CN108376310A - Building fire risk class appraisal procedure - Google Patents
Building fire risk class appraisal procedure Download PDFInfo
- Publication number
- CN108376310A CN108376310A CN201810119077.9A CN201810119077A CN108376310A CN 108376310 A CN108376310 A CN 108376310A CN 201810119077 A CN201810119077 A CN 201810119077A CN 108376310 A CN108376310 A CN 108376310A
- Authority
- CN
- China
- Prior art keywords
- building
- cluster
- fire
- fire risk
- data
- 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
- 238000000034 method Methods 0.000 title claims abstract description 75
- 238000012502 risk assessment Methods 0.000 claims abstract description 45
- 238000010276 construction Methods 0.000 claims abstract description 41
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 28
- 238000003066 decision tree Methods 0.000 claims abstract description 23
- 238000013461 design Methods 0.000 claims abstract description 17
- 238000010801 machine learning Methods 0.000 claims abstract description 17
- 238000007635 classification algorithm Methods 0.000 claims abstract description 15
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000011156 evaluation Methods 0.000 claims description 13
- 241000288113 Gallirallus australis Species 0.000 claims description 6
- 235000013399 edible fruits Nutrition 0.000 claims description 5
- 238000011524 similarity measure Methods 0.000 claims description 4
- 238000012360 testing method Methods 0.000 claims description 4
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 230000002265 prevention Effects 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 17
- 238000005516 engineering process Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 239000000463 material Substances 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000004880 explosion Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 239000000047 product Substances 0.000 description 2
- 208000001836 Firesetting Behavior Diseases 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 239000004035 construction material Substances 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 238000012407 engineering method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000004513 sizing Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000013517 stratification Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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
- G06Q50/10—Services
- G06Q50/26—Government or public services
- G06Q50/265—Personal security, identity or safety
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Entrepreneurship & Innovation (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Security & Cryptography (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Primary Health Care (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of building fire risk class appraisal procedures, including obtain the corresponding default risk assessment index of the target construction；Carry out default value pretreatment；It is cluster number of clusters according to preset design fire risk class, clustering processing acquisition cluster result is carried out to carrying out the pretreated building data by k prototypes clustering algorithm；Classification prediction is carried out on the cluster result by least one machine learning classification algorithm, determines whether the classification prediction result that the cluster result carries out above reaches preset accuracy；According to cluster result classification building data and the preset design fire risk class, graded to the fire risk of the target construction using J48 decision trees, output level of building fire risk decision rule；When there is new building to build up, the level of building fire risk of building can be assessed by decision rule, to help fire-fighting supervision department to realize Precision management, emphasis prevention and control.
Description
Technical field
The present invention relates to public fire protection security technology areas, and in particular to a kind of building fire risk class assessment side
Method.
Background technology
The management of fire safety work of most of social unit for many years be generally " responsibility on wall, checked on paper, fire it is hidden
Trouble is important but not urgent thing " state, the main reason is that the papery fire archieves due to generally using and emphasis list
The management of fire safety tools such as position residence management system can not effectively assist social unit to implement security against fire responsibility, improve
Security against fire self-service supportability.
In addition, since the scale of present Tall Office Building is increasing, construction is complicated, in construction style, style, function etc.
Aspect is different, and with new technology, the application of new material and new building structure so that building space construction area increases
Greatly, electrical equipment increases, and fire hazard becomes larger.Modern architecture and traditional architecture construction material, space size, structure type,
Auxiliary facility is very different using the various aspects of function, its fire is made to have following features：The intensity of a fire is swift and violent, and sprawling is rapid；
Building interior is complicated, and evacuating personnel is difficult；Fire fighting and rescue difficulty is big.
Therefore, for determining and elimination fire hazard, it is necessary to study fire safety evaluation technology, improve fire assessment body
System.However, current fire safety evaluation method evaluation index is single, starts with mostly from one-sided factor, do not consider
Influence of the different aspect factor to evaluation result, it is not accurate enough and comprehensive.
Currently, both at home and abroad associated mechanisms or it is personal developed many Fire risk assessment methods, from the angle of methodology,
It is broadly divided into qualitative method, semiquantitative method and sizing technique.
Qualitative method in Fire risk assessment method is mainly used for identifying that least favorable event of fire, main method have safety
Checking list method, classification of risks indicator method (Risk Category Indicator Method), risk assessment inspection table of setting on fire
Method (Arson Risk Assessment Checklist) etc..Such methods are mainly to comment with the pertinent regulations of specification or regulations
Sentence foundation, fire risk feature is determined in a simple manner decoupled, to take mandatory mode to solve the problems, such as security against fire.
Semi-quantitative method in Fire risk assessment method is mainly used for determining the relative risk of fire, also referred to as fire wind
Dangerous staging.Compare the risky value matrix method of typical method (Risk Value Matrix Method), fire Safety Assessment
System (Fire Safety Evaluation System) method, specific commercial property assessment table (Specific Commercial
Property Evaluation Schedule) method, Dow fire explosion index (Dow Fire and Explosion
Index) method, fire size class stratification (Hierarchical Approach), SIA81-Gretener methods, fire risk index
Method (Fire Risk Index Method) etc..For semiquantitative method due to the feature that its is fast and simple, structuring is strong, application is more wide
It is general.Disadvantage is however that for specific type building, process exploitation, do not have universality, especially factor selection and weight is true
It is fixed；In addition, evaluation result and the accumulation of the know-how, experience and relevant historical data of method developer etc. are closely related, tool
There is certain subjectivity.
Quantitative approach in Fire risk assessment method is mainly used for determining the practical risk of fire, by specific false
If, data and mathematical, retrospect generates quantized result and simultaneously reflects potential fire risk distribution, also referred to as probabilistic method.Than
More typical method has：
CRISP (Computation of Risk Indices by Simulation Procedures) method,
FiRECAM (Fire Risk Evaluation and Cost Assessment Model) method,
BFSEM methods (The Building Fire Safety Engineering Method),
Fire evaluate and risk assessment (Fire Evaluation and Risk Assessment) method,
Risk analysis event tree method,
β reliable indicator fire risk assessment methods etc..
Such methods advantage is that result reflects the uncertain essence of risk, the disadvantage is that need a large amount of data information and when
Between.
It is not ten thousand that advantage and disadvantage by analyzing the above all kinds of methods, which can be seen that any type Fire risk assessment method,
Can, it is all based on specific application and background and generates, there is the specific scope of application, the particular problem being directed at it and field
There is preferable applicability in scape application range.In concrete practice, in the case where existing method is unable to meet demand, evaluator
The difference according to purpose of appraisals, assessment object, evaluation stage etc. is often needed, and the moneys such as human and material resources, financial resources that can be grasped
Source situation invents suitable Fire risk assessment method.
