CN107220671A - A kind of Artificial Olfactory on-line correction sample generating method based on self organization map - Google Patents
A kind of Artificial Olfactory on-line correction sample generating method based on self organization map Download PDFInfo
- Publication number
- CN107220671A CN107220671A CN201710393408.3A CN201710393408A CN107220671A CN 107220671 A CN107220671 A CN 107220671A CN 201710393408 A CN201710393408 A CN 201710393408A CN 107220671 A CN107220671 A CN 107220671A
- Authority
- CN
- China
- Prior art keywords
- mrow
- mover
- msubsup
- msup
- neuron
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0001—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00 by organoleptic means
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24143—Distances to neighbourhood prototypes, e.g. restricted Coulomb energy networks [RCEN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
Abstract
The invention belongs to fire detection field, a kind of Artificial Olfactory on-line correction sample generating method based on self organization map, including initial training stage and online updating stage are disclosed；Initial training stage builds self-organizing organization chart neutral net sandwich construction with sample class number, carries out self organization map neutral net neuron weight initialization using training sample and regard self organization map neutral net neuron weights as initial training sample set；In the online updating stage, according to the classification results of subsequent classifier, it is adjusted with test sample localized region neuron weights.Self organization map neutral net neuron weights now are subjected to on-line correction as on-line training sample set to mode identification method.As a result show that the present invention under the conditions of working online, can improve the anti-long term drift ability of Artificial Olfactory；Calibration samples are automatically generated during can working online, realize that automatic online correction provides guarantee for Artificial Olfactory.
Description
Technical field
The invention belongs to fire detection field, more particularly to a kind of Artificial Olfactory based on self organization map are online
Calibration samples generation method.
Background technology
Artificial Olfactory is a kind of novel smell analysis means, with detect it is quick, noninvasive, easy to operate, low into
This advantages of.Artificial Olfactory is broadly divided into gas sensor array and mode identification method two parts, wherein " gas sensing
Device array " is using the inexpensive gas sensor with cross-sensitivity, to obtain smell collection of illustrative plates；" pattern-recognition " then uses people
The methods such as work intelligence, machine learning carry out qualitative and quantitative analysis to smell.Gas sensor " long term drift " is Artificial Olfactory
The unavoidable problem of system, with the extension of use time, the smell collection of illustrative plates of " gas sensor array " can occur it is slow and
Irregular change；So that the recognition correct rate of " pattern-recognition " method progressively declines with the time, most become untrustworthy at last；
The data acquisition that specific sample is periodically carried out to Artificial Olfactory is usually taken, " pattern knowledge is carried out according to the data collected
" algorithm is not corrected, and the smell collection of illustrative plates after making algorithm and drifting about is matched.But conventional correction procedure needs to stop manually
The normal work of olfactory system is, it is necessary to configure and buy special test sample, in addition it is also necessary to which professional operator is regulated and controled,
Integrated cost is higher.It is not feasible for the system that dismounting is mobile constant and needs are monitored on-line.
In summary, the problem of prior art is present be：1) for needing the equipment worked online for a long time, it is impossible to carry out normal
Rule correction, therefore system identification accuracy rate can be remarkably decreased with the working time；2) conventional correction procedure needs to prepare specific correction sample
Can not normal work, the substantial amounts of financial resources of waste, man power and material during this and equipment calibration.
The content of the invention
The problem of existing for prior art, the invention provides a kind of Artificial Olfactory based on self organization map is online
Calibration samples generation method.
The present invention is achieved in that a kind of Artificial Olfactory on-line correction sample generation side based on self organization map
Method, the Artificial Olfactory on-line correction sample generating method based on self organization map include initial training stage and it is online more
New stage；
The initial training stage will be detected each time is expressed as a p dimensional vector, comprising K class samples, then builds K layers
Self organization map neutral net, every layer N × M structure is made up of neuron two dimensional surface, K × N × M is had in neutral net refreshing
Through member, the neural network weight matrix is made to beWhereinRepresent kth layer plane coordinate for (n, m) p right-safeguarding value to
Amount, and k ∈ { 1 ..., K }, n ∈ { 1 ..., N }, m ∈ { 1 ..., M }；As certain p dimension training samples xtInto after network, carry out initial
Change；
The online updating stage is completed after initial training, into test, sample xpGeneric be it is unknown, subsequently
Grader is according to the on-line training sample in W to xpClassified, the classification results of subsequent classifier are kth class.
Further, the initialization is specifically included：
1) by training sample xtClassificationCalculate theAll neuron weights of layer and xtDistance：
Using the minimum neuron of distance as triumph neuron, the position of neuron is preserved：
Wherein,WithRespectively triumph neuron isThe two-dimensional coordinate of layer；
2) toLayer neuron weights carry out following iteration：
Wherein α ∈ (0,1) are learning rate, are defined as follows：
3) repeat 1) with 2), initial training is carried out up to training sample fully enters neutral net；
4) 1) -3 are repeated), the distribution of self organization map neural network weight is approached with training sample.
Further, the online updating stage specifically includes：
1) all neuron weights and x are calculatedpDistance：
Using the minimum neuron of distance as triumph neuron, the position of the neuron is preserved：
WhereinThe sequence number of layer where triumph neuron,WithRespectively triumph neuron isThe two-dimensional coordinate of layer；
2) following iteration is carried out to kth layer neuron weights：
Wherein α ∈ (0,1) are learning rate, are defined as follows：
3) grader is trained using neural network weight matrix W or carried out using " the on-line training sample " in W
Next subseries.
Another object of the present invention is to provide the online school of Artificial Olfactory described in a kind of application based on self organization map
The Artificial Olfactory of positive sample generation method.
Another object of the present invention is to provide the online school of Artificial Olfactory described in a kind of application based on self organization map
The gas sensor of positive sample generation method.
Advantages of the present invention and good effect are：Result in table 2 shows that using the present invention Artificial Olfactory can be made
The amendment of mode identification method is completed in the case where not carrying out conventional correction, is compared with situation about not being corrected, in multiclass
In other classification application, system identification accuracy rate rises to 48.51% by 42.56%.
Brief description of the drawings
Fig. 1 is the Artificial Olfactory on-line correction sample generating method provided in an embodiment of the present invention based on self organization map
Flow chart.
Embodiment
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to embodiments, to the present invention
It is further elaborated.It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not used to
Limit the present invention.
The application principle of the present invention is explained in detail below in conjunction with the accompanying drawings.
As shown in figure 1, the Artificial Olfactory on-line correction sample life provided in an embodiment of the present invention based on self organization map
Comprise the following steps into method：
S101：Detection is expressed as a vector each time, builds self organization map neutral net, and training sample enters nerve net
Network, is initialized；
S102：Complete after initial training, into test, subsequent classifier is according to on-line training sample to classifying.
Artificial Olfactory on-line correction sample generating method provided in an embodiment of the present invention based on self-organizing organization chart
Including two stages of initial training and online updating.
1st, initial training
It will each time detect and be expressed as a p dimensional vector, it is assumed that include the self organization map of K class samples, then one K layers of structure
(Self-organized map) neutral net, every layer N × M structure is made up of neuron two dimensional surface, therefore in neutral net
Shared K × N × M neuron, makes the neural network weight matrix beWhereinRepresent that kth layer plane coordinate is
The p dimension weight vectors of (n, m), and k ∈ { 1 ..., K }, n ∈ { 1 ..., N }, m ∈ { 1 ..., M }.
As certain p dimension training samples xtInto after network, initialization：
1) by training sample xtClassificationCalculate theAll neuron weights of layer and xtDistance：
Using the minimum neuron of distance as triumph neuron, the position of the neuron is preserved：
Wherein,WithRespectively triumph neuron isThe two-dimensional coordinate of layer.
2) toLayer neuron weights carry out following iteration：
Wherein α ∈ (0,1) are learning rate, are defined as follows：
3) repeat step 1) and step 2), until training sample fully enters neutral net and carries out initial training.
4) repeat step 1)-step 3), meet and stop after certain cycle-index.
Neural network weight matrix W now is " on-line training sample " collection of initialization, N × M nerve of kth layer
First weights are " on-line training sample " collection of kth class sample.
