a large language model, the main prediction model is obtained by fine tuning a multi-mode large model based on a multi-mode data sample, and the auxiliary optimization model adopts a random forest algorithm or a gradient lifting tree algorithm.
  11. 11. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, wherein the processor implements the method of splice cable outcome prediction of any of claims 1 to 9 when executing the computer program.
  12. 12. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the method of splice cable result prediction of any of claims 1 to 9.
  13. 13. A computer program product comprising a computer program which, when executed by a processor, implements the method of predicting the outcome of a spliced cable as claimed in any one of claims 1 to 9.

Description

Method, system, electronic equipment and storage medium for predicting cut-over optical cable result Technical Field The invention relates to the technical field of artificial intelligence, in particular to a method, a system, electronic equipment and a storage medium for predicting a cutting-over optical cable result. Background In the field of operation and maintenance of a communication network, optical cable cutting is a key operation for guaranteeing network upgrading iteration and fault repair, and continuity and stability of communication service are directly related. Currently, risk assessment and result prediction of cable splicing mainly depend on personal experience of operation and maintenance personnel or an auxiliary system based on simple rules and a single traditional machine learning algorithm. However, in the prior art, the related data of the cutover are stored in different service systems in a scattered manner, and a large amount of unstructured text and real-time sequence data cannot be integrated effectively. Meanwhile, the existing prediction model is difficult to capture complex coupling relations among multiple factors such as equipment, resources, environment, manual operation and the like, so that the prediction accuracy is limited in a complex network environment or under extreme working conditions. Disclosure of Invention The invention provides a method, a system, electronic equipment and a storage medium for predicting a cutting-over optical cable result, which are used for solving the defects in the prior art, so that high-precision cutting-over result prediction is realized. The invention provides a method for predicting a cutting optical cable result, which comprises the following steps: Extracting a plurality of basic features from multi-mode data related to the cut-over optical cable, and analyzing the multi-mode data to obtain a plurality of enhancement features; Performing feature cross combination processing on the plurality of basic features and the plurality of enhancement features to generate a plurality of associated features; Integrating the plurality of basic features, the plurality of enhancement features and the plurality of association features to generate a multi-dimensional feature set; the multi-dimensional feature set is input into an artificial intelligent model to obtain a prediction result output by the artificial intelligent model, wherein the artificial intelligent model comprises a main prediction model, an auxiliary optimization model and a large language model, the main prediction model is obtained by fine tuning a multi-mode large model based on a multi-mode data sample, and the auxiliary optimization model adopts a random forest algorithm or a gradient lifting tree algorithm. According to the method for predicting the cut-over optical cable result, the multi-mode data comprise historical cut-over data, resource state information, network topology data, equipment performance data, real-time operation and maintenance data, text data and time-sequence associated data. According to the method for predicting the cut-over optical cable result provided by the invention, the method for extracting a plurality of basic features from multi-mode data related to the cut-over optical cable comprises the following steps: extracting the type and the core number of the optical cable from the network topology data and the resource state information; extracting a planned cut duration and a buffer duration from the historical cut data; extracting the current resource occupancy rate and the historical peak value from the resource state information; and extracting the ambient temperature and humidity from the real-time operation and maintenance data. According to the method for predicting the cut-over optical cable result provided by the invention, the multi-mode data is analyzed to obtain a plurality of enhancement features, and the method comprises the following steps: constructing an association relation between the network topology data and the fault cases in the historical cut-over data by utilizing a knowledge graph technology, and calculating the similarity between the current cut-over scheme and the historical fault cases to output a fault association coefficient as a risk association characteristic; Predicting network flow and resource load in the time sequence associated data by adopting a long-short-term memory network, and generating time sequence prediction characteristics including flow peak value offset and load super-threshold probability; and carrying out semantic analysis on the text data through a large language model, and extracting operation risk keywords and experience weight values as text semantic features. According to the method for predicting the cut-over optical cable result provided by the invention, the characteristic cross combination processing is carried out on the plurality of basic characteristics and the plurality of enhancement char