How the fire risk state of building is effectively assessed, and how to evaluate building fire risk class is always fire-fighting
The key points and difficulties of management of safe operation.Therefore, it is necessary to which it is next to provide a kind of new building fire risk class appraisal procedure
Solve above-mentioned technical problem.
Invention content
The main purpose of the present invention is to provide a kind of simple and effective building fire risk class appraisal procedures.
To achieve the above object, the present invention provides a kind of building fire risk class appraisal procedure, including step：
According to the data of target construction, the corresponding default risk assessment index of the target construction is obtained；
Default value pretreatment is carried out to the default risk assessment index of acquisition, obtains pretreated building number
According to；
It is cluster number of clusters according to preset design fire risk class, by k- prototypes clustering algorithm to carrying out the pre- place
Building data after reason carry out clustering processing and obtain cluster result；
Classification prediction is carried out on the cluster result by least one machine learning classification algorithm, is determined described poly-
Whether the classification prediction result that class result carries out above reaches preset accuracy；
When the classification prediction result carried out on the cluster result reaches preset accuracy, tied according to cluster
The other building data of fruit and the preset design fire risk class, using J48 decision trees to the target structures
The fire risk of object is graded.
Further, the data according to target construction obtain the corresponding default risk of the target construction and comment
The step of estimating index, including：
According to the data dimension of the data of target construction, the default of the corresponding data dimension of the target construction is obtained
Risk assessment index.
Further, the default risk assessment index of the corresponding data dimension includes at least one in following attribute value
Kind：
8 numerical attributes or Continuous valued attributes, including：A1 builds age, A2 index beds area, the total face of A3 refuge stories
The same time can accommodate maximum number, A7 building heights and the undergrounds A8 in product, A4 aerial layers area, A5 construction areas, A6 building
Level is accumulated；And
10 symbol attributes or discrete value attribute, including：The grounds the B1 number of plies, the undergrounds the B2 number of plies, B3 Fire lifts number, B4
It takes refuge layer number, B5 buildings classification, B6 building structure, the whether voluminous power of B7, B8 fire resistance ratings, B9 construction status value and B10 rows
Administrative division domain.
Further, the default risk assessment index of described pair of acquisition carries out default value pretreatment, is pre-processed
The step of rear building data, including：
When the attribute value of the default risk assessment index is discrete type, default value is positioned as the default risk and is commented
Estimate the highest attribute value of frequency of occurrence in index；And/or
When the attribute value of the default risk assessment index is continuous type, default value is positioned as the default risk and is commented
Estimate the mean value of the attribute value in index.
Further, the preset design fire risk class is 3 grades or 4 grades.
Further, it is described according to preset design fire risk class be cluster number of clusters, pass through k- prototype clustering algorithms
The step of clustering processing obtains cluster result is carried out to carrying out the pretreated building data, including：
It is cluster number of clusters according to preset design fire risk class, adjustment k- prototypes cluster sample similarity measure formulas
In the numerical value of undetermined parameter τ that contains, pass through k- prototypes clustering algorithm and carried out to carrying out the pretreated building data
Clustering processing.
Further, it is described carry out classifying on the cluster result by least one machine learning classification algorithm it is pre-
It surveys, determines the step of whether the classification prediction result that the cluster result carries out above reaches preset accuracy, including：
The machine learning classification algorithm in 8 kinds of WEKA kits is used on the cluster result to being added to classification
The building data of information carry out class test；
Wherein, the machine learning classification algorithm includes：It is Bayesian network, random tree, naive Bayesian, SMO, random gloomy
Woods, J48, NBTree and RBF net.
Further, described to reach preset accuracy to reach 90% or more accuracy.
Further, the default risk assessment index of the corresponding data dimension includes：
5 Continuous valued attributes：A1 building ages, A2 index beds area, A5 construction areas, A6 buildings interior same time can hold
Receive maximum number, A7 building heights；And
1 discrete type attribute：B3 Fire lift numbers.
Further, [A1 build age≤15.5 year] and [building height≤64.5 meter A7] and [in A6 buildings same
Time can accommodate the people of maximum number≤5900] and [A2 index beds area≤60519 square metre] and [A5 construction areas≤
189848.5 square metres], it is determined that [fire risk=light grade]；
At [the A1 building ages>15.5] and [building height≤72.45 meter A7] and [A6 building in the same time can accommodate
The people of maximum number≤7600], it is determined that [fire risk=middle rank]；
At [the A1 building ages>15.5] and [building height≤72.45 meter A7] and [A6 building in the same time can accommodate
Maximum number>7600 people], it is determined that [fire risk=serious grade]
In [A1 builds age≤15.5 year] and [building height≤64.5 meter A7] and [the same time can accommodate in A6 buildings
Maximum number>5900 people], it is determined that [fire risk=serious grade]；
At [the A1 building ages>15.5] and [A7 building heights>72.45 meters] and [Fire lift number=1 or 2],
Then determine [fire risk=serious grade].
In the inventive solutions, it by the data according to target construction, obtains the target construction and corresponds to
Default risk assessment index；Default value pretreatment is carried out to the default risk assessment index of acquisition, after being pre-processed
Building data；It is cluster number of clusters according to preset design fire risk class, by k- prototypes clustering algorithm to carrying out institute
It states pretreated building data and carries out clustering processing acquisition cluster result；Existed by least one machine learning classification algorithm
Classification prediction is carried out above the cluster result, determines whether the classification prediction result that the cluster result carries out above reaches pre-
If accuracy；When the classification prediction result carried out on the cluster result reaches preset accuracy, according to poly-
The building data of class result classification and the preset design fire risk class, using J48 decision trees to the target
The fire risk of building is graded, and level of building fire risk decision rule is outputed；When there is new building to build up, can lead to
Decision rule is crossed to assess the level of building fire risk of building, Precision management, emphasis are anti-to help fire-fighting supervision department to realize
Control.
Description of the drawings
Fig. 1 is the flow chart of the building fire risk class appraisal procedure in one embodiment of the invention；
Fig. 2 is the corresponding k- prototypes cluster result in τ=0.0009 on building data set after pre-processing；
Fig. 3 is the corresponding k- prototypes cluster result in τ=0.0010 on building data set after pre-processing；
Fig. 4 is the decision tree that J48 algorithms generate on τ=0.0009 and=3 corresponding cluster results；
Fig. 5 is the decision tree that J48 algorithms generate on the cluster result corresponding with K=4 of τ=0.0009；
Fig. 6 is the decision tree that J48 algorithms generate on the cluster result corresponding with K=3 of τ=0.0010；
Fig. 7 is the decision tree that J48 algorithms generate on the cluster result corresponding with K=4 of τ=0.0010.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific implementation mode
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation describes, it is clear that described embodiment is only a part of the embodiment of the present invention, instead of all the embodiments.Base
Embodiment in the present invention, those of ordinary skill in the art obtained without creative efforts it is all its
His embodiment, shall fall within the protection scope of the present invention.
