CN111950810A - Multivariable time sequence prediction method and device based on self-evolution pre-training - Google Patents

Multivariable time sequence prediction method and device based on self-evolution pre-training Download PDF

Info

Publication number
CN111950810A
CN111950810A CN202010876972.2A CN202010876972A CN111950810A CN 111950810 A CN111950810 A CN 111950810A CN 202010876972 A CN202010876972 A CN 202010876972A CN 111950810 A CN111950810 A CN 111950810A
Authority
CN
China
Prior art keywords
time
prediction
model
training
univariate
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010876972.2A
Other languages
Chinese (zh)
Other versions
CN111950810B (en
Inventor
李文中
万晨
张治杰
丁望祥
叶保留
陆桑璐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nanjing University
Original Assignee
Nanjing University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nanjing University filed Critical Nanjing University
Priority to CN202010876972.2A priority Critical patent/CN111950810B/en
Publication of CN111950810A publication Critical patent/CN111950810A/en
Application granted granted Critical
Publication of CN111950810B publication Critical patent/CN111950810B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • 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

Abstract

The invention discloses a multivariate time sequence prediction method and equipment based on self-evolution pre-training, wherein the method is based on a pre-training strategy, a convolutional network and a long and short memory network depth sequence model, and is combined with univariate self-evolution information and multivariate dependency relationship information to carry out modeling, so that an optimization algorithm of multivariate time sequence prediction is realized, and meanwhile, the overall prediction precision and the prediction precision of local univariate are considered. The method has better overall prediction precision, and is superior to the existing multivariate time sequence prediction method in the aspect of prediction precision guarantee of local univariates.

Description

Multivariable time sequence prediction method and device based on self-evolution pre-training
Technical Field
The invention relates to time series prediction, in particular to a multivariate time series prediction method.
Background
The time sequence prediction method is adopted to obtain the predicted value at the future moment, and the method has basic and important significance for assisting decision, optimizing resource allocation, taking intervention measures in advance and the like. Information systems that operate in real time have a wide range of applications. Many systems generate a large amount of time series data, mainly including system performance monitoring data, and the performance prediction of the monitoring data generates many application values, such as predicting network traffic for resource planning and anomaly discovery, predicting disk failure of a data center for early replacement, and the like. Therefore, the future data for predicting the system performance provides important help for daily operation, operation and maintenance and the like of the modern information system. There has been a great deal of research in recent years focusing on the problem of time series prediction for real-time systems. Linear statistical algorithms (e.g. AR, ARIMA) have an impact on the field of time series prediction. However, since linear models lack applicability in many practical applications, nonlinear time series analysis and prediction have been proposed and increasingly applied. In addition to classical statistical methods, machine learning models have become important algorithms in the field of time series prediction. Researchers apply machine learning models such as linear regression models, support vector machines, decision trees, etc. to time series analysis. The time series prediction algorithm based on machine learning takes time series prediction as a supervised learning task, wherein the input attribute is time series data of historical observation, and the output label is future data. However, the multivariate time series prediction algorithm focusing on the global prediction target optimization at present can bring the problem of local variable prediction precision loss. For example, in the regression task of multivariate prediction, existing algorithms target minimization of the sum of the overall prediction errors (weighting) of all variables, and improve the overall prediction accuracy by optimizing such minimization targets.
Disclosure of Invention
The purpose of the invention is as follows: in order to guarantee the prediction precision of local variables, a multivariate time sequence prediction algorithm based on self-evolution pre-training is provided. Not only can realize the optimization algorithm of multivariable time series prediction, but also can give consideration to the overall prediction precision and the prediction precision of local univariate
The technical scheme is as follows: in order to predict the performance of a system running in real time and perform reasonable operation and maintenance aiming at a predicted value, the invention provides a multivariate time sequence prediction method based on self-evolution pre-training, which comprises the following steps:
s1, acquiring N-dimensional index data representing the performance of the real-time running system, forming N univariate time sequences as historical input sequences, and preprocessing;
s2, inputting the N univariate time sequences into an auto-evolution pre-training model SE based on a differential feature structure, establishing a univariate time sequence linear regression model, and explicitly fusing multi-order differential information into the linear autoregressive model to form corresponding prediction output;
s3, performing one-dimensional convolution operation with different sizes on a two-dimensional tensor composed of N univariate time sequences, and extracting dependency relationship information among variables to obtain a plurality of characteristic graphs;
s4, inputting the characteristic diagram into a long and short memory network (LSTM) to perform time sequence modeling on the multi-scale characteristic diagram so as to capture time sequence information of the multivariate dependency relationship, wherein the corresponding multivariate modeling network component forms corresponding output;
s5, performing one-to-one corresponding addition fusion of vector elements on the outputs obtained in the steps S2 and S4, constructing multivariate prediction output, and training a network to obtain a trained model;
s6, the time series to be predicted is input to the trained model obtained in step S5, and the predicted value of the time series is obtained.
In a second aspect, the invention features a computer device that includes one or more processors; memory for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to carry out the method according to the first aspect of the invention.
Has the advantages that: the invention provides a multivariate time sequence prediction method based on self-evolution pre-training, which is based on a pre-training strategy, a convolutional network and a long and short memory network depth sequence model and is combined with univariate self-evolution information and multivariate dependency relationship information to carry out modeling, thereby realizing an optimization algorithm of multivariate time sequence prediction and simultaneously considering the overall prediction precision and the prediction precision of local univariates. The method has better overall prediction precision, and is superior to the existing multivariate time sequence prediction method in the aspect of prediction precision guarantee of local univariates.
Drawings
FIG. 1 is a general flow diagram of a multivariate time series prediction method according to an embodiment of the invention;
FIG. 2 is a schematic diagram of the main processes of a multivariate time series prediction method according to an embodiment of the invention;
FIG. 3 is a process flow diagram of a model training phase according to an embodiment of the invention;
FIG. 4 is a process flow diagram of a model run phase according to an embodiment of the invention.
Detailed Description
The technical scheme of the invention is further explained by combining the attached drawings. It should be understood that the following embodiments are provided only for the purpose of thoroughly and completely disclosing the present invention and fully conveying the technical concept of the present invention to those skilled in the art, and the present invention may be embodied in many different forms and is not limited to the embodiments described herein. The terminology used in the exemplary embodiments illustrated in the accompanying drawings is not intended to be limiting of the invention.
In the embodiment of the invention, a multivariate time sequence prediction method based on self-evolution pre-training is used for predicting the time sequence. Based on a pre-training strategy, a convolutional network and a long and short memory network depth sequence model, and by combining univariate self-evolution information and multivariate dependency relationship information to carry out modeling, an optimization algorithm of multivariate time sequence prediction is realized, and meanwhile, the overall prediction precision and the prediction precision of local univariates are considered. Referring to fig. 1 and 2, the method of the present invention comprises the steps of:
step S1, collecting corresponding multivariate time series data from the real-time operating system through the monitoring system, wherein the multivariate time series data may include a plurality of key indexes such as CPU load, bandwidth utilization, bandwidth speed, memory utilization, response delay, IO read/write, and the like. In reality, data often causes problems such as data missing, data irregularity, etc., so that the data is preprocessed first, and then the preprocessed data is used as the input of the model. In addition, in order to have a better modeling effect on the time series, the periodicity of the time series also needs to be evaluated to obtain a history window T.
Preprocessing data, namely dividing a data set by a time dimension, reserving a training set for training a model, and standardizing the data to enable data values to be mapped to the same interval, so that the model learning is facilitated, such as min-max standardization:
Figure BDA0002652876580000031
wherein, minxIs the minimum value of x, maxxIs the maximum value in x. The periodicity of the time series is evaluated by the autocorrelation function ACF to obtain the historical input series length T. A plurality of key indexes such as CPU load, bandwidth utilization rate, bandwidth speed, memory utilization rate, response delay, IO read-write and the like respectively form a univariate time sequence, the length of each sequence is T, and x represents a vector value of the sequence.
And step S2, inputting the N univariate time sequence distributions of the key indexes including CPU load, bandwidth utilization rate, memory utilization rate, response time delay and the like preprocessed in the step S1 into a Self-Evolution pre-training model SE (Self-Evolution) based on a difference feature structure, and establishing a univariate time sequence linear regression model, wherein the model has the function of ensuring the prediction accuracy of the univariate time sequence generated by a real-time system. In addition, considering that the multi-order difference information can improve the applicability and robustness in the prediction of stable and non-stable sequences, the multi-order difference information is explicitly merged into the linear autoregressive model to form corresponding prediction output. The method specifically comprises the following steps:
and S2-1, inputting the N univariate time sequences into a self-evolution pre-training model SE based on the difference characteristic structure, and establishing a univariate time sequence linear regression model. For the ith sequence there are:
Figure BDA0002652876580000041
wherein
Figure BDA0002652876580000042
Representing the time series value at time t-j in the ith sequence,
Figure BDA0002652876580000043
for the corresponding linear weight, T is the length of the historical input sequence obtained by evaluating the periodicity of the time sequence by the autocorrelation function ACF.
In step S2-2, when the difference order q is 1, there is
Figure BDA0002652876580000044
Namely, it is
Figure BDA0002652876580000045
It can be seen that the predicted value under the first order difference contains
Figure BDA0002652876580000046
Handle
Figure BDA0002652876580000047
The equal difference characteristics are explicitly merged into the linear autoregressive model to form the corresponding prediction output
Figure BDA0002652876580000048
Figure BDA0002652876580000049
Figure BDA00026528765800000410
The fusion weight value which is the difference characteristic is used for training and limited to be a non-negative real number, softmax is corresponding normalization operation, and Q is the maximum difference order;
Figure BDA00026528765800000411
is a standard linear autoregressive term, xt-T：t-1Representing the time series from time T-1 to time T-T, T being the length of the history input series.
And step S3, performing one-dimensional convolution operations with different sizes on the two-dimensional tensor composed of the N univariate time sequences, and extracting the dependency relationship information among the variables to obtain a plurality of characteristic graphs.
The invention adopts the time sequence convolution of one-dimensional CNN layers of a plurality of convolution kernels to express the complex dependency relationship among the multivariate time sequences, and aims to obtain a characteristic map (feature map) containing the local dependency relationship and multilevel information of the multivariate sequence. For input sequences
Figure BDA00026528765800000412
A one-dimensional convolution operation in the variable dimension is performed, represented as follows:
C(k)＝W(k)*Xt-T：t-1
W(k)is the k-th convolution kernel, C(k)Is the convolution result, i.e. the feature map extracted from the dependency relationship between the variables. In convolutional networks, k convolution kernels W are typically employed(1)，W(2)，...，W(k)Feature map C for forming multiple channels(1)，C(2)，...，C(k). Unlike prior algorithms that set all convolution kernels to the same size, the present invention employs convolution kernels of varying sizes. In contrast to the same scale convolution kernel, the multi-scale convolution kernel explicitly describes the associated time period characteristics at different scales. C(1)，C(2)，...，C(k)Contains multi-level information of the local dependency relationship of the multi-variable time sequence.
Step S4, inputting the feature maps into a Long Short-Term Memory (LSTM) network for time sequence modeling of the multi-scale feature maps, capturing the time sequence information of the multivariable dependency relationship, and inputting each feature map into the corresponding LSTM respectively to obtain the corresponding output.
LSTM is a classical recurrent neural network that can solve the long-term dependence of sequences. The network component based on multivariate modeling in the invention is a structure formed by a plurality of LSTM basic memory units, and each LSTM correspondingly inputs a characteristic diagram. Specifically, the feature maps C of all channels of the kth convolution kernel obtained in step S3 are obtained(k)Splicing according to time dimension, wherein the number of channels is determined by the dimension of time sequence data, and the spliced characteristic diagram is used as the input of a corresponding kth long-short memory network; the hidden output of the corresponding network unit is input to the full connection layer to form an output vector h(k)∈RN×1. K identical LSTM summed outputs h (1)，h(2)，...，h(K)These outputs contain the dependency information of the sequence. Accordingly, all vectors h(1)，h(2)，...，h(K)Adding to form multivariate dependence modeling output result
Figure BDA0002652876580000051
And S5, performing one-to-one corresponding addition fusion of vector elements on the outputs obtained in the steps S2 and S4, constructing multivariate prediction output, and training a network to obtain a trained model. The method comprises the following steps:
and (4) carrying out one-to-one corresponding addition fusion of vector elements on the outputs obtained in the steps S2 and S4, wherein the fusion method comprises the following steps:
Figure BDA0002652876580000052
W(se)，W(id)to fuse the weights, by manually setting the hyper-parameters as a model,
Figure BDA0002652876580000053
representing element multiplication. Such two parts addThe sexual fusion has clear practical meaning, the interpretability of the algorithm is increased, and the loss function is minimized, so that the accuracy loss of local single variables can be controlled not to be increased, and meanwhile, the overall prediction accuracy loss can be controlled, and the purpose of the algorithm is achieved. For example, the predicted value of the ith variable
Figure BDA0002652876580000054
Then for the prediction of the variable, the relative stool and urine of the two values directly reflect the proportion of the two parts of information in the prediction. After the model construction is completed, constructing a loss function corresponding to an optimization target for training the model as follows:
Figure BDA0002652876580000055
the overall model can be optimized based on a gradient descent method. The weight of the model training is the weight of the Self-Evolution, CNN and LSTM models. The model is a single step prediction model, then the final output
Figure BDA0002652876580000061
Is a multivariate vector value.
The above steps S1-S5 are a model training stage, and as shown in fig. 3, a multivariate time series prediction model based on self-evolution pre-training is trained through historical time series data.
And step S6, predicting the time sequence to be predicted by using the trained model.
As shown in fig. 4, in the model operation stage, the time sequence to be predicted is preprocessed and then input to the trained model obtained in step S5 to obtain the predicted value of the time sequence, that is, the model finally obtains the predicted output of the real-time system.
It will be understood by those skilled in the art that all or part of the steps for implementing the above embodiments may be implemented by hardware, or may be implemented by a program instructing associated hardware, and the program may be stored in a computer-readable storage medium. In the context of the present invention, the computer-readable medium may be considered tangible and non-transitory. Non-limiting examples of a non-transitory tangible computer-readable medium include a non-volatile memory circuit (e.g., a flash memory circuit, an erasable programmable read-only memory circuit, or a mask read-only memory circuit), a volatile memory circuit (e.g., a static random access memory circuit or a dynamic random access memory circuit), a magnetic storage medium (e.g., an analog or digital tape or hard drive), and an optical storage medium (e.g., a CD, DVD, or blu-ray disc), among others.
Program code for implementing the methods of the present invention may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the program codes, when executed by the processor or controller, cause the functions/operations specified in the flowchart and/or block diagram to be performed. The program code 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.
Further, while operations are depicted in a particular order, this should be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the invention. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single implementation. Conversely, various features that are described in the context of a single implementation can also be implemented in multiple implementations separately or in any suitable subcombination.
Although the preferred embodiments of the present invention have been described in detail, the present invention is not limited to the details of the embodiments, and various equivalent modifications can be made within the technical spirit of the present invention, and the scope of the present invention is also within the scope of the present invention.

Claims (7)

1. A multivariate time sequence prediction method based on self-evolution pre-training is characterized by comprising the following steps:
s1, acquiring N-dimensional index data representing the performance of the real-time running system, forming N univariate time sequences as historical input sequences, and preprocessing;
s2, inputting the N univariate time sequences into a self-evolution pre-training model based on a differential feature structure, establishing a univariate time sequence linear regression model, explicitly integrating multi-order differential information into the linear regression model to form corresponding prediction output
Figure FDA0002652876570000011
S3, performing one-dimensional convolution operation with different sizes on a two-dimensional tensor composed of N univariate time sequences, and extracting dependency relationship information among variables to obtain a plurality of characteristic graphs;
s4, inputting the characteristic diagram into the long and short memory network to perform time sequence modeling on the multi-scale characteristic diagram to capture the time sequence information of the multivariable dependency relationship, and the corresponding multivariable modeling network components form corresponding output
Figure FDA0002652876570000012
S5, and outputting the result of the steps S2 and S4
Figure FDA0002652876570000013
And
Figure FDA0002652876570000014
carrying out one-to-one corresponding addition fusion of vector elements to formEstablishing a multivariable prediction output model, and training to obtain a trained model;
and S6, preprocessing the time series needing to be predicted and inputting the preprocessed time series into the trained model obtained in the step S5, so as to obtain the predicted value of the time series.
2. The detection method according to claim 1, wherein the step S2 includes:
s2-1, inputting the N univariate time sequences into a self-evolution pre-training model constructed based on differential features, and establishing a univariate time sequence linear regression model, wherein for the ith sequence:
Figure FDA0002652876570000015
wherein
Figure FDA0002652876570000016
Representing the time series value of time t-j in the ith sequence,
Figure FDA0002652876570000017
t is the length of the historical input sequence;
s2-2, explicitly fusing the multi-order difference information into the linear regression model to form corresponding prediction output
Figure FDA0002652876570000018
Figure FDA0002652876570000019
Figure FDA0002652876570000021
The fusion weight of the difference features is defined, softmax is corresponding normalization operation, and Q is the maximum difference order;
Figure FDA0002652876570000022
is a standard linear autoregressive term, xt-T：t-1Representing a time series from time T-1 to time T-T.
3. The prediction method according to claim 1, wherein the one-dimensional CNN convolution operation in step S3 is as follows:
C(k)＝W(k)*Xt-T：t-1
W(k)is the k-th convolution kernel, C(k)Is the result of convolution, i.e. a feature map, X, extracted from the dependencies between variablest-Tt-1A two-dimensional tensor representing a composition of N univariate time series,
Figure FDA0002652876570000023
xt-T：t-1representing the time series from time T-1 to time T-T, T being the length of the history input series.
4. The prediction method according to claim 1, wherein the step S4 includes: splicing the characteristic graphs C (k) of all channels of the kth convolution kernel obtained in the step S3 according to time dimension, and then taking the characteristic graphs C (k) as the input of the corresponding kth long-short memory network; the hidden output of the corresponding network unit is input to the full connection layer to form an output vector h(k)∈RN×1K identical LSTM total outputs h(1)，h(2)，...，h(K)All vectors h(1)，h(2)，...，h(K)Adding to form multivariate dependence modeling output result
Figure FDA0002652876570000024
5. The detecting method according to claim 1, wherein the results outputted from steps S2 and S4 are weighted and fused in step S5, and the fusing method is as follows:
Figure FDA0002652876570000025
W(se)，W(id)in order to fuse the weight values,
Figure FDA0002652876570000026
representing element multiplication.
6. The detection method according to claim 5, wherein the model is trained in step S5, and a loss function corresponding to the optimization objective is constructed as follows:
Figure FDA0002652876570000027
n is the dimension of the multivariate time series.
7. A computer device, the device comprising:
one or more processors;
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one processor, the programs when executed by the processor implementing the steps of the method of any of claims 1-6.
CN202010876972.2A 2020-08-27 2020-08-27 Multi-variable time sequence prediction method and equipment based on self-evolution pre-training Active CN111950810B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010876972.2A CN111950810B (en) 2020-08-27 2020-08-27 Multi-variable time sequence prediction method and equipment based on self-evolution pre-training

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010876972.2A CN111950810B (en) 2020-08-27 2020-08-27 Multi-variable time sequence prediction method and equipment based on self-evolution pre-training

Publications (2)

Publication Number Publication Date
CN111950810A true CN111950810A (en) 2020-11-17
CN111950810B CN111950810B (en) 2023-12-15

Family

ID=73366711

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010876972.2A Active CN111950810B (en) 2020-08-27 2020-08-27 Multi-variable time sequence prediction method and equipment based on self-evolution pre-training

Country Status (1)

Country Link
CN (1) CN111950810B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113627591A (en) * 2021-07-30 2021-11-09 马上消费金融股份有限公司 Dynamic graph data processing method and device, electronic equipment and storage medium
CN113807951A (en) * 2021-09-23 2021-12-17 中国建设银行股份有限公司 Transaction data trend prediction method and system based on deep learning
CN114039758A (en) * 2021-11-02 2022-02-11 中邮科通信技术股份有限公司 Network security threat identification method based on event detection mode

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108648457A (en) * 2018-06-28 2018-10-12 苏州大学 A kind of method, apparatus and computer readable storage medium of prediction of speed
CN108830430A (en) * 2018-08-01 2018-11-16 苏州大学张家港工业技术研究院 Convolutional neural networks multiple spot regressive prediction model for traffic flow forecasting

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108648457A (en) * 2018-06-28 2018-10-12 苏州大学 A kind of method, apparatus and computer readable storage medium of prediction of speed
CN108830430A (en) * 2018-08-01 2018-11-16 苏州大学张家港工业技术研究院 Convolutional neural networks multiple spot regressive prediction model for traffic flow forecasting

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
万晨: "一种基于自演化预训练的多变量时间序列预测算法", 计算机学报, vol. 45, no. 3, pages 513 - 525 *
张振;曾献辉;: "基于CNN-LightGBM模型的高速公路交通量预测", 信息技术与网络安全, no. 02, pages 18 - 23 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113627591A (en) * 2021-07-30 2021-11-09 马上消费金融股份有限公司 Dynamic graph data processing method and device, electronic equipment and storage medium
CN113807951A (en) * 2021-09-23 2021-12-17 中国建设银行股份有限公司 Transaction data trend prediction method and system based on deep learning
CN114039758A (en) * 2021-11-02 2022-02-11 中邮科通信技术股份有限公司 Network security threat identification method based on event detection mode

Also Published As

Publication number Publication date
CN111950810B (en) 2023-12-15

Similar Documents

Publication Publication Date Title
EP3857377B1 (en) Disk drive failure prediction with neural networks
US11593611B2 (en) Neural network cooperation
US20200097810A1 (en) Automated window based feature generation for time-series forecasting and anomaly detection
CN111950810A (en) Multivariable time sequence prediction method and device based on self-evolution pre-training
KR20210032140A (en) Method and apparatus for performing pruning of neural network
CN111091196B (en) Passenger flow data determination method and device, computer equipment and storage medium
US10902311B2 (en) Regularization of neural networks
Ayodeji et al. Causal augmented ConvNet: A temporal memory dilated convolution model for long-sequence time series prediction
US20200167660A1 (en) Automated heuristic deep learning-based modelling
US20200050982A1 (en) Method and System for Predictive Modeling for Dynamically Scheduling Resource Allocation
CN112734106A (en) Method and device for predicting energy load
US20220027739A1 (en) Search space exploration for deep learning
CN112508177A (en) Network structure searching method and device, electronic equipment and storage medium
Kim et al. Comparing cost prediction methods for apartment housing projects: CBR versus ANN
CN116665798A (en) Air pollution trend early warning method and related device
KR102480518B1 (en) Method for credit evaluation model update or replacement and apparatus performing the method
Zhou et al. Deep learning-based intelligent multilevel predictive maintenance framework considering comprehensive cost
CN112990530B (en) Regional population quantity prediction method, regional population quantity prediction device, electronic equipment and storage medium
JP7424373B2 (en) Analytical equipment, analytical methods and analytical programs
CN114239934A (en) Power grid load sequence prediction method and system based on generative countermeasure network
CN114662009A (en) Industrial internet factory collaborative recommendation algorithm based on graph convolution
KR102531291B1 (en) Method for predicting energy consumption of a building, and computing device performing the method
KR102499435B1 (en) Method for generating credit evaluation model based on artificial intelligence and apparatus performing the method
Shetty et al. A Weighted Ensemble of VAR and LSTM for Multivariate Forecasting of Cloud Resource Usage. J
Zhao et al. Neural-Network-Based Feature Learning: Recurrent Neural Network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant