CN116415840A - Automatic index early warning method and system based on machine learning model - Google Patents
Automatic index early warning method and system based on machine learning model Download PDFInfo
- Publication number
- CN116415840A CN116415840A CN202310121509.0A CN202310121509A CN116415840A CN 116415840 A CN116415840 A CN 116415840A CN 202310121509 A CN202310121509 A CN 202310121509A CN 116415840 A CN116415840 A CN 116415840A
- Authority
- CN
- China
- Prior art keywords
- early warning
- index data
- index
- data
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000010801 machine learning Methods 0.000 title claims abstract description 34
- 230000008859 change Effects 0.000 claims abstract description 159
- 238000004458 analytical method Methods 0.000 claims abstract description 106
- 238000012549 training Methods 0.000 claims abstract description 48
- 230000002159 abnormal effect Effects 0.000 claims abstract description 47
- 238000001514 detection method Methods 0.000 claims abstract description 16
- 230000001360 synchronised effect Effects 0.000 claims abstract description 15
- 238000004140 cleaning Methods 0.000 claims description 33
- 238000012795 verification Methods 0.000 claims description 20
- 230000007774 longterm Effects 0.000 claims description 18
- 238000012545 processing Methods 0.000 claims description 18
- 238000004891 communication Methods 0.000 claims description 16
- 238000005457 optimization Methods 0.000 claims description 13
- 230000000737 periodic effect Effects 0.000 claims description 13
- 230000005540 biological transmission Effects 0.000 claims description 7
- 238000009499 grossing Methods 0.000 claims description 6
- 238000013450 outlier detection Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 5
- 238000004806 packaging method and process Methods 0.000 claims description 5
- 238000000354 decomposition reaction Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 description 12
- 230000009286 beneficial effect Effects 0.000 description 10
- 230000000694 effects Effects 0.000 description 6
- 230000008569 process Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/067—Enterprise or organisation modelling
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Tourism & Hospitality (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an index automatic early warning method and system based on a machine learning model, wherein the method comprises the following steps: acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model; analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition; and determining an early warning requirement in the early warning stage, carrying out early warning operation by matching a target early warning mode from a preset early warning mode library based on the early warning requirement, and carrying out synchronous early warning notification on the management terminal. The method and the system realize intelligent and reliable early warning analysis on the real-time state of the index data, ensure that corresponding early warning operation is timely carried out when the index data is abnormal, and simultaneously send early warning notification to the management terminal, thereby ensuring the accuracy, the intelligence and the timeliness of early warning of the index.
Description
Technical Field
The invention relates to the technical field of automatic early warning, in particular to an index automatic early warning method and system based on a machine learning model.
Background
The index represents the state of a certain link of the enterprise, if the index state can be comprehensively assessed through the historical expression of index data, the early warning of the index state and the notification after the early warning can be automatically realized, and the intelligent early warning of the index can be brought with greater service value;
however, whether the index is abnormal is judged or the threshold value is set by mainly relying on manpower and expert experience at present, so that corresponding early warning operation cannot be performed in time when the index is abnormal, the early warning error is large, and the operation efficiency and the safety monitoring effect of enterprises are directly affected;
therefore, the invention provides an index automatic early warning method and system based on a machine learning model.
Disclosure of Invention
The invention provides an index automatic early warning method and system based on a machine learning model, which are used for realizing accurate and reliable construction of an index data state change prediction model by learning and training historical index data, so that intelligent and reliable early warning analysis on the index data real-time state is facilitated, corresponding early warning operation is ensured to be timely carried out when the index data is abnormal, and meanwhile early warning notification is sent to a management terminal, so that the accuracy, the intelligence and the timeliness of index early warning are ensured.
The invention provides an index automatic early warning method based on a machine learning model, which comprises the following steps:
step 1: acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model;
step 2: analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition;
step 3: and determining an early warning requirement in the early warning stage, carrying out early warning operation by matching a target early warning mode from a preset early warning mode library based on the early warning requirement, and carrying out synchronous early warning notification on the management terminal.
Preferably, in step 1, historical index data is obtained, which includes:
acquiring an index analysis request sent by a management terminal, analyzing the index analysis request, and determining a service identifier corresponding to the index analysis request;
generating a data calling request based on the service identifier, transmitting the data calling request to a preset server, analyzing the data calling request based on the preset server, and searching the service database based on the service identifier to obtain a target service database corresponding to the service identifier;
And extracting the historical index data in the target service database, packaging the historical index data and feeding back to the data receiving end.
Preferably, an automatic index early warning method based on a machine learning model packs historical index data and feeds the packed historical index data back to a data receiving end, and the method comprises the following steps:
acquiring the obtained historical index data, and clustering the historical index data to obtain a target classification result corresponding to the historical index data;
extracting isolated sample data contained in index data of each category based on a clustering result, extracting data characteristics of the index data of each category, and matching target cleaning rules from a preset data cleaning rule base based on the data characteristics;
performing first cleaning on the isolated sample data based on the target cleaning rule, determining a cleaning result of the isolated sample, and performing second cleaning on the index data of the corresponding category based on the target cleaning rule when the cleaning result meets the expected requirement;
and obtaining standard index data based on the second cleaning result, and storing the standard index data.
Preferably, in step 1, historical index data is trained based on an index analysis rule and an index outlier detection rule, and an index data state change prediction model is constructed, which comprises the following steps:
Acquiring the obtained historical index data, determining service operation attributes corresponding to the historical index data, and determining index analysis rules and outlier detection rules for the historical index data based on the service operation attributes;
dividing the obtained historical index data into a training set and a verification set, determining a logic analysis structure of the historical index data based on an index analysis rule and an abnormal value detection rule, and determining target iteration times based on the data quantity of the historical index data contained in the training set;
performing iterative training of target iterative times on the training set based on the logic analysis structure to obtain logic analysis results after each iterative training, determining a reference logic analysis result corresponding to each historical index data in the training set, and determining error data of the logic analysis results after each iterative training and the corresponding reference logic analysis results;
and determining a weight value of the logic analysis result after each iteration training based on the error data, correcting the logic analysis result after each iteration training based on the weight value to obtain a target logic analysis result, and constructing an index data state change prediction model based on the target logic analysis result.
Preferably, an automatic index early warning method based on a machine learning model constructs an index data state change prediction model based on a target logic analysis result, and the method comprises the following steps:
acquiring a constructed index data state change prediction model and a verification set obtained by splitting historical index data, and inputting the verification set into the constructed index data state change prediction model for analysis to obtain a target output result;
extracting standard index states corresponding to each historical index data in the verification set, judging that the constructed index data state change prediction model is qualified when the target output result is consistent with the standard index states, otherwise, judging that the constructed index data state change prediction model is unqualified, and determining difference data between the target output result and the standard index states;
and determining an optimization strategy for the index data state change prediction model based on the difference data, and optimizing the index data state change prediction model based on the optimization strategy.