It is to be appreciated that institute's directional instruction (such as up, down, left, right, before and after, transverse direction, diameter in the embodiment of the present invention
To, it is horizontal, vertical ...) be only used for explaining under a certain particular pose (as shown in the picture) the relative position pass between each component
System, motion conditions etc., if the particular pose changes, directionality instruction also correspondingly changes correspondingly.
In addition, the description for being such as related to " first ", " second " in the present invention is used for description purposes only, and should not be understood as
It indicates or implies its relative importance or implicitly indicate the quantity of indicated technical characteristic.Define as a result, " first ",
The feature of " second " can explicitly or implicitly include at least one of the features.In the description of the present invention, " multiple " contain
Justice is at least two, such as two, three etc., unless otherwise specifically defined.
In the present invention unless specifically defined or limited otherwise, term " connection ", " fixation " etc. shall be understood in a broad sense,
For example, " fixation " may be a fixed connection, it may be a detachable connection, or integral；It can be mechanical connection, can also be
Electrical connection；It can be directly connected, can also can be indirectly connected through an intermediary the connection inside two elements or two
The interaction relationship of a element, unless otherwise restricted clearly.It for the ordinary skill in the art, can basis
Concrete condition understands the concrete meaning of above-mentioned term in the present invention.
In addition, the technical solution between each embodiment of the present invention can be combined with each other, but must be general with this field
Logical technical staff can be implemented as basis, will be understood that when the combination of technical solution appearance is conflicting or cannot achieve this
The combination of technical solution is not present, also not the present invention claims protection domain within.
Referring to FIG. 1, to provide a kind of building fire risk class appraisal procedure 101 in one embodiment of the invention
Method flow diagram, the building fire risk class appraisal procedure 101 include the following steps：
Step S10 obtains the corresponding default risk assessment of the target construction and refers to according to the data of target construction
Mark.
It is possible to further which according to the data dimension of the data of target construction, it is corresponding to obtain the target construction
The default risk assessment index of data dimension.
The default risk assessment index of the corresponding data dimension includes at least one of following attribute value：
8 numerical attributes or Continuous valued attributes, including：A1 builds age, A2 index beds area, the total face of A3 refuge stories
The same time can accommodate maximum number, A7 building heights and the undergrounds A8 in product, A4 aerial layers area, A5 construction areas, A6 building
Level is accumulated；And
10 symbol attributes or discrete value attribute, including：The grounds the B1 number of plies, the undergrounds the B2 number of plies, B3 Fire lifts number, B4
It takes refuge layer number, B5 buildings classification, B6 building structure, the whether voluminous power of B7, B8 fire resistance ratings, B9 construction status value and B10 rows
Administrative division domain.
Specifically, in an optional embodiment, the default risk assessment index of the corresponding data dimension includes：
5 Continuous valued attributes：A1 building ages, A2 index beds area, A5 construction areas, A6 buildings interior same time can hold
Receive maximum number, A7 building heights；And
1 discrete type attribute：B3 Fire lift numbers.
Step S20 carries out default value pretreatment to the default risk assessment index of acquisition, obtains pretreated build
Build object data.
Further, it in the step 20, when the attribute value of the default risk assessment index is discrete type, will lack
Province's value is positioned as the highest attribute value of frequency of occurrence in the default risk assessment index；
When the attribute value of the default risk assessment index is continuous type, default value is positioned as the default risk and is commented
Estimate the mean value of the attribute value in index.
Due to containing a large amount of default value in the building initial data of acquisition, existing machine learning algorithm can not be direct
Effectively handle the data containing default value, it is therefore necessary to be pre-processed to original building data, to ensure data
Completeness.The building data containing default value are pre-processed using statistical method, the default value in data is carried out
Estimation and supplement.
Specifically, containing a large amount of default value, including discrete and continuous default value in original building data, herein
Discrete and continuous default value is estimated and supplemented respectively using following method.Assuming that attribute AmIt is right on data set D
The value answered is { x1m, x2m... ..., xNm, such as x1mFor a default value, work as AmFor discrete value attribute when, Wo Menling
x1m={ x2m... ..., xNmPattern (mode), i.e. { x2m... ..., xNmIn the maximum attribute value of frequency of occurrence；
Work as AmFor Continuous valued attributes when, we enable x1m={ x2m... ..., xNmMean value (mean), i.e.,
For example, for Category Attributes value：
{ #, Sunny, Sunny, Sunny, Overcast, Overcast, Rainy, Rainy, Rainy, Rainy } is used
Rainy makes up default value " # ", because the frequency that Rainy occurs in the data set is 4；
For continuous property { #, 90,78,96,80,70,65,95,70,80,70,90,75,80 }, useMake up default value " # ".
Step S30 is cluster number of clusters according to preset design fire risk class, by k- prototypes clustering algorithm to carrying out
The pretreated building data carry out clustering processing and obtain cluster result.
Further, in step s 30, it is cluster number of clusters according to preset design fire risk class, for example, setting fire
Calamity risk class is divided into three-level (light grade, middle rank and serious grade) or level Four (I grades, II grades, III level and IV grades), then corresponds to
The number of cluster be 3 or 4, the numerical value of undetermined parameter τ contained in adjustment k- prototypes cluster sample similarity measure formulas leads to
It crosses k- prototypes clustering algorithm and carries out clustering processing to carrying out the pretreated building data.
In order to determine the level of building fire risk of building, by find building data in general character contact, to building
Data carry out sub-clustering, ensure that the similitude of data in cluster is maximum, the similitude of data is minimum between cluster.This appraisal procedure is former using k-
Type cluster clusters the building data of mixed attributes, to effectively distinguish the fire risk of different building data.
Specifically, building data after obtaining pretreatment, poly- by using k- prototypes (k-prototypes)
Building data are divided into different clusters by class algorithm, and the determination of cluster number can be adjusted according to the needs of practical application, if
Determine level of building fire risk and is divided into three-level (light grade, middle rank and serious grade) or level Four (I grades, II grades, III level and IV grades), it is right
The number for the cluster answered is 3 or 4.K- prototypes are clustered containing there are one undetermined parameter τ in sample similarity measure formulas, which uses
In adjustment Continuous valued attributes and the influence measured to sample similarity of discrete value attribute, in analysis below we demonstrate τ=
0.0009 and τ=0.0010 pair k- prototype cluster results influence.
Under request in person together combine Fig. 2 and Fig. 3, different parameters on building data set after pre-processing are set forth
Corresponding k- prototypes cluster result.It is selected that age, A5 construction areas and this 3 dimensions of A7 building heights are built to show with A1
Cluster result, wherein Fig. 2-a1,2-a2,2-a3,2-a4 is corresponding (τ=0.0009, K=3) cluster result, and Fig. 2-a1 are
(A1, A5, A7) dimension cluster result schematic diagram, Fig. 2-a2 be (A1, A5) dimension cluster result schematic diagram, Fig. 2-a3 be (A1,
A7) dimension cluster result schematic diagram, Fig. 2-a4 are (A5, A7) dimension cluster result schematic diagram；Fig. 2-b1,2-b2,2-b3,2-b4
For correspondence (τ=0.0009, K=4) cluster result, Fig. 2-b1 are (A1, A5, A7) dimension cluster result schematic diagram, and Fig. 2-b2 are
(A1, A5) dimension cluster result schematic diagram, Fig. 2-b3 are (A1, A7) dimension cluster result schematic diagram, and Fig. 2-b4 tie up for (A5, A7)
Spend cluster result schematic diagram；
Fig. 3-a1,3-a2,3-a3,3-a4 are corresponding (τ=0.0010, K=3) cluster result, Fig. 3-a1 be (A1, A5,
A7) dimension cluster result schematic diagram, Fig. 3-a2 are (A1, A5) dimension cluster result schematic diagram, and Fig. 3-a3 are poly- for (A1, A7) dimension
Class result schematic diagram, Fig. 3-a4 are (A5, A7) dimension cluster result schematic diagram；Fig. 3-b1,3-b2,3-b3,3-b4 is corresponding (τ
=0.0010, K=4) cluster result, Fig. 3-b1 are (A1, A5, A7) dimension cluster result schematic diagram, and Fig. 3-b2 tie up for (A1, A5)
Cluster result schematic diagram is spent, Fig. 3-b3 are (A1, A7) dimension cluster result schematic diagram, and Fig. 3-b4 are (A5, A7) dimension cluster knot
Fruit schematic diagram.
By taking (τ=0.0010, K=3) corresponding result of Fig. 3-(a) displays as an example, it is assumed that k- prototype clustering algorithms obtain
3 clusters be respectively C1, C2 and C3, the number containing sample is respectively 10829,3143 and 397 in each cluster, clusters the time
About 3.4389 seconds, we can be found that building data are clearly gathered from (A1, A5, A7) dimension picture of Fig. 3-(a)
At 3 clusters.By the way that by the comparison of Fig. 3-(a) (τ=0.0009, K=3) cluster results corresponding with Fig. 2-(a), we can be intuitive
Ground finds that (τ=0.0010, K=3) corresponding cluster result is better than the cluster result exhibition in τ=0.0009, K=3,2 dimension
Show and clearly reflect this point, by taking the comparison in (A1, A5) dimension as an example：Sample weight in (A1, A5) dimension in Fig. 2-(a)
It is folded, and sample can clearly divide in (A1, A5) dimension in Fig. 3-(a)；Same situation is present in Fig. 2-(a) and Fig. 3-(a)
Cluster result in (A1, A7) dimension is shown.
In addition, in correspondence (τ=0.0010, K=4) cluster result that Fig. 3-(b) is shown, it is assumed that k- prototype clustering algorithms
4 obtained clusters are respectively C1, C2, C3 and C4, and the number containing sample is respectively 10522,391,279 and in each cluster
3177, the cluster time is about 9.9067 seconds；
In correspondence (τ=0.0009, K=3) cluster result that Fig. 2-(a) is shown, it is assumed that k- prototype clustering algorithms obtain
3 clusters be respectively C1, C2 and C3, the number containing sample is respectively 10757,3134 and 478 in each cluster, clusters the time
About 9.6793 seconds；
In correspondence (τ=0.0009, K=4) cluster result that Fig. 2-(b) is shown, it is assumed that k- prototype clustering algorithms obtain
3 clusters be respectively C1, C2, C3 and C4, the number containing sample is respectively 9316,2160,392 and 2501 in each cluster, is gathered
The class time is about 16.7103 seconds
Step S40 carries out classification prediction, really by least one machine learning classification algorithm on the cluster result
Whether the classification prediction result that the fixed cluster result carries out above reaches preset accuracy.
Further, it in the step S40, can be tied in the cluster by least one machine learning classification algorithm
Classification prediction is carried out above fruit, determines whether the classification prediction result that the cluster result carries out above reaches preset accuracy
The step of, including：
The machine learning classification algorithm in 8 kinds of WEKA kits is used on the cluster result to being added to classification
The building data of information carry out class test；
Wherein, the machine learning classification algorithm includes：It is Bayesian network, random tree, naive Bayesian, SMO, random gloomy
Woods, J48, NBTree and RBF net.
Further, described to reach preset accuracy to reach 90% or more accuracy.
For the cluster result obtained using k- prototype clustering algorithms, need to verify its accuracy and validity.Because according to
The cluster result obtained according to different parameters (for example, equilibrating factor etc. in the number of cluster, k- Prototype Algorithms) is to finally building
The influence of object risk assessment is very big, and accurately and effectively cluster result can ensure to obtain more credible, reliable evaluation result.
This appraisal procedure carries out classification prediction using different machine learning classification algorithms on cluster result, most using nicety of grading
High cluster data is as final cluster result.
Specifically, in a specific example：
This step use the Supervised machine learning algorithm in 8 kinds of WEKA kits to be added to classification information (although
The semantic information of classification is unknown at this time) building data carry out class test：Bayesian network, random tree, naive Bayesian,
SMO, random forest, J48, NBTree and RBF net.
Tables 1 and 2 gives classification accuracy rate (packet of this 8 kinds of sorting algorithms on the corresponding cluster result of different parameters
Include the accuracy on each cluster and the accuracy on entire data set), the mode for using 10 10 folding cross validations obtains
Experimental result.The result shown in the Tables 1 and 2 can be found that k-prototypes clusters obtain very accurate cluster and tie
Fruit, 8 graders of all selections (τ=0.0009, K=3), (τ=0.0009, K=4), (τ=0.0010, K=3), with
And (τ=0.0010, K=4) nearly all obtains 90% or more classification accuracy rate, especially random forest and J48 decision trees, it is whole
All 99% or more, this shows to cluster to the building number without classification information by k- prototypes nicety of grading on a data set
According to mark be effective, the result is that believable.
The outstanding classification performance on building data set in view of J48 decision trees, calculated (τ=0.0009, K=3),
The attribute weight of (τ=0.0009, K=4), (τ=0.0010, K=3) and (τ=0.0010, K=4) corresponding 4 data sets
It spends, as shown in table 3.
The calculating of Attribute Significance is to use the mode of information gain, can pass through the Attribute in WEKA kits
evaluator:Weka.attributeSelection.InfoGainAttributeEval is (using information gain as the list of foundation
One attribute evaluation device) function realization.
With data set D={ (xn,yn)|xn=(xn1,xn2... ..., xnM),yn∈ { 1,2 ... ..., C }, n=1,2 ... ...,
N } for, using following formula computation attribute AmThe information gain of (m ∈ { 1,2 ... ..., m })：
Gain (D, Am)=Before_Ent (D)-After_Ent (D, Am),
The comentropy of data set D is before wherein attribute Am being used to divide：
Nc indicates the number of C class samples in data set D；
Assuming that attribute AmValue beSo use AmData set D after division
Comentropy is：
| D | indicate the mould of data set D, i.e. the number containing sample in data set D,Indicate data setMould,Indicate data setIn c class samples number.
Table 1：The classification accuracy rate of different classifications device on the corresponding k-prototypes cluster results in τ=0.0009
Table 2：The classification accuracy rate of different classifications device on the corresponding k-prototypes cluster results in τ=0.0010
Table 3：The corresponding Attribute Significance of different cluster results
Step S50, when the classification prediction result carried out on the cluster result reaches preset accuracy, according to band
There are the building data of cluster result classification and the preset design fire risk class, using J48 decision trees to described
The fire risk of target construction is graded.
Further, for having gathered class and having verified the building data of cluster result accuracy, its fire how is determined
Risk class is a critical issue.This appraisal procedure extracts knowledge rule, root using J48 decision tree techniques from cluster data
It further grades to the fire risk of building according to these knowledge rules.
Specifically, please also refer to Fig. 4-7, include with the default risk assessment index of the corresponding data dimension：5
Continuous valued attributes：A1 building the age, A2 index beds area, A5 construction areas, A6 building in the same time can accommodate maximum number,
A7 building heights；And 1 discrete type attribute：B3 Fire lift numbers；For illustrate.
Fig. 4：Decision tree that J48 algorithms generate on the cluster result corresponding with K=3 of τ=0.0009 (32 nodes,
In 19 leaf nodes, contribute 0.70 second)；
Fig. 5：Decision tree that J48 algorithms generate on the cluster result corresponding with K=4 of τ=0.0009 (36 nodes,
In 21 leaf nodes, contribute 0.80 second)；
Fig. 6：Decision tree that J48 algorithms generate on the cluster result corresponding with K=3 of τ=0.0010 (22 nodes,
In 12 leaf nodes, contribute 0.79 second)；
Fig. 7：Decision tree that J48 algorithms generate on the cluster result corresponding with K=4 of τ=0.0010 (totally 23 nodes,
Wherein 12 leaf nodes are contribute 0.84 second).
After obtaining the building data with classification, the mark of class is Cluster-1, Cluster-2, Cluster-
3 or Cluster-1, Cluster-2, Cluster-3, Cluster-4, for differentiate Cluster-1, Cluster-2,
Concrete meaning representated by Cluster-3 or Cluster-4, representative level of building fire risk.This step is generated in (τ
=0.0009, K=3), (τ=0.0009, K=4), (τ=0.0010, K=3) and (τ=0.0010, K=4) corresponding 4
J48 decision trees on a data set, as shown in Figure 4-Figure 7.Using these decision trees come it is final determine Cluster-1,
Level of building fire risk representated by Cluster-2, Cluster-3 or Cluster-4.Intuitively comparison it can be found that τ=
Corresponding two decision trees are simple for 0.0010 corresponding two decision tree ratios τ=0.0009.Therefore, by taking Fig. 6 and Fig. 7 as an example in detail
Carefully illustrate the specifying information for including in these decision trees：The interior nodes wherein set indicate the conditional attribute of data set；Leaf node
Classification is represented, the triple [a, b, c] or four-tuple [a, b, c, d] on leaf node represent of different cluster samples in the node
Number；Path of each from root node to leaf node represents a classifying rules.
Decision tree shown in fig. 6 contains 22 nodes altogether, wherein 12 leaf nodes, the time of contributing is about 0.79 second, per cluster
The number of sample is respectively | C1 |=10829, | C2 |=3143 and | C3 |=397.Three are chosen from root node to representing difference
The path of the leaf node of cluster, as follows：
R1：If [building age≤15.5 year] and [building height≤64.5 meter] and [the same time can hold in building
Receive the people of maximum number≤5900] and [index bed area≤60519 square metre] and [construction area≤189848.5 square metre], that
[cluster classification=Cluster-1] (leaf node sample distribution is [10808,0,56])；
R2：If [the building age>15.5] and [building height≤72.45 meter] and [the same time can hold in building
Receive the people of maximum number≤7600], then [cluster classification=Cluster-2] (leaf node sample distribution is [0,3139,7])；
R3：If [the building age>15.5] and [building height≤72.45 meter] and [the same time can hold in building
Receive maximum number>7600 people], then [cluster classification=Cluster-3] (leaf node sample distribution is [0,1,6]).
In conjunction with fire risk common sense, may determine that according to this three rule：The light grade fire risks of Cluster-1=,
Cluster-2=middle ranks fire risk and the serious grade fire risks of Cluster-3=.The reason is as follows that：
[1], for regular R1, when again building is relatively new, height is very low, galleryful considerably less, index bed area and is built
Build area all very little when, can directly infer such building occur fire risk it is smaller, be light grade risk；
[2], regular R3 can be pushed away when building is older, the slightly lower while galleryful of height is again very much
The very risky of fire occurs for disconnected such building, is serious grade risk；
[3], after being inferred to the fire risk of light grade and serious grade, remaining sample is then only intermediate fire wind
Danger.In fact, for regular R2, when building is older, highly slightly lower while galleryful passes through again not counting when too many
With the comparison of regular R1, it is believed that the risk that fire occurs for such building is higher than light grade, for middle rank.
Decision tree shown in Fig. 7 contains 23 nodes altogether, wherein 12 leaf nodes, the time of contributing is about 0.84 second, per cluster
The number of sample is respectively | C1 |=10522, | C2 |=391, | C3 |=279 and | C4 |=3177.Four are chosen from root node
To the path for the leaf node for representing different clusters, as follows：
T1：If [building age≤14.5 year] and [building height≤64.5 meter] and [the same time can hold in building
Receive the people of maximum number≤5900] and [index bed area≤60519 square metre] and [construction area≤189848.5 square metre], that
[cluster classification=Cluster-1] (leaf node sample distribution is [10501,52,0,0])；
T2：If [14.5<Build age≤27.5 year] and [building height>70 meters] and [same time in building
Maximum number can be accommodated>550 people], then [cluster classification=Cluster-2] (leaf node sample distribution is [0,21,0,1])；
T3：If [the building age>27.5], then [cluster classification=Cluster-3] (leaf node sample distribution be [0,
0,279,0])；
T4：If [14.5<Build age≤27.5 year] and [building height≤70 meter] and [same time in building
The people of maximum number≤5500 can be accommodated], then [cluster classification=Cluster-4] (leaf node sample distribution be [0,5,0,
3168])。
In conjunction with fire risk common sense, may determine that according to this four rule：Cluster-1=I grades of fire risks,
Cluster-4=II grades of fire risks, Cluster-2=III grades of fire risks and Cluster-3=IV grades of fire risks.It is former
Because as follows：
[1], for rule T 1, when again building is relatively new, height is very low, galleryful considerably less, index bed area and is built
Build area all very little when, can directly infer such building occur fire risk it is smaller, be I grades of risks；
[2], it for rule T 4, by [building age≤14.5 year] and [building height≤64.5 meter] with T1 and [builds
The people of maximum number≤5900 can be accommodated by building the same time in object] comparison, [14.5<Build age≤27.5 year] and [building is high
≤ 70 meters of degree] and the fire risk of [the same time can accommodate the people of maximum number≤5500 in building] corresponding building want high
Some because the age of building be slightly above rule T 1, it is inferred that such building be II grades of risks；
[3], for rule T 2, pass through [14.5 years with T4<Build age≤27.5 year] and [building height≤70 meter]
And [the same time can accommodate the people of maximum number≤5500 in building] comparison, [14.5<Building age≤27.5 year] and [build
Build height>70 meters] and [the same time can accommodate maximum number in building>550 people] height of corresponding building is higher than
T4 influences a key factor of fire risk assessment, it can be considered that the corresponding buildings of T2 since " building height " is
Fire risk is higher than T4, is III level risk；
[4], after being inferred to the fire risk of I grades, II grades and III level, remaining sample is then only IV grades of fire
Risk.In fact, for rule T 3, influence fire risk to evaluate mostly important factor since " building age " is, when
[the building age>27.5] when, it is believed that the fire risk of building is IV grades.
In the present embodiment, by building fire risk class appraisal procedure, level of building fire risk judgement rule are outputed
Then.When there is new building to build up, the level of building fire risk of building can be assessed by decision rule, to help fire-fighting
Supervision department realizes Precision management, emphasis prevention and control.
In one embodiment, the default risk assessment index of the corresponding data dimension includes：
5 Continuous valued attributes：A1 building ages, A2 index beds area, A5 construction areas, A6 buildings interior same time can hold
Receive maximum number, A7 building heights；And
1 discrete type attribute：B3 Fire lift numbers；
The building fire risk class appraisal procedure specific rules include：
In [A1 builds age≤15.5 year] and [building height≤64.5 meter A7] and [the same time can accommodate in A6 buildings
The people of maximum number≤5900] and [A2 index beds area≤60519 square metre] and [construction area≤189848.5 square metre A5],
Then determine [fire risk=light grade]；
At [the A1 building ages>15.5] and [building height≤72.45 meter A7] and [A6 building in the same time can accommodate
The people of maximum number≤7600], it is determined that [fire risk=middle rank]；
At [the A1 building ages>15.5] and [building height≤72.45 meter A7] and [A6 building in the same time can accommodate
Maximum number>7600 people], it is determined that [fire risk=serious grade]；
In [A1 builds age≤15.5 year] and [building height≤64.5 meter A7] and [the same time can accommodate in A6 buildings
Maximum number>5900 people], it is determined that [fire risk=serious grade]；
At [the A1 building ages>15.5] and [A7 building heights>72.45 meters] and [Fire lift number=1 or 2],
Then determine [fire risk=serious grade].
It is understood that in the description of this specification, reference term " embodiment ", " another embodiment ", " other
The description of embodiment " or " first embodiment~N embodiments " etc. means specific spy described in conjunction with this embodiment or example
Sign, structure, material or feature are included at least one embodiment or example of the invention.In the present specification, to above-mentioned
The schematic representation of term may not refer to the same embodiment or example.Moreover, the specific features of description, structure, material
Or feature can be combined in any suitable manner in any one or more of the embodiments or examples.
It should be noted that herein, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that process, method, article or system including a series of elements include not only those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or system institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including this
There is also other identical elements in the process of element, method, article or system.
The preferred embodiment of the present invention is not intended to limit the scope of the invention, every under the design of the present invention,
Using equivalent structure transformation made by description of the invention and accompanying drawing content, or directly/it is used in other relevant technologies indirectly
Field is included in the scope of patent protection of the present invention.
Claims (10)
1. a kind of building fire risk class appraisal procedure, which is characterized in that including step：
According to the data of target construction, the corresponding default risk assessment index of the target construction is obtained；
Default value pretreatment is carried out to the default risk assessment index of acquisition, obtains pretreated building data；
Be cluster number of clusters according to preset design fire risk class, by k- prototypes clustering algorithm to carrying out the pretreatment after
Building data carry out clustering processing obtain cluster result；
Classification prediction is carried out on the cluster result by least one machine learning classification algorithm, determines the cluster knot
Whether the classification prediction result that fruit carries out above reaches preset accuracy；
When the classification prediction result carried out on the cluster result reaches preset accuracy, according to cluster result class
Other building data and the preset design fire risk class, using J48 decision trees to the target construction
Fire risk is graded.
2. building fire risk class appraisal procedure according to claim 1, which is characterized in that described to be built according to target
The data for building object, the step of obtaining the target construction corresponding default risk assessment index, including：
According to the data dimension of the data of target construction, the default risk of the corresponding data dimension of the target construction is obtained
Evaluation index.
3. building fire risk class appraisal procedure according to claim 2, which is characterized in that the corresponding data
The default risk assessment index of dimension includes at least one of following attribute value：
8 numerical attributes or Continuous valued attributes, including：A1 builds age, A2 index beds area, the A3 refuge stories gross area, A4
Aerial layer area, A5 construction areas, A6 buildings interior same time can accommodate maximum number, A7 building heights and the undergrounds A8 level
Product；And
10 symbol attributes or discrete value attribute, including：The grounds the B1 number of plies, the undergrounds the B2 number of plies, B3 Fire lifts number, B4 take refuge
Layer number, B5 buildings classification, B6 building structure, the whether voluminous power of B7, B8 fire resistance ratings, B9 construction status value and the administrative areas B10
Domain.
4. building fire risk class appraisal procedure according to claim 1, which is characterized in that the institute of described pair of acquisition
The step of default risk assessment index carries out default value pretreatment, obtains pretreated building data is stated, including：
When the attribute value of the default risk assessment index is discrete type, default value is positioned as the default risk assessment and is referred to
The highest attribute value of frequency of occurrence in mark；And/or
When the attribute value of the default risk assessment index is continuous type, default value is positioned as the default risk assessment and is referred to
The mean value of attribute value in mark.
5. building fire risk class appraisal procedure according to claim 1, which is characterized in that the preset setting
Level of building fire risk is 3 grades or 4 grades.
6. building fire risk class appraisal procedure according to claim 1, which is characterized in that described according to preset
Design fire risk class is cluster number of clusters, by k- prototypes clustering algorithm to carry out the pretreated building data into
Row clustering processing obtains the step of cluster result, including：
It is cluster number of clusters according to preset design fire risk class, contains in adjustment k- prototypes cluster sample similarity measure formulas
The numerical value of some undetermined parameter τ is clustered by k- prototypes clustering algorithm to carrying out the pretreated building data
Processing.
7. building fire risk class appraisal procedure according to claim 1, which is characterized in that described to pass through at least one
Kind machine learning classification algorithm carries out classification prediction on the cluster result, determines point that the cluster result carries out above
The step of whether class prediction result reaches preset accuracy, including：
The machine learning classification algorithm in 8 kinds of WEKA kits is used on the cluster result to being added to classification information
The building data carry out class test；
Wherein, the machine learning classification algorithm includes：Bayesian network, random tree, naive Bayesian, SMO, random forest,
J48, NBTree and RBF net.
8. building fire risk class appraisal procedure according to claim 7, which is characterized in that described to reach preset
Accuracy is to reach 90% or more accuracy.
9. the building fire risk class appraisal procedure according to any one of claim 1-8, which is characterized in that described
The default risk assessment index of corresponding data dimension includes：
5 Continuous valued attributes：A1 building ages, A2 index beds area, A5 construction areas, A6 buildings interior same time can accommodate most
Big number, A7 building heights；And
1 discrete type attribute：B3 Fire lift numbers.
10. building fire risk class appraisal procedure according to claim 9, which is characterized in that
In [A1 builds age≤15.5 year] and [building height≤64.5 meter A7] and [the same time can accommodate maximum in A6 buildings
The people of number≤5900] and [A2 index beds area≤60519 square metre] and [construction area≤189848.5 square metre A5], then really
Fixed [fire risk=light grade]；
At [the A1 building ages>15.5] and [building height≤72.45 meter A7] and [A6 building in the same time can accommodate maximum
The people of number≤7600], it is determined that [fire risk=middle rank]；
At [the A1 building ages>15.5] and [building height≤72.45 meter A7] and [A6 building in the same time can accommodate maximum
Number>7600 people], it is determined that [fire risk=serious grade]
In [A1 builds age≤15.5 year] and [building height≤64.5 meter A7] and [the same time can accommodate maximum in A6 buildings
Number>5900 people], it is determined that [fire risk=serious grade]；
At [the A1 building ages>15.5] and [A7 building heights>72.45 meters] and [Fire lift number=1 or 2], then really
Fixed [fire risk=serious grade].
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810119077.9A CN108376310A (en) | 2018-02-06 | 2018-02-06 | Building fire risk class appraisal procedure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810119077.9A CN108376310A (en) | 2018-02-06 | 2018-02-06 | Building fire risk class appraisal procedure |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108376310A true CN108376310A (en) | 2018-08-07 |
Family
ID=63017452
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810119077.9A Pending CN108376310A (en) | 2018-02-06 | 2018-02-06 | Building fire risk class appraisal procedure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108376310A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190943A (en) * | 2018-08-20 | 2019-01-11 | 公安部沈阳消防研究所 | Dynamic Fire risk assessment method, device and server based on machine learning |
| CN109242361A (en) * | 2018-10-31 | 2019-01-18 | 深圳市中电数通智慧安全科技股份有限公司 | A kind of fire-fighting methods of risk assessment, device and terminal device |
| CN109389795A (en) * | 2018-09-05 | 2019-02-26 | 深圳市中电数通智慧安全科技股份有限公司 | Dynamic Fire risk assessment method, device, server and storage medium |
| CN109389178A (en) * | 2018-10-26 | 2019-02-26 | 深圳市元征科技股份有限公司 | A kind of maintenance factory's ranking method, system and electronic equipment and storage medium |
| CN109657916A (en) * | 2018-11-19 | 2019-04-19 | 深圳市中电数通智慧安全科技股份有限公司 | A kind of Fire risk assessment method, device and server |
| CN109684714A (en) * | 2018-12-21 | 2019-04-26 | 杭州睿兴栋宇建筑科技有限公司 | A kind of architectural design method based on machine learning and BIM technology |
| CN109871613A (en) * | 2019-02-18 | 2019-06-11 | 南京林业大学 | A kind of forest fire discrimination model acquisition methods and prediction application |
| CN111145517A (en) * | 2020-01-03 | 2020-05-12 | 上海枫昱能源科技有限公司 | Artificial intelligence electric line fire risk real-time early warning method and system |
| CN111724007A (en) * | 2019-03-18 | 2020-09-29 | 马上消费金融股份有限公司 | Risk evaluation method, evaluation device, intelligent system and storage device |
| CN112215451A (en) * | 2020-07-21 | 2021-01-12 | 中国人民公安大学 | Differentiation security check method and system based on civil aviation passenger classification |
| CN112529327A (en) * | 2020-12-21 | 2021-03-19 | 北京建筑大学 | Method for constructing fire risk prediction grade model of buildings in commercial areas |
| CN114066319A (en) * | 2021-12-01 | 2022-02-18 | 清华大学 | Fireproof planning method and device for historical block, electronic equipment and storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103577700A (en) * | 2013-11-14 | 2014-02-12 | 哈尔滨工程大学 | Boat firefighting system interlock failure prediction method |
| KR101467115B1 (en) * | 2013-09-12 | 2014-12-01 | 세종대학교산학협력단 | System and method for firefighting in a building |
