CN109801685A - A kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model - Google Patents
A kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model Download PDFInfo
- Publication number
- CN109801685A CN109801685A CN201910196395.XA CN201910196395A CN109801685A CN 109801685 A CN109801685 A CN 109801685A CN 201910196395 A CN201910196395 A CN 201910196395A CN 109801685 A CN109801685 A CN 109801685A
- Authority
- CN
- China
- Prior art keywords
- water
- bayes
- discrimination model
- model
- bursting source
- 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
- Investigating Or Analyzing Non-Biological Materials By The Use Of Chemical Means (AREA)
Abstract
The water bursting source recognition methods based on PCA method Yu Bayes discrimination model that the invention discloses a kind of, comprising: each basal water Hydro-chemical component analyzes data in recent years in collection research area；Based on practical Hydro-chemical component data, the Bayes discrimination model based on PCA method is established；By training sample back substitution into discrimination model, the differentiation effect of model is obtained；Based on sample to be tested, calculated result and actual result are compared, predict water bursting source.The present invention is based on PCA methods to analyze each basal water Hydro-chemical component, using the principle of dimensionality reduction, each interionic overlapping disturbing factor is excluded, several principal components are chosen, water bursting source discrimination model is established in conjunction with Bayes method of discrimination, carries out the identification of research area's water bursting source.This method has certain novelty, and identification effect is high and precision is quasi-, provides a kind of new practical approach for the identification prediction of water bursting source.
Description
Technical field
The present invention relates to hydrogeological fields, and in particular to a kind of water bursting source based on PCA method Yu Bayes discrimination model
Recognition methods.
Background technique
In coal in China production process, water damage accident frequently occurs, and all causes seriously to personal safety, economic development
It threatens.Therefore fast and accurately judge water bursting source position, effective key message could be provided for the development of water damage control measure.
Currently, since data of water quality is compared with other data, having the characteristics that quick, accurate, economy, in water bursting source differentiation
Mainstream as water bursting source identification.Because the water quality components in each water-bearing layer are different, mine is generally differentiated using chemical measurements of water
Water bursting source.Common method has BP neural network method, clustering methodology, regression analysis etc., but does not all account for and respectively comment
Correlation between the valence factor, the case where causing the overlapping between each information and generate erroneous judgement, and principal component analysis (PCA) can disappear
Except the overlay information between water sample index, discrimination precision is improved.The present invention is analyzing the correlation between each index using PCA method
On the basis of, the strong main component of correlation is extracted, finally combines Bayes diagnostic method to establish the Bayes based on PCA method and differentiates mould
Type differentiates water bursting source.
Summary of the invention
1. the purpose of the present invention
The present invention determines the too low defect of overlong time, discrimination precision for diagnostic method traditional in art methods, will
PCA method combines the conventional method instead of Individual forecast water bursting source with Bayes diagnostic method, reduces and differentiates the time, improves differentiation
Precision.
2. technical solution of the present invention
Based on above-mentioned purpose, the method for establishing the identification water bursting source of the Bayesr discrimination model based on PCA method.This method
It include: step A, collection research area each basal water water quality type data in recent years；Step B the, to (training of practical data of water quality
Sample), establish the Bayes discrimination model based on PCA method；Step C obtains model by training sample back substitution into discrimination model
Differentiation effect；Step D is based on sample to be tested, calculated result and actual result is compared, and predicts water bursting source.
3. beneficial effects of the present invention
The present invention is based on PCA methods to analyze each aquifer water chemical constituent, using the principle of dimensionality reduction, excludes each ion
Between be overlapped disturbing factor, select several principal components, establish water bursting source discrimination model in conjunction with Bayes diagnostic method, carry out research area
The identification of water bursting source.This method has certain novelty, and identification effect is high and precision is quasi-, mentions for the identification prediction of water bursting source
A kind of new practical approach is supplied.
Detailed description of the invention
Fig. 1 is that the present invention is based on the water bursting source recognition methods flow charts of the Bayes discrimination model of PCA method.
Specific embodiment
Below with reference to attached drawing of the present invention and application example, invention is further explained.
Step A, collection research area each basal water water quality type data in recent years；
Step B establishes the Bayes discrimination model based on PCA method to practical data of water quality (training sample)；
Step C obtains the differentiation effect of model by training sample back substitution into discrimination model；
Step D is based on sample to be tested, calculated result and actual result is compared, and predicts water bursting source.
Below in conjunction with specific example, detailed step is enumerated:
Step A: certain mine water bursting source is divided into Cenozoic's unconsolidated formation water I, coal measures Sandstone Water II, Taiyuan Forma-tion limestone water III.Three
24 water samples are chosen in class water source as training sample, wherein 11, unconsolidated formation water, Sandstone Water 6, limestone water 7.It considers
The chemical constituent of different water cut layer is different, therefore chooses six big conventional ions as discriminant criterion: Ca2+(X1)、Na++K+(X2)、
Mg2+(X3)、HCO3 -(X4)、Cl-(X5)、SO4 2-(X6)。
Step B: being first standardized training sample data, then carries out principal component point to the data after standardization
Analysis, obtains the correlation matrix between each discriminant criterion, the results are shown in Table 1.It is related, Cl between each ion-With Na++K+'s
The degree of association has reached 0.782, illustrates there is overlay information between two indexes, cannot directly use, and otherwise will cause extra duplicate
Information increases its calculation amount, it is also possible to reduce the precision of water source differentiation, and cause to judge by accident.Therefore, principal component is carried out to sample
Analysis is necessary.The characteristic value, contribution rate and accumulation contribution rate (being shown in Table 2) for obtaining each Assessing parameters simultaneously, choose first 4
Principal component, preceding 4 principal components occupy 96.53% data information amount as shown in Table 2, can carry out to sample information effective general
It states.
The correlation matrix of each Hydrochemical Composition index of table 1
Each Principal Component Explanation variance rate of table 2
It recycles SPSS software to establish Bayes diagnostic method, obtains the discriminant function model at three kinds of water sources, discrimination formula is such as
Under:
Z in formula1、Z2、Z3The respectively discriminant function of Cenozoic's unconsolidated formation water, coal measures Sandstone Water and Taiyuan Forma-tion limestone water
Value；X1、X2、X3、X4For ion concentration value.The discriminant score of which water sample is bigger when differentiation, then which the water sample just belongs to
Class.
Step C: 24 training sample back substitutions being entered in the Bayes discriminant function having built up, are reclassified, if
The coincidence rate for reclassifying result and known class is very high, then the effect of discriminant function is all right.By table 3 and table 4 it is found that the first kind
The False Rate of water sample is 18.2%, and the False Rate of the second class water sample is 16.7%, and the False Rate of third class water sample is 14.3%, always
False Rate is 16.67%.It can be seen that the discrimination model that the invention is established is feasible efficient to water source differentiation.
3 back substitution of table differentiates result
4 training sample of table, which returns, sentences result statistics
Step D: the water sample to be measured of selection is carried out based on PCA method and Bayes discrimination model with what training sample was established
It predicts and is compared with direct Bayes diagnostic method, by prediction result (table 5) it is found that working as in 11 gushing water water sample prediction results
In, the misjudgement of only one Sandstone Water is limestone water, differentiates accuracy up to 90.9%；Differentiate that result is compared with direct Bayes
It is found that directly Bayes differentiates that accuracy only has 63.6%, the results showed that method of the invention differentiates more quasi- than single Bayes
Really, the influence between sample is largely eliminated, and improves differentiation accuracy rate.
5 forecast sample of table differentiates result
Each aquifer water chemical constituent is analyzed in conclusion the present invention is based on PCA methods, excludes each interionic overlapping
Disturbing factor chooses principal component, establishes water bursting source discrimination model in conjunction with Bayes diagnostic method, carries out research area's water bursting source
Prediction.This method has certain novelty, and identification effect is high and precision is quasi-, provides one kind newly for the identification prediction of water bursting source
Practical approach.
Specific example described above, to the purpose of the present invention, process and effect are described in detail, and are not limited to this hair
Bright protection scope, all within the spirits and principles of the present invention, any modification, equivalent replacement for being made etc. should be included in
Within protection scope of the present invention.
Claims (2)
1. a kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model, comprising:
Step A, collection research area each basal water water quality type data in recent years；
Step B establishes the Bayes discrimination model based on PCA method to practical data of water quality (training sample)；
Step C obtains the differentiation effect of model by training sample back substitution into discrimination model；
Step D is based on sample to be tested, calculated result and actual result is compared, and predicts water bursting source.
2. in stepb, being based on practical data of water quality (training sample), the gushing water of the Bayes discrimination model based on PCA method is established
Water source recognition methods, as follows in detail:
(1) training sample data are standardized first, then the data after standardization is led with SPSS software
Constituent analysis obtains the correlation matrix between each discriminant criterion.If the related coefficient between two indices is larger, illustrate two
Person influences each other, and there are information overlap phenomenons, cannot directly use, and otherwise will cause extra duplicate information, makes its calculation amount
Increase, it is also possible to reduce the precision of water source differentiation, and cause to judge by accident, so PCA processing must be carried out to this group of data.Simultaneously
To the characteristic value of each discriminant criterion, contribution rate and accumulation contribution rate, preceding 4 discriminant criterions that accumulation contribution rate is greater than 95% are chosen
(X1, X2, X3, X4) is used as principal component, recycles SPSS software to establish Bayes discrimination model and obtains component score coefficient matrix.
(2) according to principal component scores coefficient matrix, Bayes computation model (as follows) is established.Differentiated according to the Bayes of foundation
Model sentence to training sample, obtains the differentiation effect of model.
Z in formula1、Z2、Z3The discriminant score at respectively 3 kinds water sources.
(3) combine above-mentioned analysis, according to the Bayes discrimination model of foundation, then forecast sample is differentiated, by calculated result with
Actual result is compared, and obtains the accuracy rate of Model checking, and then differentiate to water bursting source.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910196395.XA CN109801685A (en) | 2019-03-15 | 2019-03-15 | A kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910196395.XA CN109801685A (en) | 2019-03-15 | 2019-03-15 | A kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109801685A true CN109801685A (en) | 2019-05-24 |
Family
ID=66562929
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910196395.XA Pending CN109801685A (en) | 2019-03-15 | 2019-03-15 | A kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109801685A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110261560A (en) * | 2019-07-05 | 2019-09-20 | 安徽大学 | The water source recognition methods of complex hydrologic geology water bursting in mine and system |
| CN111881974A (en) * | 2020-07-27 | 2020-11-03 | 河南理工大学 | Water inrush source identification method based on pipe-PCA-FCL discrimination model |
| CN112381117A (en) * | 2020-10-22 | 2021-02-19 | 合肥工业大学 | Coal mine water inrush water source mixing proportion calculation and dynamic monitoring method based on conventional water chemistry |
| CN112945209A (en) * | 2021-03-30 | 2021-06-11 | 淮南矿业(集团)有限责任公司 | Early warning method, system and device for water inrush of Aohu water |
| CN113255212A (en) * | 2021-05-17 | 2021-08-13 | 中国南方电网有限责任公司超高压输电公司昆明局 | Model selection method for converter valve cooling system based on PCA and Bayesian classifier |
-
2019
- 2019-03-15 CN CN201910196395.XA patent/CN109801685A/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| 琚棋定等: "基于主成分分析与贝叶斯判别法的矿井突水水源识别方法研究", 《煤炭工程》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110261560A (en) * | 2019-07-05 | 2019-09-20 | 安徽大学 | The water source recognition methods of complex hydrologic geology water bursting in mine and system |
| CN111881974A (en) * | 2020-07-27 | 2020-11-03 | 河南理工大学 | Water inrush source identification method based on pipe-PCA-FCL discrimination model |
| CN111881974B (en) * | 2020-07-27 | 2023-06-09 | 河南理工大学 | Water inrush source identification method based on Piper-PCA-FCL discrimination model |
| CN112381117A (en) * | 2020-10-22 | 2021-02-19 | 合肥工业大学 | Coal mine water inrush water source mixing proportion calculation and dynamic monitoring method based on conventional water chemistry |
| CN112381117B (en) * | 2020-10-22 | 2023-10-17 | 合肥工业大学 | Coal mine water inrush source mixing proportion calculation and dynamic monitoring method based on conventional water chemistry |
| CN112945209A (en) * | 2021-03-30 | 2021-06-11 | 淮南矿业(集团)有限责任公司 | Early warning method, system and device for water inrush of Aohu water |
| CN113255212A (en) * | 2021-05-17 | 2021-08-13 | 中国南方电网有限责任公司超高压输电公司昆明局 | Model selection method for converter valve cooling system based on PCA and Bayesian classifier |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109801685A (en) | A kind of water bursting source recognition methods based on PCA method Yu Bayes discrimination model | |
| TWI739798B (en) | Method and device for establishing data recognition model | |
| Tan et al. | Analyst following and forecast accuracy after mandated IFRS adoptions | |
| Mittal et al. | Stock prediction using twitter sentiment analysis | |
| CN108805357A (en) | A kind of Fisher discrimination model water bursting source prediction techniques based on PCA analyses | |
| CN105095238B (en) | For detecting the decision tree generation method of fraudulent trading | |
| CN109871654A (en) | A kind of water bursting source recognition methods based on PCA and Method of Gray | |
| CN109784383B (en) | Rail crack identification method based on graph domain feature and DS evidence theory fusion | |
| CN109993459A (en) | A kind of complexity multi-aquifer water bursting in mine water source recognition methods | |
| CN105334185B (en) | The near-infrared model maintaining method differentiated based on spectrum projection | |
| CN109255134A (en) | Method for acquiring fault condition of pumping well | |
| Prokić et al. | Detecting shibboleths | |
| CN107132310A (en) | Transformer equipment health status method of discrimination based on gauss hybrid models | |
| CN108108890A (en) | A kind of defective workmanship risk analysis method based on process incidence relation | |
| CN107561280A (en) | Kit for predicting breast cancer recurrence | |
| CN110379459A (en) | A kind of method and system being associated with discovery molecular marker with gene function based on transcript profile dynamic change of temporal series | |
| CN108876030A (en) | A kind of water bursting source prediction technique based on Fisher discrimination model | |
| CN109887552A (en) | A kind of water bursting source differentiation prediction technique | |
| CN106645098A (en) | Identification method for original places of jades by combining spectral flattening with multivariate statistical model | |
| RU2743836C1 (en) | Method for detecting influx of pumped water in oil well | |
| CN108389135A (en) | A kind of water bursting source method of discrimination based on Grey Relational Model | |
| CN107742256A (en) | A kind of network puppet public opinion identification method based on decision making of combinatorial optimization tree | |
| CN117034601A (en) | Soft measurement modeling method for overflow granularity in multi-station ore dressing process | |
| CN103984756B (en) | Semi-supervised probabilistic latent semantic analysis based software change log classification method | |
| CN103217374A (en) | Acute B lymphocytic leukemia initiating cell characteristic determination system and method |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190524 |