Method and system for predicting operation risks of testing and inspection institutions based on multi-source data

Abstract

The present application proposes a method and system for predicting the operational risks of testing and inspection institutions based on multi-source data, which relates to the field of risk prediction technology. First, the first testing standard is input into the first adjustment dimension determination model, so as to obtain the first and second adjustment dimensions that respectively characterize the changes in subjective requirements and objective requirements. Then, combined with the personnel change information corresponding to the testing standard adjustment time interval, the second risk information used to characterize the subjective factors is analyzed and obtained. Then, based on the second adjustment dimension and the first testing condition analysis, the third risk information used to characterize the objective factors is obtained. Finally, the second and third risk information are combined to obtain the target risk information of the testing institution. The technical solution of the present application, combined with big data time series information and machine learning models, extracts the potential correlation between testing standard changes, testing personnel changes, and testing risks, so that the risk prediction results of the testing institution are more accurate and meet actual requirements.

Description

Technical Field

[0001] The present invention belongs to the technical field of risk prediction, and in particular relates to a method and system for predicting the operational risks of testing and inspection organizations based on multi-source data. Background Art

[0002] In the field of operational risk prediction technology for testing and inspection institutions, existing methods mostly focus on equipment status monitoring or single-dimensional data analysis, and suffer from a significant lack of collaborative analysis. Traditional technical solutions usually separate testing standard update information from personnel operation data. They neither establish a dynamic coupling model for standard version evolution and personnel qualification status, nor conduct a quantitative assessment of the compatibility between standard parameter adjustments and personnel operating habits. Especially in scenarios where standards are frequently updated, existing systems cannot effectively identify typical compound risks such as "personnel who have not completed training during the implementation cycle of the new standard operate high-risk equipment", nor can they warn of hidden risks caused by "the mismatch between personnel operating experience and the complexity of the new standard". This blind spot in the analysis of standard-personnel collaborative change information directly leads to an increase in the false alarm rate of the risk prediction model in dynamic scenarios, and is unable to accurately reflect the overall risk situation of the human-machine-environment system during the standard iteration process. Summary of the Invention

[0003] The purpose of the present invention is to provide a method and system for predicting the operation risks of testing and inspection institutions based on multi-source data, so as to solve the technical problem of increased false alarm rate of risk prediction models in dynamic scenarios in the prior art.

[0004] The present invention proposes a method for predicting the operational risk of a testing and inspection organization based on multi-source data, which includes:

[0005] S1: performing a first analysis process on the first detection standard information and the first preceding detection standard information to determine a plurality of first adjustment dimensions and a plurality of second adjustment dimensions;

[0006] The first analysis process refers to comparing and analyzing the new and old detection standard information; the first adjustment dimension refers to the adjustment dimension related to subjective operation, and the second adjustment dimension refers to the adjustment dimension related to the objective detection environment;

[0007] S2: Obtaining first person information and first preceding person information according to the first detection standard information and the first preceding detection standard information respectively;

[0008] Wherein, the first detection standard information corresponds to the first personnel information, and the first preceding detection standard information corresponds to the first preceding personnel information;

[0009] S3: performing a first differential analysis on the first personnel information and the first preceding personnel information to obtain first risk information;

[0010] S4: Determine second risk information based on the first risk information and the plurality of first adjustment dimensions;

[0011] S5: Determine third risk information based on the plurality of second adjustment dimensions and the first detection condition information;

[0012] S6: Determine target risk information based on the second risk information and the third risk information.

[0013] Preferably, the S1 includes the following sub-steps:

[0014] Step S11: obtaining first difference information between the first detection standard information and the first preceding detection standard information, and determining a structured revision feature data set based on the first difference information; the structured revision feature data set includes a plurality of revision feature data items;

[0015] Step S12: performing a first operation responsibility analysis and a first human factor constraint analysis on the plurality of revision feature data items one by one to obtain a plurality of first subjective adjustment dimension information;

[0016] Step S13: performing a first numerical feature structured analysis on each of the plurality of revision feature data items in the structured revision feature data set to obtain a plurality of first objective adjustment dimension information;

[0017] Step S14: performing a first conflict analysis on the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information to obtain first priority adjustment information;

[0018] Step S15: adjusting the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information based on the first priority adjustment information to obtain a plurality of first adjustment dimensions and a plurality of second adjustment dimensions.