| WO2015079191A2 (en) * | 2013-11-05 | 2015-06-04 | Airbus Ds Sas | Method for obtaining a system for active, decentralized multi-agent situation control |
| CN106897792A (en) * | 2017-01-10 | 2017-06-27 | 广东广业开元科技有限公司 | A kind of structural fire protection risk class Forecasting Methodology and system |
| CN110781261A (en) * | 2019-10-14 | 2020-02-11 | 西南交通大学 | Building damage risk screening method and system based on geographic spatial information |
-
2018
- 2018-02-06 CN CN201810119077.9A patent/CN108376310A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101467115B1 (en) * | 2013-09-12 | 2014-12-01 | 세종대학교산학협력단 | System and method for firefighting in a building |
| WO2015079191A2 (en) * | 2013-11-05 | 2015-06-04 | Airbus Ds Sas | Method for obtaining a system for active, decentralized multi-agent situation control |
| CN103577700A (en) * | 2013-11-14 | 2014-02-12 | 哈尔滨工程大学 | Boat firefighting system interlock failure prediction method |
| CN106897792A (en) * | 2017-01-10 | 2017-06-27 | 广东广业开元科技有限公司 | A kind of structural fire protection risk class Forecasting Methodology and system |
| CN110781261A (en) * | 2019-10-14 | 2020-02-11 | 西南交通大学 | Building damage risk screening method and system based on geographic spatial information |
Non-Patent Citations (3)
| Title |
|---|
| 侯忠辉 等: "基于决策树的消防火灾等级模型研究", 《网络安全技术与应用》 * |
| 王瑛: "基于聚类分析的客户信用等级评定", 《福建电脑》 * |
| 程克非 等: "基于J48决策树算法的水质评价方法", 《计算机工程》 * |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109190943A (en) * | 2018-08-20 | 2019-01-11 | 公安部沈阳消防研究所 | Dynamic Fire risk assessment method, device and server based on machine learning |
| CN109389795A (en) * | 2018-09-05 | 2019-02-26 | 深圳市中电数通智慧安全科技股份有限公司 | Dynamic Fire risk assessment method, device, server and storage medium |
| CN109389178A (en) * | 2018-10-26 | 2019-02-26 | 深圳市元征科技股份有限公司 | A kind of maintenance factory's ranking method, system and electronic equipment and storage medium |
| CN109242361A (en) * | 2018-10-31 | 2019-01-18 | 深圳市中电数通智慧安全科技股份有限公司 | A kind of fire-fighting methods of risk assessment, device and terminal device |
| CN109657916A (en) * | 2018-11-19 | 2019-04-19 | 深圳市中电数通智慧安全科技股份有限公司 | A kind of Fire risk assessment method, device and server |
| CN109684714B (en) * | 2018-12-21 | 2023-04-18 | 杭州睿兴栋宇建筑科技有限公司 | Building design method based on machine learning and BIM technology |
| CN109684714A (en) * | 2018-12-21 | 2019-04-26 | 杭州睿兴栋宇建筑科技有限公司 | A kind of architectural design method based on machine learning and BIM technology |
| CN109871613A (en) * | 2019-02-18 | 2019-06-11 | 南京林业大学 | A kind of forest fire discrimination model acquisition methods and prediction application |
| CN109871613B (en) * | 2019-02-18 | 2023-05-19 | 南京林业大学 | Forest fire discrimination model acquisition method and prediction application |
| CN111724007A (en) * | 2019-03-18 | 2020-09-29 | 马上消费金融股份有限公司 | Risk evaluation method, evaluation device, intelligent system and storage device |
| CN111724007B (en) * | 2019-03-18 | 2022-12-06 | 马上消费金融股份有限公司 | Risk evaluation method, evaluation device, intelligent system and storage device |
| CN111145517B (en) * | 2020-01-03 | 2022-03-25 | 上海枫昱能源科技有限公司 | Artificial intelligence electric line fire risk real-time early warning method and system |
| CN111145517A (en) * | 2020-01-03 | 2020-05-12 | 上海枫昱能源科技有限公司 | Artificial intelligence electric line fire risk real-time early warning method and system |
| CN112215451A (en) * | 2020-07-21 | 2021-01-12 | 中国人民公安大学 | Differentiation security check method and system based on civil aviation passenger classification |
| CN112529327A (en) * | 2020-12-21 | 2021-03-19 | 北京建筑大学 | Method for constructing fire risk prediction grade model of buildings in commercial areas |
| CN114066319A (en) * | 2021-12-01 | 2022-02-18 | 清华大学 | Fireproof planning method and device for historical block, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108376310A (en) | Building fire risk class appraisal procedure | |
| Bedford et al. | Probabilistic risk analysis: foundations and methods | |
| Rashed et al. | Assessing vulnerability to earthquake hazards through spatial multicriteria analysis of urban areas | |
| Lee et al. | An efficient design support system based on automatic rule checking and case-based reasoning | |
| CN107886235A (en) | A kind of Fire risk assessment method for coupling certainty and uncertainty analysis | |
| CN106202665B (en) | Initial failure root primordium recognition methods based on domain mapping and weighted association rules | |
| CN103154845A (en) | Machine learning for power grids | |
| CN112529327A (en) | Method for constructing fire risk prediction grade model of buildings in commercial areas | |
| Li et al. | Real‐Time Warning and Risk Assessment of Tailings Dam Disaster Status Based on Dynamic Hierarchy‐Grey Relation Analysis | |
| CN111126853A (en) | Fuzzy FMEA-based hydraulic engineering risk early warning analysis method and system | |
| Hosseini Nourzad et al. | Vulnerability of infrastructure systems: Macroscopic analysis of critical disruptions on road networks | |
| D'Urso et al. | Multicriteria fuzzy analysis for a GIS‐based management of earthquake scenarios | |
| CN112734154A (en) | Multi-factor public opinion risk assessment method based on fuzzy number similarity | |
| CN109656904A (en) | A kind of case risk checking method and system | |
| CN107145995A (en) | Production environment safety prediction methods, devices and systems | |
| Hui | Construction of information security risk assessment model in smart city | |
| CN109214937A (en) | The anti-fraud determination method of settlement of insurance claim intelligence and system | |
| Xu | Assessment of nuclear waste repository options using the ER approach | |
| CN112950350B (en) | Loan product recommendation method and system based on machine learning | |
| CN104573374A (en) | Organization code information quality data fuzzy evaluation device | |
| González et al. | Using fuzzy logic to introduce the human factor in the failure frequency estimation of storage vessels in chemical plants | |
| CN113947303A (en) | Intelligent quantification method for fire-fighting acceptance check | |
| Amiri Shahmirani et al. | Application of fuzzy modelling to predict the earthquake damage degree of buildings based on field data | |
| Chang et al. | A deep belief network and case reasoning based decision model for emergency rescue | |
| Vahdat et al. | A risk-based system for managing the retrofitting of school buildings in seismic prone areas: A case study from Iran |
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 |
Application publication date: 20180807 |
|
| WD01 | Invention patent application deemed withdrawn after publication |