CN102495875A - Marine disaster early warning expert system based on data mining - Google Patents

Abstract

The invention provides a marine disaster early warning expert system based on data mining. The system comprises 7 modules including a marine data base, a data mining module, a marine disaster knowledge base, a knowledge acquisition interface, an inference machine, an explanation module and a marine disaster early warning display interface. Marine data are abundant and various, large-scale complicated marine data are simplified and compressed by adopting a fuzzy clustering rough set attribute reduction method, attribute factor information related to the marine disaster is retained to the largest extent, information containing a large amount of knowledge is obtained by using the data mining, the information containing the knowledge is regularly extracted and treated, and then strategy rules capable of predicting the marine disasters are found out. The strategy rules organize and form the marine disaster knowledge base and the inference machine, knowledge rules and data in the knowledge base are used for inference, heuristic method inference and judgment are conducted simulating experts in the marine area, marine disasters are evaluated, levels of the marine disasters are judged, and an explanation program is combined with an early warning threshold value to send out disaster early warning towards users in a simple and visual mode. The system effectively solves problems of shortage of expert early warning and marine knowledge personnel on the marine disasters, and provides new ways and methods for popularizing scientific marine disaster early warning technology.

Description

A kind of early warning of marine disasters expert system based on data mining

Technical field

The present invention relates to the evaluation criteria of Oceanic disasters, disaster is judged grade, reaches the expert system to early warning of marine disasters.

Background technology

China suffers one of the most serious country of Oceanic disasters in the world, and along with global warming, extreme weather events take place frequently and the fast development of coastal economy, the marine disaster prevention and reduction pressure that the coastland faces is increasing.At present domestic having begun to take shape comprises the stereopsis system that is made up of all kinds of buoys, subsurface buoy, satellite, radar, aircraft, boats and ships and bank base (island) observation website; Obtain the oceanographic data of great deal of rich; Pass through the window of international exchanges data simultaneously; Obtain the Marine Environmental Elements data in the whole world, had the Marine Sciences data and the relevant information of a large amount of preciousnesses.How to utilize the Marine Sciences data more efficiently; Improve the early warning defence of various Oceanic disasters; Not only to effectively alleviating the loss that Oceanic disasters cause; Ensure that maritime people's security of the lives and property and continuous development of society economy have very deep directive significance, and the digital ocean construction of China provided provide powerful support for, so be necessary to set up Oceanic disasters property early warning system.

Process effective, novel, potentially useful, final intelligible pattern is obtained in data mining (Data Mining) from mass data.Pass through data mining technology; Be used in the abundant oceanographic data resource; Thereby the discovery useful information is understood mutual relationship, the rule of development of various ocean essentials, and for scientific research of seas and integrated management the information decision support being provided is vast ocean science worker's a research direction.

Oceanographic data type complicacy is various, and the early warning of marine disasters that relates to also has suitable complicacy.The knowledge information that data mining is come out is that early warning of marine disasters provides decision-making, needs suitable professional knowledge just can accomplish.Because in human society, ocean knowledge aspect Expert Resources is quite rare, and expert system has been arranged, and then can make this precious expertise obtain general application.

Expert system is to utilize a large amount of expertises, and the method for utilization knowledge reasoning solves the practical problems in each specific area.In expert system application Yu Haiyang disaster; Expertise, technology, experimental data and mathematical model according to marine field; Simulation marine field expert carries out heuristic inference, judgement with regard to a certain specific Oceanic disasters problem; Can well predict Oceanic disasters, give the alarm, be public's service.

Summary of the invention

The present invention provides a kind of early warning of marine disasters expert system based on data mining.This system comprises seas and oceans database, data-mining module, and the Oceanic disasters knowledge base, the knowledge acquisition interface, inference machine, explanation module, early warning of marine disasters is showed seven modules in interface.Oceanographic data is rich and varied; The rough set attribute reduction of employing fuzzy clustering earlier method is carried out brief and compression to the oceanographic data of large-scale complex, farthest keeps the attribute factor information relevant with Oceanic disasters, utilizes data mining to obtain containing the information of a large amount of knowledge again; To these knowledge information processes; Fuzzy clustering, Rule Extraction is handled, and finds the tactful rule that can predict Oceanic disasters.Expert system with these tactful rule tissues formation; Simulation marine field expert carries out heuristic inference, judgement with regard to a certain Oceanic disasters problem; The prediction Oceanic disasters; Give the alarm, efficiently solve expert's early warning and ocean knowledge personnel problem of shortage in the Oceanic disasters, new approach and means are provided for promoting science early warning of marine disasters technology.

Description of drawings

Fig. 1 is an early warning of marine disasters expert system Organization Chart of the present invention.

Fig. 2 is the process flow diagram synoptic diagram of expert's early warning Oceanic disasters of the present invention.

Fig. 3 is ocean knowledge library structure figure of the present invention.

Embodiment

The invention discloses a kind of early warning of marine disasters expert system, embodiment is described below in conjunction with accompanying drawing.

Please with reference to Fig. 1.Fig. 1 is an early warning of marine disasters expert system Organization Chart of the present invention.Comprise seas and oceans database, data-mining module, the Oceanic disasters knowledge base, the knowledge acquisition interface, inference machine, explanation module, early warning of marine disasters is showed seven functional modules in interface.

Please with reference to Fig. 2.The groundwork flow process of early warning expert system of the present invention is following:

With the oceanographic data process data cleansing that various devices are gathered, pre-service forms the Oceanic disasters database; Data-mining module excavates the Oceanic disasters data of database, exports the knowledge of certain correlation rule, triggers inference machine; Inference machine calls that knowledge rule and data carry out reasoning in the knowledge base; Assess Oceanic disasters, judge the grade of Oceanic disasters, interpretive routine combines threshold value of warning; With the simple and direct mode, send early warning of marine disasters to the user.

Please participate in Fig. 3.The Oceanic disasters knowledge base is stored extra large early warning of marine disasters rule knowledge, is divided into static knowledge base and dynamic knowledge storehouse.It comprises typical ocean disasters such as storm tide, tsunami, disastrous wave, sea ice, red tide, typhoon at least in interior disaster Rules of Assessment storehouse; And Oceanic disasters grade decision rule storehouse, according to the early warning of marine disasters policing rule storehouse of Oceanic disasters grade decision rule storehouse, Rules of Assessment storehouse formulation.Wherein static knowledge base is stored relatively stable, the long rule of refreshing one's knowledge of time; The dynamic knowledge storehouse, the storage update time is some rules faster.The renewal of dynamic knowledge storehouse rule, main source be tap layer in process to the Oceanic disasters data mining, some new knowledges of discovery expand and the expert through knowledge acquisition interface additional to the dynamic knowledge storehouse.

1, seas and oceans database comprises proto-ocean database and the seas and oceans database of crossing through the data cleansing conversion process relevant with Oceanic disasters.The raw data library storage is the oceanographic data of acquisition and recording by all means; Comprise through the station; buoy boats and ships, oceanographic data harvesters such as satellite; a series of complicated type data that obtain, marine hydrology, maritime meteorology, marine physics, thalassochemistry, sea life, Marine Geology, ocean landform and marine geophysics etc.The oceanographic data library storage relevant with Oceanic disasters is by treated oceanographic data in the proto-ocean database.When data cleansing and deal with data; Grasp oceanographic data characteristic, the purpose of Oceanic disasters database application and the purpose of demand and Knowledge Discovery according to calendar year experience; Mainly comprise and remove redundant information, field reduction and format conversion; Processing such as time series data are in strict accordance with the design of oceanographic data standard, enforcement.

2, data-mining module adopts various mining algorithms to standardized Oceanic disasters database, and its task is that the knowledge information that contains is found.The Rules of Assessment of Oceanic disasters knowledge rule library storage; Disaster grade decision rule is the deciding grade and level standard, is difficult to effectively reflect the degree of Oceanic disasters, from this respect consideration; Data-mining module has been selected fuzzy cluster analysis; The comprehensive effect of each item ocean attribute factor in various Oceanic disasters gears to actual circumstances the estimation standard of Oceanic disasters more, and early warning is more accurate.

On the one hand, excavate the result of coming out, under the rule of Oceanic disasters static knowledge base, through inference machine identification, explanation module is shown to the user with the form that is easily understood; Utilize on the other hand and excavate out the knowledge that obtains, be converted into and the corresponding to form of the knowledge of Oceanic disasters knowledge base, the Oceanic disasters knowledge base is dynamically expanded, assist the expert to carry out knowledge acquisition with specific knowledge transformation technology.

Simultaneously because the data type complicacy of oceanographic data library storage is various; Enormous amount; Before data mining; Employing fuzzy clustering rough set attribute reduction method is carried out brief and compression to the oceanographic data of large-scale complex, farthest keeps the attribute factor information relevant with Oceanic disasters, improves data mining efficient.Described fuzzy clustering rough set attribute reduction method was divided into for three steps:

(1) obfuscation of property value.If N is the number of categories of attribute, N=1,2 ... C _NBe N classification, M (C _N) be C _NIn the property value number that comprises, C _N(i) be i property value, then property value C in N the classification _N(i) subordinate function does

$u_N\left({C_N}^{\left(i\right)}\right)=\frac{M\left(C_N\right)}{n}$ And $\underset{I=1}{\overset{C}{\mathrm{\Σ}}}{u}_{\mathrm{Ij}}=1,$ $\∀j=1,...,n$

(2) fuzzy object cluster.Utilize the subordinate function initialization to be subordinate to matrix U, calculate cluster centre

$C_i=\frac{\underset{j-1}{\overset{m}{\mathrm{\Σ}}}{u_{\mathrm{ij}}}^m{X}_j}{\underset{j-1}{\overset{m}{\mathrm{\Σ}}}{u_{\mathrm{ij}}}^m}$

Make objective function $J(U,C_1,C_2,...,C_C)=\underset{i=1}{\overset{c}{\mathrm{\Σ}}}{J}_i=\underset{i-1}{\overset{c}{\mathrm{\Σ}}}\underset{j}{\overset{n}{\mathrm{\Σ}}}{u_{\mathrm{Ij}}}^m{d_{\mathrm{Ij}}}^2$ Less than a specified threshold value then algorithm stop.(d _Ij=‖ c _i-x _j‖ is the Euclidean distance between i cluster centre and j data points, and m is a weighted index);

(3) attribute reduction.According to the result of fuzzy clustering, can obtain compressing fuzzy tables of data, make up the differentiation matrix of fuzzy object | U| * | (wherein each in the matrix is expressed as U|: C _Ij={ a ∈ A|f (x _i, a) ≠ f (x _i, a) }),

By distinguishing function $f_A\left(C_{\mathrm{Ij}}\right)=\underset{(i,j)\∈U\×U}{\mathrm{\Π}}\mathrm{\Σ}{C}_{\mathrm{IJ}}$ Try to achieve minimum attribute reduction.

3, knowledge acquisition interface towards ocean knowledge expert and technician, is used for increasing, revising, the maintenance function of enhanced system the Oceanic disasters knowledge base.Obtaining of Oceanic disasters knowledge wherein can be through the summary over the years experience that empirical tests crosses of getting off, expertise and knowledge base is had the knowledge information etc. of expansion from what data-mining module came.

4, inference machine, according to the knowledge in the knowledge base, under the effect of interpretive routine, to the knowledge information that data-mining module excavates out, the grade of reasoning and judging Oceanic disasters, the assessment disaster in conjunction with threshold value of warning, draws the process of the prediction scheme of Oceanic disasters at last.Wherein inference machine can adopt forward reasoning, backward reasoning, advanced inference method such as fuzzy reasoning.

5, explanation module is explanation and the description to Oceanic disasters prediction scheme reasoning process, and the last mode of simple and direct is as a result presented to the user.Be easily understood for the result is showed, therefrom extract characteristic attribute in the knowledge base rule, convert the alarm data of Oceanic disasters to a kind of standard format.Here design the Oceanic disasters alarm is described with a five-tuple: Alarm=(time, source, type, level, severity).Here alarm time (time) is the time of origin that shows potential Oceanic disasters; Alarm source (source) is a sign of sending alarm signal, is generally the positional information of gathering oceanographic data equipment, can represent through longitude and latitude; Alarm type (type) is set forth the physical meaning of alarm; Oceanic disasters such as alarm type storm tide, tsunami, disastrous wave, sea ice, red tide, typhoon; Oceanic disasters grade (level), the expression disaster the order of severity, wherein level setting according to country about the Oceanic disasters standard formulation.The corresponding corresponding disaster of security strategy (severity) expression, the corresponding strategy that people should take, for example evacuation route etc.

6, early warning of marine disasters is showed the interface, and the result is carried out simple displaying the directly perceived, and the object of displaying can be for government bodies, relate to extra large unit, the public etc.Early warning of marine disasters is showed the various servers of employing between interface and the inference machine explanation module, like GIS, Web, mobile information service device etc.

Claims (8)

1. the early warning of marine disasters expert system based on data mining is characterized in that, should comprise:

One seas and oceans database is responsible for the oceanographic data that various approach collect is stored;

One data-mining module, the Oceanic disasters database of being responsible for that standard treated is crossed adopts algorithm to excavate;

One Oceanic disasters knowledge base is that the assessment of storage disaster, disaster grade are judged the set of decision rule knowledge;

One knowledge acquisition interface, towards expert, knowledge engineer, increasing, revising and safeguarding the Oceanic disasters knowledge base;

One inference machine according to the rule in the ocean knowledge storehouse, carries out knowledge reasoning according to strategy;

One explanation module is in conjunction with threshold value of warning, to the explanation and the description of early warning reasoning process.

One early warning of marine disasters is showed the interface, and The reasoning results is showed.

2. the seas and oceans database in the early warning of marine disasters expert system according to claim 1 is characterized in that comprising proto-ocean database and treated Oceanic disasters database.Wherein, to proto-ocean data of database harvester, specifically comprise the station, buoy, boats and ships, satellite etc.To the cleaning of proto-ocean database data, pre-service, according to the rule criterion clean of Oceanic disasters knowledge base, form the database that constitutes by all kinds of Oceanic disasters factors, these disaster factors comprise temperature; Salinity, ocean current, wind speed; Meteorology, biology, geographic position etc. are in interior data factor.

3. according to the data-mining module in the early warning of marine disasters expert system of claim 1, the various mining algorithms of Oceanic disasters The data are carried out knowledge excavation.It is characterized in that; At first various according to the data type complicacy of claim 2 oceanographic data library storage; Enormous amount, before data mining, employing fuzzy clustering rough set attribute reduction method is carried out brief and compression to the oceanographic data of large-scale complex; Farthest keep the attribute factor information relevant, improve data mining efficient with Oceanic disasters.Secondly; Because the assessment of Oceanic disasters knowledge rule library storage, disaster grade decision rule is the deciding grade and level standard, is difficult to effectively reflect the degree of Oceanic disasters; Data-mining module according to right 3 has been selected fuzzy cluster analysis; The comprehensive effect of each item ocean attribute factor in various Oceanic disasters gears to actual circumstances the estimation standard of Oceanic disasters more, and early warning is more accurate.Data-mining module according to right 3 adopts based on the open language of XML simultaneously, and to the disaster factors data modeling, the result can outwards provide output interface with the form of standard, supplies inference machine, reaches the use of Oceanic disasters dynamic data base.

4. according to Oceanic disasters knowledge base in the early warning of marine disasters expert system of claim 1, it is characterized in that storing the disaster assessment, the disaster grade is judged, the DECISION KNOWLEDGE rule.The Oceanic disasters knowledge base comprises static knowledge base and dynamic knowledge storehouse, and the knowledge rule of static knowledge library storage mainly supplies inference machine reasoning usefulness, and the knowledge rule in dynamic knowledge storehouse is mainly accepted the new regulation that the knowledge processing process is come, and regularly upgrades to static knowledge base.

5. according to knowledge acquisition interface in the early warning of marine disasters expert system of claim 1, it is characterized in that, be used for increasing, revising and safeguarding the rule in dynamic knowledge storehouse towards expert, knowledge engineer.The data that get into from the knowledge acquisition interface all must be passed through knowledge processing, form logical organization and the data structure consistent with knowledge base, leave in the dynamic knowledge storehouse.

6. according to inference machine in the early warning of marine disasters expert system of claim 1; It is characterized in that using the rule in the ocean knowledge storehouse; Carry out knowledge reasoning in proper order according to certain reasoning, and in reasoning process, control and user session; In conjunction with interpretive routine and threshold value of warning, the result of last Oceanic disasters is showed the user.Inference machine is made up of actuator, scheduler and consistance telegon etc.Scheduler (is selected next step execution of an action confession system with the information of knowledge and arthmetic statement according to control strategy from agenda.Knowledge information in the actuator working knowledge storehouse, the action that execution scheduler is selected.The main effect of consistance telegon is when obtaining new data or newly supposing, the correlated results that has obtained is carried out the likelihood correction, to guarantee result's self-consistentency property.The inference method of inference machine can adopt forward reasoning, advanced inference methods such as backward reasoning fuzzy reasoning.

7. according to explanation module in the early warning of marine disasters expert system of claim 1, it is characterized in that combining threshold value of warning, explanation and description to the early warning reasoning process make the result show simple, intuitive.

8. show the interface according to early warning of marine disasters in the early warning of marine disasters expert system of claim 1; It is characterized in that towards the user of Oceanic disasters expert early warning system; Its authority only limits to use Oceanic disasters expert early warning system, shows the early warning result to the user.Wherein the user can be government bodies, relate to extra large unit, the public etc.Early warning of marine disasters shows that interface terminal can directly link to each other with inference machine, interpretive routine, also can link to each other with inference machine, interpretive routine through various servers.

Images