CN103886030B - Cost-sensitive decision-making tree based physical information fusion system data classification method - Google Patents
Cost-sensitive decision-making tree based physical information fusion system data classification method Download PDFInfo
- Publication number
- CN103886030B CN103886030B CN201410077291.4A CN201410077291A CN103886030B CN 103886030 B CN103886030 B CN 103886030B CN 201410077291 A CN201410077291 A CN 201410077291A CN 103886030 B CN103886030 B CN 103886030B
- Authority
- CN
- China
- Prior art keywords
- decision tree
- classification
- data
- decision
- test sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/28—Databases characterised by their database models, e.g. relational or object models
- G06F16/284—Relational databases
- G06F16/285—Clustering or classification
Abstract
Disclosed is a cost-sensitive decision-making tree based physical information fusion system data classification method. The cost-sensitive decision-making tree based physical information fusion system data classification method includes: firstly, using physical units for data acquisition; then transmitting acquired data to information units; finally, performing further analyzing and processing in the information units. The information unit is operated in the following steps that 1), the information units respectively establish decision-making trees corresponding to the physical units according to respectively stored sample data; 2), the established decision-making trees are tested, specific value of the number of misclassification to the overall number of tests of the decision-making trees is recorded, and the decision-making trees too large in the specific value are deleted; 3), classifying the data on the basis of the rest decision-making trees and combining specific value of the previously recorded number of misclassification to the overall number of tests. A tree structure can be conveniently and rapidly established to perform efficient classification on the data, which is easy for a user to know about the data, and the user does not need to be with a lot of background knowledge.
Description
Technical field
The present invention relates to a kind of Decision-Tree Method based on cost-sensitive, it is mainly used in solving information physical fusion system
In system, data efficient classification and the loss producing when classifying or cost summation are minimum problem, belong to information physical and merge system
The interleaving techniques application that system data is excavated.
Background technology
Information physical emerging system is considered as the third wave of world information technology after computer, the Internet.
Information physical emerging system can be understood as the high-efficiency network intelligent information system based on embedded device, is to have height certainly
Main perception, autonomous judge, from main regulation and capacity of self-government, be capable of virtual world and real physical world interconnection with collaborative
Intelligence system of future generation.Information physical emerging system functionally mainly considers performance optimization, is that collection calculates
(computation), communication (communication) and control (control) 3c are in the intellectual technology of one.Now, information
Physics emerging system technology has been obtained for international industrial and commercial circles and the highest attention of many large-scale international corporations, development speed pole
For rapid, it has been applied to multiple important development such as traffic, medical treatment, energy field, has had broad application prospects.
Data mining is an iterative process, and it searches valuable, unusual fresh information from substantial amounts of data,
It is people and computer result of the joint efforts；It describes between knowledge and the search capability of computer of problem and target in human expert
Seeking balance, in the hope of obtaining best result.Data mining is one of field with fastest developing speed in computer industry, and it is only in the past
It is one of computer science and statistics theme, nowadays, it has developed rapidly becomes an independent field.Data
Excavate a most powerful advantage and be that it can be the application of many methods and techniques and substantial amounts of problem set.Data mining is one
The individual non-human act carrying out on large data sets, so its maximum target market is whole data warehouse, Data Mart and determines
Plan supports industry, including industries such as retail, manufacture, telecommunications, medical treatment, insurance, transports.
Classification is a kind of important data analysis form, and it extracts the model portraying significant data class.This model is referred to as
Grader, the class label of prediction classification.Classification is generally divided into two steps: the first step, and we set up one based on the data providing
Individual disaggregated model；Second step, determines whether the accuracy rate of this model can accept, if it is then using this model to new
Data is classified.Most sorting algorithm is the algorithm of memory resident, often assumes that data volume very little.With modern technologies
Continuous development, data mining research set up on these working foundations, develop telescopic classification and Predicting Technique, can
Process the big, data of resident disk.Classification have extensive application, including fraud detection, target marketing, performance prediction, manufacture and
Medical diagnosiss etc..
