CN117574087A - Model determining method, memory fault predicting device, medium and equipment - Google Patents
Model determining method, memory fault predicting device, medium and equipment Download PDFInfo
- Publication number
- CN117574087A CN117574087A CN202311631231.8A CN202311631231A CN117574087A CN 117574087 A CN117574087 A CN 117574087A CN 202311631231 A CN202311631231 A CN 202311631231A CN 117574087 A CN117574087 A CN 117574087A
- Authority
- CN
- China
- Prior art keywords
- memory
- aggregation
- feature
- data
- initial
- 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
- 230000015654 memory Effects 0.000 title claims abstract description 350
- 238000000034 method Methods 0.000 title claims abstract description 77
- 230000002776 aggregation Effects 0.000 claims abstract description 313
- 238000004220 aggregation Methods 0.000 claims abstract description 313
- 238000012216 screening Methods 0.000 claims abstract description 71
- 238000012545 processing Methods 0.000 claims abstract description 38
- 238000012795 verification Methods 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 13
- 230000008569 process Effects 0.000 claims description 12
- 238000011156 evaluation Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000006116 polymerization reaction Methods 0.000 description 9
- 238000004891 communication Methods 0.000 description 8
- 238000012163 sequencing technique Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000009795 derivation Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000009977 dual effect Effects 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/285—Selection of pattern recognition techniques, e.g. of classifiers in a multi-classifier system
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/3003—Monitoring arrangements specially adapted to the computing system or computing system component being monitored
- G06F11/3037—Monitoring arrangements specially adapted to the computing system or computing system component being monitored where the computing system component is a memory, e.g. virtual memory, cache
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/30—Monitoring
- G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
- G06F11/3466—Performance evaluation by tracing or monitoring
- G06F11/3476—Data logging
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/211—Selection of the most significant subset of features
- G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/217—Validation; Performance evaluation; Active pattern learning techniques
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- Quality & Reliability (AREA)
- Mathematical Physics (AREA)
- Computer Hardware Design (AREA)
- Test And Diagnosis Of Digital Computers (AREA)
Abstract
The invention discloses a model determining method, a memory fault predicting device, a medium and equipment. The method for determining the memory fault prediction model comprises the following steps: acquiring memory original data, and performing multidimensional aggregation processing on the memory original data based on at least part of dimensionality, memory structure information and data type information of various aggregation parameter items to obtain initial aggregation characteristics; screening the initial aggregation characteristics based on the correlation between the initial aggregation characteristics and the memory faults to obtain screened aggregation characteristics; and evaluating the candidate models based on the screening aggregate features, determining a memory failure prediction model in the candidate models based on index data of the candidate models, and determining target aggregate features in a feature group corresponding to the memory failure prediction model. And performing feature screening on the initial aggregation features through the correlation of the initial aggregation features on the memory faults to ensure that the screened features have stronger correlation with the memory faults, and further determining a high-performance memory fault prediction model.
Description
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a method for determining a model, a method for predicting a memory failure, a device, a medium, and equipment.
Background
Faults that may exist in the device memory include restorable faults and unrecoverable faults, wherein unrecoverable faults generally cause the device to downtime, affecting the operation of the device. The prediction of the memory faults of the equipment can predict the possible irrecoverable faults, so that the operations such as backing up the data in the memory and maintaining the memory are facilitated, and the fault risk is reduced.
In the process of realizing the invention, the prior art is found to have at least the following technical problems: at present, a machine learning model is used for predicting memory faults, but the correlation between data characteristics for performing fault prediction and the memory faults is generally not high, so that the accuracy of memory prediction is affected.
Disclosure of Invention
The invention provides a method for determining a model, a method, a device, a medium and equipment for predicting memory faults, so as to improve the accuracy of memory prediction.
According to an aspect of the present invention, there is provided a method for determining a memory failure prediction model, including:
Acquiring memory original data, and performing multidimensional aggregation processing on the memory original data based on at least part of dimensionality, memory structure information and data type information of various aggregation parameter items to obtain initial aggregation characteristics;
screening the initial aggregation features based on the correlation between the initial aggregation features and the memory faults to obtain screened aggregation features, wherein different numbers of screened aggregation features form different feature groups;
and training at least one candidate model based on each feature group, performing index verification on each candidate model, determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature groups corresponding to the memory failure prediction models.
Optionally, the multi-dimensional aggregation process includes intra-structure multi-dimensional aggregation and/or inter-structure multi-dimensional aggregation;
the intra-structure multidimensional aggregation includes: based on the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items; and/or, in a specific memory structure, based on the data type information and at least part of the aggregation parameter item aggregation mode;
The inter-structure multidimensional aggregation includes: and based on any two memory structures in the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items.
Optionally, the performing multidimensional aggregation on the memory original data based on at least part of dimensions, memory structure information and data type information of the multiple aggregation parameter items to obtain an initial aggregation feature includes: and traversing optional dimension information of a plurality of dimensions in any aggregation mode, and forming an initial aggregation feature in the aggregation mode based on the optional dimension information corresponding to the dimensions.
Optionally, the memory original data includes first memory original data of the memory failure device and second memory original data of the non-memory failure device;
the method further comprises the steps of: setting a fault tag for the initial aggregation feature obtained based on the first memory original data; setting a non-fault tag for the initial aggregation feature obtained based on the second memory original data; and performing correlation calculation based on the initial aggregation characteristics and the corresponding tags, and determining the correlation between the initial aggregation characteristics and the memory faults.
Optionally, the screening the initial aggregation feature based on the correlation between the initial aggregation feature and the memory failure to obtain a screened aggregation feature, where different numbers of screened aggregation features form different feature groups, including: and sequencing the initial aggregation features based on the correlation between the initial aggregation features and the memory faults, and determining screening aggregation features based on feature screening quantity and the sequencing of the initial aggregation features, wherein the screening aggregation features of the feature screening quantity form a feature group.
Optionally, the number of the feature groups is a plurality;
the training based on each feature set to obtain at least one candidate model, performing index verification on each candidate model, determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature sets corresponding to the memory failure prediction models, including: respectively acquiring sample data based on each feature group, and respectively training the at least one initial model based on the sample data corresponding to each feature group to obtain at least one trained candidate model; respectively performing index verification on the candidate models based on preset evaluation indexes to obtain index data of the candidate models; determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining screening aggregation features in a feature group corresponding to the memory failure prediction model as target aggregation features corresponding to the memory failure prediction model.
According to another aspect of the present invention, there is provided a memory failure prediction method, including:
acquiring original memory data of equipment to be detected; performing aggregation processing on the original memory data based on target aggregation characteristics corresponding to the memory fault prediction model to obtain feature data to be processed; and carrying out prediction processing on the feature data to be processed based on the memory fault prediction model to obtain a memory fault prediction result of the equipment to be detected. The memory failure prediction model and the target aggregation characteristic are obtained based on the determination method of the memory failure prediction model provided by the embodiment of the invention.
According to another aspect of the present invention, there is provided a determination apparatus for a memory failure prediction model, including:
the feature aggregation module is used for acquiring memory original data, and carrying out multidimensional aggregation processing on the memory original data based on at least part of dimensions, memory structure information and data type information of various aggregation parameter items to obtain initial aggregation features;
the feature screening module is used for screening the initial aggregation features based on the correlation between the initial aggregation features and the memory faults to obtain screening aggregation features, and different feature groups are formed by different numbers of screening aggregation features;
The model determining module is used for obtaining at least one candidate model based on training of each feature group, carrying out index verification on each candidate model, determining a memory failure prediction model from a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature groups corresponding to the memory failure prediction models.
According to another aspect of the present invention, there is provided a memory failure prediction apparatus, including:
the data acquisition module is used for acquiring original memory data of the equipment to be detected;
the data processing module is used for carrying out aggregation processing on the original memory data based on the target aggregation characteristics corresponding to the memory fault prediction model to obtain feature data to be processed;
and the fault prediction module is used for performing prediction processing on the feature data to be processed based on the memory fault prediction model to obtain a memory fault prediction result of the equipment to be detected. The memory failure prediction model and the target aggregation characteristic are obtained based on the determination method of the memory failure prediction model provided by the embodiment of the invention.
According to another aspect of the present invention, there is provided an electronic apparatus including:
At least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of determining a memory failure prediction model and/or the method of predicting a memory failure according to any of the embodiments of the present invention.
According to another aspect of the present invention, there is provided a computer readable storage medium storing computer instructions for causing a processor to implement a method for determining a memory failure prediction model and/or a method for predicting a memory failure according to any embodiment of the present invention when executed.
According to the technical scheme provided by the embodiment of the invention, the initial aggregation characteristics are obtained by carrying out multidimensional aggregation processing on the memory original data, so that the characteristic derivation is realized, and the characteristic comprehensiveness and diversity are improved. And performing feature screening on the initial aggregation features through the correlation of the initial aggregation features on the memory faults to ensure that the screened features have stronger correlation with the memory faults, providing contribution for prediction of the memory faults, and further improving the accuracy of the memory fault prediction from the feature angle. And evaluating the candidate models through the high-correlation screening aggregation characteristics to determine a high-performance memory fault prediction model, and improving the memory prediction accuracy.
It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the invention or to delineate the scope of the invention. Other features of the present invention will become apparent from the description that follows.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for determining a memory failure prediction model according to an embodiment of the present invention;
FIG. 2 is a flowchart of a memory failure prediction method according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a determining device for a memory failure prediction model according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of a memory failure prediction apparatus according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of an electronic device implementing an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the invention described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, 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.
Fig. 1 is a flowchart of a method for determining a memory failure prediction model according to an embodiment of the present invention, where the method may be performed by a memory failure prediction model determining device, and the memory failure prediction model determining device may be implemented in hardware and/or software, and the memory failure prediction model determining device may be configured in an electronic device such as a computer, a server, a mobile phone, or the like. As shown in fig. 1, the method includes:
s110, acquiring memory original data, and performing multidimensional aggregation processing on the memory original data based on at least part of dimensions of various aggregation parameter items, memory structure information and data type information to obtain initial aggregation characteristics.
And S120, screening the initial aggregation features based on the correlation between the initial aggregation features and the memory faults to obtain screened aggregation features, wherein different numbers of screened aggregation features form different feature groups.
S130, training at least one candidate model based on each feature set, performing index verification on each candidate model, determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature sets corresponding to the memory failure prediction models.
In this embodiment, the memory failure prediction may be understood as predicting the probability of occurrence of an unrecoverable failure (Uncorrectable Memory Error, UE) of the device, and determining a memory failure prediction model for implementing the memory failure prediction based on memory raw data of the device, where the memory raw data may be historical operation data of a memory of the device, and optionally, the memory raw data may be related data of occurrence of a recoverable failure (Correctable Error, CE) of the memory, and the memory raw data includes, by way of example, a memory structure in which the recoverable failure occurs, coordinate data of the recoverable failure in the memory structure, a timestamp in which the recoverable failure occurs, and so on.
The correlation between the original memory data and the memory fault is poor, the contribution degree to the memory prediction is weak, and accordingly, the problem of low prediction accuracy exists under the condition of performing the memory prediction based on the original memory data. Aiming at the problems, the original memory data is subjected to multidimensional aggregation to obtain initial aggregation characteristics, screening is carried out based on the initial aggregation characteristics and the coherence of the memory faults to obtain aggregation characteristics with higher coherence with the memory faults, the aggregation characteristics are used for predicting the memory faults, and the prediction accuracy of the memory faults is improved from the aspect of characteristic correlation.
The initial memory data may be subjected to multidimensional aggregation based on at least a portion of dimensions of the memory structure information, the data type information, and the plurality of aggregation parameter items to obtain an initial aggregation feature. The Memory structure information is multi-level structure information of the device Memory, and includes Dual-Inline-Memory-Modules (dimm), memory block rank, physical bank, and abscissa and ordinate of the physical bank. The dual in-line memory module dimm comprises a plurality of rank groups, each rank group comprises a plurality of bank groups, and each bank group comprises an abscissa and an ordinate of a memory page mapping point. The data type information includes the number of occurrence of the restorability failure and coordinate point information of occurrence of the restorability failure. The aggregation parameter term includes parameter terms of multiple dimensions including, but not limited to, feature aggregation manner, aggregation time range, aggregation scale, alignment threshold, and threshold alignment. The feature aggregation mode includes, but is not limited to, mean, sum, maximum, minimum, variance, standard deviation and ordered preset quantile values. The polymerization time range includes a preset polymerization period, for example, a polymerization period of 15 minutes, a polymerization period of 1 hour, a polymerization period of 1 day, etc., and here, the polymerization time range may be at least one, and may be set according to polymerization requirements. The aggregate size is the aggregate size in the abscissa and ordinate dimensions of the memory structure, which may be preset ratio data, for example, in 10% of the abscissa length. The aggregate size may be at least one and may be set according to the aggregate requirements. The comparison threshold is threshold data for comparison, and at least one can be set according to the aggregation requirement; the threshold comparison method comprises greater than, less than, greater than or equal to and less than or equal to.
Optionally, the multi-dimensional aggregation process includes intra-structure multi-dimensional aggregation and/or inter-structure multi-dimensional aggregation. In the case where the memory result information includes a plurality of levels of memory structures, intra-structure multi-dimensional aggregation may be understood as multi-dimensional feature aggregation performed under any level of memory structures, and inter-structure multi-dimensional aggregation may be understood as multi-dimensional feature aggregation performed under two or more memory structures.
Wherein the intra-structure multidimensional aggregation comprises: based on the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items; and/or, in a specific memory structure, based on the data type information and at least part of the aggregation parameter item aggregation mode.
The aggregation manner based on the memory structure information, the data type information and at least part of the aggregation parameter items may specifically be an aggregation manner based on the memory structure information, the data type information, the feature aggregation manner and the aggregation time range. The aggregation process based on the aggregation mode can be to traverse the memory structure information, the data type information, the feature aggregation mode and the optional dimension information of each dimension in the aggregation time range respectively, form a feature item of an initial aggregation feature based on the optional dimension information corresponding to each of the plurality of dimensions, and perform clustering processing on the memory original data based on the feature item to obtain a feature value of the initial aggregation feature. The number of the initial aggregation features obtained based on the aggregation mode is the product of the number of the pieces of optional dimension information in each dimension in the aggregation mode, namely the product of the number of first optional dimension information in the memory structure information, the number of second optional dimension information in the data type information, the number of third optional dimension information in the feature aggregation mode and the number of fourth optional dimension information in the aggregation time range.
For example, taking memory structure information as dimm of the dual in-line storage module, data type information as the number of occurrence of restorable faults, characteristic aggregation mode as the maximum value, and aggregation time range as 1 hour as an example, and the obtained initial aggregation characteristic as the maximum value of the number of occurrence of restorable faults in all dimm within 1 hour.
In the specific memory structure, the aggregation mode based on the data type information and at least part of the aggregation parameter items may specifically be an aggregation mode based on the data type information, the aggregation scale, the feature aggregation mode and the aggregation time range in the abscissa or ordinate dimension of the physical storage bank. Correspondingly, the aggregation process based on the aggregation mode can be that data type information, an aggregation scale, a characteristic aggregation mode and optional dimension information of an aggregation time range are respectively traversed on the abscissa or ordinate dimension of the physical storage bank to form a characteristic item of an initial clustering characteristic, and the clustering processing is carried out on the memory original data based on the characteristic item to obtain the characteristic value of the initial aggregation characteristic.
Illustratively, the initial cluster features obtained based on the above manner may be: sliding on the abscissa of all banks within 1 hour by adopting 10% of the length of the abscissa as a ruler, wherein the maximum value of the number of restorable faults occurs on the ruler with the length of 10% of the length of the abscissa; alternatively, sliding is performed on the ordinate of all banks within 1 hour using a scale of 5% of the ordinate length, and the average value of the number of restorable failures occurring on the scale of 5% of the ordinate length, and the like.
Wherein the inter-structure multidimensional aggregation comprises: and based on any two memory structures in the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items. Specifically, the aggregation mode can be an aggregation mode between any two memory structures based on data type information, an aggregation time range, a comparison threshold value and a threshold value comparison method. Correspondingly, the aggregation process based on the multi-dimensional aggregation among the structures can be to randomly select two different memory structures, traverse data type information, a feature aggregation mode, an aggregation time range, a comparison threshold value and a threshold value comparison mode, determine optional dimensional information of the dimensions, form feature items of initial aggregation features, and perform clustering processing on memory original data based on the feature items to obtain feature values of the initial aggregation features. For example, two different memory structures may be dual inline memory modules dimm and memory block rank, the data type information may be the number of occurrence of restorable failures, the aggregate time range may be 1 hour, the comparison threshold may be 100, the threshold comparison may be greater than, and the corresponding initial aggregate feature may be the number of ranks in dimm where the number of restorable failures occurs within 1 hour of all ranks greater than 100.
In summary, based on at least part of dimensions, memory structure information and data type information of a plurality of aggregation parameter items, performing multidimensional aggregation processing on the memory original data to obtain initial aggregation characteristics, including: and traversing optional dimension information of a plurality of dimensions in any aggregation mode, and forming an initial aggregation feature in the aggregation mode based on the optional dimension information corresponding to the dimensions. Determining optional dimension information of each dimension, and splicing the optional dimension information of a plurality of dimensions into an initial aggregation feature, wherein the obtained initial aggregation feature has no repetition. The number of initial aggregated features obtained for each aggregation may be the product of the number of selectable dimension information in each dimension of the aggregation.
Through the multiple polymerization modes, multiple initial polymerization features can be obtained, feature derivation of memory original data is achieved, comprehensiveness and diversity of feature types are improved, and further optional features are provided for feature screening.
The initial aggregated feature may be stored in a form based on a key-value, where the key is a feature term, i.e., a feature identification (which may be, for example, a feature name), of the initial aggregated feature, and the value is a feature value of the initial aggregated feature.
On the basis of determining the initial aggregation characteristics, determining the correlation between each initial aggregation characteristic and the memory fault, and representing the correlation degree between the initial aggregation characteristics and the memory fault through a correlation value, wherein the larger the correlation value is, the higher the correlation between the initial aggregation characteristics and the memory fault is. The higher the correlation between the initial aggregation feature and the memory fault, the greater the influence of the initial aggregation feature on the memory fault, and the greater the contribution degree in the memory prediction process. Correspondingly, the initial aggregation characteristics can be screened according to the correlation between the initial aggregation characteristics and the memory faults, so that the screened aggregation characteristics for predicting the memory faults are obtained.
On the basis of the above embodiment, the memory original data includes first memory original data of the memory failure device and second memory original data of the non-memory failure device; the memory fault device is a device with an unrecoverable fault, and the non-memory fault device is a device without the unrecoverable fault.
Setting a fault tag for the initial aggregation feature obtained based on the first memory original data; setting a non-fault tag for the initial aggregation feature obtained based on the second memory original data; and performing correlation calculation based on the initial aggregation characteristics and the corresponding tags, and determining the correlation between the initial aggregation characteristics and the memory faults. The fault label and the non-fault label may be different labels, for example, the fault label may be "1", and the non-fault label may be "0".
And (3) performing aggregation processing on the memory original data corresponding to each device respectively to obtain initial aggregation characteristics, and setting corresponding labels, namely, the characteristic item of each initial aggregation characteristic can correspond to a plurality of characteristic values, and each characteristic value can correspond to a label. And forming a feature matrix and a label matrix by the feature value of the initial aggregation feature and the corresponding label, and performing correlation calculation based on the feature matrix and the label matrix to obtain correlation data of the initial aggregation feature and the memory fault. The manner in which the correlation is calculated is not limited herein, and in some embodiments, may be calculated based on a pearson correlation algorithm, which is not limited herein.
Based on the above embodiment, feature screening may be performed based on the number of feature screening, or feature screening may be performed based on a correlation threshold. In some embodiments, the initial aggregate features may be screened multiple times to obtain multiple feature sets, where each screening condition is different, each feature set includes a screened aggregate feature that is obtained by screening, and the screened aggregate features included in different feature sets are different.
In some embodiments, the screening the initial aggregation feature based on the correlation between the initial aggregation feature and the memory failure to obtain a screened aggregation feature includes: and sequencing the initial aggregation features based on the correlation between the initial aggregation features and the memory faults, and determining screening aggregation features based on feature screening quantity and the sequencing of the initial aggregation features, wherein the screening aggregation features of the feature screening quantity form a feature group.
Based on the correlation data of the initial aggregation features and the memory faults, sequencing the correlation of the initial aggregation features and the memory faults from large to small, and selecting the first n features in the sequencing as screening aggregation features to form a feature group. And under the condition that n is the feature screening number and a plurality of numerical values are taken, each feature screening number n corresponds to a feature group, and different feature screening numbers n correspond to different feature groups, namely different numbers of screening aggregation features form different feature groups.
In some embodiments, the screening the initial aggregation feature based on the correlation between the initial aggregation feature and the memory failure to obtain a screened aggregation feature includes: and taking the initial aggregation characteristics meeting the correlation threshold as screening aggregation characteristics to form a characteristic group. Specifically, the correlation data of each initial aggregation feature and the memory fault is compared with a correlation threshold, and the initial aggregation feature with the correlation data larger than the correlation threshold is determined as the initial aggregation feature. Wherein, under the condition that the correlation threshold takes different values, a plurality of characteristic groups are formed. Different correlation thresholds can be used for screening different numbers of screening aggregation features, different numbers of screening aggregation can be obtained by setting different correlation thresholds, and different numbers of screening aggregation features form different feature groups, namely a plurality of different feature groups are obtained.
The correlation between each initial aggregation feature and the memory fault is determined, a basis is provided for feature screening, and correspondingly, features in the feature group obtained through screening have strong correlation with the memory fault, so that the prediction of the memory fault is facilitated. Further, multiple screening of different screening conditions is carried out on the initial aggregation characteristics to obtain multiple different characteristic groups, so that the characteristic groups with diversity are provided, and the influence of the screening conditions on the prediction of the memory faults is avoided.
The machine learning model capable of realizing the memory fault prediction is of various types, namely, a plurality of types of initial models exist, the initial models are trained to obtain candidate models, the candidate models are subjected to index verification based on the plurality of characteristic groups so as to evaluate each candidate model, and the memory fault prediction model is determined in the plurality of candidate models. The index verification of the candidate models can be performance index verification, and accordingly, the memory fault prediction model is a model with optimal performance in a plurality of candidate models, so that the prediction precision of the memory fault prediction model is improved.
Exemplary candidate models (or initial models) include, but are not limited to, random forest models, decision tree models, support vector machine models, and XGBoost models, recurrent neural network models, convolutional neural network models, long-term memory models, and the like. Optionally, an algorithm library is maintained in advance, the algorithm library comprises a plurality of candidate models, and the algorithm library can be updated according to requirements.
Optionally, training at least one candidate model based on each feature set, performing index verification on each candidate model, determining a memory failure prediction model from a plurality of candidate models based on index data of the candidate models, and determining a target aggregate feature in a feature set corresponding to the memory failure prediction model, including: respectively acquiring sample data based on each feature group, and respectively training the at least one initial model based on the sample data corresponding to each feature group to obtain at least one trained candidate model; respectively performing index verification on the candidate models based on preset evaluation indexes to obtain index data of the candidate models; determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining screening aggregation features in a feature group corresponding to the memory failure prediction model as target aggregation features corresponding to the memory failure prediction model
For each feature set, sample data is collected based on the screening aggregate features included in the feature set, forming a sample data set. The method comprises the steps of determining a data value of a screening aggregation characteristic based on first memory original data of memory fault equipment, forming first sample data, determining a fault label corresponding to the first sample data, determining a data value of the screening aggregation characteristic based on second memory original data of non-memory fault equipment, forming second sample data, and determining a non-fault label of the second sample data. And forming a sample data set corresponding to the feature group based on the first sample data, the second sample data and the labels respectively corresponding to the first sample data and the second sample data. Each feature set corresponds to a sample dataset.
At least one type of initial model is trained through the sample data set corresponding to each feature set to obtain at least one type of candidate model, namely, each feature set can correspond to a plurality of trained candidate models.
And performing cross-validation evaluation on each trained candidate model, specifically, performing index validation on the trained candidate models based on preset evaluation indexes to obtain index data of each trained candidate model. Wherein, the preset evaluation index comprises but is not limited to F1-score, accuracy, recall rate and the like. And weighting the evaluation values of the plurality of preset evaluation indexes of the trained candidate model respectively to obtain index data of the candidate model.
Comparing index data of a plurality of candidate models, determining a candidate model of the maximum index data as a memory failure prediction model, and correspondingly, determining screening aggregation features in a feature group corresponding to the memory failure prediction model as target aggregation features.
According to the technical scheme, the initial aggregation characteristics are obtained by carrying out multidimensional aggregation processing on the memory original data, so that characteristic derivation is realized, and characteristic comprehensiveness and diversity are improved. And performing feature screening on the initial aggregation features through the correlation of the initial aggregation features on the memory faults to ensure that the screened features have stronger correlation with the memory faults, providing contribution for prediction of the memory faults, and further improving the accuracy of the memory fault prediction from the feature angle. And evaluating the candidate models through the high-correlation screening aggregation characteristics to determine a high-performance memory fault prediction model, and improving the memory prediction accuracy.
Fig. 2 is a flowchart of a memory failure prediction method provided in an embodiment of the present invention, where the embodiment is applicable to a case of performing memory failure prediction on a device to be detected based on a predetermined memory prediction model, the method may be performed by a memory failure prediction device, and the memory failure prediction device may be implemented in the form of hardware and/or software, and the memory failure prediction device may be configured in an electronic device such as a computer, a server, a mobile phone, or the like. As shown in fig. 2, the method includes:
s210, acquiring original memory data of the equipment to be detected.
And S220, performing aggregation processing on the original memory data based on the target aggregation characteristics corresponding to the memory fault prediction model to obtain feature data to be processed.
S230, predicting the feature data to be processed based on the memory failure prediction model to obtain a memory failure prediction result of the equipment to be detected.
In this embodiment, the memory failure prediction model is determined based on the determination method of the memory failure prediction model provided in the foregoing embodiment, and the target aggregate feature corresponding to the memory failure prediction model is a feature applicable to the memory failure prediction model.
And performing aggregation processing on the original memory data of the equipment to be detected, determining a characteristic value corresponding to the target aggregation characteristic, namely the characteristic data to be processed, forming an input vector by the characteristic data to be processed, and inputting the input vector into a memory failure prediction model to obtain a memory failure prediction result of the equipment to be detected.
According to the technical scheme, the accuracy of memory fault prediction is improved by calling the predetermined memory fault prediction model and the target aggregation characteristic.
Fig. 3 is a schematic structural diagram of a determining device for a memory failure prediction model according to an embodiment of the present invention. As shown in fig. 3, the apparatus includes:
the feature aggregation module 310 is configured to obtain memory original data, and perform multidimensional aggregation processing on the memory original data based on at least part of dimensions, memory structure information and data type information of a plurality of aggregation parameter items to obtain initial aggregation features;
the feature screening module 320 is configured to screen the initial aggregation feature based on the correlation between the initial aggregation feature and the memory failure to obtain a screened aggregation feature, where different numbers of screened aggregation features form different feature groups;
the model determining module 330 is configured to train to obtain at least one candidate model based on each feature set, perform index verification on each candidate model, determine a memory failure prediction model from a plurality of candidate models based on index data of the candidate models, and determine a target aggregate feature in a feature set corresponding to the memory failure prediction model.
According to the technical scheme, the initial aggregation characteristics are obtained by carrying out multidimensional aggregation processing on the memory original data, so that characteristic derivation is realized, and characteristic comprehensiveness and diversity are improved. And performing feature screening on the initial aggregation features through the correlation of the initial aggregation features on the memory faults to ensure that the screened features have stronger correlation with the memory faults, providing contribution for prediction of the memory faults, and further improving the accuracy of the memory fault prediction from the feature angle. And evaluating the candidate models through the high-correlation screening aggregation characteristics to determine a high-performance memory fault prediction model, and improving the memory prediction accuracy.
On the basis of the above embodiment, optionally, the multi-dimensional aggregation process includes intra-structure multi-dimensional aggregation and/or inter-structure multi-dimensional aggregation;
the intra-structure multidimensional aggregation includes: based on the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items; and/or, in a specific memory structure, based on the data type information and at least part of the aggregation parameter item aggregation mode;
the inter-structure multidimensional aggregation includes: and based on any two memory structures in the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items.
Optionally, the feature aggregation module 310 is configured to: and traversing optional dimension information of a plurality of dimensions in any aggregation mode, and forming an initial aggregation feature in the aggregation mode based on the optional dimension information corresponding to the dimensions.
On the basis of the above embodiment, optionally, the memory original data includes first memory original data of a memory failure device and second memory original data of a non-memory failure device;
the apparatus further comprises: the correlation determination module is used for setting fault labels for the initial aggregation characteristics obtained based on the first memory original data; setting a non-fault tag for the initial aggregation feature obtained based on the second memory original data; and performing correlation calculation based on the initial aggregation characteristics and the corresponding tags, and determining the correlation between the initial aggregation characteristics and the memory faults.
Based on the above embodiments, optionally, the feature screening module 320 is configured to: and sequencing the initial aggregation features based on the correlation between the initial aggregation features and the memory faults, and determining screening aggregation features based on feature screening quantity and the sequencing of the initial aggregation features, wherein the screening aggregation features of the feature screening quantity form a feature group.
On the basis of the above embodiment, optionally, the number of the feature groups is a plurality;
the model determination module 330 is configured to: respectively acquiring sample data based on each feature group, and respectively training the at least one initial model based on the sample data corresponding to each feature group to obtain at least one trained candidate model; respectively performing index verification on the candidate models based on preset evaluation indexes to obtain index data of the candidate models; determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining screening aggregation features in a feature group corresponding to the memory failure prediction model as target aggregation features corresponding to the memory failure prediction model.
The memory failure prediction model determining device provided by the embodiment of the invention can execute the memory failure prediction model determining method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the executing method.
Fig. 4 is a schematic structural diagram of a memory failure prediction apparatus according to an embodiment of the present invention. As shown in fig. 4, the apparatus includes:
a data acquisition module 410, configured to acquire original memory data of a device to be detected;
The data processing module 420 is configured to aggregate the original memory data based on a target aggregate feature corresponding to the memory failure prediction model, so as to obtain feature data to be processed;
and the fault prediction module 430 is configured to perform prediction processing on the feature data to be processed based on the memory fault prediction model, so as to obtain a memory fault prediction result of the device to be detected.
The memory failure prediction device provided by the embodiment of the invention can execute the memory failure prediction method provided by any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.
Fig. 5 is a schematic structural diagram of an electronic device according to an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. Electronic equipment may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
As shown in fig. 5, the electronic device 10 includes at least one processor 11, and a memory, such as a Read Only Memory (ROM) 12, a Random Access Memory (RAM) 13, etc., communicatively connected to the at least one processor 11, in which the memory stores a computer program executable by the at least one processor, and the processor 11 may perform various appropriate actions and processes according to the computer program stored in the Read Only Memory (ROM) 12 or the computer program loaded from the storage unit 18 into the Random Access Memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 may also be stored. The processor 11, the ROM 12 and the RAM 13 are connected to each other via a bus 14. An input/output (I/O) interface 15 is also connected to bus 14.
Various components in the electronic device 10 are connected to the I/O interface 15, including: an input unit 16 such as a keyboard, a mouse, etc.; an output unit 17 such as various types of displays, speakers, and the like; a storage unit 18 such as a magnetic disk, an optical disk, or the like; and a communication unit 19 such as a network card, modem, wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.
The processor 11 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various processors running machine learning model algorithms, digital Signal Processors (DSPs), and any suitable processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the XX method.
In some embodiments, the method of determining a memory failure prediction model, and/or the method of memory failure prediction may be implemented as a computer program tangibly embodied on a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and/or installed onto the electronic device 10 via the ROM 12 and/or the communication unit 19. When a computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the XX method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the method of determining the memory failure prediction model, and/or the method of memory failure prediction, in any other suitable manner (e.g., by means of firmware).
Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.
The determination method of the memory failure prediction model used to implement the present invention, and/or the computer program of the memory failure prediction method, may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions/acts specified in the flowchart and/or block diagram block or blocks to be implemented. The computer program may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.
The embodiment of the invention also provides a computer readable storage medium, wherein the computer readable storage medium stores computer instructions, and the computer instructions are used for enabling a processor to execute a method for determining a memory failure prediction model, and the method comprises the following steps:
acquiring memory original data, and performing multidimensional aggregation processing on the memory original data based on at least part of dimensionality, memory structure information and data type information of various aggregation parameter items to obtain initial aggregation characteristics; screening the initial aggregation features based on the correlation between the initial aggregation features and the memory faults to obtain screened aggregation features, wherein different numbers of screened aggregation features form different feature groups; and training at least one candidate model based on each feature group, performing index verification on each candidate model, determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature groups corresponding to the memory failure prediction models.
Alternatively, the computer instructions are for causing the processor to perform a memory failure prediction method, the method comprising:
acquiring original memory data of equipment to be detected; performing aggregation processing on the original memory data based on target aggregation characteristics corresponding to the memory fault prediction model to obtain feature data to be processed; and carrying out prediction processing on the feature data to be processed based on the memory fault prediction model to obtain a memory fault prediction result of the equipment to be detected.
In the context of the present invention, a computer-readable storage medium may be a tangible medium that can contain, or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. Alternatively, the computer readable storage medium may be a machine readable signal medium. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which a user can provide input to the electronic device. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), blockchain networks, and the internet.
The computing system may include clients and servers. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service are overcome.
It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps described in the present invention may be performed in parallel, sequentially, or in a different order, so long as the desired results of the technical solution of the present invention are achieved, and the present invention is not limited herein.
The above embodiments do not limit the scope of the present invention. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of the present invention.
Claims (11)
1. The method for determining the memory fault prediction model is characterized by comprising the following steps of:
acquiring memory original data, and performing multidimensional aggregation processing on the memory original data based on at least part of dimensionality, memory structure information and data type information of various aggregation parameter items to obtain initial aggregation characteristics;
screening the initial aggregation features based on the correlation between the initial aggregation features and the memory faults to obtain screened aggregation features, wherein different numbers of screened aggregation features form different feature groups;
And training at least one candidate model based on each feature group, performing index verification on each candidate model, determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature groups corresponding to the memory failure prediction models.
2. The method of claim 1, wherein the multi-dimensional aggregation process comprises intra-structure multi-dimensional aggregation and/or inter-structure multi-dimensional aggregation;
the intra-structure multidimensional aggregation includes: based on the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items; and/or, in a specific memory structure, based on the data type information and at least part of the aggregation parameter item aggregation mode;
the inter-structure multidimensional aggregation includes: and based on any two memory structures in the memory structure information, the data type information and at least part of the aggregation mode of the aggregation parameter items.
3. The method according to claim 2, wherein the performing multidimensional aggregation on the memory raw data based on at least a portion of dimensions of the plurality of aggregation parameter items, memory structure information and data type information to obtain an initial aggregation feature includes:
And traversing optional dimension information of a plurality of dimensions in any aggregation mode, and forming an initial aggregation feature in the aggregation mode based on the optional dimension information corresponding to the dimensions.
4. The method of claim 1, wherein the memory raw data comprises first memory raw data of a memory failed device and second memory raw data of a non-memory failed device;
the method further comprises the steps of:
setting a fault tag for the initial aggregation feature obtained based on the first memory original data;
setting a non-fault tag for the initial aggregation feature obtained based on the second memory original data;
and performing correlation calculation based on the initial aggregation characteristics and the corresponding tags, and determining the correlation between the initial aggregation characteristics and the memory faults.
5. The method of claim 1, wherein the screening the initial aggregate feature based on the correlation of the initial aggregate feature with memory failures to obtain a screened aggregate feature comprises:
and sorting the initial aggregation features based on the correlation between the initial aggregation features and the memory faults, and determining the screening aggregation features based on the feature screening quantity and the sorting of the initial aggregation features.
6. The method of claim 5, wherein the number of feature sets is a plurality;
the training based on each feature set to obtain at least one candidate model, performing index verification on each candidate model, determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining target aggregate features in feature sets corresponding to the memory failure prediction models, including:
respectively acquiring sample data based on each feature group, and respectively training the at least one initial model based on the sample data corresponding to each feature group to obtain at least one trained candidate model;
respectively performing index verification on the candidate models based on preset evaluation indexes to obtain index data of the candidate models;
determining a memory failure prediction model in a plurality of candidate models based on index data of the candidate models, and determining screening aggregation features in a feature group corresponding to the memory failure prediction model as target aggregation features corresponding to the memory failure prediction model.
7. The memory fault prediction method is characterized by comprising the following steps of:
Acquiring original memory data of equipment to be detected;
performing aggregation processing on the original memory data based on target aggregation characteristics corresponding to a memory failure prediction model to obtain feature data to be processed, wherein the memory failure prediction model and the target aggregation characteristics are obtained based on the determination method of the memory failure prediction model according to any one of claims 1-6;
and carrying out prediction processing on the feature data to be processed based on the memory fault prediction model to obtain a memory fault prediction result of the equipment to be detected.
8. A memory failure prediction model determining apparatus, comprising:
the feature aggregation module is used for acquiring memory original data, and carrying out multidimensional aggregation processing on the memory original data based on at least part of dimensions, memory structure information and data type information of various aggregation parameter items to obtain initial aggregation features;
the feature screening module is used for screening the initial aggregation features based on the correlation between the initial aggregation features and the memory faults to obtain screening aggregation features, and different feature groups are formed by different numbers of screening aggregation features;
the model determining module is used for obtaining at least one candidate model based on training of each feature group, carrying out index verification on each candidate model, determining a memory failure prediction model from a plurality of candidate models based on index data of the candidate models, and determining target aggregation features in feature groups corresponding to the memory failure prediction models.
9. A memory failure prediction apparatus, comprising:
the data acquisition module is used for acquiring original memory data of the equipment to be detected;
the data processing module is used for carrying out aggregation processing on the original memory data based on the target aggregation characteristics corresponding to the memory fault prediction model to obtain feature data to be processed;
and the fault prediction module is used for performing prediction processing on the feature data to be processed based on the memory fault prediction model to obtain a memory fault prediction result of the equipment to be detected.
10. An electronic device, the electronic device comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores a computer program executable by the at least one processor to enable the at least one processor to perform the method of determining a memory failure prediction model of any one of claims 1-6 and/or the method of memory failure prediction of claim 7.
11. A computer readable storage medium storing computer instructions for causing a processor to implement the method of determining a memory failure prediction model according to any one of claims 1-6 and/or the method of memory failure prediction according to claim 7 when executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311631231.8A CN117574087A (en) | 2023-11-30 | 2023-11-30 | Model determining method, memory fault predicting device, medium and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311631231.8A CN117574087A (en) | 2023-11-30 | 2023-11-30 | Model determining method, memory fault predicting device, medium and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117574087A true CN117574087A (en) | 2024-02-20 |
Family
ID=89886142
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311631231.8A Pending CN117574087A (en) | 2023-11-30 | 2023-11-30 | Model determining method, memory fault predicting device, medium and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117574087A (en) |
-
2023
- 2023-11-30 CN CN202311631231.8A patent/CN117574087A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116049146B (en) | Database fault processing method, device, equipment and storage medium | |
| CN113837596A (en) | Fault determination method and device, electronic equipment and storage medium | |
| CN115409419A (en) | Value evaluation method and device of business data, electronic equipment and storage medium | |
| CN115333783A (en) | API call abnormity detection method, device, equipment and storage medium | |
| CN111783883A (en) | Abnormal data detection method and device | |
| CN115481694B (en) | Data enhancement method, device and equipment for training sample set and storage medium | |
| CN115391160B (en) | Abnormal change detection method, device, equipment and storage medium | |
| CN116431505A (en) | Regression testing method and device, electronic equipment, storage medium and product | |
| CN115794473A (en) | Root cause alarm positioning method, device, equipment and medium | |
| CN116226628A (en) | Feature optimization method, device, equipment and medium | |
| CN115630708A (en) | Model updating method and device, electronic equipment, storage medium and product | |
| CN117574087A (en) | Model determining method, memory fault predicting device, medium and equipment | |
| CN115600607A (en) | Log detection method and device, electronic equipment and medium | |
| CN115794578A (en) | Data management method, device, equipment and medium for power system | |
| CN114999665A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN113807391A (en) | Task model training method and device, electronic equipment and storage medium | |
| CN114037058B (en) | Pre-training model generation method and device, electronic equipment and storage medium | |
| CN115994147B (en) | Track library construction method and device, electronic equipment and medium | |
| CN114037057B (en) | Pre-training model generation method and device, electronic equipment and storage medium | |
| CN115576850B (en) | Data index testing method and device, electronic equipment and storage medium | |
| CN115455019A (en) | Search intention identification method, device and equipment based on user behavior analysis | |
| CN117608896A (en) | Transaction data processing method and device, electronic equipment and storage medium | |
| CN117217777A (en) | Evaluation method, device, equipment and medium based on contrast learning | |
| CN116628491A (en) | Training method and device for classification model, electronic equipment and storage medium | |
| CN115437895A (en) | Database health state determination method and device, computer equipment and storage medium |
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 |