Preferably, in step 2, the method analyzes the index data based on the index data state change prediction model to determine the state change trend of the index data, including:
Acquiring index data to be analyzed, and visually displaying the index data to be analyzed in a preset rectangular coordinate system to obtain long-term change trend information corresponding to the index data to be analyzed;
removing abnormal values in the long-term change trend information, smoothing the long-term change trend information after removing the abnormal values, and periodically decomposing the long-term change trend information based on a smoothing result;
obtaining periodic variation data sequences based on periodic decomposition results, determining target influence factors influencing the state change of index data, respectively determining the correlation between different periodic variation data sequences and the target influence factors, and carrying out average value processing on the obtained multiple groups of correlations to obtain target association relation between the index data to be analyzed and the target influence factors;
the method comprises the steps of obtaining an index data state change prediction model, inputting index data to be analyzed and a target association relation into the index data state change prediction model, processing the index data to be analyzed and the target association relation to obtain random change information generated by target influence factors in a preset time period of the index data to be analyzed, and predicting values of the index data to be analyzed in the preset time period based on the random change information to obtain a state change trend of the index data.
Preferably, in step 2, when the state change trend meets the early warning condition, an early warning stage for the index data is determined, which includes:
acquiring service characteristics of a target service corresponding to the index data, setting a normal amplitude interval of the index data based on the service characteristics, acquiring a prediction result of a state change trend of the index data, and determining an amplitude change value of the index data based on the state change trend obtained by prediction;
comparing the amplitude variation value with a preset amplitude variation threshold value, and judging that the early warning condition is met when the amplitude variation value is larger than the preset amplitude variation threshold value;
and determining a target difference value between a peak value in the state change trend of the index data and a normal amplitude interval based on the judging result, matching the target difference value with an early warning stage difference value comparison table, and determining an early warning stage of the index data based on the matching result.
Preferably, in step 3, an early warning requirement of an early warning stage is determined, and an early warning operation is performed by matching a target early warning mode from a preset early warning mode library based on the early warning requirement, including:
Acquiring an early warning stage of index data, extracting early warning characteristics of the early warning stage, and determining an early warning grade and an early warning type based on the early warning characteristics;
determining an early warning requirement for an early warning stage based on the early warning grade and the early warning type, and performing first matching on the early warning type in the early warning requirement and a preset early warning mode class in a preset early warning mode library to obtain a target early warning mode class;
and performing second matching on the early warning level in the early warning requirement and the early warning parameters of the preset early warning modes in the preset early warning mode class, obtaining a target early warning mode based on a matching result, and performing early warning operation based on the target early warning mode.
Preferably, in step 3, a method for automatically pre-warning indexes based on a machine learning model performs synchronous pre-warning notification on a management terminal, including:
acquiring a first communication address of a management terminal and a second communication address of an early warning terminal, and constructing a data transmission link based on the first communication address and the second communication address;
acquiring an early warning operation corresponding to the early warning stage, converting the early warning operation into corresponding text content, and determining a target keyword corresponding to the early warning operation based on the text content;
and marking the target keywords, and transmitting the marking result and the text content to the management terminal based on the data transmission link to perform synchronous early warning notification.
The invention provides an index automatic early warning system based on a machine learning model, which comprises:
the model construction module is used for acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model;
the index analysis module is used for analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition;
and the early warning notification module is used for determining early warning requirements in the early warning stage, carrying out early warning operation by matching target early warning modes from a preset early warning mode library based on the early warning requirements, and carrying out synchronous early warning notification on the management terminal.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims thereof as well as the appended drawings.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention. In the drawings:
FIG. 1 is a flow chart of an automatic index early warning method based on a machine learning model in an embodiment of the invention;
FIG. 2 is a flowchart of step 1 in an automatic index early warning method based on a machine learning model according to an embodiment of the present invention;
fig. 3 is a block diagram of an automatic index early warning system based on a machine learning model in an embodiment of the invention.
Detailed Description
The preferred embodiments of the present invention will be described below with reference to the accompanying drawings, it being understood that the preferred embodiments described herein are for illustration and explanation of the present invention only, and are not intended to limit the present invention.
Example 1:
the embodiment provides an automatic index early warning method based on a machine learning model, as shown in fig. 1, including:
step 1: acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model;
step 2: analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition;
Step 3: and determining an early warning requirement in the early warning stage, carrying out early warning operation by matching a target early warning mode from a preset early warning mode library based on the early warning requirement, and carrying out synchronous early warning notification on the management terminal.
In this embodiment, the historical index data refers to index data generated during the operation process of a certain business or during the production process of a certain product in a past period of time, and the value change condition of the index data in the past period of time, wherein the index data represents the state of a certain link of an enterprise.
In this embodiment, the index analysis rule and the index abnormal value detection rule are known in advance, and are set according to the operation characteristics of the service, specifically, a manner or a method for analyzing the current state of the index data.
In this embodiment, the index abnormal value detection rule refers to a strategy for performing abnormal analysis on the value of index data, so as to reject abnormal data in the index data, and different types of index data correspond to different index abnormal detection rules.
In this embodiment, training the historical index data refers to analyzing the value change condition of the historical index data, so as to determine an analysis strategy for the index data, and finally, to construct an index data state change prediction model, so as to perform intelligent analysis on the index data, thereby being convenient for timely finding out an abnormal condition and performing corresponding early warning operation when the index data is abnormal.
In this embodiment, the index data state change prediction model is obtained after training the historical index data, and is used for analyzing the real-time state of the index data and predicting the relevant trend of the index data in a period of time in the future, so that the abnormal situation of the index data can be found in time.
In this embodiment, the state change trend is used to characterize the value change condition of the index data, so that corresponding early warning operation is conveniently performed when the value of the index data is not within the set normal value range.
In this embodiment, the early warning condition is set in advance, specifically may be a normal value range of the set index data, and when abnormal fluctuation occurs in the index data or the value is not in the set normal value range, corresponding early warning operation is performed.
In this embodiment, the early warning stage is used for representing the degree that the index data meets the early warning condition, and specifically may be primary early warning, intermediate early warning, advanced early warning, etc., when the deviation between the value of the index data and the set normal value range is larger, the early warning stage is higher, and the early warning threshold value corresponding to each stage is set in advance, or may be adjusted.
In this embodiment, the warning requirement characterizes the warning level required in different warning stages or the warning level of the warning.
In this embodiment, the preset early warning mode library is set in advance, and is used for storing different early warning modes, and early warning modes corresponding to different early warning stages are different.