[0019] Preferably, the S2 includes the following sub-steps:

[0020] S21: Acquire a first time interval of the first detection standard information, and acquire a second time interval of the first preceding detection standard information;

[0021] S22: Determine first personnel information and first preceding personnel information according to the first time interval and the second time interval.

[0022] Preferably, the first personnel information and the first preceding personnel information include basic personnel information and historical illegal operation information.

[0023] Preferably, S3 includes the following sub-steps:

[0024] S31: Determine a first person risk vector and a first predecessor person risk vector based on the first person information and the first predecessor person information;

[0025] S32: Inputting the first personnel risk vector and the first preceding personnel risk vector into a first risk information determination model to output first risk information.

[0026] Preferably, the S4 includes:

[0027] Similarity matching is performed between the first risk information and the plurality of first adjustment dimensions, and the plurality of first adjustment dimensions that meet similarity requirements are determined as second risk information.

[0028] Preferably, the S5 includes:

[0029] Each of the second adjustment dimensions is matched with the first detection condition information one by one for similarity, and third risk information is determined based on at least one of the second adjustment dimensions that meets the similarity requirement.

[0030] Preferably, the first conflict analysis includes:

[0031] determining a plurality of first detection contradiction items according to the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information;

[0032] The plurality of first detection contradiction items are used to represent contradiction information between at least one piece of the first subjective adjustment dimension information and at least one piece of the first objective adjustment dimension information.

[0033] This application also proposes a testing and inspection agency operation risk prediction system based on multi-source data, which is used to implement the above-mentioned testing and inspection agency operation risk prediction method based on multi-source data.

[0034] The present application proposes a method and system for predicting the operational risks of testing and inspection institutions based on multi-source data, which relates to the field of risk prediction technology. First, the first testing standard after the latest adjustment is input into the first adjustment dimension determination model, so as to obtain the first and second adjustment dimensions that respectively characterize the changes in subjective requirements and objective requirements, and then, combined with the personnel change information corresponding to the time interval of the testing standard adjustment, the second risk information for characterizing the subjective factors is analyzed and obtained, and then, based on the second adjustment dimension and the first testing condition analysis, the third risk information for characterizing the objective factors is obtained, and finally, the second and third risk information are combined to obtain the target risk information of the testing institution. The technical solution of the present application, combined with big data time series information and machine learning models, extracts the potential correlation between changes in testing standards-changes in testing personnel-testing risks, so that the risk prediction results of the testing institution are more accurate and meet actual requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other implementation drawings based on the provided drawings without inventive effort.

[0036] Figure 1 It is an execution flow chart of the method for predicting the operation risk of a testing and inspection organization based on multi-source data in the present invention.

[0037] Figure 2 This is an example of the time interval used for the new and old detection standards in the present invention. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. The exemplary embodiments and descriptions are only used to explain the present invention but are not intended to limit the present invention.

[0040] The following is a detailed description of the method and system for predicting operational risks of testing and inspection organizations based on multi-source data of the present invention.

[0041] There is a close relationship between the testing standards and testing conditions of a testing organization and the latest testing standards. Under different testing standards, the requirements for the testing process and testing conditions of the same testing organization will be quite different.

[0042] There are multiple sources of risk in key links of the detection process (such as sample collection and processing, instrument calibration and maintenance, test method execution, data analysis and interpretation, and environmental conditions). These mainly include: insufficient sample representativeness or contamination / damage leading to biased results; instrument miscalibration or equipment failure affecting measurement accuracy; improper test method selection or operational errors (such as human error, uncontrolled conditions); data processing errors or misinterpretation of results (such as calculation errors, ignoring outliers); and the impact of environmental factors (temperature and humidity changes) and cross-contamination on high-sensitivity detection.

[0043] The above risk factors are analyzed from both subjective and objective perspectives. Subjective factors may include: contamination during sample collection, failure to select appropriate testing methods, human error during testing, data processing errors and misjudgments due to insufficient human experience, etc. Objective factors may include: sample collection volume not meeting testing standards, equipment calibration or maintenance cycles not meeting testing standards, and environmental factors and cross-contamination during testing not meeting testing standards.

