CN109272033B - Online soft interval kernel learning algorithm based on step length control - Google Patents

Online soft interval kernel learning algorithm based on step length control Download PDF

Info

Publication number
CN109272033B
CN109272033B CN201811037902.7A CN201811037902A CN109272033B CN 109272033 B CN109272033 B CN 109272033B CN 201811037902 A CN201811037902 A CN 201811037902A CN 109272033 B CN109272033 B CN 109272033B
Authority
CN
China
Prior art keywords
online
kernel
sample
function
decision function
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.)
Active
Application number
CN201811037902.7A
Other languages
Chinese (zh)
Other versions
CN109272033A (en
Inventor
宋允全
李月菱
于琪
雷鹤杰
梁锡军
渐令
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.)
China University of Petroleum East China
Original Assignee
China University of Petroleum East China
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 China University of Petroleum East China filed Critical China University of Petroleum East China
Priority to CN201811037902.7A priority Critical patent/CN109272033B/en
Publication of CN109272033A publication Critical patent/CN109272033A/en
Application granted granted Critical
Publication of CN109272033B publication Critical patent/CN109272033B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06F18/24—Classification techniques

Abstract

The invention relates to an online soft interval kernel learning algorithm (OSKL) based on step length control. A nonlinear classifier is constructed by introducing a kernel function, soft interval parameters are introduced to control the influence of noise data, and an online kernel learning algorithm with robustness is designed based on a basic framework of an online gradient descent algorithm. The algorithm can reduce the storage space of the model, effectively control the noise influence, and has the advantages of only O (1) of the calculation complexity of model updating, strong real-time performance, easy realization and the like, thereby being a natural tool for processing and analyzing the data stream problem. The online learning algorithm solves the problem that the traditional classification method based on batch processing technology cannot process data stream efficiently, and also solves the problem that the existing online learning algorithms such as Kernel Perceptron and Pegasos cannot effectively inhibit noise influence.

Description

Online soft interval kernel learning algorithm based on step length control
Technical Field
The invention belongs to the field of data mining and machine learning, relates to a method for data mining and data processing, and particularly relates to an online soft interval kernel learning algorithm (OSKL) based on step length control.
Background
The classification problem is a classic problem in the field of data mining and machine learning. The traditional classification method based on the batch processing technology firstly collects data, builds a learning model based on the collected data, and selects an optimization algorithm to solve the model to obtain a classifier. With the rapid development of technologies such as e-commerce, social media, mobile internet, internet of things, and the like, more and more application scenarios require real-time processing of large-scale data streams. The traditional classification method based on the batch processing technology has the defects of high calculation complexity, low model updating efficiency and the like when processing a large-scale data stream problem. The online learning is based on a basic frame of point-by-point learning, data information is learned point-by-point through a dynamic updating model, the calculation complexity of the model updating once is only O (1), and the online learning method has the advantages of low calculation complexity, high model updating efficiency, strong real-time performance and the like, and is a natural tool for processing and analyzing data flow problems. In addition, some error labels are inevitable in large-scale label data, and the error labels can seriously influence the construction and the effect of the classifier. Therefore, it is desirable to design a data stream mining algorithm with fault tolerance.
Disclosure of Invention
The invention aims to provide an online soft interval kernel learning algorithm based on step length control, which can reduce model storage space, effectively control noise influence, remarkably improve model updating efficiency and meet the real-time requirement of practical application problems, aiming at the problem that the existing classification method based on batch processing technology cannot efficiently process data stream classification and the online learning algorithm cannot inhibit noise influence.
According to an embodiment of the present invention, an online soft interval kernel learning algorithm based on step size control is provided, which includes the following steps:
initializing model parameters, a decision function and a model kernel function.
And (II) collecting the data stream, and predicting the class label of the data stream sample by using a classification decision function.
And (III) acquiring a sample real label, appointing a loss function, and calculating a sample loss value.
And (IV) calculating the updating step size of the decision function of the classifier.
And (V) updating a classifier decision function based on the basic framework of the online gradient descent algorithm.
In the learning algorithm according to the embodiment of the present invention, in step (one), the specific steps of model initialization are:
determining a training sample set and a test sample set, initializing a model threshold parameter C, and initializing a decision function f of a binary problem 00, a gaussian kernel function is designated as a model kernel function k (·, ·).
In the learning algorithm according to the embodiment of the present invention, in the step (two), the specific steps of predicting the class label of the data stream sample by using the classification decision function are as follows:
collecting data stream in the form of one-by-one { (x)t,yt)}t＝1,2,…，xtDenotes the t-th sample input, ytRepresenting the t-th sample output (class label). Using a decision function ft-1Predicting the t-th sample in a data stream
Figure BDA0001791348020000021
The label of (2):
Figure BDA0001791348020000022
in the learning algorithm according to the embodiment of the present invention, in step (three), the specific operation flow for calculating the sample loss is as follows:
the sample point (x) is calculated by assigning the most common change function of the two-classification problem as the loss functiont,yt) Change loss of (c):
Figure BDA0001791348020000023
in the learning algorithm according to the embodiment of the present invention, in step (four), the update step τ is calculatedtThe specific operation flow is as follows: determining the update step τ based on the following two considerationst: firstly, realizing current sample x with highest possible confidencetCorrect classification of points, i.e. reaching zero loss (l)t0); secondly, the stability of the algorithm is ensured as much as possible, namely the fluctuation of the decision function in the updating process is reduced. Optimum step length tautIs a solution to the following optimization problem:
Figure BDA0001791348020000024
on the other hand, large-scale sampling data inevitably contains a large amount of error label data, and the error labels can seriously influence the construction of the decision function and the effect of a corresponding classifier. To this end, we introduce soft interval threshold parameter controlSystem update step τtC is less than or equal to C, so that the influence of error label data on the model is limited, and the stability of the classifier is ensured. Based on the sample points (x) calculated in step (three)t,yt) 1 change loss ltAnd a step size control parameter C, determining an update step size tautComprises the following steps:
Figure BDA0001791348020000025
in the learning algorithm according to the embodiment of the present invention, in step (v), the specific operation flow for updating the classifier decision function is as follows:
based on the update step τ calculated in step (IV)tUnder the basic framework of the online gradient descent algorithm, a decision function f is settPerform the update
Figure BDA0001791348020000026
Obtaining a new decision function ft。
The invention relates to an online soft interval kernel learning algorithm based on step length control. An online kernel learning classifier is established by introducing a change loss function, a Gaussian kernel function and a soft interval threshold parameter C, so that online prediction of data flow is realized. The method adopts soft interval threshold parameters to enable the updating of the decision function of the classifier to be smoother and has robustness. Compared with classical online learning algorithms Kernel Perceptron and Pegesoso, the proposed algorithm OSKL significantly improves the classification accuracy. The online classification algorithm OSKL can flexibly process the classification problem in a data flow scene, and compared with the traditional static classification mode based on batch processing technology, the online classification algorithm OSKL greatly reduces the calculation complexity and the model operation time.
Drawings
FIG. 1 is a schematic diagram of an online soft interval kernel learning algorithm based on step control
FIG. 2 is a schematic diagram showing classification accuracy comparison of three algorithms on a reference data set
FIG. 3 is a schematic diagram showing the comparison of the classification accuracy of the average test of three algorithms on a noisy label data set ijcnn
FIG. 4 is a schematic diagram showing the comparison of the classification accuracy of the average test of three algorithms on the noise-containing tag data set codna
FIG. 5 is a schematic diagram showing comparison of average test classification accuracy of three algorithms on noise-containing tag data set eegeye
Detailed Description
The specific steps of the present invention are explained below with reference to the drawings.
The first embodiment is as follows: the online classification experiments on the original reference data sets ijcnn, codrna, eegeye are taken as examples for explanation. Fig. 1 is a schematic diagram of an online soft interval kernel learning algorithm based on step size control according to an embodiment of the present invention, where the online learning algorithm includes the following steps:
the method comprises the following steps: model parameters, decision functions and model kernel functions are initialized. The method comprises the following specific steps:
initializing a model threshold parameter C to 0.05, and initializing a binary problem decision function f0When 0, a gaussian kernel function is designated as a model kernel function, i.e., k (x)i,xj)＝exp(-‖xi-xj‖2D), where d is taken as the dimension of the sample input x.
Step two: and collecting the data stream, and predicting the class label of the data stream sample by using a classification decision function. The method comprises the following specific steps: collecting data stream in the form of one-by-one { (x)t,yt)}t＝1,2,…，xtDenotes the t-th sample input, ytRepresenting the t-th sample output (class label). Using a decision function ft-1Predicting the t-th sample in a data stream
Figure BDA0001791348020000031
The label of (2):
Figure BDA0001791348020000032
step three: obtainingAnd (4) a sample real label appoints a loss function and calculates a sample loss value. The method comprises the following specific steps: the sample point (x) is calculated by assigning the most common change function of the two-classification problem as the loss functiontYt) loss of change:
Figure BDA0001791348020000033
step four: and calculating the updating step length of the decision function of the classifier. The method comprises the following specific steps: introducing soft interval threshold parameter to control update step length tautC is less than or equal to C, so that the influence of error label data on the model is limited, and the stability of the classifier is ensured. Based on the sample points (x) calculated in step (three)t,yt) 1 change loss ltAnd step size control parameter C, determining the updating step size tau of the t steptComprises the following steps:
Figure BDA0001791348020000041
step five: and updating a decision function of the classifier based on a basic framework of an online gradient descent algorithm. The method comprises the following specific steps: based on the update step τ calculated in step (IV)tUnder the basic framework of the online gradient descent algorithm, a decision function f is settPerform the update
Figure BDA0001791348020000042
Obtaining a new decision function ft。
FIG. 2 is a schematic diagram showing the comparison of the average online test accuracy of predictions made by the online learning algorithm of the present invention and the existing online learning algorithms Kernel Perceptron and Pegasos in the reference data set ijcnn, the reference data set codrna and the reference data set eegeye. As can be seen from FIG. 2, the average test accuracy of the online learning algorithm of the present invention on the above 3 reference data sets is better than that of the other two methods.
Example two: on the basis of the original reference data sets ijcnn, codrna and eegeye, a noise label is added, and an online classifier is trained on the data set containing the noise label. Unlike the first embodiment, in the first embodiment, 30% of the data set is randomly selected as the test set, and the rest of the data is added with the noise label to construct the training set. Specifically, we modulo 20, modulo 10, modulo 5 the sample index, and multiply the sample point label with a remainder of 0 by-1 to obtain noise label data.
Fig. 3-5 are graphs of the average classification performance (average test accuracy, ACA) on the original 30% dataset noiseless test dataset with training on the online classifiers Kernel permeatron, pegaso and OSKL on the noisy labeled datasets ijcnn, codna, eegeye. Experimental results show that the classification precision of the three algorithms is reduced with the increase of noise of training samples indexed by mod20, mod10 and mod5, but the OSKL algorithm provided by I can effectively control the noise influence under the condition of containing noise, and the classification effect is obviously higher than that of online classifiers Kernel Perception and Pegasos.
The above-described embodiments are intended to illustrate rather than to limit the invention, and any modifications and variations of the present invention are possible within the spirit and scope of the claims.

Claims (6)

1. An online soft interval kernel learning method based on step length control is characterized by comprising the following steps:
initializing model parameters, a decision function and a model kernel function; adding a noise label on the basis of the original reference data set ijcnn, codrna and eegeye, and training an online classifier on a data set containing the noise label; randomly selecting 30% of the data set as a test set, and adding noise labels into the rest data to construct a training set;
(II) collecting data stream in the form of one-by-one, and utilizing decision function ft-1For data stream sample xtPredicting the label of (2);
(III) obtaining the real label ytThen, change loss of the sample point is calculated
Figure FDA0003283364400000012
(IV) calculating the update step τt: if ltWhen the value is 0, then τt0; if ltIf greater than 0, then
Figure FDA0003283364400000011
(V) updating the classifier decision function ft＝ft-1+τtytk(xt,·)。
2. The method for learning online soft interval core based on step size control as claimed in claim 1, wherein in step (one), the specific operation flow is: determining a training sample set and a test sample set, initializing a model threshold parameter C, and initializing a decision function f of a binary problem00, a gaussian kernel function is chosen as the model kernel function k (·, · k), i.e. k (x)i,xj)＝exp(-||xi-xj||2D), where d is taken as the dimension of the sample input x.
3. The on-line soft interval kernel learning method based on step size control as claimed in claim 1, wherein: in the step (two), the specific steps of predicting the data stream sample class label by using the classification decision function are as follows: collecting data stream in the form of one-by-one { (x)t,yt)}t＝1,2,...，xtDenotes the t-th sample input, ytRepresents the t-th sample output (class label); using a decision function ft-1Predicting the label of the t-th sample in the data stream:
Figure FDA0003283364400000021
4. the method for learning online soft interval core based on step size control as claimed in claim 1, wherein in step (iii), the specific operation procedure for calculating the sample loss is as follows: the sample point (x) is calculated by assigning the most common change function of the two-classification problem as the loss functiont,yt) Change loss of (c):
Figure FDA0003283364400000023
5. the on-line soft interval kernel learning method based on step size control as claimed in claim 1, wherein in step (IV), the update step size τ is calculatedtThe specific operation flow is as follows: determining the update step τ based on the following two considerationst(ii) a Firstly, realizing current sample x with highest possible confidencetCorrect classification of points, i.e. reaching zero loss (l)t0); secondly, the stability of the algorithm is ensured as much as possible, namely the fluctuation of the decision function in the updating process is reduced; the optimal step length tau can be obtainedtComprises the following steps:
Figure FDA0003283364400000022
on the other hand, a large amount of error label data is inevitable in large-scale sampling data, and the error labels can seriously influence the structure of a decision function and the effect of a corresponding classifier; for this purpose, a soft interval threshold parameter is introduced to control the update step length tautC is less than or equal to C, so that the influence of error label data on the model is limited, and the stability of the classifier is ensured; based on the sample points (x) calculated in step (three)t,yt) 1 change loss ltAnd a step size control parameter C, determining an update step size tautComprises the following steps:
Figure FDA0003283364400000031
6. the step-size-control-based online soft interval kernel learning method as claimed in claim 1, wherein in step (five), the specific operation flow for updating the classifier decision function is: based on the update step τ calculated in step (IV)tUnder the basic framework of the online gradient descent algorithm, a decision function f is settPerform the update
Figure FDA0003283364400000032
Obtaining a new decision function ft；
The step length control-based online soft interval kernel learning algorithm establishes an online kernel learning classifier by introducing a change loss function, a Gaussian kernel function and a soft interval threshold parameter C, so that online prediction of a data stream is realized; the method adopts soft interval threshold parameters to enable the updating of the decision function of the classifier to be smoother and to have robustness; compared with classical online learning algorithms Kernel Perceptron and Pegesoso, the provided algorithm OSKL remarkably improves the classification precision; the online soft interval kernel learning algorithm based on step length control can flexibly process the classification problem under the scene of data flow, and compared with the traditional static classification mode based on batch processing technology, the algorithm greatly reduces the calculation complexity and reduces the model operation time.
CN201811037902.7A 2018-09-06 2018-09-06 Online soft interval kernel learning algorithm based on step length control Active CN109272033B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811037902.7A CN109272033B (en) 2018-09-06 2018-09-06 Online soft interval kernel learning algorithm based on step length control

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811037902.7A CN109272033B (en) 2018-09-06 2018-09-06 Online soft interval kernel learning algorithm based on step length control

Publications (2)

Publication Number Publication Date
CN109272033A CN109272033A (en) 2019-01-25
CN109272033B true CN109272033B (en) 2022-03-08

Family

ID=65187352

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811037902.7A Active CN109272033B (en) 2018-09-06 2018-09-06 Online soft interval kernel learning algorithm based on step length control

Country Status (1)