Decision tree is a kind of tree construction of similar flow chart, and it is a kind of typical sorting technique.It enters to data first
Row is processed, and generates readable rule using inductive algorithm and sets up decision tree, then using decision-making, new data is analyzed.This
In matter, decision tree is process data classified by series of rules.In 20 century 70 later stages and 80 years 20th century
For the initial stage, machine learning academy j.rossquinlan develops decision Tree algorithms, the referred to as riffle of iteration
(iterativedichotomiser, id3).The follow-up c4.5 that quinlan had also been proposed id3 later becomes new supervision
Practise the performance comparision benchmark of algorithm.1984, multidigit statistician published
" classificationandregressiontrees " (cart), describes the generation of binary decision tree.Traditional decision tree
Calculate and mostly employ greedy method, and employ the method construct tree construction of dividing and ruling of top-down recurrence.
The prototype of cost-sensitive (cost-sensitivelearning, csl) classification problem is medical diagnosiss problem.At this
In problem, doctor must be weighed for the probability of patient's a variety of medical treatment test, test cost and desired test effect
Weighing apparatus.The learning method of cost-sensitive mainly considers in classification, when different classification errors can lead to different punishment dynamics
How to train grader.For example in medical treatment, " being the cost of Healthy People by patient's mistaken diagnosis " and " Healthy People mistaken diagnosis is patient's
Cost " is different；Usurp in detection in financial credit card, " by usurping the cost being mistakenly considered normally to use " and will " normally use by mistake
Think the cost usurped " also different.It is not difficult to find out, probability erroneous judgement is very little, but if can not correctly judge
One potential erroneous judgement, it will lead to a series of loss, thus more have realistic meaning using final loss as weighing target.
Since the cost study based on cost-sensitive is suggested, cause the concern of a lot of experts it is proposed that a lot
Novel method.As a whole, there are two class basic ideas: a kind of method is exactly the sorting technique not changing classics, and simply right
Original data set makees certain process, or perhaps the one layer of algorithm of outer wrapping in classical sorting algorithm, makes up to pre-
The fixed purpose to cost-sensitive；Another kind of thinking is then on classical taxonomy algorithm basis, adds some other factorses, reaches
Purpose to cost sensitive learning.
Content of the invention
Technical problem: it is an object of the invention to provide a kind of information physical emerging system number based on cost-sensitive decision tree
According to sorting technique, the method increased the consideration of cost-sensitive during decision tree foundation, to solve information physical fusion
Data efficient classification in system, and in classification, the loss of generation or cost summation are down to the problem of minimum.
Technical scheme: the information physical emerging system data classification side based on cost-sensitive decision tree of the present invention
Method, user first builds many decision trees in information physical emerging system, then these decision trees are formed a combination decision tree.
Under the consideration based on cost-sensitive, by the calculating to combination decision tree error rate, data is classified.
Information physical emerging system of the present invention is made up of some physical locations and an information unit, wherein said
Physical location be used for collecting data；Described information unit is used for receiving and store the data of physical location transmission, and to this
It is analyzed and process.
Comprised the following steps based on the information physical emerging system data classification method of cost-sensitive decision tree:
The component units of information physical emerging system are divided into multiple physical locations and an information list by step 1) user
Unit；Described physical location is used for collecting data, and described information unit is used for analysis and processing data；
Step 2) training sample and test sample placed in systems by user in advance, described training sample and test sample
In each sample include catalogue number(Cat.No.), Property Name, corresponding property value and classification；
Step 3) user starts each physical location and collects training sample data；
Each physical location of step 4) sends the training sample collected data to information unit, and information unit is to difference
The training sample data that physical location sends is stored respectively；
Step 5) information unit is respectively these physical locations according to the sample data storing respectively and sets up corresponding decision-making
Tree, described decision tree is the grader of one of machine learning classification algorithm tree, and grader is a kind of computer journey
Sequence, effect is automatically data can be assigned to known class；
The decision tree obtaining is placed in a list step 6) information unit, forms a decision tree list, described
List is according to linear precedence, the set of the data item of arrangement in data structure, can carry out in this data structure
Basic operation include to element lookup, insertion and deletion；
Step 7) information unit is tested to each decision tree.Specifically comprise the following steps that
Step 7.1) one group of ready test sample is input in the decision tree that each establishes；
Step 7.2) test sample can obtain corresponding classification results through the decision-making of decision tree；
Step 7.3) classification of the classification results obtaining and test sample itself to be compared, identical then classification is correct,
Difference then classification error；The sample number of misregistration classification；
Step 7.4) if the test sample number of mistake classification and the ratio of total test sample number are more than 10%, certainly
This decision tree is deleted, if being otherwise retained in decision tree list in plan tree list；Wherein by the test sample of mistake classification
The ratio of number and total test sample number is designated as wi, and in information unit, i represents i-th decision tree in decision tree list to record；
Step 8) is after above-mentioned steps, if decision tree list is sky, chooses that minimum decision tree of wi as
Whole decision tree；
If step 9) decision tree list comprises only a decision tree, this class decision tree is final decision tree；
If step 10) decision tree list contains the decision tree of two or more than two, by determining in decision tree list
The grader that plan tree forms for multiple decision trees as combination decision tree, described combination decision tree, each decision tree has accordingly
Weights, this weights is the test sample number of mistake classification and the ratio of total test sample number；
Step 11) calculates wrong classification rate h of combination decision tree, returns corresponding classification during h value minimum as data
Classification results；Wherein h passes throughCalculate and obtain, i represents i-th decision tree；J represents classification；ci,j
Represent whether the classification results obtaining after i-th decision tree are classification j, if then ci,jIt is worth for 1, if otherwise ci,jIt is worth for 0；
wiThe ratio of the test sample number of mistake classification and total test sample number in i-th decision tree when testing before expression；N represents
In combination decision tree, the classification of classification results is a tree of the decision tree of j, and argmin represents selectionIn minimum
Value is as the value of final h.
Beneficial effect: the present invention, when information physical emerging system data is classified, employs data mining
The decision Tree algorithms of middle sorting algorithm, and introduce process with regard to cost-sensitive to solve data in information physical emerging system
Efficiently classify, and in classification, the loss of generation or cost summation are down to the problem of minimum.Specifically, of the present invention
Following beneficial effect is had based on the information physical emerging system data classification method of cost-sensitive decision tree:
(1) decision tree should be readily appreciated that and realizes, user in learning process not it should be understood that a lot of background knowledges, only
Need by explanation after it is possible to understand that the meaning expressed by decision tree at once.
(2) decision tree can directly embody the feature of data, and tree person easy to use observes and understands.
(3) for decision tree, the preparation of data is often simple, and can process discrete data or continuous simultaneously
Type data, can make feasible and respond well result within the relatively short time to large data source.
(4) decision tree is easy to by test, model be evaluated and tested, can be with rating model credibility.If given observe
Data, then be easy to release corresponding classifying ruless according to produced decision tree.
(5) cost sensitive learning introducing is by training dataset, trains and obtains examining of minimal error classification cost
Disconnected learning system, is finally to pursue mistake classification cost minimization, reducing the probability of mistake classification.So it is no longer
Classical classification and prediction, but the extension of classical taxonomy.
Brief description
Fig. 1 decision tree sets up process schematic,
The information physical emerging system data classification method flow chart based on cost-sensitive decision tree for the Fig. 2.
Specific embodiment
The present invention, using the decision Tree algorithms with reference to cost-sensitive, the classification of optimization information physics emerging system data, subtracts
The loss that few mistake classification brings.For a more detailed description to the present invention below according to drawings and Examples.
The present invention is specifically described according to weather sample data, weather sample data include catalogue number(Cat.No.), Property Name and
Corresponding property value, classification.Wherein attribute includes rain or shine (corresponding property value has sunny, overcast, rainy), temperature
(corresponding property value has hot, mild, cool), humidity (corresponding property value has high, normal), (corresponding attribute of blowing
Value has true, false), classification is play (corresponding classification has yes, no) of whether going out.
1st, the component units of information physical emerging system are divided into multiple physical locations and an information unit by user.Institute
State physical location to be used for collecting weather data；Described information unit is used for analysis and processing data.
2nd, weather training sample and weather test sample are placed in systems by user in advance.Described weather training sample and
Each sample in weather test sample includes catalogue number(Cat.No.), Property Name, corresponding property value, classification.
3rd, user starts each physical location collection weather training sample data.
4th, each physical location sends the weather collected training sample data to information unit.Information unit is to difference
The weather training sample data that physical location sends is stored respectively.
5th, information unit is respectively these physical locations according to the weather sample data storing respectively and sets up corresponding decision-making
Tree.Described decision tree is the grader of one of machine learning classification algorithm tree, wherein root node comprise all of
Input sample data, leafy node is the result of classification, and remaining node represents the test to certain attribute.Grader is a kind of meter
Calculation machine program, effect is automatically data can be assigned to known class.It is as follows that each decision tree specific sets up process:
5.1st, set up a node as the root node of decision tree
5.2nd, training of judgement concentrates whether only one of which classification, if then node is labeled as leafy node, and terminates to determine
The construction process of plan tree；If otherwise whether judgment sample is with the presence of attribute.
If 5.3 samples do not have attribute, it is leafy node by having the most category label of number in training set, and ties
Bundle construction process；If otherwise beginning look for best splitting rule.
5.4th, calculate information gain, that chooses maximum sets up branch, and the deletion from sample data as Split Attribute
This attribute.Described information gain is a kind of criterion, is to see to bring how much information for classification by attribute (information herein is
Refer to the impact for final classification result), the information brought more (impact is bigger), this attribute is more important.Wherein information gain
Formula beD is training set, piRepresent the nonzero probability belonging to class i in training set.Described point
Split attribute to refer to select certain attribute as the condition of decision tree tree branch, if property value is discrete, typically
Comprise the branch of this attribute be possible to property value number；If serial number attribute, then typically judge this numerical value
Or whether more than the constant less than or equal to certain predefined, provide a y-bend division.
5.5th, each branch is carried out with the operation of above-mentioned Step d, and carries out the judgement of b and step c.
6th, the decision tree obtaining is placed in a list information unit, forms a decision tree list.Described list
Can carry out basic in data structure according to linear precedence, the set of the data item of arrangement in this data structure
Operation include to element lookup, insertion, and delete.
7th, information unit is tested to each decision tree.Specifically comprise the following steps that
7.1st, one group of ready weather test sample is input in the decision tree that each establishes；
7.2nd, weather test sample can obtain corresponding classification results through the decision-making of decision tree；
7.3rd, the classification results obtaining and the weather test sample classification of itself are compared, identical then classification is correct,
Difference then classification error；The sample number of misregistration classification；
If the ratio of the weather test sample number of 7.4 mistake classification and total weather test sample number is more than 10%,
This decision tree is deleted in decision tree list；If it is not, being then retained in decision tree list.Wherein by mistake point class testing
The ratio of sample number and total test sample number is designated as wi, in information unit, wherein i represents i-th in decision tree list record
Individual decision tree；
8th, judge whether decision tree list is empty, if then choosing that minimum decision tree of wi is final decision tree.
9 if it is determined that decision tree list is not empty, then continue to judge whether decision tree list comprises only a decision tree,
If then this class decision tree is final decision tree.If it is not, then using the decision tree in decision tree list as combination decision tree.Institute
State the grader that combination decision tree forms for multiple decision trees, each decision tree has corresponding weights (herein for mistake classification
Test sample number and the ratio of total test sample number), data needs the classification results of each decision tree comprehensive can obtain
Classification afterwards.
10th, calculate wrong classification rate h of combination decision tree, return corresponding classification dividing as data during h value minimum
Class result.H passes throughCalculate and obtain, wherein i represents i-th decision tree；J represents classification；ci,jTable
Show whether the classification results obtaining after i-th decision tree are classification j, if then ci,jIt is worth for 1, if otherwise ci,jIt is worth for 0；wi
The ratio of the test sample number of mistake classification and total test sample number in i-th decision tree when testing before expression；N expression group
In conjunction decision tree, the classification of classification results is a tree of the decision tree of j；Argmin represents selectionIn minima
Value as final h.
Claims (1)
1. a kind of information physical emerging system data classification method based on cost-sensitive decision tree is it is characterised in that the method bag
Include following steps:
Step 1) component units of information physical emerging system are divided into multiple physical locations and an information unit by user；Institute
State physical location and be used for collecting data, described information unit is used for analysis and processing data；
Step 2) training sample and test sample placed in systems, in described training sample and test sample by user in advance
Each sample includes catalogue number(Cat.No.), Property Name, corresponding property value and classification；
Step 3) user start each physical location collect training sample data；
Step 4) each physical location sends the training sample collected data to information unit, and information unit is to different physics
The training sample data that unit sends is stored respectively；
Step 5) information unit is respectively these physical locations according to the sample data storing respectively and sets up corresponding decision tree, institute
State the grader that decision tree is one of machine learning classification algorithm tree, grader is a kind of computer program, make
With being automatically data can be assigned to known class；
Step 6) decision tree obtaining is placed in a list information unit, forms a decision tree list, described list
Can carry out basic in data structure according to linear precedence, the set of the data item of arrangement in this data structure
Operation include to element lookup, insertion and deletion；
Step 7) information unit tests to each decision tree, specifically comprises the following steps that
Step 7.1) one group of ready test sample is input in the decision tree that each establishes；
Step 7.2) test sample can obtain corresponding classification results through the decision-making of decision tree；
Step 7.3) classification of the classification results obtaining and test sample itself is compared, identical correct, difference of then classifying
Then classification error；The sample number of misregistration classification；
Step 7.4) if the test sample number of mistake classification and the ratio of total test sample number are more than 10%, in decision tree
This decision tree is deleted, if being otherwise retained in decision tree list in list；Wherein by mistake classification test sample number and
The ratio of total test sample number is designated as wi, in information unit, i represents i-th decision tree in decision tree list to record；
Step 8) after above-mentioned steps, if decision tree list is sky, choose wiThat minimum decision tree is determined as final
Plan tree；
Step 9) if decision tree list comprises only a decision tree, this decision tree is final decision tree；
Step 10) if decision tree list contains the decision tree of more than two, using the decision tree in decision tree list as group
Close decision tree, the grader that described combination decision tree forms for multiple decision trees, each decision tree has corresponding weights, this weights
Test sample number for mistake classification and the ratio of total test sample number；
Step 11) calculate wrong classification rate h of combination decision tree, corresponding classification dividing as data when returning h value minimum
Class result；Wherein h passes throughCalculate and obtain, i represents i-th decision tree；J represents classification；ci,jRepresent
Whether the classification results being obtained after i-th decision tree are classification j, if then ci,jIt is worth for 1, if otherwise ci,jIt is worth for 0；wiTable
The ratio of the test sample number of mistake classification and total test sample number in i-th decision tree during test before showing；N represents combination
In decision tree, the classification of classification results is a tree of the decision tree of j, and argmin represents selectionIn minima make
Value for final h.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410077291.4A CN103886030B (en) | 2014-03-05 | 2014-03-05 | Cost-sensitive decision-making tree based physical information fusion system data classification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410077291.4A CN103886030B (en) | 2014-03-05 | 2014-03-05 | Cost-sensitive decision-making tree based physical information fusion system data classification method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103886030A CN103886030A (en) | 2014-06-25 |
| CN103886030B true CN103886030B (en) | 2017-01-25 |
Family
ID=50954922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410077291.4A Active CN103886030B (en) | 2014-03-05 | 2014-03-05 | Cost-sensitive decision-making tree based physical information fusion system data classification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103886030B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104657610B (en) * | 2015-02-13 | 2017-11-17 | 南京邮电大学 | A kind of information physical emerging system sequential logic robustness appraisal procedure |
| CN104778250B (en) * | 2015-04-14 | 2017-09-29 | 南京邮电大学 | Information physical emerging system data classification method based on genetic planning decision tree |
| CN104915608B (en) * | 2015-05-08 | 2018-07-27 | 南京邮电大学 | A kind of information physical emerging system secret protection type data classification method |
| CN106611181A (en) * | 2016-05-30 | 2017-05-03 | 四川用联信息技术有限公司 | Method for constructing cost-sensitive two-dimensional decision tree |
| CN108763221B (en) * | 2018-06-20 | 2022-05-17 | 科大讯飞股份有限公司 | Attribute name representation method and device |
| CN109358608A (en) * | 2018-11-28 | 2019-02-19 | 广东电网有限责任公司 | A kind of transformer state methods of risk assessment and device based on integrated study |
| CN110389952B (en) * | 2019-06-06 | 2021-09-07 | 口碑(上海)信息技术有限公司 | Dish data processing method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2012108889A (en) * | 2010-10-22 | 2012-06-07 | Intec Inc | Distributed database system and distributed data processing system |
| CN102999582A (en) * | 2012-11-15 | 2013-03-27 | 南京邮电大学 | Lightweight rule-based WoT (Web of Things) monitoring system |
| CN103092918A (en) * | 2012-12-21 | 2013-05-08 | 东莞中国科学院云计算产业技术创新与育成中心 | Method for construction and query of knowledge base of emergency aid decision-making system based on location information |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2001325276A (en) * | 2000-05-12 | 2001-11-22 | Internatl Business Mach Corp <Ibm> | Information processing system, information processing method, computer program, storage medium and program transmitter |
-
2014
- 2014-03-05 CN CN201410077291.4A patent/CN103886030B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2012108889A (en) * | 2010-10-22 | 2012-06-07 | Intec Inc | Distributed database system and distributed data processing system |
| CN102999582A (en) * | 2012-11-15 | 2013-03-27 | 南京邮电大学 | Lightweight rule-based WoT (Web of Things) monitoring system |
| CN103092918A (en) * | 2012-12-21 | 2013-05-08 | 东莞中国科学院云计算产业技术创新与育成中心 | Method for construction and query of knowledge base of emergency aid decision-making system based on location information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103886030A (en) | 2014-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103886030B (en) | Cost-sensitive decision-making tree based physical information fusion system data classification method | |
| CN110597735B (en) | Software defect prediction method for open-source software defect feature deep learning | |
| CN104992184B (en) | A kind of multiclass image classification method based on semi-supervised extreme learning machine | |
| CN104573669A (en) | Image object detection method | |
| CN110188209A (en) | Cross-module state Hash model building method, searching method and device based on level label | |
| CN103412888B (en) | A kind of point of interest recognition methods and device | |
| CN108549954A (en) | Risk model training method, risk identification method, device, equipment and medium | |
| CN107506793A (en) | Clothes recognition methods and system based on weak mark image | |
| CN106897821A (en) | A kind of transient state assesses feature selection approach and device | |
| CN111008337B (en) | Deep attention rumor identification method and device based on ternary characteristics | |
| CN106874963B (en) | A kind of Fault Diagnosis Method for Distribution Networks and system based on big data technology | |
| CN111754345A (en) | Bit currency address classification method based on improved random forest | |
| CN107392241A (en) | A kind of image object sorting technique that sampling XGBoost is arranged based on weighting | |
| CN105654196A (en) | Adaptive load prediction selection method based on electric power big data | |
| CN111914162B (en) | Method for guiding personalized learning scheme based on knowledge graph | |
| CN111292195A (en) | Risk account identification method and device | |
| CN107918657A (en) | The matching process and device of a kind of data source | |
| CN112732921B (en) | False user comment detection method and system | |
| CN106649266A (en) | Logical inference method for ontology knowledge | |
| CN106407482B (en) | A kind of network academic report category method based on multi-feature fusion | |
| CN107977670A (en) | Accident classification stage division, the apparatus and system of decision tree and bayesian algorithm | |
| CN108762503A (en) | A kind of man-machine interactive system based on multi-modal data acquisition | |
| CN112756759A (en) | Spot welding robot workstation fault judgment method | |
| CN104850862B (en) | A kind of sorting technique based on unit cost income sensitivity decision tree | |
| CN107423697A (en) | Activity recognition method based on non-linear fusion depth 3D convolution description |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20180705 Address after: 210003 A, block 1502, Nanjing University of Finances and Economics science and Technology Park, 128 North Railway Street, Nanjing, Jiangsu. Patentee after: Nanjing Hengtong Mdt InfoTech Ltd Address before: 210023 9 Wen Yuan Road, Ya Dong new town, Nanjing, Jiangsu. Patentee before: Nanjing Post & Telecommunication Univ. |