2nd, online updating
After completion initial training, into test, sample xpGeneric is unknown, and subsequent classifier is according in W
" on-line training sample " is to xpClassified, if the classification results of subsequent classifier are kth class, step is as follows：
1) all neuron weights and x are calculatedpDistance：
Using the minimum neuron of distance as triumph neuron, the position of the neuron is preserved：
WhereinThe sequence number of layer where triumph neuron,WithRespectively triumph neuron isThe two dimension seat of layer
Mark.
2) following iteration is carried out to kth layer neuron weights：
Wherein α ∈ (0,1) are learning rate, are defined as follows：
3) grader is trained using neural network weight matrix W or carried out using " the on-line training sample " in W
Next subseries.
Online updating makes neuron weights (i.e. " on-line training sample " collects) and synchronous, the Ke Yiwei of current sample changed holding
The real-time update of subsequent classifier provides support, grader is matched with producing the Artificial Olfactory response after drift.
The application effect of the present invention is explained in detail with reference to detection.
Using the Artificial Olfactory for including 16 metal-oxide gas transducers, respectively to ammonia, acetaldehyde, acetone,
Six kinds of ethene, ethanol and toluene etc. (i.e. K=6) gas is detected.Above carried from each sensor response curve in detection every time
2 steady state characteristics and 6 behavioral characteristics are taken, each detection is characterized as to the vector of one 128 dimension (p=128).Drift suppression
System, detection time span is 36 months and concentrates all detection data preparations to 10 data that each data set includes number
1 is shown in Table according to detection time corresponding relation.N=M=10 is taken, using the data in data set 1 as initial training sample, by data
Collect the data in 2-10 as test sample, sorting technique uses nearest neighbor method (KNN), chooses using the inventive method and does not adopt
Contrasted with two kinds of situations of the inventive method, as a result as shown in table 2.Result in table 2 shows that the present invention can be in online work
Under the conditions of work, the anti-long term drift ability of Artificial Olfactory is improved.
The data set of table 1 and detection time
| Data set number | Detection time |
| Data set 1 | The 1-2 months |
| Data set 2 | The 3-10 months |
| Data set 3 | The 11-13 months |
| Data set 4 | The 14-15 months |
| Data set 5 | 16th month |
| Data set 6 | The 17-20 months |
| Data set 7 | 21st month |
| Data set 8 | The 22-23 months |
| Data set 9 | The 24-35 months |
| Data set 10 | 36th month |
The testing result of table 2
The foregoing is merely illustrative of the preferred embodiments of the present invention, is not intended to limit the invention, all essences in the present invention
Any modifications, equivalent substitutions and improvements made within refreshing and principle etc., should be included in the scope of the protection.
Claims (5)
1. a kind of Artificial Olfactory on-line correction sample generating method based on self organization map, it is characterised in that described to be based on
The Artificial Olfactory on-line correction sample generating method of self organization map includes initial training stage and online updating stage；
The initial training stage will be detected each time is expressed as a p dimensional vector, comprising K class samples, then build K layers from group
Knit figure neutral net, every layer N × M structure is made up of neuron two dimensional surface, K × N × M is had in neutral net neural
Member, makes the neural network weight matrix beWhereinRepresent kth layer plane coordinate for (n, m) p right-safeguarding value to
Amount, and k ∈ { 1 ..., K }, n ∈ { 1 ..., N }, m ∈ { 1 ..., M }；As certain p dimension training samples xtInto after network, carry out initial
Change；
The online updating stage is completed after initial training, into test, sample xpGeneric is unknown, subsequent classification
Device is according to the on-line training sample in W to xpClassified, the classification results of subsequent classifier are kth class.