Country Link
CN (1) CN109272033B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105975907A (en) * 2016-04-27 2016-09-28 江苏华通晟云科技有限公司 SVM model pedestrian detection method based on distributed platform
CN106911580A (en) * 2017-01-24 2017-06-30 中国科学院声学研究所 A kind of TCP jamming control methods and system
CN107247968A (en) * 2017-07-24 2017-10-13 东北林业大学 Based on logistics equipment method for detecting abnormality under nuclear entropy constituent analysis imbalance data
CN107991877A (en) * 2017-12-20 2018-05-04 东南大学 A kind of Dynamic Model Identification method and system based on Recognition with Recurrent Neural Network

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8374442B2 (en) * 2008-11-19 2013-02-12 Nec Laboratories America, Inc. Linear spatial pyramid matching using sparse coding
KR101613971B1 (en) * 2009-12-30 2016-04-21 삼성전자주식회사 Method for transforming program code

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105975907A (en) * 2016-04-27 2016-09-28 江苏华通晟云科技有限公司 SVM model pedestrian detection method based on distributed platform
CN106911580A (en) * 2017-01-24 2017-06-30 中国科学院声学研究所 A kind of TCP jamming control methods and system
CN107247968A (en) * 2017-07-24 2017-10-13 东北林业大学 Based on logistics equipment method for detecting abnormality under nuclear entropy constituent analysis imbalance data
CN107991877A (en) * 2017-12-20 2018-05-04 东南大学 A kind of Dynamic Model Identification method and system based on Recognition with Recurrent Neural Network

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
Continuous query scheduler based on operators clustering;M.Sami Soliman 等;《Journal of Central South University of Technology》;20110331;782-790 *
基于正则化路径的支持向量机近似模型选择;丁立中 等;《计算机研究与发展》;20120615;1248-1255 *

Also Published As

Publication number Publication date
CN109272033A (en) 2019-01-25

Similar Documents

Publication Publication Date Title
CN108510012B (en) Target rapid detection method based on multi-scale feature map
CN111639710B (en) Image recognition model training method, device, equipment and storage medium
WO2020063715A1 (en) Method and system for training binary quantized weight and activation function for deep neural networks
Xiang et al. TUMK-ELM: a fast unsupervised heterogeneous data learning approach
CN109871542B (en) Text knowledge extraction method, device, equipment and storage medium
CN110532555B (en) Language evaluation generation method based on reinforcement learning
CN112364242B (en) Graph convolution recommendation system for context awareness
CN111984791B (en) Attention mechanism-based long text classification method
US11450096B2 (en) Systems and methods for progressive learning for machine-learned models to optimize training speed
CN113077388A (en) Data-augmented deep semi-supervised over-limit learning image classification method and system
CN112529071B (en) Text classification method, system, computer equipment and storage medium
WO2019138897A1 (en) Learning device and method, and program
Han et al. L-Net: lightweight and fast object detector-based ShuffleNetV2
CN111078895A (en) Remote supervision entity relation extraction method based on denoising convolutional neural network
Zhang et al. Improvement of dynamic hand gesture recognition based on HMM algorithm
CN115630164A (en) Remote supervision relation extraction method based on positive and negative direction joint learning and prototype representation
CN115062727A (en) Graph node classification method and system based on multi-order hypergraph convolutional network
CN109409529A (en) A kind of event cognitive analysis method, system and storage medium
CN113326884A (en) Efficient learning method and device for large-scale abnormal graph node representation
CN109272033B (en) Online soft interval kernel learning algorithm based on step length control
CN112668543A (en) Isolated word sign language recognition method based on hand model perception
CN110347853B (en) Image hash code generation method based on recurrent neural network
CN117033961A (en) Multi-mode image-text classification method for context awareness
Wang et al. Q-YOLO: Efficient inference for real-time object detection
CN109359677B (en) Noise-resistant online multi-classification kernel learning algorithm

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
CB03 Change of inventor or designer information

Inventor after: Song Yunquan

Inventor after: Li Yueling

Inventor after: Yu Qi

Inventor after: Lei Hejie

Inventor after: Liang Xijun

Inventor after: Gradual order

Inventor before: Song Yunquan

Inventor before: Lei Hejie

Inventor before: Lv Cong

Inventor before: Liang Xijun

Inventor before: Gradual order

CB03 Change of inventor or designer information
GR01 Patent grant
GR01 Patent grant