In this embodiment, the target early warning mode refers to a mode suitable for early warning in the current early warning stage, and is one of preset early warning mode libraries.
In this embodiment, the synchronous early warning notification of the management terminal refers to that when the index data is abnormal, the abnormal condition of the index data is transmitted to the management terminal for display and reminding, so that the manager can find the abnormal condition of the index data in time.
The beneficial effects of the technical scheme are as follows: by learning and training the historical index data, an index data state change prediction model is accurately and reliably constructed, intelligent and reliable early warning analysis on the real-time state of the index data is facilitated, corresponding early warning operation is timely carried out when the index data is abnormal, and meanwhile early warning notification is sent to a management terminal, so that the accuracy, the intelligence and the timeliness of early warning of the index are guaranteed.
Example 2:
on the basis of embodiment 1, the present embodiment provides an automatic index early warning method based on a machine learning model, as shown in fig. 2, in step 1, historical index data is obtained, including:
acquiring an index analysis request sent by a management terminal, analyzing the index analysis request, and determining a service identifier corresponding to the index analysis request;
generating a data calling request based on the service identifier, transmitting the data calling request to a preset server, analyzing the data calling request based on the preset server, and searching the service database based on the service identifier to obtain a target service database corresponding to the service identifier;
and extracting the historical index data in the target service database, packaging the historical index data and feeding back to the data receiving end.
In this embodiment, the index analysis request is generated by the management terminal, and is used for retrieving index data corresponding to the index analysis requirement, including the type of index data to be retrieved, the data amount of index data to be retrieved, and the like.
In this embodiment, the service identifier is a service type tag corresponding to an index to be analyzed, and by using the service identifier, the type and the data of the index data to be called can be accurately and reliably determined.
In this embodiment, the data call request is sent to the preset server, and is used to characterize the type and the amount of the index data to be called to the preset server.
In this embodiment, the preset server is set in advance, and is used for storing index data corresponding to different types and different services.
In this embodiment, the service database is an intermediary in the preset server for storing index data of different types of services, and is set in advance, and the service database in the preset server is not unique.
In this embodiment, the target service database refers to a database corresponding to the preset server meeting the current index data calling request, and corresponding index data is stored in the database.
The beneficial effects of the technical scheme are as follows: by analyzing the index analysis request, the historical index data corresponding to the index analysis request is accurately and effectively acquired from the preset server, convenience and guarantee are provided for constructing the index data state change prediction model, and therefore intelligent and reliable early warning analysis on the real-time state of the index data is facilitated.
Example 3:
on the basis of embodiment 2, the present embodiment provides an automatic index early warning method based on a machine learning model, wherein the method includes packaging historical index data and feeding back the packaged historical index data to a data receiving end, and includes:
Acquiring the obtained historical index data, and clustering the historical index data to obtain a target classification result corresponding to the historical index data;
extracting isolated sample data contained in index data of each category based on a clustering result, extracting data characteristics of the index data of each category, and matching target cleaning rules from a preset data cleaning rule base based on the data characteristics;
performing first cleaning on the isolated sample data based on the target cleaning rule, determining a cleaning result of the isolated sample, and performing second cleaning on the index data of the corresponding category based on the target cleaning rule when the cleaning result meets the expected requirement;
and obtaining standard index data based on the second cleaning result, and storing the standard index data.
In this embodiment, the target classification result refers to each type of index data obtained after clustering the obtained historical index data.
In this embodiment, the isolated sample data refers to index data having a larger average value deviation of the data value existing in each type of index data from that of the type of index data, and is at least one data.
In this embodiment, the data feature is a parameter for characterizing the type of data, the range of values, and the like of the isolated sample data.
In this embodiment, the preset data cleansing rule base is set in advance, and is used for storing data cleansing rules corresponding to different data types.
In this embodiment, the target cleansing rule refers to a data cleansing rule applicable to cleansing the index data of the current type, and is at least one of preset data rule bases.
In this embodiment, the first cleansing refers to cleansing of isolated sample data in each category of index data by a target cleansing rule.
In this embodiment, the expected requirements are set in advance for measuring the minimum cleaning criteria for the isolated sample data, such as whether the cleaning is clean or not.
In this embodiment, the second cleaning means cleaning the index data of each category, that is, cleaning the index data of each category as a whole, and excluding the data missing segments therein.
In this embodiment, the standard index data refers to index data without interference data or abnormal data obtained after cleaning the index data.
The beneficial effects of the technical scheme are as follows: the index data is clustered, and the index data of each category is cleaned according to the clustering result, so that the accuracy of the finally obtained index data is guaranteed, the accuracy and the reliability of early warning analysis of the index data are guaranteed, and the early warning analysis effect of the index data is improved.
Example 4:
on the basis of embodiment 1, the present embodiment provides an automatic index early warning method based on a machine learning model, in step 1, historical index data is trained based on an index analysis rule and an index outlier detection rule, and an index data state change prediction model is constructed, including:
acquiring the obtained historical index data, determining service operation attributes corresponding to the historical index data, and determining index analysis rules and outlier detection rules for the historical index data based on the service operation attributes;
dividing the obtained historical index data into a training set and a verification set, determining a logic analysis structure of the historical index data based on an index analysis rule and an abnormal value detection rule, and determining target iteration times based on the data quantity of the historical index data contained in the training set;
performing iterative training of target iterative times on the training set based on the logic analysis structure to obtain logic analysis results after each iterative training, determining a reference logic analysis result corresponding to each historical index data in the training set, and determining error data of the logic analysis results after each iterative training and the corresponding reference logic analysis results;
And determining a weight value of the logic analysis result after each iteration training based on the error data, correcting the logic analysis result after each iteration training based on the weight value to obtain a target logic analysis result, and constructing an index data state change prediction model based on the target logic analysis result.
In this embodiment, the service operation attribute refers to an operation characteristic, an operation mode, a final operation purpose of a service, and the like of the service corresponding to the history index data in the operation process.
In this embodiment, the training set and the verification set refer to an index data set obtained by splitting the history index data into two parts, one part is used for training a model, and the other part is used for verifying the model obtained after training.
In this embodiment, the logic analysis structure is a logic for characterizing analysis of the history index data, and specifically may be a logic for characterizing analysis of the history index data one by one in the order from front to back, or the like.
In this embodiment, the target iteration number refers to the training number that needs to be determined by the data amount of the historical index data when the model is constructed, so as to ensure the accuracy and reliability of the finally obtained index data state change prediction model.
In this embodiment, the logical analysis result refers to an analysis result of index data corresponding to data included in the training set determined after the index data is analyzed according to the logical analysis structure, and specifically may be a composition component of the index data, an association relationship between components, and the like.
In this embodiment, the reference logical analysis result refers to a standard structure corresponding to the history index data, an interaction relationship between standard structures, and the like.
In this embodiment, the error data is used to characterize a difference between a logic analysis result obtained by analyzing the history index data through the logic analysis structure and a reference logic analysis result corresponding to the history index data, and may specifically be a difference between structures or the like.
In this embodiment, the target logic analysis result refers to a final logic analysis result obtained by correcting the obtained logic analysis result according to the weight value of the logic analysis result obtained after each iteration training, and the final logic analysis result can be directly used for constructing an index data state change prediction model.
The beneficial effects of the technical scheme are as follows: the method comprises the steps of analyzing historical index data, accurately and reliably determining index analysis rules and abnormal value detection rules corresponding to the historical index data, determining a logic analysis structure of the historical index data through the index analysis rules and the abnormal value detection rules, training the target iteration times of the historical index data according to the logic analysis structure, accurately and effectively correcting each iteration training result according to error data between a logic analysis result obtained by each iteration training and a reference logic analysis result, and accurately and effectively constructing an index data state change prediction model, so that intelligent and reliable early warning analysis of the real-time state of the index data is facilitated, and corresponding early warning operation is ensured to be carried out in time when the index data is abnormal.
Example 5:
on the basis of embodiment 4, the present embodiment provides an automatic index early warning method based on a machine learning model, and constructs an index data state change prediction model based on a target logic analysis result, including:
acquiring a constructed index data state change prediction model and a verification set obtained by splitting historical index data, and inputting the verification set into the constructed index data state change prediction model for analysis to obtain a target output result;
extracting standard index states corresponding to each historical index data in the verification set, judging that the constructed index data state change prediction model is qualified when the target output result is consistent with the standard index states, otherwise, judging that the constructed index data state change prediction model is unqualified, and determining difference data between the target output result and the standard index states;
and determining an optimization strategy for the index data state change prediction model based on the difference data, and optimizing the index data state change prediction model based on the optimization strategy.
In this embodiment, the target output result refers to an analysis result of the verification set by the index data state change prediction model obtained after the verification set is input into the index data state change prediction model for analysis.
In this embodiment, the standard indicator state refers to an actual state corresponding to the historical indicator data contained in the verification set, that is, an actual value corresponding to the historical indicator data contained in the verification set, a corresponding actual change trend, and the like.
In this embodiment, the difference data is used to represent the difference between the target output result and the standard index state, and may specifically be the difference between the values, the difference of the change trend of the index data state, and the like.
In this embodiment, the optimization policy refers to data that needs to be optimized and specific steps of optimization when the prediction of the index data by the index data state change prediction model is not qualified, where the mode of optimizing the index data state change prediction model is determined according to the prediction result.
In this embodiment, optimizing the index data state change prediction model based on the optimization strategy refers to performing parameter adjustment on holes in the index data state change prediction model according to the optimization strategy, so as to perfect the index data state change prediction model, and ensure that the index data state change prediction model can accurately and effectively analyze index data.
The beneficial effects of the technical scheme are as follows: the obtained index data state change prediction model is verified through the obtained verification set, and the verification result is compared with the standard index state corresponding to the verification set, so that when the prediction result is unqualified, an optimization strategy of the index data state change prediction model is accurately and effectively formulated, the index data state change prediction model is perfected through the optimization strategy, the accuracy and the reliability of the constructed index data state change prediction model are ensured, intelligent and reliable early warning analysis of the real-time state of index data is facilitated, and corresponding early warning operation is ensured in time when the index data is abnormal.
Example 6:
on the basis of embodiment 1, the present embodiment provides an automatic index early warning method based on a machine learning model, in step 2, analysis is performed on index data based on an index data state change prediction model, and a state change trend of the index data is determined, including:
acquiring index data to be analyzed, and visually displaying the index data to be analyzed in a preset rectangular coordinate system to obtain long-term change trend information corresponding to the index data to be analyzed;
removing abnormal values in the long-term change trend information, smoothing the long-term change trend information after removing the abnormal values, and periodically decomposing the long-term change trend information based on a smoothing result;
obtaining periodic variation data sequences based on periodic decomposition results, determining target influence factors influencing the state change of index data, respectively determining the correlation between different periodic variation data sequences and the target influence factors, and carrying out average value processing on the obtained multiple groups of correlations to obtain target association relation between the index data to be analyzed and the target influence factors;
the method comprises the steps of obtaining an index data state change prediction model, inputting index data to be analyzed and a target association relation into the index data state change prediction model, processing the index data to be analyzed and the target association relation to obtain random change information generated by target influence factors in a preset time period of the index data to be analyzed, and predicting values of the index data to be analyzed in the preset time period based on the random change information to obtain a state change trend of the index data.
In this embodiment, the index data to be analyzed refers to index data that needs to be subjected to index state prediction, and the index data is not unique.
In this embodiment, the preset rectangular coordinate system is set in advance, and is used for visually displaying the obtained index data to be analyzed.
In this embodiment, the long-term change trend information refers to the change condition of the value of the index data from the beginning to the end of the index data to be analyzed.
In this embodiment, the abnormal value refers to index data whose value in the long-term change trend information differs greatly from the average value or is invalid.
In this embodiment, the periodic decomposition refers to splitting the index data to be analyzed into segments with periodicity, so as to facilitate prediction of the trend of the index data in a future time period.
In this embodiment, the periodic variation data sequence refers to index data corresponding to each period obtained by periodically decomposing the long-term variation trend information of the index data, a value corresponding to the index data, and the like.
In this embodiment, the target influencing factor refers to a cause that influences the state change of the index data, so as to facilitate prediction of the change trend of the index data.
In this embodiment, the correlation is used to characterize the degree of correlation between the periodic variation data sequence and the resulting target impression factor, so as to facilitate determination of the trend of the index data under the influence of the target influence factor.
In this embodiment, the target association relationship is an interaction relationship between the target influence factor and the target data to be analyzed, which is used for characterizing the final result.
In this embodiment, the preset time period is set in advance, and is used to characterize the predicted time span of the index data.
In this embodiment, the random variation information refers to an interference condition that the target influencing factor affects the state of the index data to be analyzed in a preset period, that is, a specific numerical value affecting the state of the index data, and the like.
