CN112434887A - Water supply network risk prediction method combining network kernel density estimation and SVM - Google Patents
Water supply network risk prediction method combining network kernel density estimation and SVM Download PDFInfo
- Publication number
- CN112434887A CN112434887A CN202011497723.9A CN202011497723A CN112434887A CN 112434887 A CN112434887 A CN 112434887A CN 202011497723 A CN202011497723 A CN 202011497723A CN 112434887 A CN112434887 A CN 112434887A
- Authority
- CN
- China
- Prior art keywords
- network
- water supply
- risk
- supply network
- event
- 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
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 73
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000012706 support-vector machine Methods 0.000 claims abstract description 29
- 238000002790 cross-validation Methods 0.000 claims abstract description 9
- 238000013058 risk prediction model Methods 0.000 claims abstract description 4
- 230000004044 response Effects 0.000 claims abstract description 3
- 238000012549 training Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 10
- 238000012423 maintenance Methods 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 238000000926 separation method Methods 0.000 claims description 2
- 230000002265 prevention Effects 0.000 abstract 1
- 238000004880 explosion Methods 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 4
- 239000000463 material Substances 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000009933 burial Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 201000010099 disease Diseases 0.000 description 1
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000013077 scoring method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Geometry (AREA)
- General Business, Economics & Management (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computational Mathematics (AREA)
- Public Health (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Computer Hardware Design (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a water supply network risk prediction method combining network kernel density estimation and SVM, which is characterized in that a network-point event network topology is established by matching event points and diseased pipelines based on event data such as pipeline leakage, rupture and pipe burst; calculating the nuclear density of each pipe section unit event point by adopting a NetKDE method, quantitatively representing the structure risk level of the pipe section, and dividing the risk condition grade of the pipe section by adopting a natural boundary method; constructing a water supply network risk prediction model based on an SVM (support vector machine) by taking the risk condition grade as a response vector, wherein the model adopts a Radial Basis Function (RBF) kernel function, and determines optimal model parameters through cross validation and grid search; and predicting the risk condition of the pipe section structure of the water supply network by using the optimized model to form a water supply network risk distribution map. The invention combines the network nuclear density estimation and the support vector machine method, realizes the event-driven water supply network risk dynamic prediction, and opens up a new technical approach for water supply network risk identification and accident disaster active prevention and control.
Description
Technical Field
The invention relates to the technical field of water supply Network risk prediction, in particular to a water supply Network risk prediction method combining Network Kernel Density Estimation (NetKDE) and SVM (support Vector machine).
Background
With the continuous expansion of the scale of urban water supply network facilities and the increase of service life, the problems of pipeline aging and operation safety are paid more and more attention. The water supply network is used as one of important urban infrastructure and lifeline systems, and safe operation of the water supply network is very important for guaranteeing sustainable development of economy and society. Carry out scientific and effective discernment and prediction to water supply network risk, help pipe network facility safe operation and calamity risk active management.
At present, index scoring methods are used for more prediction and evaluation of urban water supply network risk conditions, indexes of the methods are comprehensive, but in practical application, part of indexes often lack background information support, subjective influence factors exist in a weight assignment process, and quantitative description is lacked for pipeline risks. The artificial intelligence and machine learning methods are introduced into the pipe network risk prediction evaluation, the association mode of the pipe risk condition and the associated characteristic variable can be established through data sample supervised learning, and the method has a good application prospect in the aspect of pipe explosion prediction diagnosis at present. Konstantinos et al combine evolutionary polynomial regression with K-means clustering to provide a data-driven water supply pipeline fault prediction model; francis et al establish a water supply system pipe explosion prediction model based on a Bayesian network, and discuss risk factors such as pipe age, pipe diameter, material, fracture history and uncertainty thereof; the method comprises the following steps that a water supply network pipe burst diagnosis model is established on the basis of water pressure monitoring data of a water supply network and a PSO-SVM algorithm; wangbang uses 5 factors of pipe material, running pressure, pipe age, pipe diameter and road load as independent variables and uses the explosion of pipeline as dependent variables, and utilizes random forest algorithm to build pipe explosion prediction model.
Careful analysis shows that although a plurality of pipe explosion detection researches exist, a pipe network risk prediction method widely applied by a water department does not exist at present, and a quantitative prediction evaluation technology for the structural condition of a buried pipeline is lacked.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a water supply network risk prediction method combining network nuclear density estimation and SVM.
The purpose of the invention can be realized by the following technical scheme:
a water supply network risk prediction method combining network kernel density estimation and SVM comprises the following steps:
step 1: collecting events and maintenance records of a water supply network, vertically mapping event points to corresponding damaged pipelines according to the coordinates of the event point positions and the recorded information, and establishing a pipe network-event topological model;
step 2: calculating the nuclear density of each pipe section unit event point by adopting a network nuclear density estimation method aiming at a pipe network-event topological model, representing the pipe section structure risk level, and further dividing the risk condition grade of each pipe section by adopting a natural boundary method;
and step 3: establishing a risk prediction model based on the SVM by taking the risk condition grade as a response vector, and determining the optimal parameters of the classifier by adopting a cross validation and grid search method;
and 4, step 4: and forecasting the structural risk condition of the pipe section of the water supply network by using the SVM risk forecasting model adopting the optimal parameters to form a water supply network risk distribution map.
Further, the step 1 comprises the following sub-steps:
step 101: vertically mapping the pipe network event points to corresponding damaged pipelines according to a spatial proximity principle and characteristic description of the damaged pipelines in the maintenance record;
step 102: the water supply network abstraction is defined as a planar undirected graph N ═ (V, L), where V and L represent node sets and link sets, respectively, and P ═ P is defined1,…,pmAnd the distance between any two event points on the N is measured by a Dijkstra network shortest path algorithm, and the establishment of the pipe network-event topology model is finished.
Further, the water supply network event and maintenance records in the step 1 comprise water supply network GIS data, water supply pipeline leakage, rupture and pipe burst event data and relevant maintenance records.
Further, the step 2 comprises the following sub-steps:
step 201: performing space discrete subdivision on the pipe network-event topological model established in the step 1 to obtain a sequence of linear pipe section units, and taking the name of the linear pipe section units as the line elements;
step 202: counting the number of event points on each line element, and calculating the kernel density value of each line element event point by adopting a network kernel density estimation method;
step 203: and according to the nuclear density value of each element, adopting a natural boundary method to divide different risk condition grades, wherein the risk levels are respectively marked as good, common and poor from low to high.
Further, in the step 202, a network kernel density estimation method is adopted to calculate the kernel density value of each line element event point, and a corresponding calculation formula is as follows:
where λ(s) is the event-point kernel density of the line element s, k (·) is the Gaussian kernel function, r is the bandwidth, d isisIs the shortest distance of the network of the thread i and the thread s, ciThe number of event points on line i, and n is the number of event points within the bandwidth r.
Further, the value of the bandwidth r is determined by k-order nearest neighbor distance between event points, and the corresponding calculation formula is as follows:
in the formula (d)ijIs the k-th nearest neighbor distance.
Further, the step 3 comprises the following sub-steps:
step 301: establishing a water supply network pipe section sample database based on the pipe network core density estimation and risk grade division results and associating the pipeline attribute characteristics;
step 302: training and learning the SVM classifier by using a sample training set in the database, and optimizing SVM classifier parameters by using a K-fold cross validation and network search method.
Further, the step 301 specifically includes: and aiming at the results of core density estimation and risk grade division based on the water supply network, carrying out digital processing on discrete features in the data associated with the pipeline attribute features by adopting independent thermal coding, carrying out normalization processing on the continuous features, dividing a sample training set and a test set in proportion after comprehensive processing is finished, and combining to form a water supply network pipe section sample database.
Further, the SVM classifier in step 302 is a multi-classification support vector machine based on radial basis function; the process of optimizing SVM classifier parameters by using K-fold cross validation and network search method specifically comprises the following steps:
taking values of the penalty coefficient C and the kernel function parameter gamma in a set range; training a classifier under the set C and gamma, verifying the classification accuracy, and selecting the C and gamma with the highest cross verification classification accuracy as the optimal parameter combination of the classifier; and when the multiple groups of C and gamma reach the expected classification accuracy rate at the same time, selecting the group of C and gamma with smaller C value as the SVM classifier parameters.
Further, the risk distribution map of the water supply network in the step 4 is displayed by color separation according to the risk condition of the structure of the water supply network pipe sections, namely the prediction result of the risk condition grade of each pipe section.
Compared with the prior art, the invention has the following advantages:
(1) according to the method, the nuclear density values of the event points on different linens are calculated through water supply network-event point mapping and network nuclear density estimation, so that the structural risk level of each pipe section can be described more accurately and quantitatively;
(2) the method combines the advantages of the network kernel density estimation and the support vector machine method, can more fully mine and analyze the pipe network event and the historical maintenance record, has more accurate and reasonable model prediction result, and supports event-driven pipe network risk dynamic prediction and evaluation.
Drawings
FIG. 1 is a block flow diagram of a water supply network risk prediction method incorporating network kernel density estimation and a support vector machine in accordance with the present invention;
FIG. 2 is a water supply network and event space distribution plot for an embodiment;
FIG. 3 is a nuclear density profile of an event in a water supply network according to an embodiment;
FIG. 4 is a diagram illustrating parameter optimization of an SVM classifier in an embodiment;
fig. 5 is a graph of a ROC model for predicting risk of a pipe network in an embodiment, where fig. 5(a) is a graph of a good ROC with a good risk level, fig. 5(b) is a graph of a general ROC with a general risk level, fig. 5(c) is a graph of a poor ROC with a poor risk level, and fig. 5(d) is a graph of a model ROC.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be obtained by a person skilled in the art without any inventive step based on the embodiments of the present invention, shall fall within the scope of protection of the present invention.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
A water supply network risk prediction method combining network nuclear density estimation and a support vector machine uses municipal water supply network data in Yangpu district of Shanghai city for example description.
The flow chart of the method is shown in figure 1.
Pipe network-event matching and network topology model construction:
the water supply network abstraction is defined as a planar undirected graph N ═ (V, L), where V and L represent node sets and link sets, respectively, and P ═ { P ═ is defined1,…,pmAnd the distance between any two event points on the N is measured by a Dijkstra network shortest path algorithm.
And collecting water supply network and event record data, including water supply network GIS data, event data of water supply pipeline leakage, rupture, pipe burst and the like and related maintenance records. And vertically mapping the event points to corresponding disease pipelines according to the event point position coordinates and the record information. FIG. 2 shows the results of water supply network-event matching in an embodiment.
Estimating the nuclear density of pipe network events and grading risk:
calculating the nuclear density of the water supply network event by adopting a network nuclear density estimation method, quantitatively describing the structural risk level of each thread unit,
wherein λ(s) is the event point kernel density of the line element s, the greater the density value, the higher the structural risk level of the line element, k (·) is the Gaussian kernel function, r is the bandwidth, disIs the shortest distance of the network of the thread i and the thread s, ciThe number of event points on line i, and n is the number of event points within the bandwidth r.
The value of the line element bandwidth r is determined by the nearest distance of k order between the event points,
in the formula (d)ijThe nearest distance of k order, namely the mean value of network distances from an event point to the kth nearest point; the k value determines the smoothness degree of the density distribution, the larger the k value is, the larger the bandwidth r is, and the smoother the nuclear density distribution of the event points of the water supply network is; and in the calculation process, a better bandwidth parameter is obtained by adjusting the k value.
Obtaining a sequence of the line elements (pipe sections) with the event nuclear density value through network nuclear density estimation, further dividing the risk level of each pipe section by adopting a Natural boundary method (Jenk's Natural Break), and respectively marking the values as good (1), general (2) and poor (3) according to the nuclear density value from low to high. FIG. 3 shows the results of nuclear density estimation and risk classification for water supply networks according to the embodiments.
Constructing a water supply network risk prediction model:
and establishing a water supply network pipe section sample database based on the pipe network core density estimation and risk grade classification results and associated pipeline attribute characteristics. Carrying out digital processing on discrete characteristics of a sample, such as pipeline material, joint form, road grade and the like by adopting One-Hot Encoder; and (4) carrying out normalization treatment on the continuous characteristics such as the age, the burial depth, the pipe diameter, the coordinates of the starting point and the stopping point and the like. According to the following steps: and 3, dividing the sample training set and the test set in a ratio.
Based on the basic theory of a Support Vector Machine (SVM), a multi-classification support vector machine is established by adopting a Radial Basis Function (RBF), and the nonlinear relation between characteristic variables of pipe sections of a water supply network and the risk conditions of the pipe sections is fitted. The SVM is a machine learning algorithm based on a statistical learning theory and proposed by Vapnik in 1995, can effectively solve the problem of high-dimensional model construction of data, and has good generalization capability. Training the SVM classifier by using a sample training set, and optimizing SVM classifier parameters including a penalty coefficient C and a kernel function parameter gamma by using a K-fold cross validation and network search method.
Taking values of the penalty coefficient C and the kernel function parameter gamma in a set range; training a classifier under the set C and gamma, verifying the classification accuracy, and selecting the C and gamma with the highest cross verification classification accuracy as the optimal parameter combination of the classifier; and when the multiple groups of C and gamma reach the expected classification accuracy rate at the same time, selecting the group of C and gamma with smaller C value as the optimal parameters of the SVM classifier.
As shown in fig. 4, in this embodiment, C is 100, γ is 10, and the accuracy of cross validation classification can reach 90% or more.
Predicting the dynamic risk distribution of the water supply network based on the model:
and predicting the risk condition of the water supply network by using the model after parameter optimization, and realizing event-driven dynamic risk prediction. Table 1 gives the test set classification results in the form of a confusion matrix. Wherein, 3335 pipelines with good structural condition accurately predict 3169 pipelines with the accuracy rate of 0.949 and the recall rate of 0.95; the number of pipes 1147 with general structural conditions is accurately predicted 932, the accuracy rate is 0.823, and the recall rate is 0.813; the structural condition is 494 poor pipelines, 436 pipelines are accurately predicted, the precision rate is 0.862, and the recall rate is 0.883. The average accuracy of the model was 0.878 and the recall was 0.882.
TABLE 1
And further analyzing classification prediction results of samples with different risk condition grades by adopting a ROC (Receiver operating characteristics curve). The Area under the ROC curve (AUC) can quantify the ROC curve, and the closer the AUC value is to 1, the better the classification effect is. Fig. 5(a), 5(b), 5(c) and 5(d) give ROC curves for different risk status levels. The AUC values of good and bad conditions are 0.97, the AUC value of general conditions is 0.94, the overall AUC value of sample data is 0.96, and the method has good classification prediction performance on the whole.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A water supply network risk prediction method combining network kernel density estimation and SVM is characterized by comprising the following steps:
step 1: collecting events and maintenance records of a water supply network, vertically mapping event points to corresponding damaged pipelines according to the coordinates of the event point positions and the recorded information, and establishing a pipe network-event topological model;
step 2: calculating the nuclear density of each pipe section unit event point by adopting a network nuclear density estimation method aiming at a pipe network-event topological model, representing the pipe section structure risk level, and further dividing the risk condition grade of each pipe section by adopting a natural boundary method;
and step 3: establishing a risk prediction model based on the SVM by taking the risk condition grade as a response vector, and determining the optimal parameters of the classifier by adopting a cross validation and grid search method;
and 4, step 4: and forecasting the structural risk condition of the pipe section of the water supply network by using the SVM risk forecasting model adopting the optimal parameters to form a water supply network risk distribution map.
2. A method for water supply network risk prediction in combination with network kernel density estimation and SVM as claimed in claim 1 wherein said step 1 comprises the sub-steps of:
step 101: vertically mapping the pipe network event points to corresponding damaged pipelines according to a spatial proximity principle and characteristic description of the damaged pipelines in the maintenance record;
step 102: the water supply network abstraction is defined as a planar undirected graph N ═ (V, L), where V and L represent node sets and link sets, respectively, and P ═ P is defined1,…,pmAnd the distance between any two event points on the N is measured by a Dijkstra network shortest path algorithm, and the establishment of the pipe network-event topology model is finished.
3. The method of claim 1, wherein the water supply network event and maintenance records of step 1 comprise water supply network GIS data, water supply pipe leakage, rupture, pipe burst event data, and associated maintenance records.
4. A method for water supply network risk prediction in combination with network kernel density estimation and SVM as claimed in claim 1 wherein said step 2 comprises the sub-steps of:
step 201: performing space discrete subdivision on the pipe network-event topological model established in the step 1 to obtain a sequence of linear pipe section units, and taking the name of the linear pipe section units as the line elements;
step 202: counting the number of event points on each line element, and calculating the kernel density value of each line element event point by adopting a network kernel density estimation method;
step 203: and according to the nuclear density value of each element, adopting a natural boundary method to divide different risk condition grades, wherein the risk levels are respectively marked as good, common and poor from low to high.
5. The method for predicting risk of a water supply network in combination with a network kernel density estimation and SVM of claim 4, wherein the step 202 of calculating the kernel density value of each line element event point by using the network kernel density estimation method corresponds to the following calculation formula:
where λ(s) is the event-point kernel density of the line element s, k (·) is the Gaussian kernel function, r is the bandwidth, d isisIs the shortest distance of the network of the thread i and the thread s, ciThe number of event points on line i, and n is the number of event points within the bandwidth r.
6. The method for predicting risk of a water supply network by combining network kernel density estimation and SVM as claimed in claim 5, wherein the value of the bandwidth r is determined by k-th order nearest neighbor distance between event points, and the corresponding calculation formula is as follows:
in the formula (d)ijIs the k-th nearest neighbor distance.
7. A method for water supply network risk prediction in combination with network kernel density estimation and SVM as claimed in claim 1 wherein said step 3 comprises the sub-steps of:
step 301: establishing a water supply network pipe section sample database based on the pipe network core density estimation and risk grade division results and associating the pipeline attribute characteristics;
step 302: training and learning the SVM classifier by using a sample training set in the database, and optimizing SVM classifier parameters by using a K-fold cross validation and network search method.
8. The method for predicting risk of a water supply network in combination with a network kernel density estimation and SVM of claim 7, wherein said step 301 comprises: and aiming at the results of core density estimation and risk grade division based on the water supply network, carrying out digital processing on discrete features in the data associated with the pipeline attribute features by adopting independent thermal coding, carrying out normalization processing on the continuous features, dividing a sample training set and a test set in proportion after comprehensive processing is finished, and combining to form a water supply network pipe section sample database.
9. The water supply network risk prediction method combining network kernel density estimation and SVM as claimed in claim 7 wherein the SVM classifier of step 302 is a radial basis function based multi-classification support vector machine; the process of optimizing SVM classifier parameters by using K-fold cross validation and network search method specifically comprises the following steps:
taking values of the penalty coefficient C and the kernel function parameter gamma in a set range; training a classifier under the set C and gamma, verifying the classification accuracy, and selecting the C and gamma with the highest cross verification classification accuracy as the optimal parameter combination of the classifier; and when the multiple groups of C and gamma reach the expected classification accuracy rate at the same time, selecting the group of C and gamma with smaller C value as the SVM classifier parameters.
10. The method as claimed in claim 1, wherein the risk profile of the water supply network in step 4 is displayed by color separation according to the prediction results of the risk level of each pipe section.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011497723.9A CN112434887B (en) | 2020-12-17 | 2020-12-17 | Water supply network risk prediction method combining network kernel density estimation and SVM |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011497723.9A CN112434887B (en) | 2020-12-17 | 2020-12-17 | Water supply network risk prediction method combining network kernel density estimation and SVM |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112434887A true CN112434887A (en) | 2021-03-02 |
CN112434887B CN112434887B (en) | 2023-04-07 |
Family
ID=74696672
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011497723.9A Active CN112434887B (en) | 2020-12-17 | 2020-12-17 | Water supply network risk prediction method combining network kernel density estimation and SVM |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112434887B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113191599A (en) * | 2021-04-12 | 2021-07-30 | 国家石油天然气管网集团有限公司华南分公司 | Pipeline risk level evaluation method and device based on support vector machine |
CN116467551A (en) * | 2023-06-20 | 2023-07-21 | 成都同飞科技有限责任公司 | Water supply pipe network leakage positioning method and system based on correlation coefficient |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040034612A1 (en) * | 2002-03-22 | 2004-02-19 | Nick Mathewson | Support vector machines for prediction and classification in supply chain management and other applications |
CN110288048A (en) * | 2019-07-02 | 2019-09-27 | 东北大学 | A kind of submarine pipeline methods of risk assessment of SVM directed acyclic graph |
CN110458313A (en) * | 2018-05-07 | 2019-11-15 | 中国科学院沈阳自动化研究所 | A kind of power consumer credit risk early-warning method and system based on SVM |
-
2020
- 2020-12-17 CN CN202011497723.9A patent/CN112434887B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20040034612A1 (en) * | 2002-03-22 | 2004-02-19 | Nick Mathewson | Support vector machines for prediction and classification in supply chain management and other applications |
CN110458313A (en) * | 2018-05-07 | 2019-11-15 | 中国科学院沈阳自动化研究所 | A kind of power consumer credit risk early-warning method and system based on SVM |
CN110288048A (en) * | 2019-07-02 | 2019-09-27 | 东北大学 | A kind of submarine pipeline methods of risk assessment of SVM directed acyclic graph |
Non-Patent Citations (2)
Title |
---|
林平 等: "基于路网的风险评估模型在警务实践中的应用", 《情报探索》 * |
闫宏明: "基于支持向量机的矿井瓦斯事故风险预测技术研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113191599A (en) * | 2021-04-12 | 2021-07-30 | 国家石油天然气管网集团有限公司华南分公司 | Pipeline risk level evaluation method and device based on support vector machine |
CN116467551A (en) * | 2023-06-20 | 2023-07-21 | 成都同飞科技有限责任公司 | Water supply pipe network leakage positioning method and system based on correlation coefficient |
CN116467551B (en) * | 2023-06-20 | 2023-08-25 | 成都同飞科技有限责任公司 | Water supply pipe network leakage positioning method and system based on correlation coefficient |
Also Published As
Publication number | Publication date |
---|---|
CN112434887B (en) | 2023-04-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108280553B (en) | Mountain torrent disaster risk zoning and prediction method based on GIS-neural network integration | |
WO2022135265A1 (en) | Failure warning and analysis method for reservoir dispatching rules under effects of climate change | |
CN105117602B (en) | A kind of metering device running status method for early warning | |
CN106650767B (en) | Flood forecasting method based on cluster analysis and real-time correction | |
CN109583520B (en) | State evaluation method of cloud model and genetic algorithm optimization support vector machine | |
CN104200288B (en) | A kind of equipment fault Forecasting Methodology based on dependency relation identification between factor and event | |
CN103793854A (en) | Multiple combination optimization overhead transmission line operation risk informatization assessment method | |
CN112434887B (en) | Water supply network risk prediction method combining network kernel density estimation and SVM | |
CN106934237A (en) | Radar cross-section redaction measures of effectiveness creditability measurement implementation method | |
CN105069537A (en) | Constructing method of combined air quality forecasting model | |
CN111611524A (en) | Gas risk assessment and safety supervision resource matching method and device | |
CN115063020B (en) | Multi-dimensional safety scheduling device and method for cascade hydropower station based on risk monitoring fusion | |
CN111525587B (en) | Reactive load situation-based power grid reactive voltage control method and system | |
CN112116160A (en) | Important power transmission channel disaster monitoring method based on optimized neural network improved cellular automaton | |
Liu et al. | A comprehensive risk analysis of transportation networks affected by rainfall‐induced multihazards | |
CN105912857A (en) | Selection and configuration method of distribution equipment state monitoring sensors | |
CN109214604A (en) | The haze calamity risk forecast method and system of power transmission and transforming equipment based on Adaboost | |
CN116579768B (en) | Power plant on-line instrument operation and maintenance management method and system | |
CN111680712B (en) | Method, device and system for predicting oil temperature of transformer based on similar time in day | |
CN108615098A (en) | Water supply network pipeline burst Risk Forecast Method based on Bayesian survival analysis | |
Zhou et al. | Structural health monitoring of offshore wind power structures based on genetic algorithm optimization and uncertain analytic hierarchy process | |
Wumaier et al. | Short-term forecasting method for dynamic traffic flow based on stochastic forest algorithm | |
Zhang et al. | Real-time burst detection based on multiple features of pressure data | |
CN113486950B (en) | Intelligent pipe network water leakage detection method and system | |
CN113191582B (en) | Road torrential flood vulnerability evaluation method based on GIS and machine learning |
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 |