[0044] Updates to testing standards primarily focus on improving accuracy and quality, and include: stricter instrument calibration and verification requirements (e.g., more frequent or precise calibration); adjustments to test parameters (environmental conditions, processing time, etc.); changes to data processing methods and report formats; addition or upgrade of equipment functions to support new methods; personnel training and skills improvement required to cope with new standards; changes to sample preparation procedures (size, pretreatment, etc.); and enhanced compliance and documentation requirements to ensure traceability and conformity.

[0045] The above-mentioned types of changes in testing standards can be further divided into changes in objective requirements and changes in subjective requirements. The changes in objective requirements refer to objectively setting higher standards for equipment, testing processes or testing environments, while the changes in subjective requirements refer to setting higher standards for operators' familiarity with new software or new analysis methods.

[0046] An analysis of the causes of historical testing risks reveals that the primary reason may be a failure to adapt the testing environment and personnel capabilities to significant changes in testing standards. Further analysis reveals that the combined influencing factors of different types of testing standard changes and different types of operators lead to corresponding changes in the probability of different types of testing risks.

[0047] The present invention addresses the above technical problems and proposes a technical solution to explore the potential relationship between changes in testing standards, testing personnel capabilities, and testing risks, so as to accurately predict the testing risks of testing organizations within a preset time period in the future.

[0048] This embodiment proposes a method for predicting the operation risk of a testing and inspection organization based on multi-source data. The specific process is as follows: Figure 1 shown.

[0049] S1: Performing a first analysis process on the first detection standard information and the first preceding detection standard information to determine a plurality of first adjustment dimensions and a plurality of second adjustment dimensions.

[0050] The first adjustment dimension refers to an adjustment dimension related to the operator's subjective operation, and the second adjustment dimension refers to an objective operation dimension related to the operating environment, etc. The first analysis and processing involves comparing and analyzing the new and old test standard information, and obtaining a plurality of first and second adjustment dimensions for representing subjective and objective changes through semantic analysis.

[0051] The first preceding detection standard information refers to the detection standard information before the first detection standard information is adjusted, that is, the first detection standard information is obtained by adjusting on the basis of the first preceding detection standard information.

[0052] Said S1 is the basis for the technical solution of this application to accurately predict the detection risk, because only by conducting a comprehensive and accurate analysis of the newly revised first detection standard information can the adjustments of subjective and objective factors brought about by the new revision be accurately extracted, thereby providing basic support for subsequent detection risk prediction.

[0053] The S1 includes the following sub-steps:

[0054] Step S11: obtaining first difference information between the first detection standard information and the first preceding detection standard information, and determining a structured revision feature data set based on the first difference information.

[0055] The structured revision feature data set includes a plurality of revision feature data items.

[0056] The S11 specifically includes:

[0057] 1) Input the old and new versions of the standard document (the first test standard information is the latest version), use the pre-trained document structure parsing model (CNN based on the ResNet-50 architecture) to extract the hierarchical structure features of the standard document and identify the topological relationship between chapters and clauses;

[0058] 2) Compare the text sequences of the new and old versions of the standard documents to generate a revision content index table with position identifiers, including: revision type (new / deleted / modified), clause number, revision text fragment, and associated context location coordinates;

[0059] 3) Outputting a structured revision feature data set, wherein the structured revision feature data set is composed of a plurality of revision feature data items.

[0060] An exemplary revision feature data item is as follows: {Revision ID: C001, Type: Parameter modification, Location: Section 5.2.3, Old text: "Temperature range 20-25°C", New text: "Temperature control 23±0.5°C", Context keywords: [Constant oven calibration, Environmental monitoring]}.

[0061] 4) Performing dual-dimensional semantic labeling on the structured revision feature dataset.

[0062] Each revision feature data item in the structured revision feature dataset is input into the semantic classification model one by one: the semantic classification model uses a BERT-based fine-tuning model to identify statements and numerical parameters containing the responsibilities of the operating subject and physical constraints, and outputs a probability value (i.e., confidence) P_subj and a tag type (such as personnel qualifications, operating procedures, decision-making authority, equipment parameters, environmental indicators, and material properties).

[0063] Generate semantic labels with confidence scores: {Revision ID: C001, Subjective dimension label: Empty, Objective dimension label: Temperature control parameter (confidence 0.92)}.

[0064] Each of the semantic tags is a revision feature data item, and the structured revision feature data set can be obtained according to a plurality of the revision feature data items.

[0065] Step S12: performing a first operation responsibility analysis and a first human factor constraint analysis on the plurality of revision feature data items in the structured revision feature data set one by one to obtain a plurality of first subjective adjustment dimension information.

[0066] The S12 specifically includes:

[0067] 1) Operational responsibility relationship modeling:

[0068] A plurality of subjective dimension labels are screened out from the first semantic label set outputted in step S11 and used as input, and an operation responsibility network is constructed through a dependency parser (based on an improved version of Stanford CoreNLP).

[0069] Extract the triple relationship: <operating subject, action verb, controlled object>. For example, from the clause "Inspectors must complete calibration under the witness of a supervisor", extract the following: <inspector, complete, calibration>, <supervisor, witness, calibration>;

[0070] Establish information about the first subjective adjustment dimension caused by operational constraints: Convert conditional adverbial clauses into logical expressions. For example, "If data anomalies occur, retesting is required after approval by the technical director" → retest trigger condition = data anomaly ∩ approval status = true;

[0071] 2) Modeling of human factors constraint rules:

[0072] Convert semantic analysis results into executable rules:

[0073] Use the Drools rule engine syntax to generate a decision table. For example, convert "Only senior technicians can operate X-type equipment" into a rule: IF equipment type = X THEN operator, qualification level ≥ 3;

[0074] Construct an operation timing constraint matrix: for example, encode "the interval between two detections ≥ 4 hours" as a prohibited interval [t_start, t_start+4h) on the time axis;

[0075] Output the first subjective adjustment dimension information caused by the standard upgrade, which may specifically include the aforementioned operation permission constraints or operation time interval constraints. For example, {Adjustment Dimension ID: S001, Type: Operation Permission Change, Affected Object: X-Type Equipment Operation Permission, Constraint: Qualification Level ≥ 3}.

[0076] Step S13: performing a first numerical feature structured analysis on the plurality of revision feature data items in the structured revision feature data set one by one to obtain a plurality of first objective adjustment dimension information.

[0077] The S13 may specifically include:

[0078] To screen out multiple objective dimension labels from the first semantic label set output in step S11, a multi-level parsing strategy is adopted:

[0079] Level 1 parsing: Regular expressions match numerical patterns (e.g., "23±0.5℃" → {reference value: 23, fluctuation range: ±0.5, unit: ℃});

[0080] Secondary parsing: Contextual analysis to determine parameter scope (e.g., "test chamber temperature" is associated with the temperature control module with device number Lab-201);

[0081] Level 3 analysis: Dimension consistency check, automatic conversion of units to the reference system (e.g., unifying "50kPa" to "0.05MPa");

[0082] Generate multiple pieces of first objective adjustment dimension information. For example, it can be {Parameter ID: P032, Parameter Name: Temperature Control Accuracy, Old Value: ±1.0°C, New Value: ±0.5°C, Affected Devices: [Constant Incubator HWS-201, Incubator PQX-305]}.

[0083] Step S14: performing a first conflict analysis on the plurality of pieces of the first subjective adjustment dimension information and the plurality of pieces of the first objective adjustment dimension information to obtain first priority adjustment information.

[0084] The first detection standard information includes multiple subjective adjustment information and multiple objective adjustment information, and there may be contradictions and conflicts between different subjective adjustment information and objective adjustment information. Therefore, it is necessary to first perform a first conflict analysis on the above information to adaptively adjust the priorities of the multiple first subjective adjustment dimension information and the multiple first objective adjustment dimension information.

[0085] The S14 may specifically include:

[0086] 1) Dual-dimensional adjustment feature fusion and conflict detection:

[0087] Perform correlation analysis on the multiple pieces of the first subjective adjustment dimension information in step S12 and the multiple pieces of the first objective adjustment dimension information in step S13:

[0088] Establish a human-machine-environment constraint relationship matrix to detect conflicting items (e.g., a new standard requires "shortening test time to 2 hours," but equipment parameter adjustments result in a single test taking 2.5 hours).

[0089] Identify implicit dependencies (e.g., "increasing the frequency of inspector reviews" requires a corresponding increase in the capacity of the data storage server).