2. the Artificial Olfactory on-line correction sample generating method as claimed in claim 1 based on self organization map, its feature
It is, the initialization is specifically included：
1) by training sample xtClassificationCalculate theAll neuron weights of layer and xtDistance：
<mrow>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msubsup>
<mo>=</mo>
<mo>|</mo>
<mo>|</mo>
<msup>
<mi>x</mi>
<mi>t</mi>
</msup>
<mo>-</mo>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msubsup>
<mo>|</mo>
<mo>|</mo>
<mo>;</mo>
</mrow>
Using the minimum neuron of distance as triumph neuron, the position of neuron is preserved：
<mrow>
<mo>(</mo>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
<mo>,</mo>
<mover>
<mi>n</mi>
<mo>~</mo>
</mover>
<mo>,</mo>
<mover>
<mi>m</mi>
<mo>~</mo>
</mover>
<mo>)</mo>
<mo>=</mo>
<munder>
<mrow>
<mi>arg</mi>
<mi>min</mi>
</mrow>
<msup>
<mi>w</mi>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msup>
</munder>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msubsup>
<mo>;</mo>
</mrow>
Wherein,WithRespectively triumph neuron isThe two-dimensional coordinate of layer；
2) toLayer neuron weights are iterated：
<mrow>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msubsup>
<mo>=</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>)</mo>
</mrow>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msubsup>
<mo>+</mo>
<mi>&alpha;</mi>
<mo>&CenterDot;</mo>
<msup>
<mi>x</mi>
<mi>t</mi>
</msup>
<mo>;</mo>
</mrow>
Wherein α ∈ (0,1) are learning rate, are defined as follows：
<mrow>
<mi>&alpha;</mi>
<mo>=</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mfrac>
<mrow>
<mo>|</mo>
<mo>|</mo>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>,</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mover>
<mi>n</mi>
<mo>~</mo>
</mover>
<mo>,</mo>
<mover>
<mi>m</mi>
<mo>~</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>|</mo>
</mrow>
<mrow>
<munder>
<mi>max</mi>
<msup>
<mi>w</mi>
<mover>
<mi>k</mi>
<mo>~</mo>
</mover>
</msup>
</munder>
<mo>|</mo>
<mo>|</mo>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>,</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mover>
<mi>n</mi>
<mo>~</mo>
</mover>
<mo>,</mo>
<mover>
<mi>m</mi>
<mo>~</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>|</mo>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
3) repeat 1) with 2), initial training is carried out up to training sample fully enters neutral net；
4) 1) -3 are repeated), meet and stop after certain cycle-index.
3. the Artificial Olfactory on-line correction sample generating method as claimed in claim 1 based on self organization map, its feature
It is, the online updating stage specifically includes：
1) all neuron weights and x are calculatedpDistance：
<mrow>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mi>k</mi>
</msubsup>
<mo>=</mo>
<mo>|</mo>
<mo>|</mo>
<msup>
<mi>x</mi>
<mi>p</mi>
</msup>
<mo>-</mo>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mi>k</mi>
</msubsup>
<mo>|</mo>
<mo>|</mo>
<mo>;</mo>
</mrow>
Using the minimum neuron of distance as triumph neuron, the position of the neuron is preserved：
<mrow>
<mo>(</mo>
<mover>
<mi>k</mi>
<mo>^</mo>
</mover>
<mo>,</mo>
<mover>
<mi>n</mi>
<mo>^</mo>
</mover>
<mo>,</mo>
<mover>
<mi>m</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
<mo>=</mo>
<munder>
<mrow>
<mi>arg</mi>
<mi>min</mi>
</mrow>
<msup>
<mi>w</mi>
<mover>
<mi>k</mi>
<mo>^</mo>
</mover>
</msup>
</munder>
<msubsup>
<mi>d</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mover>
<mi>k</mi>
<mo>^</mo>
</mover>
</msubsup>
<mo>;</mo>
</mrow>
WhereinThe sequence number of layer where triumph neuron,WithRespectively triumph neuron isThe two-dimensional coordinate of layer；
2) following iteration is carried out to kth layer neuron weights：
<mrow>
<mfenced open = "{" close = "">
<mtable>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mi>k</mi>
</msubsup>
<mo>=</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>-</mo>
<mi>&alpha;</mi>
<mo>)</mo>
</mrow>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mi>k</mi>
</msubsup>
<mo>+</mo>
<mi>&alpha;</mi>
<mo>&CenterDot;</mo>
<msup>
<mi>x</mi>
<mi>p</mi>
</msup>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>k</mi>
<mo>&NotEqual;</mo>
<mover>
<mi>k</mi>
<mo>^</mo>
</mover>
</mrow>
</mtd>
</mtr>
<mtr>
<mtd>
<mrow>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mi>k</mi>
</msubsup>
<mo>=</mo>
<msubsup>
<mi>w</mi>
<mrow>
<mi>n</mi>
<mi>m</mi>
</mrow>
<mi>k</mi>
</msubsup>
</mrow>
</mtd>
<mtd>
<mrow>
<mi>k</mi>
<mo>=</mo>
<mover>
<mi>k</mi>
<mo>^</mo>
</mover>
</mrow>
</mtd>
</mtr>
</mtable>
</mfenced>
<mo>;</mo>
</mrow>
1
Wherein α ∈ (0,1) are learning rate, are defined as follows：
<mrow>
<mi>&alpha;</mi>
<mo>=</mo>
<msup>
<mi>e</mi>
<mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mfrac>
<mrow>
<mo>|</mo>
<mo>|</mo>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>,</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mover>
<mi>n</mi>
<mo>^</mo>
</mover>
<mo>,</mo>
<mover>
<mi>m</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>|</mo>
</mrow>
<mrow>
<munder>
<mi>max</mi>
<msup>
<mi>w</mi>
<mi>k</mi>
</msup>
</munder>
<mo>|</mo>
<mo>|</mo>
<mrow>
<mo>(</mo>
<mi>n</mi>
<mo>,</mo>
<mi>m</mi>
<mo>)</mo>
</mrow>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mover>
<mi>n</mi>
<mo>^</mo>
</mover>
<mo>,</mo>
<mover>
<mi>m</mi>
<mo>^</mo>
</mover>
<mo>)</mo>
</mrow>
<mo>|</mo>
<mo>|</mo>
</mrow>
</mfrac>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>;</mo>
</mrow>
3) grader is trained using neural network weight matrix W or carried out next time using the on-line training sample in W
Classification.
4. the Artificial Olfactory on-line correction sample based on self organization map described in a kind of application claims 1 to 3 any one
The Artificial Olfactory of generation method.
5. the Artificial Olfactory on-line correction sample based on self organization map described in a kind of application claims 1 to 3 any one
The gas sensor of generation method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710393408.3A CN107220671B (en) | 2017-05-27 | 2017-05-27 | Artificial olfaction system online correction sample generation method based on self-organizing map |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710393408.3A CN107220671B (en) | 2017-05-27 | 2017-05-27 | Artificial olfaction system online correction sample generation method based on self-organizing map |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107220671A true CN107220671A (en) | 2017-09-29 |
| CN107220671B CN107220671B (en) | 2020-07-14 |
Family
ID=59946994
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710393408.3A Active CN107220671B (en) | 2017-05-27 | 2017-05-27 | Artificial olfaction system online correction sample generation method based on self-organizing map |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107220671B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109540978A (en) * | 2018-12-13 | 2019-03-29 | 清华大学 | Odor identification equipment |
| CN111914082A (en) * | 2019-05-08 | 2020-11-10 | 天津科技大学 | Online knowledge aggregation method based on SOM neural network algorithm |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100198521A1 (en) * | 2007-07-24 | 2010-08-05 | Technion Research And Development Foundation Ltd. | Chemically sensitive field effect transistors and uses thereof in electronic nose devices |
| CN102507676A (en) * | 2011-11-01 | 2012-06-20 | 重庆大学 | On-line drift compensation method of electronic nose based on multiple self-organizing neural networks |
| CN103489033A (en) * | 2013-09-27 | 2014-01-01 | 南京理工大学 | Incremental type learning method integrating self-organizing mapping and probability neural network |
| CN103499664A (en) * | 2013-07-30 | 2014-01-08 | 中国标准化研究院 | Method for extracting and researching atlas information representing differences of honey quality |
| CN105823801A (en) * | 2016-03-03 | 2016-08-03 | 重庆大学 | Deep belief network characteristic extraction-based electronic nose drift compensation method |
| CN105913079A (en) * | 2016-04-08 | 2016-08-31 | 重庆大学 | Target domain migration extreme learning-based electronic nose heterogeneous data identification method |
| CN106226464A (en) * | 2016-07-04 | 2016-12-14 | 重庆大学 | A kind of Electronic Nose gas detecting system with defencive function |
-
2017
- 2017-05-27 CN CN201710393408.3A patent/CN107220671B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100198521A1 (en) * | 2007-07-24 | 2010-08-05 | Technion Research And Development Foundation Ltd. | Chemically sensitive field effect transistors and uses thereof in electronic nose devices |
| CN102507676A (en) * | 2011-11-01 | 2012-06-20 | 重庆大学 | On-line drift compensation method of electronic nose based on multiple self-organizing neural networks |
| CN103499664A (en) * | 2013-07-30 | 2014-01-08 | 中国标准化研究院 | Method for extracting and researching atlas information representing differences of honey quality |
| CN103489033A (en) * | 2013-09-27 | 2014-01-01 | 南京理工大学 | Incremental type learning method integrating self-organizing mapping and probability neural network |
| CN105823801A (en) * | 2016-03-03 | 2016-08-03 | 重庆大学 | Deep belief network characteristic extraction-based electronic nose drift compensation method |
| CN105913079A (en) * | 2016-04-08 | 2016-08-31 | 重庆大学 | Target domain migration extreme learning-based electronic nose heterogeneous data identification method |
| CN106226464A (en) * | 2016-07-04 | 2016-12-14 | 重庆大学 | A kind of Electronic Nose gas detecting system with defencive function |
Non-Patent Citations (2)
| Title |
|---|
| TAO LIU, KADRI CHAIBOU, ZHIYONG HUANG.: ""A Novel Retraining Method of Multiple Self-Organizing Maps for Gas Sensor Drift Compensation. Sensors and Materials"", 《SENSORS AND MATERIALS》 * |
| 刘涛，黄智勇: ""一种基于多重自组织图的电子鼻漂移抑制方法"", 《仪器仪表学报》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109540978A (en) * | 2018-12-13 | 2019-03-29 | 清华大学 | Odor identification equipment |
| CN111914082A (en) * | 2019-05-08 | 2020-11-10 | 天津科技大学 | Online knowledge aggregation method based on SOM neural network algorithm |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107220671B (en) | 2020-07-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101794396B (en) | System and method for recognizing remote sensing image target based on migration network learning | |
| CN106202177B (en) | A kind of file classification method and device | |
| CN103258214B (en) | Based on the Classifying Method in Remote Sensing Image of image block Active Learning | |
| CN106920007B (en) | PM based on second-order self-organizing fuzzy neural network2.5Intelligent prediction method | |
| CN108961235A (en) | A kind of disordered insulator recognition methods based on YOLOv3 network and particle filter algorithm | |
| CN108173708A (en) | Anomalous traffic detection method, device and storage medium based on incremental learning | |
| CN106022528A (en) | Short-term power prediction method of photovoltaic power plant based on density peak hierarchical clustering | |
| CN105938116A (en) | Gas sensor array concentration detection method based on fuzzy division and model integration | |
| CN109242049A (en) | Water supply network multiple spot leakage loss localization method and its device based on convolutional neural networks | |
| CN105574547B (en) | Adapt to integrated learning approach and device that dynamic adjusts base classifier weight | |
| CN105653450A (en) | Software defect data feature selection method based on combination of modified genetic algorithm and Adaboost | |
| CN103049651A (en) | Method and device used for power load aggregation | |
| CN104317681A (en) | Behavioral abnormality automatic detection method and behavioral abnormality automatic detection system aiming at computer system | |
| CN104680542A (en) | Online learning based detection method for change of remote-sensing image | |
| CN104679860B (en) | A kind of sorting technique of unbalanced data | |
| CN101271625A (en) | Method for detecting freeway traffic event by integration supporting vector machine | |
| CN103093643A (en) | Public parking lot berth quantity confirming method | |
| CN106991049A (en) | A kind of Software Defects Predict Methods and forecasting system | |
| CN103500342A (en) | Human behavior recognition method based on accelerometer | |
| CN116229380A (en) | Method for identifying bird species related to bird-related faults of transformer substation | |
| CN108490115A (en) | A kind of air quality method for detecting abnormality based on distributed online principal component analysis | |
| CN107220671A (en) | A kind of Artificial Olfactory on-line correction sample generating method based on self organization map | |
| CN108256473A (en) | Methods, devices and systems, computer readable storage medium for data amplification | |
| CN102842043A (en) | Particle swarm classifying method based on automatic clustering | |
| CN105241665A (en) | Rolling bearing fault diagnosis method based on IRBFNN-AdaBoost classifier |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |