CN115204278A - Abnormal sample detection method and device, electronic device and storage medium - Google Patents
Abnormal sample detection method and device, electronic device and storage medium Download PDFInfo
- Publication number
- CN115204278A CN115204278A CN202210739714.9A CN202210739714A CN115204278A CN 115204278 A CN115204278 A CN 115204278A CN 202210739714 A CN202210739714 A CN 202210739714A CN 115204278 A CN115204278 A CN 115204278A
- Authority
- CN
- China
- Prior art keywords
- sample
- abnormal
- prediction
- target
- abnormal sample
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The embodiment of the application provides an abnormal sample detection method and device, electronic equipment and a storage medium, and belongs to the technical field of artificial intelligence, wherein the abnormal sample detection method comprises the following steps: determining a prediction time; acquiring a training sample set and a prediction sample set according to the prediction time; obtaining an initial regression model, and performing model training on the initial regression model according to a training sample set to obtain a target regression model and a first commission predicted value corresponding to the training sample set; inputting the prediction sample set into a target regression model, and predicting the prediction sample set according to the target regression model to obtain a second commission prediction value; determining a target abnormal sample output threshold value according to the first commission predicted value and the second commission predicted value; and screening out the target abnormal sample from the prediction sample set according to the target abnormal sample output threshold value. The abnormal sample detection method provided by the embodiment of the application can generate a large amount of abnormal sample data, and solves the problem that the abnormal sample data accounts for a small amount or is missing.
Description
Technical Field
The present disclosure relates to the field of artificial intelligence technologies, and in particular, to a method and an apparatus for detecting an abnormal sample, an electronic device, and a storage medium.
Background
In the related art, the detection of abnormal sample data is performed according to the classification model, but the abnormal sample data occupies a small percentage and even is missing due to the fact that the abnormality belongs to a small-probability event, and the abnormal sample data cannot be provided for training of the classification model.
Disclosure of Invention
The embodiment of the application mainly aims to provide an abnormal sample detection method and device, electronic equipment and a storage medium, which can automatically generate a large amount of abnormal sample data and solve the problem that the abnormal sample data occupies a small amount or is missing.
In order to achieve the above object, a first aspect of an embodiment of the present application provides an abnormal sample detection method, where the method includes:
determining a prediction time;
acquiring a training sample set and a prediction sample set according to the prediction time;
obtaining an initial regression model, and performing model training on the initial regression model according to the training sample set to obtain a target regression model and a first commission predicted value corresponding to the training sample set;
inputting the prediction sample set into the target regression model, and predicting the prediction sample set according to the target regression model to obtain a second commission prediction value;
determining a target abnormal sample output threshold value according to the first commission predicted value and the second commission predicted value;
and screening out a target abnormal sample from the prediction sample set according to the target abnormal sample output threshold value.
In some embodiments, the determining a target exception sample output threshold from the first commission predicted value and the second commission predicted value comprises:
acquiring a first commission trial value of the training sample set;
obtaining a first deviation according to the first commission trial value and the first commission predicted value;
obtaining a second commission trial value for the prediction sample set;
obtaining a second deviation according to the second commission trial value and the second commission predicted value;
determining a target abnormal sample output threshold value according to the first deviation and the second deviation.
In some embodiments, said determining a target abnormal sample output threshold from said first deviation and said second deviation comprises:
calculating a first mean and a first standard deviation of the first deviation of the set of training samples;
determining a first abnormal sample output threshold according to the first mean value and the first standard deviation;
calculating a second mean and a second standard deviation of the second deviation of the set of prediction samples;
determining a second abnormal sample output threshold according to the second mean value and the second standard deviation;
and determining a target abnormal sample output threshold according to the first abnormal sample output threshold and the second abnormal sample output threshold.
In some embodiments, said determining a target abnormal sample output threshold from said first abnormal sample output threshold and said second abnormal sample output threshold comprises:
and if the first abnormal sample output threshold value is smaller than the second abnormal sample output threshold value, taking the first abnormal sample output threshold value as the target abnormal sample output threshold value.
In some embodiments, said determining a target outlier sample output threshold from said first variance and said second variance comprises:
acquiring a first confidence coefficient of the training sample set;
determining a first abnormal sample output threshold according to the first deviation and the first confidence;
obtaining a second confidence of the prediction sample set;
determining a second abnormal sample output threshold according to the second deviation and the second confidence;
and determining a target abnormal sample output threshold according to the first abnormal sample output threshold and the second abnormal sample output threshold.
In some embodiments, said screening out a target outlier sample from said set of predicted samples according to said outlier sample output threshold comprises:
screening out samples with second deviation larger than or equal to the target abnormal sample output threshold value from the prediction sample set to obtain initial abnormal samples;
acquiring a preset sample quantity threshold;
and screening out a target abnormal sample from the initial abnormal samples according to the sample quantity threshold value.
In some embodiments, the screening out target abnormal samples from the initial abnormal samples according to the sample number threshold comprises:
if the sample number of the initial abnormal sample is smaller than or equal to the sample number threshold, taking the initial abnormal sample as the target abnormal sample;
and if the number of the samples of the initial abnormal samples is greater than the sample number threshold, randomly selecting the samples with the number equal to the sample number threshold from the initial abnormal samples to obtain the target abnormal samples.
A second aspect of the embodiments of the present application provides an abnormal sample detection apparatus, including:
a first obtaining module for determining a prediction time;
the second acquisition module is used for acquiring a training sample set and a prediction sample set according to the prediction time;
the model training module is used for obtaining an initial regression model, performing model training on the initial regression model according to the training sample set and obtaining a target regression model and a first commission predicted value corresponding to the training sample set;
the model prediction module is used for inputting the prediction sample set into the target regression model, predicting the commission of the prediction sample set according to the target regression model and obtaining a second commission prediction value;
the threshold value calculation module is used for determining a target abnormal sample output threshold value according to the first commission prediction value and the second commission prediction value;
and the abnormal sample detection module is used for screening out the target abnormal sample from the prediction sample set according to the target abnormal sample output threshold value.
A third aspect of embodiments of the present application provides an electronic device, which includes a memory and a processor, where the memory stores a program, and the processor is configured to execute the method for detecting an abnormal sample according to any one of the embodiments of the first aspect of the present application when the program is executed by the processor.
A fourth aspect of the embodiments of the present application provides a storage medium, which is a computer-readable storage medium, where computer-executable instructions are stored in the storage medium, and the computer-executable instructions are configured to cause a computer to execute the abnormal sample detection method according to any one of the embodiments of the first aspect of the present application.
According to the abnormal sample detection method and device, the electronic device and the storage medium, prediction time is determined, a training sample set and a prediction sample set are obtained according to the prediction time, an initial regression model is obtained, model training is conducted on the initial regression model according to the training sample set, a first commission predicted value corresponding to a target regression model and the training sample set is obtained, the prediction sample set is input into the target regression model, the prediction sample set is predicted according to the target regression model, a second commission predicted value is obtained, a target abnormal sample output threshold value is determined according to the first commission predicted value and the second commission predicted value, a target abnormal sample is screened out from the prediction sample set according to the target abnormal sample output threshold value, a large amount of abnormal sample data can be generated, and the problem that the abnormal sample data occupies less or is lost is solved.
Drawings
FIG. 1 is a first flowchart of an abnormal sample detection method provided by an embodiment of the present application;
FIG. 2 is a flowchart of step S150 in FIG. 1;
FIG. 3 is a first flowchart of step S250 in FIG. 2;
FIG. 4 is a second flowchart of step S250 in FIG. 2;
fig. 5 is a flowchart of step S160 in fig. 1;
FIG. 6 is a flowchart of step S530 in FIG. 5;
fig. 7 is a block diagram of a module structure of an abnormal sample detection apparatus according to an embodiment of the present application;
fig. 8 is a schematic hardware structure diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of and not restrictive on the broad application.
It is noted that while functional block divisions are provided in device diagrams and logical sequences are shown in flowcharts, in some cases, steps shown or described may be performed in sequences other than block divisions within devices or flowcharts. The terms first, second and the like in the description and in the claims, and the drawings described above, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
In the embodiments of the present application, when data related to the identity or the characteristic of a user, such as user information, user behavior data, user history data, and user location information, is processed, permission or approval of the user is obtained, and the data collection, use, and processing, etc., comply with relevant laws and regulations and standards of relevant countries and regions. In addition, when the embodiment of the present application needs to acquire sensitive personal information of a user, individual permission or individual consent of the user is obtained through a pop-up window or a jump to a confirmation page, and after the individual permission or individual consent of the user is definitely obtained, necessary user-related data for enabling the embodiment of the present application to operate normally is acquired.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the application.
Furthermore, the described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the embodiments of the disclosure can be practiced without one or more of the specific details, or with other methods, components, devices, steps, and the like. In other instances, well-known methods, devices, implementations, or operations have not been shown or described in detail to avoid obscuring aspects of the disclosure.
The block diagrams shown in the figures are functional entities only and do not necessarily correspond to physically separate entities. I.e. these functional entities may be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and/or processor means and/or microcontroller means.
The flowcharts shown in the figures are illustrative only and do not necessarily include all of the contents and operations/steps, nor do they necessarily have to be performed in the order described. For example, some operations/steps may be decomposed, and some operations/steps may be combined or partially combined, so that the actual execution sequence may be changed according to the actual situation.
First, several terms referred to in the present application are resolved:
artificial Intelligence (AI): is a new technical science for researching and developing theories, methods, technologies and application systems for simulating, extending and expanding human intelligence; artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produces a new intelligent machine that can react in a manner similar to human intelligence, and research in this field includes robotics, language recognition, image recognition, natural language processing, and expert systems, among others. The artificial intelligence can simulate the information process of human consciousness and thinking. Artificial intelligence is also a theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results.
Gradient Boosting Decision Tree (GBDT): the method is a model in machine learning, and the main idea is to use a decision tree for iterative training, train a tree by reducing residual errors during each iteration, add all the trees to obtain an optimal model, and the optimal model is usually used for tasks such as multi-classification, click rate prediction, search sequencing and the like.
Lightweight Gradient elevator (Light Gradient Boosting Machine, lightGBM): the GBDT algorithm is optimized from three dimensions of feature quantity, split point quantity and sample quantity, the histogram algorithm is adopted to reduce the number of split points, the unilateral gradient sampling algorithm is adopted to reduce the number of samples, and the mutually exclusive feature bundling algorithm is adopted to reduce the number of features.
In the related art, the detection of abnormal sample data is performed according to the classification model, but the abnormal sample data occupies a small proportion and even is missing due to the fact that the abnormality belongs to a small-probability event, and the abnormal sample data cannot be provided for training the classification model.
Based on this, the embodiment of the application provides an abnormal sample detection method and apparatus, an electronic device, and a storage medium, wherein a first commission predicted value of a training sample set and a second commission predicted value of a prediction sample set are calculated through a regression model, a target abnormal sample output threshold value is determined according to the first commission predicted value and the second commission predicted value, and a target abnormal sample is screened out from the prediction sample set according to the target abnormal sample output threshold value, so that a large amount of abnormal sample data can be generated, and the problem that the abnormal sample data accounts for less or is missing is solved.
The method and apparatus for detecting an abnormal sample, the electronic device, and the storage medium provided in the embodiments of the present application are specifically described in the following embodiments, and first, the method for detecting an abnormal sample in the embodiments of the present application is described.
The embodiment of the application provides an abnormal sample detection method, and relates to the technical field of artificial intelligence. The abnormal sample detection method provided by the embodiment of the application can be applied to a terminal, a server side and software running in the terminal or the server side. In some embodiments, the terminal may be a smartphone, tablet, laptop, desktop computer, smart watch, or the like; the server side can be configured into an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, and cloud servers for providing basic cloud computing services such as cloud service, a cloud database, cloud computing, cloud functions, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN (content delivery network) and big data and artificial intelligence platforms; the software may be an application or the like that implements an abnormal sample detection method, but is not limited to the above form.
Embodiments of the application are operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like. The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
Referring to fig. 1, the abnormal sample detection method according to the embodiment of the first aspect of the embodiment of the present application includes, but is not limited to, step S110 to step S160.
Step S110, determining the prediction time;
step S120, acquiring a training sample set and a prediction sample set according to the prediction time;
step S130, an initial regression model is obtained, model training is carried out on the initial regression model according to a training sample set, and a first commission predicted value corresponding to a target regression model and the training sample set is obtained;
step S140, inputting the prediction sample set into a target regression model, and predicting the prediction sample set according to the target regression model to obtain a second commission prediction value;
step S150, determining a target abnormal sample output threshold value according to the first commission predicted value and the second commission predicted value;
and step S160, screening out the target abnormal samples from the prediction sample set according to the target abnormal sample output threshold value.
In step S110 of some embodiments, in the commission trial-calculation scenario of the insurance agent, the commission trial-calculation release date is 20 days per month, and in order to predict the commission amount that should be released per month by the insurance agent, the prediction time is in months. If it is desired to predict the amount of commissions that the insurance agent issued in month 5 of 2022, the predicted time may be determined to be month 5 of 2022.
In step S120 of some embodiments, a training sample set and a prediction sample set are obtained according to the prediction time, where the training sample set and the prediction sample set are both history sample data of all the insurance agents under employment less than the prediction time, and specifically, the history sample data includes the monthly policy sales volume, the cumulative policy volume, the comprehensive product sales volume, and the like. If the prediction time is 2022 year 5 month, the training sample set is history sample data of all insurance agents who are on duty in 2022 year 3 month, and the prediction sample set is history sample data of all insurance agents who are on duty in 2022 year 4 month.
In step S130 of some embodiments, the initial regression model may be LightGBM, if the training sample set obtained in step S120 is historical sample data of all insurance agents working in 4 months in 2022, the training sample set in 4 months in 2022 is preprocessed to remove noise samples, the preprocessed training sample set is divided into 70% of training sets and 30% of validation sets, and model training is performed on the LightGBM model according to the training sets and the validation sets to obtain the target regression model and the first commission predicted value corresponding to the training sample set, where the noise sample is a sample whose sample data feature value in the training sample set is null or does not satisfy a preset value range, and it can be understood that the preset value range can be defined according to business requirements. The target regression model is a trained LightGBM model, and the first commission prediction value is a prediction value of commission due to the insurance agent in 3 months of 2022 and the insurance agent in 4 months of 2022.
In step S140 of some embodiments, if the prediction sample set obtained in step S120 is historical sample data of all insurance agents who are on duty in 4 months in 2022, the prediction sample set in 4 months in 2022 is input to the trained LightGBM model, and the commission amount corresponding to the prediction sample set is predicted according to the trained LightGBM model to obtain a second commission predicted value, where the second commission predicted value is a predicted value of commission that the insurance agents in 4 months in 2022 should issue commissions in 5 months in 2022.
In steps S150 to S160 of some embodiments, a target abnormal sample output threshold value is determined according to the first commission prediction value obtained in step S130 and the second commission prediction value obtained in step S140, and a sample having a deviation value greater than the target abnormal sample output threshold value is screened from the prediction sample set as a target abnormal sample.
According to the abnormal sample detection method provided by the embodiment of the application, the prediction time is determined, the training sample set and the prediction sample set are obtained according to the prediction time, the initial regression model is obtained, model training is performed on the initial regression model according to the training sample set, the first commission predicted value corresponding to the target regression model and the training sample set is obtained, the prediction sample set is input into the target regression model, the prediction sample set is predicted according to the target regression model, the second commission predicted value is obtained, the target abnormal sample output threshold value is determined according to the first commission predicted value and the second commission predicted value, the target abnormal sample is screened out from the prediction sample set according to the target abnormal sample output threshold value, in the scene that the abnormal sample data accounts for less or is missing, sample abnormal detection can still be performed, meanwhile, a large amount of abnormal sample data can be generated, and the problem that the classification model cannot be trained due to the abnormal sample data accounts for less or is missing is solved.
In some embodiments, as shown in fig. 2, step S150 specifically includes, but is not limited to, step S210 to step S250.
Step S210, acquiring a first commission trial value of a training sample set;
step S220, obtaining a first deviation according to the first commission trial value and the first commission predicted value;
step S230, acquiring a second commission trial value of the prediction sample set;
step S240, obtaining a second deviation according to the second commission trial value and the second commission predicted value;
and step S250, determining a target abnormal sample output threshold according to the first deviation and the second deviation.
In step S210 of some embodiments, if the training sample set is history sample data of 3 months in 2022, the first commission value corresponding to the training sample set is an actual commission value that should be issued by an insurance agent who is on duty in 3 months in 2022 in 4 months in 2022, and the model training is performed on the initial LightGBM model according to the training sample set and the corresponding target variable.
In step S220 of some embodiments, if the first commission trial value is represented as F tr-real The first commission prediction value is represented by F tr-pred And the first deviation is expressed as Δ tr, the calculation method of the first deviation is shown in formula (1).
Δtr＝|F tr-real -F tr-pred | (1)
In step S230 of some embodiments, if the prediction sample set is historical sample data of 4 months in 2022, the second commission trial value corresponding to the prediction sample set is the actual commission value that should be issued by the insurance agent who is on duty in 4 months in 2022 in 5 months.
In step S240 of some embodiments, if the second commission trial-calculated value is represented as F te-try The second commission prediction value is represented as F te-pred And the second deviation is expressed as Δ te, thenThe calculation method of the second deviation is shown in formula (2).
Δte＝|F te-try -F te-pred | (2)
In step S250 of some embodiments, a target abnormal sample output threshold is determined according to the first deviation obtained in step S220 and the second deviation obtained in step S240, and a sample having a second deviation value greater than the target abnormal sample output threshold is selected from the prediction sample set as the target abnormal sample.
In steps S210 to S250 of some embodiments, the target abnormal sample is screened from the prediction sample set based on the target abnormal sample output threshold, and the sample abnormality detection can still be performed in a scene where the abnormal sample accounts for less or is missing, so that the problems that the classification model cannot be trained and the abnormal sample detection can be performed according to the classification model in the scene are avoided, and meanwhile, the screened target abnormal sample can be used for training the classification model.
In some embodiments, as shown in fig. 3, step S250 specifically includes, but is not limited to, step S310 to step S350.
Step S310, calculating a first mean value and a first standard deviation of a first deviation of a training sample set;
step S320, determining a first abnormal sample output threshold value according to the first mean value and the first standard deviation;
step S330, calculating a second mean and a second standard deviation of a second deviation of the prediction sample set;
step S340, determining a second abnormal sample output threshold according to the second mean value and the second standard deviation;
step S350, determining a target abnormal sample output threshold according to the first abnormal sample output threshold and the second abnormal sample output threshold.
In step S310 of some embodiments, assuming that the first deviation Δ tr of each sample in the training sample set satisfies a normal distribution, the first deviations of the samples are added to obtain a first deviation sum, a first sample number in the training sample set is obtained, a first mean value is obtained according to a ratio of the first deviation sum and the first sample number, and the first deviation sum and the first sample number in the training sample set are comparedAnd carrying out subtraction operation on the first average value to obtain a first deviation difference value, calculating the sum of squares of the first deviation difference value of each sample, and obtaining a first standard deviation according to the ratio of the sum of squares to the number of the first samples. For example, the first number of samples in the training sample set is N, and the first deviations corresponding to the samples are Δ tr respectively 1 ,…,Δtr N If the first deviation satisfies the normal distribution, the first mean value mu tr Is given by equation (3), the first standard deviation σ tr The calculation method of (2) is shown in formula (4).
In step S320 of some embodiments, a first confidence level is obtained, a first coefficient is determined according to the first confidence level, and a first abnormal sample output threshold is determined according to the first mean, the first standard deviation, and the first coefficient. The first coefficient varies with the first confidence level, for example, when the first deviation satisfies the positive distribution, the first coefficient is 3 if the first confidence level is 0.99, the first coefficient is 2 if the first confidence level is 0.95, and the first coefficient is 1 if the first confidence level is 0.65. If the first coefficient is denoted as a 1 The first anomalous sample output threshold is expressed as ε tr Then the calculation method of the first abnormal sample output threshold is as shown in equation (5).
ε tr ＝μ tr +a 1 σ tr (5)
In step S330 of some embodiments, assuming that the second deviation Δ te of each sample in the prediction sample set satisfies the normal distribution, the calculation method of the second mean and the second standard deviation is the same as the calculation method of the first mean and the first standard deviation, and is not repeated herein.
In step S340 of some embodiments, a second confidence is obtained when the second deviation satisfies the normal distributionAnd determining a second coefficient according to the second confidence level, and determining a second abnormal sample output threshold according to the second mean value, the second standard deviation and the second coefficient. Note that the second coefficient is calculated in the same manner as the first coefficient. If the second mean value is expressed as mu te The second standard deviation is expressed as σ te The second coefficient is represented as a 2 The second abnormal sample output threshold is expressed as ε te Then, the calculation method of the second abnormal sample output threshold is as shown in equation (6).
ε te ＝μ te +a 2 σ te (6)
In step S350 of some embodiments, the first and second abnormal sample output thresholds are compared, and the minimum of the first and second abnormal sample output thresholds is taken as the target abnormal sample output threshold. The target anomaly sample output threshold is represented as ε, then ε = min (ε) tr ,ε te ). When the first abnormal sample output threshold value is smaller than the second abnormal sample output threshold value, taking the first abnormal sample output threshold value as a target abnormal sample output threshold value; when the first abnormal sample output threshold value is equal to the second abnormal sample output threshold value, taking the first abnormal sample output threshold value or the second abnormal sample output threshold value as a target abnormal sample output threshold value; and when the first abnormal sample output threshold value is larger than the second abnormal sample output threshold value, taking the second abnormal sample output threshold value as a target abnormal sample output threshold value.
In some embodiments, if the first confidence level is 0.99, the first coefficient a is determined 1 Is 3 according to a first coefficient a 1 First mean value mu tr And a first standard deviation σ tr Obtaining a first abnormal sample output threshold value epsilon tr Namely: epsilon tr ＝μ tr +3σ tr (ii) a If the second confidence level is 0.99, determining a second coefficient a 2 Is 3 according to a second coefficient a 2 Second mean value mu te And a second standard deviation σ te Obtaining a second abnormal sample output threshold value epsilon te Namely: epsilon te ＝μ te +3σ te Will epsilon tr And ε te The minimum value of the two is used as a target abnormal sample output threshold value epsilon.
In some embodiments, if the first deviation of the training sample set satisfies the normal distribution and the second deviation of the prediction sample set satisfies the normal distribution, by performing steps S310 to S350, the first abnormal sample output threshold and the second abnormal sample output threshold can be calculated, and the target abnormal sample output threshold is determined according to the first abnormal sample output threshold and the second abnormal sample output threshold, so that the target abnormal sample can be output according to the target abnormal sample output threshold.
In some embodiments, as shown in fig. 4, step S250 specifically includes, but is not limited to, steps S410 to S450.
Step S410, obtaining a first confidence coefficient of a training sample set;
step S420, determining a first abnormal sample output threshold according to the first deviation and the first confidence coefficient;
step S430, acquiring a second confidence of the prediction sample set;
step S440, determining a second abnormal sample output threshold according to the second deviation and the second confidence coefficient;
step S450, determining a target abnormal sample output threshold according to the first abnormal sample output threshold and the second abnormal sample output threshold.
In steps S410 to S420 of some embodiments, if the first deviation of the removed training sample set satisfies the assumption condition of normal distribution, the method for calculating the first abnormal sample output threshold includes: the method comprises the steps of obtaining a first confidence coefficient and a first sample number of a training sample set, sequencing first deviations corresponding to samples in the training sample set from large to small to obtain a first deviation sequence, calculating a first position of the first deviation sequence according to the first confidence coefficient and the first sample number, and taking the first deviation corresponding to the first position as a first abnormal sample output threshold value. For example, if the first confidence is 0.99 and the first sample number is 100, the first deviations corresponding to the 100 samples are sorted from large to small to obtain a first deviation sequence, the first probability is 0.01 according to the first confidence, the first position is 1 by multiplying the first sample number by the first probability, and the first deviation corresponding to the 1 st position in the first deviation sequence is used as the first abnormal sample output threshold.
In steps S430 to S440 of some embodiments, if the second deviation of the prediction sample set is removed and satisfies the assumption condition of normal distribution, the second abnormal sample output threshold is calculated by: and obtaining a second confidence coefficient and a second sample number of the prediction sample set, sequencing second deviations corresponding to the samples in the prediction sample set from large to small to obtain a second deviation sequence, calculating a second position of the second deviation sequence according to the second confidence coefficient and the second sample number, and taking the second deviation corresponding to the second position in the second deviation sequence as a second abnormal sample output threshold value. For example, if the second confidence is 0.9 and the number of the second samples is 100, the second deviations corresponding to the 100 samples are sorted from large to small to obtain a second deviation sequence, a second probability is 0.1 according to the second confidence, the number of the second samples is multiplied by the second probability to obtain a second position 10, and the second deviation corresponding to the 10 th position in the second deviation sequence is used as a second abnormal sample output threshold.
In step S450 of some embodiments, the minimum of the first and second abnormal sample output thresholds is taken as the target abnormal sample output threshold.
In some embodiments, if the first deviation does not satisfy the normal distribution and the second deviation does not satisfy the normal distribution, the target abnormal sample output threshold may be determined according to steps S410 to S450. If the first deviation satisfies the normal distribution and the second deviation also satisfies the normal distribution, the target abnormal sample output threshold may be determined according to steps S310 to S350, or may be determined according to steps S410 to S450.
In some embodiments, as shown in fig. 5, step S160 specifically includes, but is not limited to, step S510 to step S530.
Step S510, screening out samples with second deviation larger than or equal to a target abnormal sample output threshold value from the prediction sample set to obtain initial abnormal samples;
step S520, acquiring a preset sample quantity threshold;
step S530, screening out a target abnormal sample from the initial abnormal samples according to the sample quantity threshold value.
In step S510 of some embodiments, a second deviation of a plurality of samples in the prediction sample set is traversed, and if Δ te ≧ epsilon, a sample corresponding to the second deviation is used as an initial abnormal sample, where the initial abnormal sample is a high-probability abnormal sample in the prediction sample set.
In step S520 of some embodiments, a preset sample number threshold is obtained, where the sample number threshold is used to limit the number of target abnormal samples that are finally output.
In step S530 of some embodiments, a target abnormal sample with a sample number less than or equal to the sample number threshold is screened from the initial abnormal samples according to the sample number threshold. It should be noted that, in order to ensure the accuracy of the anomaly detection, the target anomaly sample may be sent to the service end for manual verification.
By performing steps S510 to S530, a target abnormal sample whose number of samples satisfies the sample number threshold can be output from the prediction sample set.
In some embodiments, as shown in fig. 6, step S530 specifically includes, but is not limited to, step S610 to step S620.
Step S610, if the sample number of the initial abnormal sample is less than or equal to the sample number threshold, taking the initial abnormal sample as a target abnormal sample;
in step S620, if the number of samples of the initial abnormal sample is greater than the sample number threshold, randomly selecting a number of samples equal to the sample number threshold from the initial abnormal sample to obtain a target abnormal sample.
In step S610 of some embodiments, if the number of samples of the initial abnormal sample is less than or equal to the sample number threshold, which indicates that the number of samples of the initial abnormal sample satisfies the manual detection measurement range, the initial abnormal sample is used as a target abnormal sample, and the target abnormal sample is output for manual detection.
In step S620 of some embodiments, if the number of samples of the initial abnormal sample is greater than the sample number threshold, which indicates that the number of samples of the initial abnormal sample exceeds the manual detection measurement range, the initial abnormal sample is randomly sampled, so as to randomly screen a specific number of samples from the initial abnormal sample as the target abnormal sample, where a value of the specific number is equal to the sample number threshold.
Another embodiment of the present application provides an abnormal sample detection method, including: determining a prediction time; acquiring a training sample set and a prediction sample set according to the prediction time; obtaining an initial regression model, performing model training on the initial regression model according to a training sample set to obtain a target regression model and a first commission predicted value corresponding to the training sample set, obtaining a first commission trial value of the training sample set, and obtaining a first deviation according to the first commission trial value and the first commission predicted value; inputting the prediction sample set into a target regression model, predicting the prediction sample set according to the target regression model to obtain a second commission predicted value, obtaining a second commission trial value of the prediction sample set, and obtaining a second deviation according to the second commission trial value and the second commission predicted value; calculating a first mean value and a first standard deviation of the first deviation, obtaining a first confidence level, determining a first coefficient according to the first confidence level, determining a first abnormal sample output threshold according to the first mean value, the first standard deviation and the first coefficient, calculating a second mean value and a second standard deviation of the second deviation, obtaining a second confidence level, determining a second coefficient according to the second confidence level, and determining a second abnormal sample output threshold according to the second mean value, the second standard deviation and the second coefficient; judging whether the first abnormal sample output threshold value is smaller than the second abnormal sample output threshold value; if so, taking the first abnormal sample output threshold as a target abnormal sample output threshold; if the judgment result is negative, taking the second abnormal sample output threshold value as a target abnormal sample output threshold value; screening out samples with the second deviation larger than or equal to the target abnormal sample output threshold value from the prediction sample set to obtain initial abnormal samples; judging whether the number of samples of the initial abnormal samples is less than or equal to a sample number threshold value; if the judgment result is yes, taking the initial abnormal sample as a target abnormal sample; if the judgment result is negative, randomly selecting samples with the number equal to the sample number threshold value from the initial abnormal samples to obtain target abnormal samples;
another embodiment of the present application provides a method for detecting an abnormal sample, including: determining a prediction time; acquiring a training sample set and a prediction sample set according to the prediction time; obtaining an initial regression model, performing model training on the initial regression model according to a training sample set to obtain a target regression model and a first commission predicted value corresponding to the training sample set, obtaining a first commission trial value of the training sample set, and obtaining a first deviation according to the first commission trial value and the first commission predicted value; acquiring a first confidence coefficient and a first sample number of a training sample set, acquiring a first probability according to the first confidence coefficient, multiplying the first sample number and the first probability to obtain a first position, sequencing first deviations of samples in the training sample set from large to small to obtain a first deviation sequence, and taking the first deviation corresponding to the first position in the first deviation sequence as a first abnormal sample output threshold; inputting the prediction sample set into a target regression model, predicting the prediction sample set according to the target regression model to obtain a second commission prediction value, obtaining a second commission trial value of the prediction sample set, and obtaining a second deviation according to the second commission trial value and the second commission prediction value; obtaining a second confidence coefficient and a second sample number of the prediction sample set, obtaining a second probability according to the second confidence coefficient, multiplying the second sample number and the second probability to obtain a second position, sequencing second deviations of the samples in the prediction sample set from large to small to obtain a second deviation sequence, and taking the second deviation corresponding to the second position in the second deviation sequence as a second abnormal sample output threshold; taking the minimum value of the first abnormal sample output threshold value and the second abnormal sample output threshold value as a target abnormal sample output threshold value; screening out samples with the second deviation larger than or equal to the target abnormal sample output threshold value from the prediction sample set to obtain initial abnormal samples; judging whether the sample number of the initial abnormal samples is less than or equal to a sample number threshold value or not; if so, taking the initial abnormal sample as a target abnormal sample; if the judgment result is negative, randomly selecting samples with the number equal to the sample number threshold value from the initial abnormal samples to obtain target abnormal samples.
An embodiment of the present application further provides an abnormal sample detection device, as shown in fig. 7, which can implement the abnormal sample detection method described above, and the device includes: a first acquisition module 710, a second acquisition module 720, a model training module 730, a model prediction module 740, a threshold calculation module 750, and an abnormal sample detection module 760. The first obtaining module 710 is configured to determine a predicted time; the second obtaining module 720 is configured to obtain a training sample set and a prediction sample set according to the prediction time; the model training module 730 is used for obtaining an initial regression model, performing model training on the initial regression model according to a training sample set, and obtaining a first commission predicted value corresponding to the target regression model and the training sample set; the model prediction module 740 is configured to input the prediction sample set into the target regression model, and predict commissions of the prediction sample set according to the target regression model to obtain a second commissioned forecasting value; the threshold value calculation module 750 is used for determining a target abnormal sample output threshold value according to the first commission prediction value and the second commission prediction value; the abnormal sample detection module 760 is configured to filter out the target abnormal sample from the prediction sample set according to the target abnormal sample output threshold.
The abnormal sample detection device provided by the embodiment of the application determines prediction time through the first acquisition module, the second acquisition module acquires a training sample set and a prediction sample set according to the prediction time, the model training module is used for acquiring an initial regression model, model training is performed on the initial regression model according to the training sample set, a first commission predicted value corresponding to a target regression model and the training sample set is obtained, the model prediction module is used for inputting the prediction sample set into the target regression model, the prediction sample set is predicted according to the target regression model, a second commission predicted value is obtained, the threshold calculation module determines a target abnormal sample output threshold according to the first commission predicted value and the second commission predicted value, the abnormal sample detection module screens out target abnormal samples from the prediction sample set according to the target abnormal sample output threshold, a large amount of abnormal sample data can be generated, and the problem that the abnormal sample data occupies less or is lost is solved.
The abnormal sample detection device in the embodiment of the present application is used to execute the abnormal sample detection method in the above embodiment, and the specific processing procedure of the abnormal sample detection device is the same as that of the abnormal sample detection method in the above embodiment, and is not described here any more.
An embodiment of the present application further provides an electronic device, including:
at least one processor, and,
a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,
the memory stores instructions for execution by the at least one processor to cause the at least one processor, when executing the instructions, to implement a method of anomaly sample detection as in any one of the embodiments of the first aspect of the present application.
The electronic equipment provided by the embodiment of the application obtains the training sample set and the prediction sample set according to the prediction time by determining the prediction time, obtains the initial regression model, performs model training on the initial regression model according to the training sample set, obtains the target regression model and the first commission predicted value corresponding to the training sample set, inputs the prediction sample set into the target regression model, predicts the prediction sample set according to the target regression model to obtain the second commission predicted value, determines the target abnormal sample output threshold value according to the first commission predicted value and the second commission predicted value, screens out the target abnormal sample from the prediction sample set according to the target abnormal sample output threshold value, can still perform sample abnormal detection to screen out the target abnormal sample under the scene that the abnormal sample proportion is small or missing, can generate a large amount of abnormal sample data, and solves the problem that the classification model cannot be trained due to the small proportion or missing of the abnormal sample data.
The hardware structure of the electronic device is described in detail below with reference to fig. 8. The electronic device includes: a processor 810, a memory 820, an input/output interface 830, a communication interface 840, and a bus 850.
The processor 810 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an Application Specific Integrated Circuit (ASIC), or one or more Integrated circuits, and is configured to execute a related program to implement the technical solution provided in the embodiment of the present Application;
the Memory 820 may be implemented in a ROM (Read Only Memory), a static Memory device, a dynamic Memory device, or a RAM (Random Access Memory). The memory 820 may store an operating system and other application programs, and when the technical solution provided by the embodiments of the present disclosure is implemented by software or firmware, the relevant program codes are stored in the memory 820 and called by the processor 810 to execute the abnormal sample detection method according to the embodiments of the present disclosure;
an input/output interface 830 for implementing information input and output;
the communication interface 840 is used for realizing communication interaction between the device and other devices, and can realize communication in a wired manner (for example, USB, network cable, etc.) or in a wireless manner (for example, mobile network, WIFI, bluetooth, etc.); and
a bus 850 that transfers information between the various components of the device (e.g., the processor 810, the memory 820, the input/output interface 830, and the communication interface 840);
wherein processor 810, memory 820, input/output interface 830, and communication interface 840 are communicatively coupled to each other within the device via bus 850.
The embodiment of the present application also provides a storage medium, which is a computer-readable storage medium, where computer-executable instructions are stored in the computer-readable storage medium, and the computer-executable instructions are used to enable a computer to execute the abnormal sample detection method according to the embodiment of the present application.
The storage medium provided by the embodiment of the application obtains a training sample set and a prediction sample set according to prediction time by determining prediction time, obtains an initial regression model, performs model training on the initial regression model according to the training sample set to obtain a first commission predicted value corresponding to a target regression model and the training sample set, inputs the prediction sample set into a target regression model, predicts the prediction sample set according to the target regression model to obtain a second commission predicted value, determines a target abnormal sample output threshold according to the first commission predicted value and the second commission predicted value, screens out target abnormal samples from the prediction sample set according to the target abnormal sample output threshold, and can still perform sample abnormal detection to screen out the target abnormal samples under the scene that the abnormal sample data is low in occupation ratio or missing, and can generate a large amount of abnormal sample data at the same time, thereby solving the problem that the training classification model cannot be performed due to the low occupation ratio or missing of the abnormal sample data.
The memory, which is a non-transitory computer readable storage medium, may be used to store non-transitory software programs as well as non-transitory computer executable programs. Further, the memory may include high speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory located remotely from the processor, and these remote memories may be connected to the processor through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation to the technical solutions provided in the embodiments of the present application, and it is obvious to those skilled in the art that the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems with the evolution of technology and the emergence of new application scenarios.
Those skilled in the art will appreciate that the solutions shown in fig. 1 to 6 do not constitute a limitation on the embodiments of the present application, and may include more or less steps than those shown, or combine some steps, or different steps.
The above-described embodiments of the apparatus are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, i.e. may be located in one place, or may also be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment.
One of ordinary skill in the art will appreciate that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof.
The terms "first," "second," "third," "fourth," and the like in the description of the application and the above-described figures, if any, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the application described herein are capable of operation in sequences other than those illustrated or described herein. Moreover, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be understood that in the present application, "at least one" means one or more, "a plurality" means two or more. "and/or" for describing an association relationship of associated objects, indicating that there may be three relationships, e.g., "a and/or B" may indicate: only A, only B and both A and B are present, wherein A and B may be singular or plural. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. "at least one of the following" or similar expressions refer to any combination of these items, including any combination of single item(s) or plural items. For example, at least one (one) of a, b, or c, may represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
In the several embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one type of logical functional division, and other divisions may be realized in practice, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be substantially implemented or contributed to by the prior art, or all or part of the technical solution may be embodied in a software product, which is stored in a storage medium and includes multiple instructions for enabling an electronic device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: various media capable of storing programs, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The preferred embodiments of the present application have been described above with reference to the accompanying drawings, and the scope of the claims of the embodiments of the present application is not limited thereto. Any modifications, equivalents and improvements that may occur to those skilled in the art without departing from the scope and spirit of the embodiments of the present application are intended to be within the scope of the claims of the embodiments of the present application.
Claims (10)
1. A method for detecting an abnormal sample, the method comprising:
determining a prediction time;
acquiring a training sample set and a prediction sample set according to the prediction time;
obtaining an initial regression model, and performing model training on the initial regression model according to the training sample set to obtain a target regression model and a first commission predicted value corresponding to the training sample set;
inputting the prediction sample set into the target regression model, and predicting the prediction sample set according to the target regression model to obtain a second commission prediction value;
determining a target anomaly sample output threshold value according to the first commission prediction value and the second commission prediction value;
and screening out target abnormal samples from the prediction sample set according to the target abnormal sample output threshold value.
2. The abnormal sample detection method of claim 1, wherein said determining a target abnormal sample output threshold from the first commission prediction value and the second commission prediction value comprises:
obtaining a first commission trial value for the training sample set;
obtaining a first deviation according to the first commission trial value and the first commission predicted value;
obtaining a second commission trial value for the prediction sample set;
obtaining a second deviation according to the second commission trial value and the second commission predicted value;
determining a target abnormal sample output threshold value according to the first deviation and the second deviation.
3. The abnormal sample detection method of claim 2, wherein said determining a target abnormal sample output threshold from said first deviation and said second deviation comprises:
calculating a first mean and a first standard deviation of the first deviation of the set of training samples;
determining a first abnormal sample output threshold according to the first mean value and the first standard deviation;
calculating a second mean and a second standard deviation of the second deviation of the set of prediction samples;
determining a second abnormal sample output threshold according to the second mean value and the second standard deviation;
and determining a target abnormal sample output threshold according to the first abnormal sample output threshold and the second abnormal sample output threshold.
4. The abnormal sample detection method of claim 3, wherein said determining a target abnormal sample output threshold value from the first abnormal sample output threshold value and the second abnormal sample output threshold value comprises:
and if the first abnormal sample output threshold is smaller than the second abnormal sample output threshold, taking the first abnormal sample output threshold as the target abnormal sample output threshold.
5. The abnormal sample detection method of claim 2, wherein said determining a target abnormal sample output threshold from said first deviation and said second deviation comprises:
obtaining a first confidence of the training sample set;
determining a first abnormal sample output threshold value according to the first deviation and the first confidence coefficient;
obtaining a second confidence of the prediction sample set;
determining a second abnormal sample output threshold according to the second deviation and the second confidence;
and determining a target abnormal sample output threshold according to the first abnormal sample output threshold and the second abnormal sample output threshold.
6. The abnormal sample detection method according to any one of claims 1 to 5, wherein the screening out the target abnormal sample from the prediction sample set according to the abnormal sample output threshold value comprises:
screening out samples with second deviation larger than or equal to the target abnormal sample output threshold value from the prediction sample set to obtain initial abnormal samples;
acquiring a preset sample quantity threshold;
and screening a target abnormal sample from the initial abnormal samples according to the sample quantity threshold value.
7. The abnormal sample detection method of claim 6, wherein the screening of the target abnormal sample from the initial abnormal samples according to the sample number threshold comprises:
if the sample number of the initial abnormal sample is smaller than or equal to the sample number threshold, taking the initial abnormal sample as the target abnormal sample;
and if the number of the samples of the initial abnormal samples is larger than the sample number threshold, randomly selecting samples with the number equal to the sample number threshold from the initial abnormal samples to obtain the target abnormal samples.
8. An abnormal sample detection device, comprising:
a first obtaining module for determining a predicted time;
the second acquisition module is used for acquiring a training sample set and a prediction sample set according to the prediction time;
the model training module is used for obtaining an initial regression model, carrying out model training on the initial regression model according to the training sample set, and obtaining a target regression model and a first commission predicted value corresponding to the training sample set;
the model prediction module is used for inputting the prediction sample set into the target regression model, predicting the commission of the prediction sample set according to the target regression model and obtaining a second commission prediction value;
a threshold value calculation module used for determining a target abnormal sample output threshold value according to the first commission predicted value and the second commission predicted value;
and the abnormal sample detection module is used for screening out the target abnormal sample from the prediction sample set according to the target abnormal sample output threshold value.
9. An electronic device, comprising:
at least one memory;
at least one processor;
at least one program;
the program is stored in the memory, and the processor executes the at least one program to:
the abnormal sample detection method according to any one of claims 1 to 7.
10. A storage medium that is a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions for causing a computer to perform:
the abnormal sample detection method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210739714.9A CN115204278A (en) | 2022-06-28 | 2022-06-28 | Abnormal sample detection method and device, electronic device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210739714.9A CN115204278A (en) | 2022-06-28 | 2022-06-28 | Abnormal sample detection method and device, electronic device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115204278A true CN115204278A (en) | 2022-10-18 |
Family
ID=83579020
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210739714.9A Pending CN115204278A (en) | 2022-06-28 | 2022-06-28 | Abnormal sample detection method and device, electronic device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115204278A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116302661A (en) * | 2023-05-15 | 2023-06-23 | 合肥联宝信息技术有限公司 | Abnormality prediction method and device, electronic equipment and storage medium |
-
2022
- 2022-06-28 CN CN202210739714.9A patent/CN115204278A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116302661A (en) * | 2023-05-15 | 2023-06-23 | 合肥联宝信息技术有限公司 | Abnormality prediction method and device, electronic equipment and storage medium |
| CN116302661B (en) * | 2023-05-15 | 2023-10-13 | 合肥联宝信息技术有限公司 | Abnormality prediction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110516910B (en) | Insurance policy and insurance model training method and insurance risk assessment method based on big data | |
| CN111178380B (en) | Data classification method and device and electronic equipment | |
| CN109471783B (en) | Method and device for predicting task operation parameters | |
| CN111368887A (en) | Training method of thunderstorm weather prediction model and thunderstorm weather prediction method | |
| CN115641538A (en) | Equipment abnormity detection method and device, computer equipment and storage medium | |
| CN114912537A (en) | Model training method and device, behavior prediction method and device, equipment and medium | |
| CN111598338B (en) | Method, apparatus, medium, and electronic device for updating prediction model | |
| CN115204278A (en) | Abnormal sample detection method and device, electronic device and storage medium | |
| CN111159481B (en) | Edge prediction method and device for graph data and terminal equipment | |
| CN115081641A (en) | Model training method, estimation result prediction method, device and storage medium | |
| CN115545331A (en) | Control strategy prediction method and device, equipment and storage medium | |
| CN114997322A (en) | Semi-supervised model-based classification prediction method, equipment and storage medium | |
| WO2019192101A1 (en) | Customer protection analysis method, electronic device, and computer readable storage medium | |
| CN113592593A (en) | Training and application method, device, equipment and storage medium of sequence recommendation model | |
| CN116523622A (en) | Object risk prediction method and device, electronic equipment and storage medium | |
| CN115965439A (en) | Data recall method, data recall device, electronic equipment and storage medium | |
| CN115099344A (en) | Model training method and device, user portrait generation method and device, and equipment | |
| CN110704614A (en) | Information processing method and device for predicting user group type in application | |
| CN115828153A (en) | Task prediction method, device, equipment and medium based on artificial intelligence | |
| CN113220947A (en) | Method and device for encoding event characteristics | |
| CN113420789A (en) | Method, device, storage medium and computer equipment for predicting risk account | |
| CN113343577A (en) | Parameter optimization method and device, computer equipment and readable storage medium | |
| CN109840535B (en) | Method and device for realizing terrain classification | |
| CN111008749A (en) | Demand forecasting method and device | |
| CN111309706A (en) | Model training method and device, readable storage medium and electronic equipment |
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 |