CN104883278A - Method for classifying network equipment by utilizing machine learning - Google Patents
Method for classifying network equipment by utilizing machine learning Download PDFInfo
- Publication number
- CN104883278A CN104883278A CN201410510787.6A CN201410510787A CN104883278A CN 104883278 A CN104883278 A CN 104883278A CN 201410510787 A CN201410510787 A CN 201410510787A CN 104883278 A CN104883278 A CN 104883278A
- Authority
- CN
- China
- Prior art keywords
- equipment
- packet
- machine learning
- disaggregated model
- characteristic value
- 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
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a method for classifying network equipment by utilizing machine learning. The method establishes a classifying model for the network equipment by utilizing a supervised machine learning technology or a non-supervised machine learning technology, so as to classify known or newly added equipment in a network. The method has the technical advantages of analyzing behavior of the equipment through data packets transmitted and received by industrial control network equipment, and further classifying the equipment automatically. The understanding of types of the equipment is conducive to auditing network communication data. Unlike the traditional security specification for individual equipment, the security specification for equipment types is simpler and easy to manage, and the user can define corresponding security specifications for newly added equipment nodes easily.
Description
Technical field
The present invention relates to information security of computer network field, more specifically, relate to a kind of method utilizing machine learning to classify to the network equipment.
Background technology
In a network, different equipment has different functions.Such as programmable logic controller (PLC) is in order to control industrial control equipment, ftp server for file data storage and transmission.From network security angle, the classification understanding equipment is conducive to auditing to network communication data, is the communication setting specification between distinct device.FTP packet such as between engineer work station to programmable logic controller (PLC) is considered to illegal, may represent that invalid data leaks.From user perspective, a safety regulation can be defined: [engineer work station]->[programmable logic controller (PLC)]: FTP reports to the police.That is safety regulation is not aimed at certain several specific equipment, and is aimed at some device class.After new equipment enters user network, as long as know its classification, existing safety regulation just can be used and without the need to formulating/adding new safety regulation for new equipment.
In some cases, by the information such as manufacturer, model of equipment, user roughly can understand the classification of equipment.But such information often can not meet the needs of network security.Such as same ftp server, also must be not quite similar towards different user (intra-company, company is outside) its safety standards needed.Trace sth. to its source, the classification of equipment should not determined by equipment software and hardware, and should be determined by the behavior of equipment.Even if the manufacturer of equipment is different, if its network behavior is similar, same classification can be classified as.Otherwise its network behavior of same equipment is different, also need point different classes of.
Summary of the invention
In order to overcome the deficiencies in the prior art, the object of the invention is to classify to it according to network equipment behavior, object of the present invention is achieved through the following technical solutions:
A kind of method utilizing machine learning to classify to the network equipment, first the packet that collecting device is relevant, and be the characteristic value with a series of attribute that can quantize by its summary, then machine learning method judgment device classification is utilized according to characteristic value, wherein machine learning method comprises, supervised machine learning techniques or non-supervisory formula machine learning techniques, described supervised machine learning refers to that the classification and network behavior thereof that utilize known device are as training data train classification models; Described non-supervisory formula machine learning is mainly used in when user is to the classification uncertain condition of equipment, and it is based on device-dependent network packet, and equipment similar for behavior is also classified as a class by the similarity degree between computing equipment behavior.
Preferably, described supervised machine-learning process comprises:
Step one, exercise equipment disaggregated model
(a1) collect packet computation of characteristic values, collecting and be recorded in the packet that in certain time period, multiple equipment of user receive or send, for each equipment, is a series of attributes that can quantize by device-dependent packet summary;
(a2) sorting machine study, using different sorting machine learning algorithms as plug-in unit to set up disaggregated model;
Step 2, use device class model
(a3) collect new equipment related data packets computation of characteristic values, systematic collection and be recorded in certain time period interior packet relevant to new equipment, and utilize characteristic value computational algorithm to be a series of attributes that can quantize by its summary;
(a4) disaggregated model is used.In the disaggregated model that described characteristic value substitution step one is generated, predict device classification.
More preferably, in step (a1), for each equipment, the packet attribute that can quantize comprises following information:
All packets sum, all communicate nodes, distribution with each node communication amount, the agreement sum of packet, the agreement mainly used, unit type, device fabrication manufacturer;
If user does not have abundant equipment---enough representative data can not be generated, the technology can sampled by bootstrapping---certain device-dependent packet is sampled, uses the multiple characteristic value of same set of data genaration.
In step (a2), the construction step of disaggregated model is as follows:
1) characteristic value of known class equipment is utilized to build disaggregated model
2) disaggregated model generated stores in systems in which, waits for called.
In step (a3), new equipment related data packets can the attribute type that uses with exercise equipment disaggregated model of quantified property consistent.
More preferably, described disaggregated model is decision forest, and each decision forest is made up of several decision trees, and the structure of each decision tree is based on the subset of all training datas, and thus each tree is not identical, and final disaggregated model is defined jointly by all trees.
The building process of described decision forest is:
I) first utilize the technology of bootstrapping sampling to sample to training data, using each sampling as training data, build a decision tree, in building process, can sample to characteristic value;
Ii) repeat i) to build multiple decision tree;
Iii) decision forest is formed by all decision trees.
Each decision tree can predict the classification of new equipment, and final predicting the outcome is chosen in a vote by all decision trees, and the distribution of ballot defines the confidential interval of prediction.
Preferably, described non-supervisory formula machine-learning process comprises:
(b1) collect packet computation of characteristic values, collecting and be recorded in the packet that in certain time period, multiple equipment of user receive or send, will be a series of attributes that can quantize with each device-dependent packet summary;
(b2) use cluster machine learning algorithm, equipment is classified;
Described clustering algorithm is the Unsupervised Clustering algorithm based on K-means, described k-means clustering algorithm is a kind of clustering algorithm of feature based value similarity, the result of k-means is relevant with the center position of initial setting, and different initial central points can be used repeatedly to run clustering algorithm.
The technical advantage of method of the present invention is the behavior of the data packet analysis equipment received and dispatched by the network equipment, and then to equipment automatic classification.The classification of understanding equipment is conducive to auditing to network communication data.Different with traditional safety standard for individual equipment, the safety standard for device class is more simple, is easy to management.User is easy to define corresponding safety regulation to the device node newly added simultaneously.
Accompanying drawing explanation
Fig. 1 is the flow chart that the present invention utilizes supervised machine learning and classifies to the network equipment;
Fig. 2 is the flow chart that the present invention utilizes non-supervisory formula machine learning and classifies to the network equipment.
Embodiment
In order to make object of the present invention, technical scheme and advantage clearly understand, below in conjunction with drawings and Examples, the present invention is further elaborated.Should be appreciated that specific embodiment described herein only in order to explain the present invention, be not intended to limit the present invention.
The flow chart utilizing supervised machine learning to classify to the network equipment as shown in Figure 1, the method comprises exercise equipment disaggregated model and uses the large step of disaggregated model two.
The process of exercise equipment disaggregated model as shown in Figure 1, comprises the steps:
1) collect packet computation of characteristic values, collect and be recorded in the packet that in certain time period, multiple equipment of user receive or send.For each equipment, device-dependent packet summary is a series of attributes that can quantize by system, and these attributes include but not limited to following information:
All packet sums
All nodes communicated
With the distribution of each node communication amount
The agreement sum of packet
The agreement of main use
Unit type
Device fabrication manufacturer
If user does not have abundant equipment namely can not generate enough representative data, then by the technology of bootstrapping sampling, certain device-dependent packet can be sampled, uses same set of packet to generate multiple characteristic value.
2) sorting machine study, the plurality of devices characteristic value of calculating and the basis of plurality of devices classification carry out sorting machine study, use different sorting machine learning algorithms as plug-in unit to set up disaggregated model, store the disaggregated model generated, wait for called.
Below for decision forest, the foundation of disaggregated model is described.
Each decision forest is made up of several decision trees, and the structure of each decision tree is based on the subset of all training datas, and thus each tree is not identical, and final disaggregated model is defined jointly by all trees.
The building course of decision forest is as follows:
The technology of bootstrapping sampling is utilized to sample to training data;
Using each sampling as training data, build a decision tree;
In building process, can sample to characteristic value;
Decision forest is generated by all decision trees.
The process using device class model as shown in Figure 1, comprises the steps:
1) collect new equipment related data packets computation of characteristic values, collect and be recorded in certain time period with certain device-dependent packet, and utilize characteristic value computational algorithm to be a series of attributes that can quantize by its summary;
2) in the disaggregated model characteristic value of above-mentioned calculating substitution step one generated, model prediction device class.If such as use decision forest disaggregated model, each decision tree can predict the classification of new equipment, and final predicting the outcome is chosen in a vote by all decision trees, and the distribution of ballot defines the confidential interval of prediction, such as 90%[programmable logic controller (PLC)], 10%[FTP server].
The flow chart utilizing non-supervisory formula and clustering algorithm machine learning to classify to the network equipment as shown in Figure 2, the method comprises the steps:
1) collect packet computation of characteristic values, collect and be recorded in the packet that in certain time period, multiple equipment of user receive or send.For each equipment, device-dependent packet summary is a series of attributes that can quantize by system;
2) use cluster machine learning to clustering devices analysis, use the Unsupervised Clustering machine learning algorithm based on k-means, apparatus characteristic value is classified, because the result of k-means is relevant with the center position of initial setting, different initial central points can be used repeatedly to run clustering algorithm.Meanwhile, in order to select best categorical measure, different categorical measures can be selected, calculation criterion functional value.According to the definition of criterion function, categorical measure is more, and functional value is lower.By finding the flex point of this function, determine best categorical measure.User also can manual definition initial center point and categorical measure.
3) user determines, after user obtains cluster result, can adjust the result of cluster, can name for classification simultaneously.
4) can also as required, by cluster result apparatus for establishing disaggregated model
From the result of clustering algorithm, can apparatus for establishing disaggregated model further.This model can predict the type of new interpolation equipment, or predicts known device and obtain confidential interval.
More than describe the preferred embodiments of the present invention in detail, but the present invention is not limited to these embodiments, can various change be carried out in application range of the present invention.Although just illustrate the preferred embodiments of the present invention above, in the very clear scope substantially not departing from novelty of the present invention and advantage of person of ordinary skill in the field, various amendment can be carried out to exemplary embodiment.
Claims (10)
1. the method utilizing machine learning to classify to the network equipment, it is characterized in that, first the packet that collecting device is relevant, and be the characteristic value with a series of attribute that can quantize by its summary, then machine learning method judgment device classification is utilized according to characteristic value, wherein machine learning method comprises, supervised machine learning techniques or non-supervisory formula machine learning techniques
Described supervised machine learning refers to and utilizes the classification of known device and network behavior thereof as training data train classification models and use this disaggregated model to classify to the new equipment added;
Described non-supervisory formula machine learning is mainly used in when the uncertain situation of the classification of user to equipment, and it is based on device-dependent network packet, and equipment similar for behavior is also classified as a class by the similarity degree between computing equipment behavior.
2. method according to claim 1, is characterized in that, described supervised machine-learning process comprises:
Step one, exercise equipment disaggregated model
(a1) collect packet computation of characteristic values, collecting and be recorded in the packet that in certain time period, multiple equipment of user receive or send, for each equipment, is a series of attributes that can quantize by device-dependent packet summary;
(a2) sorting machine study, using different sorting machine learning algorithms as plug-in unit to set up disaggregated model;
Step 2, use device class model
(a3) collect new equipment related data packets computation of characteristic values, systematic collection and be recorded in certain time period interior packet relevant to new equipment, and utilize characteristic value computational algorithm to be a series of attributes that can quantize by its summary;
(a4) disaggregated model is used.In the disaggregated model that described characteristic value substitution step one is generated, predict device classification.
3. method according to claim 2, is characterized in that, in step (a1), for each equipment, the packet attribute that can quantize comprises following information:
All packets sum, all communicate nodes, distribution with each node communication amount, the agreement sum of packet, the agreement mainly used, unit type, device fabrication manufacturer;
If user does not have abundant equipment---enough representative data can not be generated, the technology can sampled by bootstrapping---certain device-dependent packet is sampled, uses the multiple characteristic value of same set of data genaration.
4. method according to claim 2, is characterized in that, in step (a2), the construction step of disaggregated model is as follows:
1) characteristic value of known class equipment is utilized to build disaggregated model
2) disaggregated model generated stores in systems in which, waits for called.
5. method according to claim 2, is characterized in that, in step (a3), new equipment related data packets can the attribute type that uses with exercise equipment disaggregated model of quantified property consistent.
6. method according to claim 2, it is characterized in that, described disaggregated model is decision forest, each decision forest is made up of several decision trees, the structure of each decision tree is based on the subset of all training datas, thus each tree is not identical, and final disaggregated model is defined jointly by all trees.
7. method according to claim 6, is characterized in that, the building process of described decision forest is:
I) first utilize the technology of bootstrapping sampling to sample to training data, using each sampling as training data, build a decision tree, in building process, can sample to characteristic value;
Ii) repeat i) to build multiple decision tree;
Iii) decision forest is formed by all decision trees.
8. method according to claim 7, is characterized in that, each decision tree can predict the classification of new equipment, and final predicting the outcome is chosen in a vote by all decision trees, and the distribution of ballot defines the confidential interval of prediction.
9. method according to claim 1, is characterized in that, described non-supervisory formula machine-learning process comprises:
(b1) collect packet computation of characteristic values, collecting and be recorded in the packet that in certain time period, multiple equipment of user receive or send, will be a series of attributes that can quantize with each device-dependent packet summary;
(b2) use cluster machine learning algorithm, equipment is classified.
10. method according to claim 9, it is characterized in that, described clustering algorithm is the Unsupervised Clustering algorithm based on K-means, described k-means clustering algorithm is a kind of clustering algorithm of feature based value similarity, the result of k-means is relevant with the center position of initial setting, and different initial central points can be used repeatedly to run clustering algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410510787.6A CN104883278A (en) | 2014-09-28 | 2014-09-28 | Method for classifying network equipment by utilizing machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410510787.6A CN104883278A (en) | 2014-09-28 | 2014-09-28 | Method for classifying network equipment by utilizing machine learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104883278A true CN104883278A (en) | 2015-09-02 |
Family
ID=53950618
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410510787.6A Pending CN104883278A (en) | 2014-09-28 | 2014-09-28 | Method for classifying network equipment by utilizing machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104883278A (en) |
Cited By (23)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105512285A (en) * | 2015-12-07 | 2016-04-20 | 南京大学 | Self-adaption web crawler method based on machine learning |
CN106502234A (en) * | 2016-10-17 | 2017-03-15 | 重庆邮电大学 | Industrial control system method for detecting abnormality based on double skeleton patterns |
CN106570014A (en) * | 2015-10-09 | 2017-04-19 | 阿里巴巴集团控股有限公司 | Method and device for determining home attribute information of user |
CN106714225A (en) * | 2016-12-29 | 2017-05-24 | 北京酷云互动科技有限公司 | Method and system for identifying network device and intelligent terminal |
CN107770049A (en) * | 2017-10-23 | 2018-03-06 | 林楚莲 | A kind of invited user obtains the method and system of group information |
CN108038374A (en) * | 2017-12-26 | 2018-05-15 | 郑州云海信息技术有限公司 | It is a kind of to detect the method threatened in real time |
CN108156018A (en) * | 2017-11-27 | 2018-06-12 | 上海观安信息技术股份有限公司 | Electric power networks equipment topology identification method, electronic equipment and computer storage media |
WO2018160136A1 (en) * | 2017-03-02 | 2018-09-07 | Singapore University Of Technology And Design | Method and apparatus for determining an identity of an unknown internet-of-things (iot) device in a communication network |
CN108700872A (en) * | 2016-02-29 | 2018-10-23 | 三菱电机株式会社 | Machine sort device |
CN109067778A (en) * | 2018-09-18 | 2018-12-21 | 东北大学 | A kind of industry control scanner fingerprint identification method based on sweet network data |
CN109167798A (en) * | 2018-11-01 | 2019-01-08 | 四川长虹电器股份有限公司 | A kind of household internet of things equipment DDoS detection method based on machine learning |
US10218726B2 (en) | 2016-03-25 | 2019-02-26 | Cisco Technology, Inc. | Dynamic device clustering using device profile information |
CN110019809A (en) * | 2018-01-02 | 2019-07-16 | ***通信有限公司研究院 | A kind of classification determines method, apparatus and the network equipment |
CN110034966A (en) * | 2018-01-11 | 2019-07-19 | 中国科学院声学研究所 | A kind of method for classifying data stream and system based on machine learning |
US10404727B2 (en) | 2016-03-25 | 2019-09-03 | Cisco Technology, Inc. | Self organizing learning topologies |
US10454776B2 (en) | 2017-04-20 | 2019-10-22 | Cisco Technologies, Inc. | Dynamic computer network classification using machine learning |
WO2020022953A1 (en) * | 2018-07-26 | 2020-01-30 | Singapore Telecommunications Limited | System and method for identifying an internet of things (iot) device based on a distributed fingerprinting solution |
CN111385818A (en) * | 2018-12-28 | 2020-07-07 | ***通信集团重庆有限公司 | Method, device and equipment for optimizing wireless network parameters |
CN112381559A (en) * | 2020-10-14 | 2021-02-19 | 浪潮软件股份有限公司 | Tobacco retailer segmentation method based on unsupervised machine learning algorithm |
WO2021051536A1 (en) * | 2019-09-18 | 2021-03-25 | 平安科技(深圳)有限公司 | Security warning method, related device, and computer-readable storage medium |
CN112602304A (en) * | 2018-08-21 | 2021-04-02 | 慧与发展有限责任合伙企业 | Identifying device types based on behavioral attributes |
CN112607240A (en) * | 2019-10-05 | 2021-04-06 | 山西集智数据服务有限公司 | Intelligent garbage classification throwing device |
WO2022156492A1 (en) * | 2021-01-20 | 2022-07-28 | 华为技术有限公司 | Method for determining type of terminal device and related device |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102271090A (en) * | 2011-09-06 | 2011-12-07 | 电子科技大学 | Transport-layer-characteristic-based traffic classification method and device |
CN102571486A (en) * | 2011-12-14 | 2012-07-11 | 上海交通大学 | Traffic identification method based on bag of word (BOW) model and statistic features |
CN103020582A (en) * | 2012-09-20 | 2013-04-03 | 苏州两江科技有限公司 | Method for computer to identify vehicle type by video image |
US20140053261A1 (en) * | 2012-08-15 | 2014-02-20 | Qualcomm Incorporated | On-Line Behavioral Analysis Engine in Mobile Device with Multiple Analyzer Model Providers |
-
2014
- 2014-09-28 CN CN201410510787.6A patent/CN104883278A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102271090A (en) * | 2011-09-06 | 2011-12-07 | 电子科技大学 | Transport-layer-characteristic-based traffic classification method and device |
CN102571486A (en) * | 2011-12-14 | 2012-07-11 | 上海交通大学 | Traffic identification method based on bag of word (BOW) model and statistic features |
US20140053261A1 (en) * | 2012-08-15 | 2014-02-20 | Qualcomm Incorporated | On-Line Behavioral Analysis Engine in Mobile Device with Multiple Analyzer Model Providers |
CN103020582A (en) * | 2012-09-20 | 2013-04-03 | 苏州两江科技有限公司 | Method for computer to identify vehicle type by video image |
Non-Patent Citations (1)
Title |
---|
王涛,余顺争: ""基于机器学习的网络流量分类研究进展"", 《小型微型计算机***》 * |
Cited By (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106570014A (en) * | 2015-10-09 | 2017-04-19 | 阿里巴巴集团控股有限公司 | Method and device for determining home attribute information of user |
CN105512285B (en) * | 2015-12-07 | 2018-11-06 | 南京大学 | Adaptive network reptile method based on machine learning |
CN105512285A (en) * | 2015-12-07 | 2016-04-20 | 南京大学 | Self-adaption web crawler method based on machine learning |
CN108700872B (en) * | 2016-02-29 | 2021-08-06 | 三菱电机株式会社 | Machine sorting device |
CN108700872A (en) * | 2016-02-29 | 2018-10-23 | 三菱电机株式会社 | Machine sort device |
US11240259B2 (en) | 2016-03-25 | 2022-02-01 | Cisco Technology, Inc. | Self organizing learning topologies |
US10404727B2 (en) | 2016-03-25 | 2019-09-03 | Cisco Technology, Inc. | Self organizing learning topologies |
US10218726B2 (en) | 2016-03-25 | 2019-02-26 | Cisco Technology, Inc. | Dynamic device clustering using device profile information |
CN106502234B (en) * | 2016-10-17 | 2018-11-23 | 重庆邮电大学 | Industrial control system method for detecting abnormality based on double skeleton patterns |
CN106502234A (en) * | 2016-10-17 | 2017-03-15 | 重庆邮电大学 | Industrial control system method for detecting abnormality based on double skeleton patterns |
CN106714225A (en) * | 2016-12-29 | 2017-05-24 | 北京酷云互动科技有限公司 | Method and system for identifying network device and intelligent terminal |
WO2018160136A1 (en) * | 2017-03-02 | 2018-09-07 | Singapore University Of Technology And Design | Method and apparatus for determining an identity of an unknown internet-of-things (iot) device in a communication network |
US10454776B2 (en) | 2017-04-20 | 2019-10-22 | Cisco Technologies, Inc. | Dynamic computer network classification using machine learning |
CN107770049A (en) * | 2017-10-23 | 2018-03-06 | 林楚莲 | A kind of invited user obtains the method and system of group information |
CN108156018A (en) * | 2017-11-27 | 2018-06-12 | 上海观安信息技术股份有限公司 | Electric power networks equipment topology identification method, electronic equipment and computer storage media |
CN108038374A (en) * | 2017-12-26 | 2018-05-15 | 郑州云海信息技术有限公司 | It is a kind of to detect the method threatened in real time |
CN110019809A (en) * | 2018-01-02 | 2019-07-16 | ***通信有限公司研究院 | A kind of classification determines method, apparatus and the network equipment |
CN110019809B (en) * | 2018-01-02 | 2021-11-19 | ***通信有限公司研究院 | Classification determination method and device and network equipment |
CN110034966A (en) * | 2018-01-11 | 2019-07-19 | 中国科学院声学研究所 | A kind of method for classifying data stream and system based on machine learning |
WO2020022953A1 (en) * | 2018-07-26 | 2020-01-30 | Singapore Telecommunications Limited | System and method for identifying an internet of things (iot) device based on a distributed fingerprinting solution |
CN112602304A (en) * | 2018-08-21 | 2021-04-02 | 慧与发展有限责任合伙企业 | Identifying device types based on behavioral attributes |
CN109067778B (en) * | 2018-09-18 | 2020-07-24 | 东北大学 | Industrial control scanner fingerprint identification method based on honeynet data |
CN109067778A (en) * | 2018-09-18 | 2018-12-21 | 东北大学 | A kind of industry control scanner fingerprint identification method based on sweet network data |
CN109167798A (en) * | 2018-11-01 | 2019-01-08 | 四川长虹电器股份有限公司 | A kind of household internet of things equipment DDoS detection method based on machine learning |
CN109167798B (en) * | 2018-11-01 | 2020-03-17 | 四川长虹电器股份有限公司 | Household Internet of things device DDoS detection method based on machine learning |
CN111385818A (en) * | 2018-12-28 | 2020-07-07 | ***通信集团重庆有限公司 | Method, device and equipment for optimizing wireless network parameters |
CN111385818B (en) * | 2018-12-28 | 2023-04-14 | ***通信集团重庆有限公司 | Method, device and equipment for optimizing wireless network parameters |
WO2021051536A1 (en) * | 2019-09-18 | 2021-03-25 | 平安科技(深圳)有限公司 | Security warning method, related device, and computer-readable storage medium |
CN112607240A (en) * | 2019-10-05 | 2021-04-06 | 山西集智数据服务有限公司 | Intelligent garbage classification throwing device |
CN112381559A (en) * | 2020-10-14 | 2021-02-19 | 浪潮软件股份有限公司 | Tobacco retailer segmentation method based on unsupervised machine learning algorithm |
WO2022156492A1 (en) * | 2021-01-20 | 2022-07-28 | 华为技术有限公司 | Method for determining type of terminal device and related device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104883278A (en) | Method for classifying network equipment by utilizing machine learning | |
CN106104398B (en) | Distributed big data in Process Control System | |
CN107181724B (en) | Identification method and system of cooperative flow and server using method | |
CN102647292B (en) | Intrusion detecting method based on semi-supervised neural network | |
JP2021527288A (en) | Detection of machine learning model suitability for datasets | |
CN109873779B (en) | LSTM-based hierarchical wireless signal modulation type identification method | |
CN103870751A (en) | Method and system for intrusion detection | |
CN107038167A (en) | Big data excavating analysis system and its analysis method based on model evaluation | |
JP2019513246A (en) | Training method of random forest model, electronic device and storage medium | |
CN111176953B (en) | Abnormality detection and model training method, computer equipment and storage medium | |
US7716152B2 (en) | Use of sequential nearest neighbor clustering for instance selection in machine condition monitoring | |
CN103532949A (en) | Self-adaptive trojan communication behavior detection method on basis of dynamic feedback | |
CN110365503A (en) | A kind of Index and its relevant device | |
WO2022247955A1 (en) | Abnormal account identification method, apparatus and device, and storage medium | |
CN103258027A (en) | Context awareness service platform based on intelligent terminal | |
CN111770190A (en) | Block chain distributed real-time fire fighting monitoring alarm system based on 5G and Internet of things mode | |
CN110851422A (en) | Data anomaly monitoring model construction method based on machine learning | |
CN112860676B (en) | Data cleaning method applied to big data mining and business analysis and cloud server | |
CN113556768A (en) | Sensor data anomaly detection method and system | |
CN109948650A (en) | A kind of smart home device type decision method based on message characteristic | |
CN108156018A (en) | Electric power networks equipment topology identification method, electronic equipment and computer storage media | |
CN113704389A (en) | Data evaluation method and device, computer equipment and storage medium | |
CN105446242A (en) | Intelligent equipment automation identification system | |
Sestito et al. | A general optimization-based approach to the detection of real-time Ethernet traffic events | |
CN111368060B (en) | Self-learning method, device and system for conversation robot, electronic equipment and medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
EXSB | Decision made by sipo to initiate substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150902 |
|
RJ01 | Rejection of invention patent application after publication |