In this embodiment, predicting the value of the index data to be analyzed within a preset period of time based on the random variation information includes:
obtaining a prediction result of index data in a preset time period, and determining the data quantity of the index data obtained after prediction based on the prediction result;
calculating the prediction accuracy of the index data in a preset time period based on the data amount of the index data obtained after prediction, and calculating the comprehensive evaluation value of the index data processing based on the prediction accuracy, wherein the specific steps comprise:
Calculating the prediction accuracy of the index data in a preset time period according to the following formula:
wherein, eta represents the prediction accuracy of the index data in a preset time period, and the value range is (0, 1); λ represents an error factor and the range of values is (0.01,0.03); m represents the data amount of index data obtained after prediction; m represents the data quantity of abnormal index data which is removed when abnormal value detection is carried out on index data obtained after prediction, and the value is smaller than M; s represents the index data quantity which is mispredicted in the index data corresponding to the calculation prediction accuracy after the index data of the abnormal value is removed, and the value range is smaller than M-M;
the comprehensive evaluation value for the index data processing is calculated according to the following formula:
wherein,,
representing a comprehensive evaluation value for the index data processing; omega represents a weight value corresponding to the prediction accuracy in the index data processing process; η represents the prediction accuracy of the index data in a preset time period, and the value range is (0, 1); τ represents the accuracy of detecting the abnormal value of the index data obtained by prediction, and the value range is (0, 1);
comparing the calculated comprehensive evaluation value with a preset evaluation threshold value;
If the calculated comprehensive evaluation value is greater than or equal to a preset evaluation threshold value, judging that the processing of the index data is qualified, and completing automatic early warning of the index data;
otherwise, judging that the processing of the index data is unqualified, and determining a prediction vulnerability of the index data state change prediction model based on a prediction result;
optimizing the prediction loopholes of the index data state change prediction model, and predicting the values of the index data within a preset time period again based on the optimization result until the finally obtained comprehensive evaluation value is greater than or equal to a preset evaluation threshold value.
The comprehensive evaluation value is used for representing the processing effect on the index data when the index is automatically pre-warned, and the larger the value is, the better the processing effect on the index data is.
The preset evaluation threshold is set in advance, and is the lowest standard for measuring whether the index data processing meets the processing effect.
The prediction vulnerability refers to a defect of the state change prediction model of the index data when predicting the state change trend of the index data.
The beneficial effects of the technical scheme are as follows: the method comprises the steps of visually displaying index data to be analyzed, obtaining long-term change trend information of the index data to be analyzed according to a display result, removing abnormal values in the long-term change trend information, periodically decomposing the long-term change trend information after removing the abnormal values, finally, inputting an obtained periodic change data sequence and target influence factors influencing the state change of the index data into an index data state change prediction model to predict the state change trend of the index data, so that the state change trend of the index data is accurately and reliably predicted, intelligent and reliable early warning analysis on the real-time state of the index data is conveniently realized, corresponding early warning operation is timely performed when the index data is abnormal, meanwhile, the comprehensive evaluation value of index data processing is effectively calculated through calculation, the processing effect of the index data is accurately and effectively mastered according to a calculation result, the index data state change prediction model is also conveniently and timely perfected according to the calculation result, and the accuracy and the timeliness of automatic early warning on the index data are guaranteed.
Example 7:
on the basis of embodiment 1, the present embodiment provides an automatic index early warning method based on a machine learning model, in step 2, when a state change trend meets an early warning condition, an early warning stage for index data is determined, including:
acquiring service characteristics of a target service corresponding to the index data, setting a normal amplitude interval of the index data based on the service characteristics, acquiring a prediction result of a state change trend of the index data, and determining an amplitude change value of the index data based on the state change trend obtained by prediction;
comparing the amplitude variation value with a preset amplitude variation threshold value, and judging that the early warning condition is met when the amplitude variation value is larger than the preset amplitude variation threshold value;
and determining a target difference value between a peak value in the state change trend of the index data and a normal amplitude interval based on the judging result, matching the target difference value with an early warning stage difference value comparison table, and determining an early warning stage of the index data based on the matching result.
In this embodiment, the target service refers to a service type or the like corresponding to the index data.
In this embodiment, the service characteristics refer to the service type of the target service, the service operation mode in the operation process, and the like.
In this embodiment, the normal amplitude interval refers to the normal value corresponding to the index data of the target service in the normal operation process.
In this embodiment, the magnitude change value refers to a difference between a maximum value and a minimum value of the predicted index data within a preset time period.
In this embodiment, the preset amplitude change threshold is set in advance, and is used to characterize the maximum value change range of the allowable index data.
In this embodiment, the peak points refer to the maximum value point and the minimum value point in the state change trend of the index data.
In this embodiment, the target difference refers to the difference between the maximum value point and the minimum value point and the normal copy interval, respectively.
In this embodiment, the difference comparison table of the early warning stages is set in advance, and is used for representing the difference intervals corresponding to different early warning stages, specifically, when the target difference is larger, the early warning stage is higher.
The beneficial effects of the technical scheme are as follows: the method comprises the steps of determining the service characteristics of target service corresponding to index data, locking a normal amplitude interval of the index data, accurately and effectively judging an amplitude change value of the index data through a prediction result, and finally comparing the amplitude change value with a preset amplitude change threshold value, so that when an early warning condition is met, accurate and effective determination of an early warning stage is realized according to a peak point value in a state change trend of the index data and a target difference value of the normal amplitude interval, and corresponding early warning operation is conveniently carried out according to the early warning stage in a corresponding early warning mode, and timeliness and accuracy of early warning are guaranteed.
Example 8:
on the basis of embodiment 1, the embodiment provides an automatic index early warning method based on a machine learning model, in step 3, early warning requirements of an early warning stage are determined, and early warning operation is performed by matching a target early warning mode from a preset early warning mode library based on the early warning requirements, including:
acquiring an early warning stage of index data, extracting early warning characteristics of the early warning stage, and determining an early warning grade and an early warning type based on the early warning characteristics;
determining an early warning requirement for an early warning stage based on the early warning grade and the early warning type, and performing first matching on the early warning type in the early warning requirement and a preset early warning mode class in a preset early warning mode library to obtain a target early warning mode class;
and performing second matching on the early warning level in the early warning requirement and the early warning parameters of the preset early warning modes in the preset early warning mode class, obtaining a target early warning mode based on a matching result, and performing early warning operation based on the target early warning mode.
In this embodiment, the early warning features refer to early warning intensity, early warning characteristics, and the like corresponding to the early warning stage.
In this embodiment, the preset early warning modes are set in advance, and are used for storing early warning modes corresponding to different early warning types, and the early warning modes are not unique, and each early warning mode corresponds to a corresponding early warning intensity.
In this embodiment, the first matching refers to matching the pre-warning mode class corresponding to the pre-warning type from the pre-set pre-warning mode library.
In this embodiment, the target early warning mode class is an early warning mode class consistent with the early warning type, and all early warning modes corresponding to the early warning type are stored in the early warning mode class.
In this embodiment, the preset early warning mode is set in advance and is not unique.
In this embodiment, the early warning parameters refer to early warning intensities, early warning timeliness and the like corresponding to different early warning modes in early warning.
In this embodiment, the second matching refers to matching out the final needed early warning mode from the target early warning mode class according to the early warning level.
The beneficial effects of the technical scheme are as follows: by determining the corresponding early warning characteristics of the early warning stage, the early warning grade and the early warning type are accurately and effectively locked according to the early warning characteristics, the early warning requirements of the early warning stage are locked according to the early warning grade and the early warning type, and finally, the target early warning mode is determined according to the early warning requirements, so that the early warning mode of the early warning stage is accurately and reliably locked, and the early warning pertinence and the early warning timeliness of index data are ensured.
Example 9:
on the basis of embodiment 1, the present embodiment provides an automatic index early warning method based on a machine learning model, and in step 3, the method for performing synchronous early warning notification on a management terminal includes:
acquiring a first communication address of a management terminal and a second communication address of an early warning terminal, and constructing a data transmission link based on the first communication address and the second communication address;
acquiring an early warning operation corresponding to the early warning stage, converting the early warning operation into corresponding text content, and determining a target keyword corresponding to the early warning operation based on the text content;
and marking the target keywords, and transmitting the marking result and the text content to the management terminal based on the data transmission link to perform synchronous early warning notification.
In this embodiment, the first communication address characterizes an IP address of the management terminal.
In this embodiment, the second communication address refers to the IP address of the early warning terminal.
In this embodiment, the text content refers to converting the early warning operation into a corresponding code form or text form, so as to obtain the corresponding text content.
In this embodiment, the target keyword refers to an early warning purpose, an early warning condition, an early warning strength, an early warning type, and the like of the early warning operation.
The beneficial effects of the technical scheme are as follows: the method has the advantages that the data transmission link is accurately and reliably constructed by determining the first communication address of the management terminal and the second communication address of the early warning terminal, the early warning operation is converted into corresponding text content, the target keywords in the text content are marked, and the marked text content is transmitted to the management terminal for synchronous early warning notification, so that the accuracy, the intelligence and the timeliness of index early warning are guaranteed.
Example 10:
the embodiment provides an automatic index early warning system based on a machine learning model, as shown in fig. 3, including:
the model construction module is used for acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model;
the index analysis module is used for analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition;
and the early warning notification module is used for determining early warning requirements in the early warning stage, carrying out early warning operation by matching target early warning modes from a preset early warning mode library based on the early warning requirements, and carrying out synchronous early warning notification on the management terminal.
The beneficial effects of the technical scheme are as follows: by learning and training the historical index data, an index data state change prediction model is accurately and reliably constructed, intelligent and reliable early warning analysis on the real-time state of the index data is facilitated, corresponding early warning operation is timely carried out when the index data is abnormal, and meanwhile early warning notification is sent to a management terminal, so that the accuracy, the intelligence and the timeliness of early warning of the index are guaranteed.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (10)
1. An automatic index early warning method based on a machine learning model is characterized by comprising the following steps:
step 1: acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model;
step 2: analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition;
Step 3: and determining an early warning requirement in the early warning stage, carrying out early warning operation by matching a target early warning mode from a preset early warning mode library based on the early warning requirement, and carrying out synchronous early warning notification on the management terminal.
2. The automatic index early warning method based on a machine learning model according to claim 1, wherein in step 1, obtaining historical index data includes:
acquiring an index analysis request sent by a management terminal, analyzing the index analysis request, and determining a service identifier corresponding to the index analysis request;
generating a data calling request based on the service identifier, transmitting the data calling request to a preset server, analyzing the data calling request based on the preset server, and searching the service database based on the service identifier to obtain a target service database corresponding to the service identifier;
and extracting the historical index data in the target service database, packaging the historical index data and feeding back to the data receiving end.
3. The automatic index early warning method based on a machine learning model according to claim 2, wherein the step of packaging the historical index data and feeding the packaged historical index data back to the data receiving end comprises the steps of:
Acquiring the obtained historical index data, and clustering the historical index data to obtain a target classification result corresponding to the historical index data;
extracting isolated sample data contained in index data of each category based on a clustering result, extracting data characteristics of the index data of each category, and matching target cleaning rules from a preset data cleaning rule base based on the data characteristics;
performing first cleaning on the isolated sample data based on the target cleaning rule, determining a cleaning result of the isolated sample, and performing second cleaning on the index data of the corresponding category based on the target cleaning rule when the cleaning result meets the expected requirement;
and obtaining standard index data based on the second cleaning result, and storing the standard index data.
4. The automatic index early warning method based on a machine learning model according to claim 1, wherein in step 1, historical index data is trained based on an index analysis rule and an index outlier detection rule, and an index data state change prediction model is constructed, comprising:
acquiring the obtained historical index data, determining service operation attributes corresponding to the historical index data, and determining index analysis rules and outlier detection rules for the historical index data based on the service operation attributes;
Dividing the obtained historical index data into a training set and a verification set, determining a logic analysis structure of the historical index data based on an index analysis rule and an abnormal value detection rule, and determining target iteration times based on the data quantity of the historical index data contained in the training set;
performing iterative training of target iterative times on the training set based on the logic analysis structure to obtain logic analysis results after each iterative training, determining a reference logic analysis result corresponding to each historical index data in the training set, and determining error data of the logic analysis results after each iterative training and the corresponding reference logic analysis results;
and determining a weight value of the logic analysis result after each iteration training based on the error data, correcting the logic analysis result after each iteration training based on the weight value to obtain a target logic analysis result, and constructing an index data state change prediction model based on the target logic analysis result.
5. The automatic index early warning method based on a machine learning model according to claim 4, wherein the construction of the index data state change prediction model based on the target logic analysis result comprises:
Acquiring a constructed index data state change prediction model and a verification set obtained by splitting historical index data, and inputting the verification set into the constructed index data state change prediction model for analysis to obtain a target output result;
extracting standard index states corresponding to each historical index data in the verification set, judging that the constructed index data state change prediction model is qualified when the target output result is consistent with the standard index states, otherwise, judging that the constructed index data state change prediction model is unqualified, and determining difference data between the target output result and the standard index states;
and determining an optimization strategy for the index data state change prediction model based on the difference data, and optimizing the index data state change prediction model based on the optimization strategy.
6. The automatic index early warning method based on a machine learning model according to claim 1, wherein in step 2, the state change trend of the index data is determined by analyzing the index data based on the state change prediction model of the index data, and the method comprises the following steps:
acquiring index data to be analyzed, and visually displaying the index data to be analyzed in a preset rectangular coordinate system to obtain long-term change trend information corresponding to the index data to be analyzed;
Removing abnormal values in the long-term change trend information, smoothing the long-term change trend information after removing the abnormal values, and periodically decomposing the long-term change trend information based on a smoothing result;
obtaining periodic variation data sequences based on periodic decomposition results, determining target influence factors influencing the state change of index data, respectively determining the correlation between different periodic variation data sequences and the target influence factors, and carrying out average value processing on the obtained multiple groups of correlations to obtain target association relation between the index data to be analyzed and the target influence factors;
the method comprises the steps of obtaining an index data state change prediction model, inputting index data to be analyzed and a target association relation into the index data state change prediction model, processing the index data to be analyzed and the target association relation to obtain random change information generated by target influence factors in a preset time period of the index data to be analyzed, and predicting values of the index data to be analyzed in the preset time period based on the random change information to obtain a state change trend of the index data.
7. The automatic early warning method for indexes based on a machine learning model according to claim 1, wherein in step 2, when the state change trend satisfies the early warning condition, the early warning stage for the index data is determined, which comprises:
Acquiring service characteristics of a target service corresponding to the index data, setting a normal amplitude interval of the index data based on the service characteristics, acquiring a prediction result of a state change trend of the index data, and determining an amplitude change value of the index data based on the state change trend obtained by prediction;
comparing the amplitude variation value with a preset amplitude variation threshold value, and judging that the early warning condition is met when the amplitude variation value is larger than the preset amplitude variation threshold value;
and determining a target difference value between a peak value in the state change trend of the index data and a normal amplitude interval based on the judging result, matching the target difference value with an early warning stage difference value comparison table, and determining an early warning stage of the index data based on the matching result.
8. The automatic early warning method for indexes based on a machine learning model according to claim 1, wherein in step 3, the early warning requirement of the early warning stage is determined, and the early warning operation is performed by matching the target early warning mode from a preset early warning mode library based on the early warning requirement, and the method comprises the following steps:
acquiring an early warning stage of index data, extracting early warning characteristics of the early warning stage, and determining an early warning grade and an early warning type based on the early warning characteristics;
Determining an early warning requirement for an early warning stage based on the early warning grade and the early warning type, and performing first matching on the early warning type in the early warning requirement and a preset early warning mode class in a preset early warning mode library to obtain a target early warning mode class;
and performing second matching on the early warning level in the early warning requirement and the early warning parameters of the preset early warning modes in the preset early warning mode class, obtaining a target early warning mode based on a matching result, and performing early warning operation based on the target early warning mode.
9. The automatic index early warning method based on the machine learning model according to claim 1, wherein in step 3, the synchronous early warning notification is performed on the management terminal, and the method comprises the following steps:
acquiring a first communication address of a management terminal and a second communication address of an early warning terminal, and constructing a data transmission link based on the first communication address and the second communication address;
acquiring an early warning operation corresponding to the early warning stage, converting the early warning operation into corresponding text content, and determining a target keyword corresponding to the early warning operation based on the text content;
and marking the target keywords, and transmitting the marking result and the text content to the management terminal based on the data transmission link to perform synchronous early warning notification.
10. An automatic index early warning system based on a machine learning model is characterized by comprising:
the model construction module is used for acquiring historical index data, training the historical index data based on an index analysis rule and an index abnormal value detection rule, and constructing an index data state change prediction model;
the index analysis module is used for analyzing the index data based on the index data state change prediction model, determining the state change trend of the index data, and determining an early warning stage of the index data when the state change trend meets the early warning condition;
and the early warning notification module is used for determining early warning requirements in the early warning stage, carrying out early warning operation by matching target early warning modes from a preset early warning mode library based on the early warning requirements, and carrying out synchronous early warning notification on the management terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310121509.0A CN116415840B (en) | 2023-02-02 | 2023-02-02 | Automatic index early warning method and system based on machine learning model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310121509.0A CN116415840B (en) | 2023-02-02 | 2023-02-02 | Automatic index early warning method and system based on machine learning model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116415840A true CN116415840A (en) | 2023-07-11 |
| CN116415840B CN116415840B (en) | 2023-12-05 |
Family
ID=87053971
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310121509.0A Active CN116415840B (en) | 2023-02-02 | 2023-02-02 | Automatic index early warning method and system based on machine learning model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116415840B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116703245A (en) * | 2023-08-02 | 2023-09-05 | 宜兴市旭航电子有限公司 | System and method for monitoring and early warning of flow production process |
| CN116827658A (en) * | 2023-07-17 | 2023-09-29 | 青岛启弘信息科技有限公司 | AI intelligent application security situation awareness prediction system and method |
| CN117240594A (en) * | 2023-10-31 | 2023-12-15 | 深圳市常行科技有限公司 | Multi-dimensional network security operation and maintenance protection management system and method |
| CN117252446A (en) * | 2023-11-15 | 2023-12-19 | 青岛海信信息科技股份有限公司 | Method and system for process index extraction and index intelligent operation |
| CN117319440A (en) * | 2023-09-26 | 2023-12-29 | 深圳达普信科技有限公司 | Data acquisition and analysis method and system based on Internet of vehicles |
| CN117336055A (en) * | 2023-10-08 | 2024-01-02 | 深圳市马博士网络科技有限公司 | Network abnormal behavior detection method and device, electronic equipment and storage medium |