[0090] 2) Adjust priority calculation:

[0091] Construct a priority evaluation model based on the analytic hierarchy process (AHP):

[0092] Evaluation indicators include: standard enforcement level, number of affected equipment, difficulty of personnel training, and implementation cost;

[0093] Calculate the comprehensive weight of each adjustment dimension. For example, the priority score of the objective dimension P032 is 8.7 (out of 10);

[0094] Output the priority adjustment task sequence as the first priority adjustment information. The example is as follows:

[0095] {Task ID: T001, Dimension Type: Objective, Parameter ID: P032, Priority: Urgent, Latest Implementation Date: 2023-12-01},

[0096] {Task ID: T002, Dimension Type: Subjective, Adjustment ID: S001, Priority: High, Associated Training Plan: TR-2023-09}

[0097] ].

[0098] Step S15: adjusting the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information based on the first priority adjustment information to obtain a plurality of first adjustment dimensions and a plurality of second adjustment dimensions.

[0099] In this step, adjustment information that meets preset priority requirements can be selected from multiple pieces of the first subjective adjustment dimension information and multiple pieces of the first objective adjustment dimension information based on the first priority adjustment information to serve as multiple first adjustment dimensions and multiple second adjustment dimensions.

[0100] Preferably, the specific priority screening requirement can be set according to the tolerance for detection risk. The higher the tolerance, the lower the priority threshold can be set.

[0101] S2: Obtain first personnel information and first preceding personnel information according to the first detection standard information and the first preceding detection standard information.

[0102] In S1, multiple first adjustment dimensions and multiple second adjustment dimensions, respectively representing subjective and objective dimensions, are analyzed from the first detection standard information. To further analyze and derive the correlation between these adjustment dimensions, personnel changes, and detection risks, this step requires obtaining the first personnel information and the first preceding personnel information corresponding to the old and new detection standards.

[0103] The S2 specifically includes:

[0104] S21: Acquire a first time interval of the first detection standard information, and acquire a second time interval of the first preceding detection standard information.

[0105] The first detection standard information and the first preceding detection standard information are published and used successively, so each detection standard has a corresponding usage time interval, namely the first time interval and the second time interval.

[0106] like Figure 2 As an example of the time interval for using the new and old detection standards, the usage interval of the first preceding detection standard information is 2023.03.01-2024.03.01, and the usage interval of the first detection standard information is 2024.03.02-2025.03.01.

[0107] S22: Determine first personnel information and first preceding personnel information according to the first time interval and the second time interval.

[0108] In two adjacent time intervals, the personnel information of the same testing organization may change, so in this step, the corresponding first personnel information and the first preceding personnel information need to be determined respectively.

[0109] The first personnel information and the first preceding personnel information include basic information and historical illegal operation information of the corresponding personnel.

[0110] S3: Perform a first differential analysis on the first personnel information and the first preceding personnel information to obtain first risk information.

[0111] The first differential analysis refers to obtaining the first risk information through personnel changes, and the first risk information refers to corresponding changes in detection risk information of subjective reasons caused by changes in operating personnel.

[0112] S3 is mainly used to analyze and profile the first personnel information and the first preceding personnel information, so as to comprehensively evaluate the subjective risk factors brought about by the personnel team changes. S3 specifically includes the following sub-steps:

[0113] S31: Determine a first personnel risk vector and a first predecessor personnel risk vector based on the first personnel information and the first predecessor personnel information.

[0114] In this step, the characteristics of the first personnel information and the first preceding personnel information are extracted in the form of a risk vector through the first risk portrait determination model.

[0115] Specifically, the information dimensions contained in the first personnel risk vector and the first predecessor personnel risk vector should be the same to ensure that similar comparisons can be made later. The specific information dimensions in the risk vector can be set according to specific needs and should correspond to the multiple first adjustment dimensions in step S1 as much as possible. For example, it can be set to [gender, years of experience, historical operational responsibility violation information 1, historical operational responsibility violation information 2, historical human factor constraint violation information 1, historical human factor constraint violation information 2]. In the format of the above risk vector, the basic information of the detection personnel is recorded, and the risk detection behavior caused by operational responsibility violations and human factor constraint rule violations is recorded.

[0116] The above is only an example, and the information dimension of the subjective reasons in S1 and this step may be adjusted according to specific needs.

[0117] Since both the first personnel information and the first predecessor personnel information contain multiple natural persons, in this step it is necessary to generate a risk vector for each natural person in a specified format, and after clustering, extract the first personnel risk vector and the first predecessor personnel risk vector that can be used to characterize the first personnel information or the first predecessor personnel information as a whole.

[0118] S32: Inputting the first personnel risk vector and the first preceding personnel risk vector into a first risk information determination model to output first risk information.

[0119] The first risk information determination model is obtained based on machine learning model training, and is mainly used to analyze the difference between the first personnel risk vector and the first predecessor personnel risk vector, and finally obtain the first risk information based on the difference.

[0120] Preferably, the first risk information determination model can be obtained based on convolutional neural network model training, and the model training is completed using historical data belonging to the same standard testing agency.

[0121] The output first risk information includes at least one risk dimension included in the risk vector, and the risk dimensions are all caused by adjustments made by the detection personnel.

[0122] S4: Determine second risk information based on the first risk information and the plurality of first adjustment dimensions.

[0123] In this step, corresponding second risk information is determined from the first risk information based on the multiple first adjustment dimensions associated with the first detection standard information. In other words, if at least a portion of the first risk information has a high correlation with the multiple first adjustment dimensions, this portion of information is determined as the second risk information.

[0124] During the specific implementation process, each piece of the first risk information can be matched against multiple first adjustment dimensions for similarity. If a first adjustment dimension with similarity that meets a preset value is found, the first risk information can be determined as the second risk information. Through the above steps, the second risk information with high incidence can be determined based on the two dimensions of detection standard adjustment and personnel adjustment.

[0125] S5: Determine third risk information based on the plurality of second adjustment dimensions and the first detection condition information.

[0126] The first detection condition information refers to the target detection mechanism's environmental conditions, equipment conditions, and layout information of detection instruments during the detection process. If the above detection conditions do not match the multiple second adjustment dimensions included in the first detection standard information, the third risk information will be generated.

[0127] During the specific implementation process, each of the second adjustment dimensions can be matched with the first detection condition information one by one for similarity, and a first similarity can be determined for each of the second adjustment dimensions. Third risk information is determined based on at least one of the second adjustment dimensions whose first similarity is greater than a preset value.

[0128] S6: Determine target risk information based on the second risk information and the third risk information.

[0129] The target risk information integrates the objective requirement adjustment and subjective requirement adjustment information reflected in the first testing standard information, thereby being able to more comprehensively and accurately reflect the target risk information of the testing agency within a preset time interval in the future.

[0130] Preferably, the target risk information can be obtained by mechanically superimposing the second risk information and the third risk information, or can be obtained by screening after setting priorities based on preset rules, which is not specifically limited here.

[0131] This application also proposes a multi-source data-based testing and inspection agency operation risk prediction system for executing the above-mentioned multi-source data-based testing and inspection agency operation risk prediction method.

[0132] The present application proposes a method and system for predicting the operational risks of testing and inspection institutions based on multi-source data, which relates to the field of risk prediction technology. First, the first testing standard after the latest adjustment is input into the first adjustment dimension determination model, so as to obtain the first and second adjustment dimensions that respectively characterize the changes in subjective requirements and objective requirements, and then, combined with the personnel change information corresponding to the time interval of the testing standard adjustment, the second risk information for characterizing the subjective factors is analyzed and obtained, and then, based on the second adjustment dimension and the first testing condition analysis, the third risk information for characterizing the objective factors is obtained, and finally, the second and third risk information are combined to obtain the target risk information of the testing institution. The technical solution of the present application, combined with big data time series information and machine learning models, extracts the potential correlation between changes in testing standards-changes in testing personnel-testing risks, so that the risk prediction results of the testing institution are more accurate and meet actual requirements.

[0133] The above description is only a preferred embodiment of the present invention. Therefore, any equivalent changes or modifications made according to the structure, characteristics and principles described in the scope of the patent application of the present invention are included in the scope of the patent application of the present invention.

Claims

1. A method for predicting the operational risk of a testing and inspection organization based on multi-source data, characterized in that: The method includes: S1: performing a first analysis process on the first detection standard information and the first preceding detection standard information to determine a plurality of first adjustment dimensions and a plurality of second adjustment dimensions; The first analysis process refers to comparing and analyzing the new and old detection standard information; the first adjustment dimension refers to the adjustment dimension related to subjective operation, and the second adjustment dimension refers to the adjustment dimension related to the objective detection environment; The S1 includes the following sub-steps: Step S11: obtaining first difference information between the first detection standard information and the first preceding detection standard information, and determining a structured revision feature data set based on the first difference information; the structured revision feature data set includes a plurality of revision feature data items; Step S12: performing a first operation responsibility analysis and a first human factor constraint analysis on the plurality of revision feature data items one by one to obtain a plurality of first subjective adjustment dimension information; Step S13: performing a first numerical feature structured analysis on each of the plurality of revision feature data items in the structured revision feature data set to obtain a plurality of first objective adjustment dimension information; Step S14: performing a first conflict analysis on the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information to obtain first priority adjustment information; Step S15: adjusting the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information based on the first priority adjustment information to obtain a plurality of first adjustment dimensions and a plurality of second adjustment dimensions; S2: Obtaining first person information and first preceding person information according to the first detection standard information and the first preceding detection standard information respectively; Wherein, the first detection standard information corresponds to the first personnel information, and the first preceding detection standard information corresponds to the first preceding personnel information; S3: performing a first differential analysis on the first personnel information and the first preceding personnel information to obtain first risk information; The first risk information refers to the change in the detection risk information of subjective reasons caused by the change of the operator; S4: Determine second risk information based on the first risk information and the plurality of first adjustment dimensions; performing similarity matching between the first risk information and the plurality of first adjustment dimensions one by one, and determining the first risk information as the second risk information if a first adjustment dimension with similarity meeting a preset value is found; S5: Determine third risk information based on the plurality of second adjustment dimensions and the first detection condition information; The first detection condition information refers to the environmental conditions, equipment conditions, and layout information of the detection instruments of the target detection mechanism during the detection process; S6: Determine target risk information based on the second risk information and the third risk information.

2. The method for predicting the operation risk of a testing and inspection institution based on multi-source data according to claim 1 is characterized in that: The method for predicting operational risks of inspection and testing institutions based on multi-source data according to claim 1 is characterized in that S2 comprises the following sub-steps: S21: Acquire a first time interval of the first detection standard information, and acquire a second time interval of the first preceding detection standard information; S22: Determine first personnel information and first preceding personnel information according to the first time interval and the second time interval.

3. The method for predicting the operation risk of a testing and inspection institution based on multi-source data according to claim 2 is characterized in that: The first personnel information and the first preceding personnel information include basic personnel information and historical illegal operation information.

4. The method for predicting the operation risk of a testing and inspection institution based on multi-source data according to claim 3 is characterized in that: The S3 includes the following sub-steps: S31: Determine a first person risk vector and a first predecessor person risk vector based on the first person information and the first predecessor person information; S32: Inputting the first personnel risk vector and the first preceding personnel risk vector into a first risk information determination model to output first risk information.

5. The method for predicting the operation risk of a testing and inspection institution based on multi-source data according to claim 4 is characterized in that: The S4 includes: Similarity matching is performed between the first risk information and the plurality of first adjustment dimensions, and the plurality of first adjustment dimensions that meet similarity requirements are determined as second risk information.

6. The method for predicting the operation risk of a testing and inspection institution based on multi-source data according to claim 5 is characterized in that: The S5 includes: Each of the second adjustment dimensions is matched with the first detection condition information one by one for similarity, and third risk information is determined based on at least one of the second adjustment dimensions that meets the similarity requirement.

7. The method for predicting the operation risk of a testing and inspection institution based on multi-source data according to claim 6, characterized in that: The first conflict analysis includes: determining a plurality of first detection contradiction items according to the plurality of first subjective adjustment dimension information and the plurality of first objective adjustment dimension information; The plurality of first detection contradiction items are used to represent contradiction information between at least one piece of the first subjective adjustment dimension information and at least one piece of the first objective adjustment dimension information.

8. The risk prediction system for testing and inspection institutions based on multi-source data is characterized by: Used to implement the method for predicting the operational risks of testing and inspection institutions based on multi-source data in claims 1-7 above.

Images