CN103425890A - Landscape water quality analysis algorithm - Google Patents

Abstract

The invention discloses a landscape water quality analysis algorithm. The method comprises the following steps: firstly, establishing a water quality analysis index set, and establishing a standard sample water set and a to-be-evaluated sample water set on the index set; secondly, performing Vague processing on original non-negative single value data to obtain Vague set data, and calculating the distance between the to-be-evaluated sample water set and the standard sample water set; lastly, judging the water quality of a to-be-evaluated sample by using a mode identification method based on a distance Vague set. According to the method disclosed by the invention, landscape water quality can be detected and analyzed effectively, and a very good effect is achieved; in the algorithm, the influence of the hesitancy in the Vague set on a final result is considered, and a novel measure is provided for the water quality analysis and daily management of non-conventional water resources.

Description

A kind of Landscape Water Quality analytical algorithm

Affiliated technical field

The present invention relates to a kind of Landscape Water Quality analytical algorithm, be specifically related to utilize the range formula of the Vague collection of emphasizing degree of hesitation directly to carry out distance calculating to sample to be detected and master sample, thereby draw the water quality of sample to be assessed according to the distance method of pattern-recognition.Algorithm.Belong to environmental engineering water quality inspection technique field.

Background technology

Landscape water quality is for the vital impact of having of urban ecological environment, urban landscape water has become the emphasis of Construction of Eco-urban, water analysis detects can provide foundation for the sustainable development of ecologic environment, but because the water source sample does not have strict attribute, they exist uncertain aspect form and generic, also for analyzing and testing, bring certain difficulty.Due to uncertainty and the ambiguity of data, make the detection method based on fuzzy theory be widely used in water analysis.Such as people such as Ma Jianqin in document 1 " urban landscape water water quality Variable Fuzzy estimate " based on GIS (HYDROELECTRIC ENERGY science, 2011,29(8)) take GIS as platform, landscape water quality comprehensive evaluation model based on variable fuzzy sets theory has been proposed, and take landscape water as example, water quality and spatial distribution state thereof are estimated.The people such as Zhu Qiongyao are at document 2 " the water analysis early warning technology research based on rough set and evidence theory " (journal of Zhejiang university (agricultural and life science version), 2012, 38 (6)) the water quality early-warning algorithm based on rough set and D-S evidence theory use in conjunction has been proposed, the introducing of rough set better solves the time non-continuous event that may exist in time series data, eliminate the data collision in historical data, the problems such as data and Attribute Redundancy in the yojan database, thereby realize subduing data noise, the purpose of yojan data volume, and the associating of rough set and D-S evidence theory has reduced calculated amount to a certain extent, obtain better water analysis prediction effect.

Summary of the invention

The purpose of this invention is to provide a kind of Landscape Water Quality analytical algorithm, this algorithm adopts Vague collection theoretical research water analysis method, according to the water quality sample to be detected extracted, selected characteristic is set up index set, and Criterion sample water collection obtains original Vague collection data on the index set of choosing, build the Vague environment, and carry out pattern-recognition by the range formula between the Vague collection of emphasizing degree of hesitation, differentiate the water quality situation of sample to be detected.Algorithm is taken into account the certainty degree of Vague collection simultaneously, negates degree, the hesitation degree, and the core factor, have stronger dirigibility in application.

In order to achieve the above object, the present invention adopts following technical scheme:

A kind of Landscape Water Quality analytical algorithm comprises the following steps:

(1) set up index set X={x ₁, x ₂..., x _n;

(2) Criterion sample water collection E={E on index set X ₁, E ₂..., E _h, water quality E wherein ₁For I class water quality, E ₂For II class water quality, E ₃For III class water quality, E ₄For IV class water quality, E ₅For V class water quality.Pollution level should be clean, uncontaminated, light pollution, middle pollution and heavily contaminated mutually, sets up sample water collection C={C to be assessed ₁, C ₂..., C _kObtain raw data, wherein C _iThe data that collect for sampled point i;

(3) original non-negative single-value data is changed into to Vague collection data, build the Vague environment;

(4) calculate the sample C of water to be assessed _j(j=1,2 ..., Vague collection k) and master sample E _i(i=1,2 ..., the distance between Vague collection h);

(5) the Vague integrated mode of distance-based is identified, and differentiates the water quality of sample to be assessed.

The index set of the described step of algorithm (1) is chosen 4 indexs, is respectively x ₁For chemical oxygen demand (COD) (COD), x ₂For ammonia nitrogen (NH3-N), x ₃For total phosphorus (TP), x ₄For total nitrogen (TN), unit: mg/L. is as index.

The conversion formula that original non-negative single-value data is changed into to the Vague Value Data of the described step of algorithm (3) is

$x_{\mathrm{ij}}=[t_{\mathrm{ij}},1-f_{\mathrm{ij}}]=(\frac{x_{\mathrm{ij}}}{x_{j\max }},1-{\left(\frac{x_{\mathrm{ij}}}{x_{j\max }}\right)}^{\frac{1}{2}}),$

X wherein _Jmax=max{x _J1, x _J2, x _J3, x _J4, x _J5, x _J6, x _J7, x _J8, j=1,2,3,4.

Distance between the sample to be assessed of the described step of algorithm (4) and the Vague collection of master sample is calculated according to following formula

$d^{\left(m\right)}(E,C)=\frac{1}{12n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(3\right|t_{\mathrm{xi}}^m-t_{\mathrm{yi}}^{\left(m\right)}|+3|f_{\mathrm{xi}}^{\left(m\right)}-f_{\mathrm{yi}}^{\left(m\right)}|+3|S_{\mathrm{xi}}^{\left(m\right)}-S_{\mathrm{yi}}^{\left(m\right)}|+2{\mathrm{\π}}_{\mathrm{xi}}^{\left(m\right)}+2{\mathrm{\π}}_{\mathrm{yi}}^{\left(m\right)})$

（m＝0,1,2,…），

Wherein:

(1) the Vague value x=[t in A for element x _x, 1-f _x], t _x, f _x, π _x=1-t _x-f _x, S _x=t _x-f _xBeing called as respectively the certainty degree of element x to Vague collection A, negates degree, hesitation degree, core.And definition

$t_x^{\left(m\right)}=t_x(1+{\mathrm{\π}}_x+{\mathrm{\π}}_x^2+\·\·\·+{\mathrm{\π}}_x^m),f_x^{\left(m\right)}=f_x(1+{\mathrm{\π}}_x+\·\·\·+{\mathrm{\π}}_x^m),{\mathrm{\π}}_x^{\left(m\right)}={\mathrm{\π}}_x^{(m+1)},m=\mathrm{1,2},\·\·\·..$

(2) Vague collection E and C are at discrete domain X={x ₁, x ₂..., x _nOn be defined as,

E＝([t _E(x ₁),1-f _E(x ₁)],[t _E(x ₂),1-f _E(x ₂)],…,[t _E(x _n),1-f _E(x _n)]),

C＝([t _C(x ₁),1-f _C(x ₁)],[t _C(x ₂),1-f _C(x ₂)],…,[t _C(x _n),1-f _C(x _n)])

The Vague integrated mode recognition methods of the distance-based of the described step of algorithm (5) is: by step 4 calculate the numerical values recited sequence of distance $d(C_j,E_{i_1})\≤d(C_j,E_{i_2})\≤\·\·\·\≤d(C_j,E_{i_k})(j=\mathrm{1,2},\·\·\·,k)$ Sample C to be assessed _jBelong to master sample

, i wherein ₁, i ₂..., i _h1,2 ..., one of h without the full arrangement repeated.

The accompanying drawing explanation

Fig. 1 is a kind of Landscape Water Quality analytical algorithm process flow diagram.

Embodiment:

Fig. 1 is a kind of Landscape Water Quality analytical algorithm process flow diagram; A kind of Landscape Water Quality analytical algorithm of the present invention comprises the following steps:

(1) set up index set X={x ₁, x ₂..., x _n, choose x ₁For chemical oxygen demand (COD) (COD), x ₂For ammonia nitrogen (NH3-N), x ₃For total phosphorus (TP), x ₄For total nitrogen (TN), unit: mg/L. is as index.

(2) Criterion sample water collection E={E on index set X ₁, E ₂..., E _h, water quality E wherein ₁For I class water quality, E ₂For II class water quality, E ₃For III class water quality, E ₄For IV class water quality, E ₅For V class water quality.Pollution level should be clean, uncontaminated, light pollution, middle pollution and heavily contaminated mutually, sets up sample water collection C={C to be assessed ₁, C ₂..., C _kObtain raw data, wherein C _iThe data that collect for sampled point i; Table 1 is sample data to be assessed and China's water environment quality standard sample data, chooses altogether 3 sampling number certificates.

Table 1 China water environment quality standard sample and sample to be assessed (raw data)

(3) build the Vague environment; Raw data is pressed to formula

Change into Vague collection data;

X wherein _Jmax=max{x _J1, x _J2, x _J3, x _J4, x _J5, x _J6, x _J7, x _J8, j=1,2,3,4.It is as shown in table 2 that table 1 raw data is converted to the Vague Value Data.

Table 2 China water environment quality standard sample and sample to be evaluated (Vague Value Data)

(4) the Vague value x=[t in A for element x _x, 1-f _x], t _x, f _x, π _x=1-t _x-f _x, S _x=t _x-f _xBeing called as respectively the certainty degree of element x to Vague collection A, negates degree, hesitation degree, core.And definition $t_x^{\left(m\right)}=t_x(1+{\mathrm{\π}}_x+{\mathrm{\π}}_x^2+\·\·\·+{\mathrm{\π}}_x^m),f_x^{\left(m\right)}=f_x(1+{\mathrm{\π}}_x+\·\·\·+{\mathrm{\π}}_x^m),{\mathrm{\π}}_x^{\left(m\right)}={\mathrm{\π}}_x^{(m+1)},m=\mathrm{1,2},\·\·\·.$ Vague collection E and C are at discrete domain X={x ₁, x ₂..., x _nOn be defined as,

E＝([t _E(x ₁),1-f _E(x ₁)],[t _E(x ₂),1-f _E(x ₂)],…,[t _E(x _n),1-f _E(x _n)]),

C＝([t _C(x ₁),1-f _C(x ₁)],[t _C(x ₂),1-f _C(x ₂)],…,[t _C(x _n),1-f _C(x _n)])。

Calculate the sample C of water to be assessed _j(j=1,2 ..., Vague collection k) and master sample E _i(i=1,2 ..., the range formula between Vague collection h) is: $d^{\left(m\right)}(E,C)=\frac{1}{12n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\left(3\right|t_{\mathrm{xi}}^m-t_{\mathrm{yi}}^{\left(m\right)}|+3|f_{\mathrm{xi}}^{\left(m\right)}-f_{\mathrm{yi}}^{\left(m\right)}|+3|S_{\mathrm{xi}}^{\left(m\right)}-S_{\mathrm{yi}}^{\left(m\right)}|+2{\mathrm{\π}}_{\mathrm{xi}}^{\left(m\right)}+2{\mathrm{\π}}_{\mathrm{yi}}^{\left(m\right)})$ ?（m＝0,1,2,…）。

Here get parameter m=2, reckoner 2 sample C to be assessed _iThe Vague collection of (i=1,2,3) and master sample E _iDistance between the Vague collection of (i=1,2,3,4,5), result is as follows:

d ⁽²⁾(C ₁,E ₁)＝0.207,d ⁽²⁾(C ₁,E ₂)＝0.104,d ⁽²⁾(C ₁,E ₃)＝0.250,

d ⁽²⁾(C ₁,E ₄)＝0.451,d ⁽²⁾(C ₁,E ₅)＝0.603.

d ⁽²⁾(C ₂,E ₁)＝0.212,d ⁽²⁾(C ₂,E ₂)＝0.140,d ⁽²⁾(C ₂,E ₃)＝0.229,

d ⁽²⁾(C ₂,E ₄)＝0.461,d ⁽²⁾(C ₂,E ₅)＝0.599.

d ⁽²⁾(C ₃,E ₁)＝0.225,d ⁽²⁾(C ₃,E ₂)＝0.160,d ⁽²⁾(C ₃,E ₃)＝0.267

d ⁽²⁾(C ₃,E ₄)＝0.448,d ⁽²⁾(C ₃,E ₅)＝0.600

(5) step 4 is calculated the numerical values recited sequence of distance $d(C_j,E_{i_1})\≤d(C_j,E_{i_2})\≤\·\·\·\≤d(C_j,E_{i_k})(j=\mathrm{1,2},\·\·\·,k),$ The water quality result that is sample to be assessed that distance is minimum, sample C to be assessed _jBelong to master sample

, i wherein ₁i ₂I _h1,2 ..., the full arrangement without repetition of h, algorithm finishes.

The result of calculation of step (4) is sorted:

d ⁽²⁾(C ₁,E ₂)＜d ⁽²⁾(C ₁,E ₁)＜d ⁽²⁾(C ₁,E ₃)＜d ⁽²⁾(C ₁,E ₄)＜d ⁽²⁾(C ₁,E ₅)

d ⁽²⁾(C ₂,E ₂)＜d ⁽²⁾(C ₂,E ₁)＜d ⁽²⁾(C ₂,E ₃)＜d ⁽²⁾(C ₂,E ₄)＜d ⁽²⁾(C ₂,E ₅)

d ⁽²⁾(C ₃,E ₂)＜d ⁽²⁾(C ₃,E ₁)＜d ⁽²⁾(C ₃,E ₃)＜d ⁽²⁾(C ₃,E ₄)＜d ⁽²⁾(C ₃,E ₅)

So can obtain:

C ₁Belong to master sample E ₂, be the II class, uncontaminated water quality.

C ₂Belong to master sample E ₂, be the II class, uncontaminated water quality.

C ₃Belong to master sample E ₂, be the II class, uncontaminated water quality.

Claims (5)

1. a Landscape Water Quality analytical algorithm comprises the following steps:

(1) set up index set X={x ₁, x ₂..., x _n;

(2) Criterion sample water collection E={E on index set X ₁, E ₂..., E _h, water quality E wherein ₁For I class water quality, E ₂For II class water quality, E ₃For III class water quality, E ₄For IV class water quality, E ₅For V class water quality, pollution level should be clean, uncontaminated, light pollution, middle pollution and heavily contaminated mutually, sets up sample water collection C={C to be assessed ₁, C ₂..., C _kObtain raw data, wherein C _iThe data that collect for sampled point i;

(3) original non-negative single-value data is changed into to Vague collection data, build the Vague environment;

(4) calculate the sample C of water to be assessed _j(j=1,2 ..., Vague collection k) and master sample E _i(i=1,2 ..., the distance between Vague collection h);

(5) the Vague integrated mode of distance-based is identified, and differentiates the water quality of sample to be assessed.

2. a kind of Landscape Water Quality analytical algorithm as claimed in claim 1, is characterized in that, the index set of described step (1) is chosen 4 indexs, is respectively x ₁For chemical oxygen demand (COD) (COD), x ₂For ammonia nitrogen (NH3-N), x ₃For total phosphorus (TP), x ₄For total nitrogen (TN), unit: mg/L. is as index.

3. a kind of Landscape Water Quality analytical algorithm as claimed in claim 1, is characterized in that, the conversion formula that original non-negative single-value data is changed into to the Vague Value Data of described step (3) is

X wherein _Jmax=max{x _J1, x _J2, x _J3, x _J4, x _J5, x _J6, x _J7, x _J8, j=1,2,3,4.

4. a kind of Landscape Water Quality analytical algorithm as claimed in claim 1, is characterized in that, the distance between the sample to be assessed of described step (4) and the Vague collection of master sample is calculated according to following formula:

Wherein:

(1) the Vague value x=[t in A for element x _x, 1-f _x], t _x, f _x, π _x=1-t _x-f _x, S _x=t _x-f _xBeing called as respectively the certainty degree of element x to Vague collection A, negates degree, hesitation degree, core.And definition

(2) Vague collection E and C are at discrete domain X={x ₁, x ₂..., x _nOn be defined as,

E＝([t _E(x ₁),1-f _E(x ₁)],[t _E(x ₂),1-f _E(x ₂)],…,[t _E(x _n),1-f _E(x _n)]),

C＝([t _C(x ₁),1-f _C(x ₁)],[t _C(x ₂),1-f _C(x ₂)],…,[t _C(x _n),1-f _C(x _n)])。

5. a kind of Landscape Water Quality analytical algorithm as claimed in claim 1, is characterized in that, the Vague integrated mode recognition methods of the distance-based of described step (5) is: by step (4) calculate the numerical values recited sequence of distance

, sample C to be assessed _jBelong to master sample

, i wherein ₁, i ₂..., i _h1,2 ..., one of h without the full arrangement repeated.

Images