| CN117453805A (en) * | 2023-12-22 | 2024-01-26 | 石家庄学院 | Visual analysis method for uncertainty data |
| CN117540369A (en) * | 2023-11-13 | 2024-02-09 | 深圳市马博士网络科技有限公司 | Numerical behavior safety base line generation method and device for safety analysis |
| CN117633451A (en) * | 2023-10-20 | 2024-03-01 | 深圳达普信科技有限公司 | Intelligent cabin health data analysis method based on data mining and deep learning |
| CN118016202A (en) * | 2024-04-10 | 2024-05-10 | 华能山东发电有限公司白杨河发电厂 | Chemical equipment operation analysis method and system based on steam-water quality |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109784471A (en) * | 2018-10-24 | 2019-05-21 | 浙江华云信息科技有限公司 | A kind of prediction of 95598 traffic work orders and unusual fluctuation method for early warning based on more prediction models |
| CN111459778A (en) * | 2020-03-12 | 2020-07-28 | 平安科技（深圳）有限公司 | Operation and maintenance system abnormal index detection model optimization method and device and storage medium |
| US20210158213A1 (en) * | 2019-11-22 | 2021-05-27 | Canon Kabushiki Kaisha | Learning model management system, learning model management method, and recording medium |
| US20210311166A1 (en) * | 2015-07-17 | 2021-10-07 | Chenshu Wu | Method, apparatus, and system for wireless micro motion monitoring |
| CN113901105A (en) * | 2021-09-24 | 2022-01-07 | 中冶长天国际工程有限责任公司 | Rock mass instability early warning method and system based on stress monitoring data change |
| CN114116397A (en) * | 2021-11-29 | 2022-03-01 | 深圳壹账通智能科技有限公司 | Early warning attribution method, device, equipment and storage medium for monitoring indexes |
-
2023
- 2023-02-02 CN CN202310121509.0A patent/CN116415840B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20210311166A1 (en) * | 2015-07-17 | 2021-10-07 | Chenshu Wu | Method, apparatus, and system for wireless micro motion monitoring |
| CN109784471A (en) * | 2018-10-24 | 2019-05-21 | 浙江华云信息科技有限公司 | A kind of prediction of 95598 traffic work orders and unusual fluctuation method for early warning based on more prediction models |
| US20210158213A1 (en) * | 2019-11-22 | 2021-05-27 | Canon Kabushiki Kaisha | Learning model management system, learning model management method, and recording medium |
| CN111459778A (en) * | 2020-03-12 | 2020-07-28 | 平安科技（深圳）有限公司 | Operation and maintenance system abnormal index detection model optimization method and device and storage medium |
| CN113901105A (en) * | 2021-09-24 | 2022-01-07 | 中冶长天国际工程有限责任公司 | Rock mass instability early warning method and system based on stress monitoring data change |
| CN114116397A (en) * | 2021-11-29 | 2022-03-01 | 深圳壹账通智能科技有限公司 | Early warning attribution method, device, equipment and storage medium for monitoring indexes |
Non-Patent Citations (1)
| Title |
|---|
| 张弛: "基于先行指标的湖北省短期经济预测研究", 中国优秀硕士学位论文全文数据库 (经济与管理科学辑), no. 01, pages 147 - 1198 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116827658A (en) * | 2023-07-17 | 2023-09-29 | 青岛启弘信息科技有限公司 | AI intelligent application security situation awareness prediction system and method |
| CN116827658B (en) * | 2023-07-17 | 2024-01-16 | 青岛启弘信息科技有限公司 | AI intelligent application security situation awareness prediction system and method |
| CN116703245B (en) * | 2023-08-02 | 2023-10-27 | 宜兴市旭航电子有限公司 | System and method for monitoring and early warning of flow production process |
| CN116703245A (en) * | 2023-08-02 | 2023-09-05 | 宜兴市旭航电子有限公司 | System and method for monitoring and early warning of flow production process |
| CN117319440A (en) * | 2023-09-26 | 2023-12-29 | 深圳达普信科技有限公司 | Data acquisition and analysis method and system based on Internet of vehicles |
| CN117336055B (en) * | 2023-10-08 | 2024-07-23 | 深圳市马博士网络科技有限公司 | Network abnormal behavior detection method and device, electronic equipment and storage medium |
| CN117336055A (en) * | 2023-10-08 | 2024-01-02 | 深圳市马博士网络科技有限公司 | Network abnormal behavior detection method and device, electronic equipment and storage medium |
| CN117633451A (en) * | 2023-10-20 | 2024-03-01 | 深圳达普信科技有限公司 | Intelligent cabin health data analysis method based on data mining and deep learning |
| CN117633451B (en) * | 2023-10-20 | 2024-07-05 | 深圳达普信科技有限公司 | Intelligent cabin health data analysis method based on data mining and deep learning |
| CN117240594A (en) * | 2023-10-31 | 2023-12-15 | 深圳市常行科技有限公司 | Multi-dimensional network security operation and maintenance protection management system and method |
| CN117240594B (en) * | 2023-10-31 | 2024-06-18 | 深圳市常行科技有限公司 | Multi-dimensional network security operation and maintenance protection management system and method |
| CN117540369A (en) * | 2023-11-13 | 2024-02-09 | 深圳市马博士网络科技有限公司 | Numerical behavior safety base line generation method and device for safety analysis |
| CN117540369B (en) * | 2023-11-13 | 2024-06-21 | 深圳市马博士网络科技有限公司 | Numerical behavior safety base line generation method and device for safety analysis |
| CN117252446B (en) * | 2023-11-15 | 2024-02-13 | 青岛海信信息科技股份有限公司 | Method and system for process index extraction and index intelligent operation |
| CN117252446A (en) * | 2023-11-15 | 2023-12-19 | 青岛海信信息科技股份有限公司 | Method and system for process index extraction and index intelligent operation |
| CN117453805A (en) * | 2023-12-22 | 2024-01-26 | 石家庄学院 | Visual analysis method for uncertainty data |
| CN118016202A (en) * | 2024-04-10 | 2024-05-10 | 华能山东发电有限公司白杨河发电厂 | Chemical equipment operation analysis method and system based on steam-water quality |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116415840B (en) | 2023-12-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116415840B (en) | Automatic index early warning method and system based on machine learning model | |
| WO2021184630A1 (en) | Method for locating pollutant discharge object on basis of knowledge graph, and related device | |
| CN105868373B (en) | Method and device for processing key data of power business information system | |
| CN113935497A (en) | Intelligent operation and maintenance fault processing method, device and equipment and storage medium thereof | |
| CN109740838B (en) | Provider service evaluation method based on big data and related equipment | |
| CN114116397A (en) | Early warning attribution method, device, equipment and storage medium for monitoring indexes | |
| CN115409395B (en) | Quality acceptance inspection method and system for hydraulic construction engineering | |
| CN113111095B (en) | Intelligent information management method and system | |
| CN111539493A (en) | Alarm prediction method and device, electronic equipment and storage medium | |
| US9740986B2 (en) | System and method for deducing user interaction patterns based on limited activities | |
| CN118011990B (en) | Industrial data quality monitoring and improving system based on artificial intelligence | |
| CN111984442A (en) | Method and device for detecting abnormality of computer cluster system, and storage medium | |
| CN115204536A (en) | Building equipment fault prediction method, device, equipment and storage medium | |
| US20090276390A1 (en) | Modeling support system, modeling support method, and modeling support program | |
| CN114138601A (en) | Service alarm method, device, equipment and storage medium | |
| CN117240527B (en) | Network security risk prevention system and method | |
| Basili et al. | Domain analysis for the reuse of software development experiences | |
| CN117667585A (en) | Operation and maintenance efficiency evaluation method and system based on operation and maintenance quality management database | |
| CN117592656A (en) | Carbon footprint monitoring method and system based on carbon data accounting | |
| CN110113428B (en) | System and method for acquiring and analyzing data of wifi product on production line | |
| CN114266483B (en) | Dangerous waste monitoring system based on Internet of things | |
| CN115098336A (en) | Method, system, equipment and storage medium for monitoring warehouse tasks | |
| CN112215514A (en) | Operation analysis report generation method and system | |
| Azzalini et al. | Data Quality and Data Ethics: Towards a Trade-off Evaluation. | |
| US8682817B1 (en) | Product testing process using diploid evolutionary method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |