CN111144910B - Bidding 'series bid, companion bid' object recommendation method and device based on fuzzy entropy mean shadow album - Google Patents
Bidding 'series bid, companion bid' object recommendation method and device based on fuzzy entropy mean shadow album Download PDFInfo
- Publication number
- CN111144910B CN111144910B CN201911384753.6A CN201911384753A CN111144910B CN 111144910 B CN111144910 B CN 111144910B CN 201911384753 A CN201911384753 A CN 201911384753A CN 111144910 B CN111144910 B CN 111144910B
- Authority
- CN
- China
- Prior art keywords
- bidding
- bid
- objects
- fuzzy entropy
- fuzzy
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
- G06Q30/0185—Product, service or business identity fraud
-
- 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
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0609—Buyer or seller confidence or verification
-
- 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
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/08—Auctions
Landscapes
- Business, Economics & Management (AREA)
- Finance (AREA)
- Accounting & Taxation (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of computer science, in particular to a method and a device for recommending 'string bid and accompany bid' objects based on a fuzzy entropy mean shadow album, which comprises the following steps: selecting a pre-survey bidding object and acquiring bidding data; preprocessing data; extracting attribute indexes with positive correlation and attribute indexes with negative correlation with the 'string mark and the accompanying mark', respectively calculating characteristic values under different attribute indexes, and performing normalization processing; constructing a fuzzy entropy mean shadow set model; carrying out three-branch approximate division on the extracted bid inviting and bidding objects under different relevant attributes; fusing the three approximate division results; and outputting the recommended surveyed objects. The fuzzy entropy mean negative album model adopted by the invention is used for constructing and calculating the decision division threshold value pair completely from the aspect of entropy loss, so that subjective errors and irrationality caused by expert experience are avoided, and the model can be used for more effectively approximating and dividing the bidding objects with uncertainty.
Description
Technical Field
The invention relates to the technical field of computer science, in particular to a method and a device for recommending 'string bid and accompany bid' objects based on a fuzzy entropy mean shadow album.
Background
When public bidding is carried out in various industries in the society, a phenomenon that a large company is communicated with other small companies in advance to improve the bid-winning probability for winning a bid exists, namely, accompanying and bidding behaviors exist in the bidding process, but due to data imperfection and investigation sensitivity, investigators often cannot effectively and accurately find companies suspected of participating in bidding and bidding. In the face of the problem, a valuable problem is how to help investigators to find out companies which can carry out the cosmesis and the cross bidding in the bidding process through a useful technical means. The object recommendation method for 'cross bid and accompany bid' behaviors in the existing bidding cases mostly adopts artificial judgment and recommendation, and has strong subjectivity and uncertainty and the defect of inaccurate recommendation.
Disclosure of Invention
In order to solve the problems, the invention provides a 'string bid and companion bid' object recommendation method and device based on a fuzzy entropy mean negative album.
A 'series bid, accompany bid' object recommendation method based on fuzzy entropy mean shadow album includes the following steps:
s1, selecting a bidding target of survey, and collecting bidding data related to the bidding target;
s2, preprocessing data, namely arranging a bidding condition table, a bid winning condition table and a bid amount table according to the bid attracting data;
s3, extracting relevant attribute indexes of the bidding objects, including extracting attribute indexes which are positively correlated with the 'serial bid and companion' and attribute indexes which are negatively correlated from a bidding condition table, a successful bid condition table and a bidding amount table, and respectively calculating the characteristic values of the bidding objects under different attribute indexes;
s4, normalizing the characteristic values of the objects to be bid under different attribute indexes to obtain membership values mu (x) of the objects to be bid under different attribute indexes, wherein mu (x) is more than or equal to 0 and less than or equal to 1;
s5, constructing a fuzzy entropy mean shadow set model, reducing uncertainty difference between a fuzzy entropy mean shadow set and a fuzzy set A by adopting a decision division method, calculating a fuzzy entropy loss function of each bidding object according to decision division actions adopted by each bidding object, and calculating decision division threshold value pairs (alpha, beta) under different attribute indexes by minimizing a total fuzzy entropy loss function;
s6, based on the fuzzy entropy mean value negative album model, membership values mu (x) of the bidding objects under different attribute indexes and decision division threshold value pairs (alpha, beta) under different attribute indexes, performing three-branch approximate division on the bidding objects under different attribute indexes, and respectively dividing the bidding objects under different attribute indexes into a POS domain, a BND domain and a NEG domain to obtain three-branch approximate division results of the bidding objects under different attribute indexes;
s7, fusing three approximate division results of the bidding objects under different attribute indexes to obtain a fused division result;
s8, outputting recommended bidding objects possibly having bidding 'string bid, accompanied bid' behaviors according to the fused division result, wherein the recommendation mode comprises the following steps: preferentially recommending bidding objects classified to a POS domain for bidding objects under attribute indexes positively correlated with the 'string bid and companion bid'; for bidding targets under the attribute indexes having negative correlation with the "string bid and the" accompanied bid ", the bidding targets classified into the NEG field are preferentially recommended.
The utility model provides a move bid "string tender, accompany tender" object processing apparatus based on fuzzy entropy mean value shadow album, includes interconnect's processor, memory and output port, its characterized in that: the memory is used for acquiring bidding data and storing data processed by the processor, and the output port is used for setting and outputting recommended survey objects to the terminal; the processor calculates recommended survey objects by adopting the bid string and co-bid object recommendation method based on the fuzzy entropy mean negative album, and sends the recommended survey objects to an output port.
The invention has the beneficial effects that:
1. the fuzzy entropy mean negative album model adopted by the invention does not depend on any artificial given parameter, and the decision division threshold value pairs (alpha, beta) for three-branch approximate division are constructed and obtained according to the bid data from the viewpoint of uncertainty loss of the bid object in the division process, so that the subjectivity error and irrationality caused by expert experience are avoided, and the model can more effectively approximate the bid object with uncertainty.
2. According to the method, three division results of a to-be-investigated bidding company set are directly given to bidding companies (objects in a positive domain) needing preferential investigation, bidding companies (objects in a negative domain) needing no investigation and bidding companies (objects in a shadow region) needing further information acquisition for decision division according to the fuzzy entropy mean negative album model, all investigated bidding companies do not need to be analyzed one by one, and the working efficiency is greatly improved.
Drawings
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.
Fig. 1 is a flowchart of a bid string and bid accompanying object recommendation method based on a fuzzy entropy mean shadow set according to an embodiment of the present invention;
FIG. 2 shows company c according to an embodiment of the present invention 1 Normalizing the corresponding characteristic value;
FIG. 3 is a fuzzy entropy mean shadow set model according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
Fig. 1 is a flowchart of a bidding 'bid-running and bid-accompanying' object recommendation method based on a fuzzy entropy mean negative album according to an embodiment of the present invention, and the method can be directly applied to survey and prevention monitoring of 'bid-running and bid-accompanying', and the method includes, but is not limited to, the following steps:
s1, selecting a bidding target of survey, collecting bidding data related to the bidding target, and sorting out a company target set to be surveyed, which may have the behavior of accompanying and cross bidding;
s2, preprocessing data, namely respectively sorting out a bidding condition table, a bid winning condition table and a bid amount table according to the bid attracting data;
s3, extracting related attribute indexes, namely extracting attribute indexes positively correlated with the 'bid running and the associated bid' and attribute indexes negatively correlated with the 'bid running and the associated bid' from a bid condition table, a bid condition table and a bid amount table, and respectively calculating characteristic values of bidding objects under different attribute indexes, wherein the 'positive correlation' indicates that the attribute indexes have higher values, the probability of the associated bid running and the associated bid is higher, and the 'negative correlation' indicates that the attribute indexes have lower values, the probability of the associated bid running and the associated bid is higher;
s4, normalizing the characteristic values of the bidding objects under different attribute indexes to obtain membership values mu (x) (0 is more than or equal to mu (x) and less than or equal to 1) of the bidding objects under different attribute indexes, wherein the membership value of the bidding object represents the degree that the object belongs to or does not belong to a company with the companion and the string bid, and under the attribute indexes with positive correlation, the higher the membership value is, the higher the possibility that the bidding object belongs to the company with the companion and the string bid is, and the opposite is realized under the attribute indexes with negative correlation;
s5, constructing a fuzzy entropy mean shadow set model, reducing uncertainty difference between a fuzzy entropy mean shadow set and a fuzzy set A by adopting a decision division method, calculating a fuzzy entropy loss function of each bidding object according to decision division actions adopted by each bidding object, and solving decision division threshold value pairs (alpha, beta) under different attribute indexes by minimizing a total fuzzy entropy loss function, wherein the 'fuzzy entropy mean shadow set model' is a decision division model aiming at a company object set to be investigated; the fuzzy set A refers to a set of company objects to be investigated;
s6, based on the fuzzy entropy mean value negative album model, membership values mu (x) of the bidding objects under different attribute indexes and decision division threshold value pairs (alpha, beta) under different attribute indexes, performing three-branch approximate division on the bidding objects under different attribute indexes, and respectively dividing the bidding objects under different attribute indexes into a POS domain, a BND domain and a NEG domain to obtain three-branch approximate division results of the bidding objects under different attribute indexes;
s7, fusing three approximate division results of the bidding objects under different attribute indexes to obtain a fused division result;
s8, outputting recommended bidding objects possibly having bidding 'string bid, accompanied bid' behaviors according to the fused division result, wherein the recommendation mode comprises the following steps: preferentially recommending bidding objects classified to a POS domain for bidding objects under attribute indexes positively correlated with the 'string bid and companion bid'; for bidding targets under the attribute index having a negative correlation with the "string bid and the accompanied bid", bidding targets classified in the NEG domain are preferentially recommended.
In order to make the embodiments of the present invention clearer and more complete, the steps of the present invention will be described in detail.
Due to the particularity and confidentiality of bidding data in the 'string bid and accompany bid' investigation process and legal regulation problems caused by the leakage of related confidentiality information. In the process of displaying the steps of the embodiment of the invention in detail, the used data are subjected to decryption processing, no specific related information is involved, the bidding data are extracted by experts based on real cases, and the data information subjected to decryption processing does not influence the steps of displaying the embodiment of the invention and the effect of the method created by the invention in actual investigation of cases.
Step 1: suppose that company A invests in selected surveysThe bidding target is subject to the participation of the bidding items P = { P } in the designated investigated time period 1 ,p 2 ,......p i },(i∈N + ) Wherein p is i Indicating an ith bid item; the set of companies participating in bidding together with company a is C = { C = { (C) 1 ,c 2 ,......c i },(i∈N + ) Wherein c is i Indicating the ith company that participated in the bid together with company a.
For clearly explaining the specific process of the present invention in different steps, a set of bid items P = { P } after undergoing a decryption process is now given 1 ,p 2 ,p 3 ,p 4 ,p 5 ,p 6 ,p 7 And the set of companies participating in the bidding C = { C = } 1 ,c 2 ,c 3 ,c 4 And gives specific data information in step 2.
Step 2: and arranging a bidding condition table, a bid winning condition table and a bidding amount table according to the bidding data.
TABLE 1 participating Bidding situation table
In Table 1, "set P" represents all bid items in which bid object A of the selected survey participates during the pre-survey period, where P i (1 ≦ i ≦ 7) representing the ith bid item; "set C" represents the set of all companies participating in bidding with bidding object A for the selected survey, where C i (1. Ltoreq. I.ltoreq.4) represents the ith company participating in the bid of invitation together with company A; a "0" indicates that the subject in set C participated in the bid, whereas a "1" indicates that the subject in set C did not participate in the bid.
As can be seen from Table 1, the bidding object A of the selected survey participated in 7 bidding items during the pre-survey period, and 4 companies participated in some of the 7 bidding items, for example, company c 1 Participated in project p with company A 2 ，p 3 ，p 5 ，p 6 And p 7 。
TABLE 2 table of winning bid
In Table 2, "set P" represents all bid items in which bid object A of the selected survey participates during the pre-survey period, where P i (1 ≦ i ≦ 7) representing the ith bid item; "set C" represents the set of all companies participating in bidding with the bidding object A of the selected survey, where C i (1. Ltoreq. I.ltoreq.4) represents the ith company participating in the bid of invitation together with company A; in the table, "0" represents the bid amount of the company, and "1" represents the bid amount of the company; the "-" in the table indicates that the company is not involved in the bid, whereas the company is involved in the bid.
As can be seen from Table 2, the bid object A of the selected survey participated in 7 bid items during the pre-survey period, and among the 7 items, the item p was bid 4 、p 5 And p 7 . In addition, 4 companies are involved in the partial bidding process, e.g., company c 3 Participated in project p with company A 1 ，p 2 And p 3 And company c 3 The 3 bid items are bid.
TABLE 3 Bidding amount table (Unit: wanyuan)
"set P" represents all bid items participated in by bid object A of the selected survey during the pre-survey period, where P i (1 ≦ i ≦ 7) representing the ith bid item; "set C" represents the set of all companies participating in bidding with bidding object A for the selected survey, where C i (1. Ltoreq. I.ltoreq.4) represents the ith company participating in the bid of invitation together with company A; the non-zero value in the table represents the amount of bid offered by the company to participate in the project, and vice versaA value of 0 indicates that the company is not participating in the bid for the project; the non-zero number in the table indicates that the company has bid for the bid item under the bid amount, whereas the non-zero number indicates that the company has not bid for the bid under the bid amount.
As can be seen from Table 3, the bidding target A of the selected survey participated in 7 bidding items during the pre-survey period, and among the 7 items, the item p was bid 4 、p 5 And p 7 The winning sums are 780 ten thousand, 880 ten thousand and 580 ten thousand respectively. In addition, 4 companies are involved in a partial bid, e.g., company c 3 Participated in project p with company A 1 ，p 2 And p 3 And company c 3 The 3 bid items of the average bidder have bid amounts of 500 ten thousand, 690 ten thousand and 25 ten thousand respectively.
And step 3: extracting related attribute indexes: and extracting attribute indexes positively correlated with the 'bid running and the accompanied bid' and attribute indexes negatively correlated with the 'bid running and the accompanied bid' from the bid condition table, the bid winning condition table and the bid amount table, and respectively calculating characteristic values of the bid objects under different attribute indexes, wherein the attribute indexes positively correlated with the attribute indexes are more likely to accompany the bid running and the bid running, and the attribute indexes negatively correlated with the bid running and the bid winning are more likely to accompany the bid running and the bid running.
Further, the extracted attribute indexes include: the system comprises an attribute index M, an attribute index N, an attribute index P, an attribute index Q and an attribute index Z, wherein the attribute indexes M, N and P are in positive correlation with a coscript and a cross-bid, and the attribute indexes Q and Z are in negative correlation with the coscript and the cross-bid.
Wherein the attribute index M is based on the "bidding condition table" assuming that the set C = { C = 1 ,c 2 ,c 3 ,c 4 Each company c in i The total number of the participated bid items is M) 1 is less than or equal to M and less than or equal to 5), and the larger the value of M is, the company c i The higher the probability of the accompanying target and the serial target is, the more positive the index M is in positive correlation with the accompanying target and the serial target.
Wherein the attribute index N is based on the "bid condition table" assuming that the set C = { C = { (C) } 1 ,c 2 ,c 3 ,c 4 Each company c in i The highest number of continuous participation items is N (1 is more than or equal to N is less than or equal to 5), and the larger the value of N is, the company c i The larger the probability of the co-landmark and the cross-landmark is, the positive correlation between the index N and the co-landmark and the cross-landmark is.
Here, the attribute index P is based on the "winning bid condition table" and the set C = { C } is assumed on the premise that the bid-attracting object a of the survey is selected to win a bid 1 ,c 2 ,c 3 ,c 4 Each company c in i The total number of the participated items is P (1 is less than or equal to P is less than or equal to k), wherein k represents the winning bid number of the bidding object A of the selected survey, and the larger the value of P is, the company c i The larger the probability of the co-landmark and the cross-landmark is, the positive correlation between the index P and the co-landmark and the cross-landmark is.
Wherein the attribute index Q is based on "winning bid condition table", assuming the set C = { C = { C = } 1 ,c 2 ,c 3 ,c 4 Each company c in the} i The highest frequency of the continuous bid-winning items is Q (P is more than or equal to 1 and less than or equal to 5), and the larger the Q value is, the company c i The smaller the probability of the cosmesis and the cross-reference is, the negative correlation between the index Q and the cosmesis and the cross-reference is.
Wherein the attribute index Z is based on the "bid amount table" assuming that the set C = { C = 1 ,c 2 ,c 3 ,c 4 Each company c in i The rate of difference between the bid amount of (b) and the bid amount of company A is Z (0. Ltoreq. Z.ltoreq.1), and the larger the value of Z, the company c i The smaller the probability of the coscript and the cross-linked logo is, on the contrary, the smaller the Z value is, the larger the probability of the coscript and the cross-linked logo is, namely, the negative correlation between the index Z and the coscript and the cross-linked logo is.
Furthermore, the calculation formula of the attribute index Z (Z is more than or equal to 0 and less than or equal to 1) is as follows:
wherein, Z i Represents the rate of difference, pro, between the ith company and the company A to be investigated i Representing a set of items bid by the ith company in common with a company A to be investigated, A j Shows company A to be investigated is atBid amount of jth item, com ij Indicates the bid amount, X, of the ith company in the set C in the jth project j Indicating the magnitude of the bid amount for the jth item.
The five attribute indexes M, N, P, Q and Z are shown, wherein M, N and P are in positive correlation with the probability of the 'coscript and logogram', namely the probability of the 'coscript and logogram' is higher when the value is larger; q and Z are in negative correlation with the probability of the 'coscript, logotype', i.e. the probability of the 'coscript, logotype' is smaller when the value thereof is larger.
In order to more clearly and completely describe each step of the present technical solution, a calculation process of the five attribute indexes is given as an example, and is as follows:
based on the definition of the attribute index M, the characteristic value of the bidding object under the attribute index M can be directly obtained from the table 1 participating in the bidding condition table: c. C 1 ＝6，c 2 ＝c 3 ＝c 4 =3; based on the definition of the attribute index N, the characteristic value of the object to be bid under the attribute index N can be obtained directly from 'Table 1, participated bid condition table': c. C 1 ＝6，c 2 ＝1，c 3 ＝3，c 4 =2; based on the definition of the attribute index P, the characteristic value of the bidding object under the attribute index P can be directly obtained from "table 2. Winning bid condition table": c. C 1 ＝3，c 2 ＝2，c 3 ＝0，c 4 =1; based on the definition of the attribute index Q, the characteristic value of the bidding object under the attribute index Q can be directly obtained from a table 2, a winning bid condition table: c. C 1 ＝0，c 2 ＝0，c 3 ＝3，c 4 =1; based on the definition of the attribute index Z, the bidding object c under the attribute index Z is given 1 The calculation process of the characteristic value is as follows:
TABLE 4 company A and Bidding object c 1 Amount of bid (unit: ten thousand yuan)
Then, according to Table 4, X j ＝{100,10,100,100,100}(1≤j≤7)，A j ＝{580,680,22,780,880,900,580}(1≤j≤7)，pro 1 ＝{p 2 ,p 3 ,p 4 ,p 5 ,p 6 ,p 7 }，com 1j = {0,670,22,740,900,970,590} (1 ≦ j ≦ 7), so company a and bidding object c 1 Rate of difference Z between bid amounts 1 Comprises the following steps:
thus, company A and bid object c 1 Rate of difference Z between bid amounts 1 =0.25, Z can be obtained based on the same processing method 2 ＝0.37，Z 3 ＝0.40，Z 3 ＝0.70。
Based on the analysis, the bidding objects c under different attribute indexes can be obtained i The characteristic values corresponding to each are shown in table 5.
TABLE 5 Bidding object c under different attribute indexes i Corresponding characteristic value
And 4, step 4: and carrying out normalization processing on the eigenvalues obtained under different attribute indexes to obtain normalized eigenvalues, namely the membership value mu (x) of the bidding object under different attribute indexes. The membership degree mu of the bidding objects in the bidding object set is obtained in the process of carrying out three-branch approximate division through the fuzzy entropy mean shadow set A (x) Needs to satisfy 0 ≤ μ A (x) 1, so for different attribute indices in Table 5, object c is bid i The corresponding characteristic values need to be normalized, so that the characteristic values under different attribute indexes have the same dimension, and the step 6 is convenient to performThree approximate divisions, as shown in FIG. 2, are the bidding objects c under different attribute indexes 1 And normalizing the corresponding characteristic value.
Furthermore, for the characteristic values of the objects to be bid under different attribute indexes, the invention adopts a dispersion standardization normalization method to normalize the characteristic values, namely, converting the characteristic values of all the objects to be bid under a certain attribute index to unit size through linear transformation, normalizing the processed data to the interval [0,1], wherein the conversion function is as follows:
wherein, mu (x) represents the membership value of the bidding object under different attribute indexes, and mu (x) belongs to [0,1]]，x i Representing the ith characteristic value in the bid object set,
representing the minimum value of the characteristic values in the set of bidding objects,
representing the maximum value of the feature values in the set of bidding objects.
C can be derived based on the above calculation method and the data in Table 5 1 Normalized eigenvalue under attribute index M
I.e. the 1 st membership value mu (x) 1 ) =1, similarly obtainable c 2 ,c 3 And c 4 The membership values under different attribute indexes are shown in table 6.
TABLE 6 membership values of the bidding objects under different attribute indexes
And 5, constructing a fuzzy entropy mean value shadow set model.
As shown in fig. 3, which is a fuzzy entropy mean shadow set model according to an embodiment of the present invention, a rule for constructing the fuzzy entropy mean shadow set model includes: membership mu of bidding target A (x) The membership degree of the bidding objects is increased to 1, and the variation Area is defined as elongated Area; membership mu of bidding target A (x) Less than or equal to threshold beta, reducing the membership of the bidding objects to 0, and defining the change Area as Reduced Area; membership mu of bidding target A (x) Greater than beta and less than alpha, and converting the membership of the bidding objects into delta * The change Area is defined as Shadowed Area.
Further, the calculation mode of the fuzzy entropy mean shadow set model comprises the following steps:
wherein the content of the first and second substances,
expressing a polynomial of a fuzzy entropy mean value shadow set for carrying out three-branch division on an alpha and beta bidding object according to a decision division threshold value, wherein alpha and beta are two real numbers and satisfy that beta is more than or equal to 0 and less than or equal to delta * ≤α≤1，μ A (x) A membership value indicating the number of bidding targets in the set A, i.e., the degree to which the bidding target belongs to or does not belong to the company "partner, cross bid", δ * And representing membership values of the bidding objects which are divided into shadow areas under different attribute indexes, wherein the shadow areas represent areas which cannot be judged according to the current information.
In order to reduce the uncertainty difference between the fuzzy entropy mean negative album model and the bidding object set A, the invention provides the following decision division method for the bidding objects in the bidding object set A based on the angle of the fuzzy entropy:
1. if a certain bidding object x in the set A to be investigated satisfies mu A (x)≥δ * Then the bidding object x has three decisions of (E1), (R1) and (S1), and the division method is as follows:
(E1) The method comprises the following steps If El (a) e |x)≤El(a r |x)∧El(a e |x)≤El(a s↓ | x), then the object takes action a e I.e. by
(R1): if El (a) r |x)≤El(a e |x)∧El(a r |x)≤El(a s↓ | x), then the object takes action a r I.e. by
(S1): if El (a) s↓ |x)≤El(a e |x)∧El(a s↓ |x)≤El(a r | x), then the object takes action a s↓ I.e. by
2. If a bidding object x in the bidding object set A satisfies mu A (x)≤δ * Then the bidding object x has three decisions of (E2), (R2) and (S2), and the division method is as follows:
(E2) The method comprises the following steps If El (a) e |x)≤El(a r |x)∧El(a e |x)≤El(a s↓ | x), then the object takes action a e I.e. by
(R2): if El (a) r |x)≤El(a e |x)∧El(a r |x)≤El(a s↓ | x), then the object takes action a r I.e. by
(S2): if El (a) s↑ |x)≤El(a e |x)∧El(a s↑ |x)≤El(a r | x), then the object takes action a s↓ I.e. by
Wherein, a e ,a r ,a s↑ ,a s↓ Respectively representing four partitioning operations, el, of the fuzzy entropy mean shadow set for the bidding objects e 、El r 、El s↑ 、El s↓ The fuzzy entropy losses, i.e. the uncertainty losses, caused by different decision partitioning actions are represented respectively. According to the decision division action taken by each bidding object x ∈ U, the fuzzy entropy loss function can be obtained as follows:
El(a|x)＝|E e -E b |
where El (a | x) represents a fuzzy entropy loss function of a bidding object, E b Indicating the fuzzy entropy of the bidding object x at the beginning, E e Indicating that the bidding object x conducts a e Fuzzy entropy after decision division, a = { a = { a } e ,a r ,a s↑ ,a s↓ Denotes a decision division action.
Further, the overall fuzzy entropy loss for all bidding objects in the domain of discourse U can be obtained as follows:
where El (α, β) represents the overall fuzzy entropy loss function for the bidding target by minimizing the overall fuzzy entropy loss function, i.e.
Thus, the decision partition threshold value pair (α, β) for three approximate partitions of the bidding object set is obtained as follows:
wherein, delta * Representing postings divided into shadow areas under different attribute indexesConstant value of membership, delta, of the target object * Having two values, i.e.
And is
And
satisfy the requirement of
Therefore, in the calculation of the decision partition threshold value pair (alpha, beta), no matter what choice is made
Or alternatively
The resulting threshold results were consistent.
Further, the invention solves the delta by fuzzy entropy mean value of the bidding object set, namely uncertainty mean value * Calculating delta * The method comprises the following specific steps:
i. calculating the average fuzzy entropy of all bidding objects in the bidding object set A
Wherein card (·) represents the number of objects with membership not 0 or 1 in the bidding object set, and card (·) ≠ 0.
For the continuum domain of discourse U,
for a discrete universe of discourse U,
wherein the content of the first and second substances,
mean fuzzy entropy of all bidding targets in the bidding target set A, x represents one bidding target, mu A (x) Denotes a membership value of a bidding target in the bidding target set, card (-) denotes the number of targets having a membership of not 0 or 1 in the bidding target set, and S (. Mu.), A ) And a set of objects having a membership degree of not 0 or 1 among the set of bidding objects representing the bidding objects.
Solving formula based on fuzzy entropy proposed by Liang, in average fuzzy entropy
On the basis of (1) calculating an entropy mean value
Corresponding constant value delta of membership degree * ，
Wherein, aiming at the continuous domain U and the discrete domain U, the fuzzy entropy solving formula is as follows:
for a discrete type of domain of discourse U,
for the continuum domain of discourse U,
wherein e (A) represents the fuzzy entropy of the bidding object set A, and i represents the ith bidding pair in the setWherein n represents the total number of bidding objects in the set, x represents one bidding object in the set, and μ A (x) Representing a membership value of the bidding objects in the set.
In summary, δ can be obtained for constructing the fuzzy entropy mean shadow set model * And a decision partition threshold pair (α, β).
Step 6: based on the fuzzy entropy mean shadow set model and the membership values of the bidding objects under different attribute indexes, delta of the fuzzy entropy mean shadow set under the attribute indexes M, N, P, Q and Z can be respectively solved * And a decision partition threshold pair (α, β), as follows:
the membership value under the attribute index M is only 0 and 1, so that a decision division threshold value pair does not need to be obtained, three-branch division can be directly carried out, and bidding objects with the membership values of 1 and 0 are respectively divided into a POS domain and an NEG domain.
An attribute index of N, δ * Taking 0.7236 or 0.2764, deciding the partition threshold value pair (α, β) = (0.8873, 0.1127);
property index P, δ * Taking 0.67 or 0.33, decision partition threshold pair (α, β) = (0.8734, 0.1266);
an attribute index of Q, δ * Taking 0.67 or 0.33, the decision partition threshold pair (α, β) = (0.8734, 0.1266);
attribute index Z, delta * Taking 0.7015 or 0.2985, the decision partition threshold pair (α, β) = (0.8812, 0.1188).
In summary, three approximate partitions can be performed on all bidding objects under different attribute indexes according to the decision partition threshold value pairs under the indexes and the membership values of the bidding objects, so as to obtain the three approximate partitions as shown in table 7 below.
Table 7. Three-branch division result of bidding objects based on fuzzy entropy mean shadow set under different attribute indexes
The fuzzy set divided by the fuzzy entropy mean shadow set model is essentially a company set which is likely to have the behaviors of the coscript and the logotype, and the positive domain, the negative domain and the shadow region formed on the basis of the decision division threshold value pair respectively represent a high-possibility region, a low-possibility region and a region which cannot be judged according to the current information and belong to the company of the coscript and the coscript.
And 7: and fusing the three approximate division results to obtain a fused division result.
In the present invention, M, N, and P have positive correlation with "bid for string and bid for co" and Q and Z have negative correlation with "bid for string and bid for co" so that the bid for co has a greater probability of "bid for string and bid for co" in the M, N, and P attribute indexes, and the bid for NEG has a greater probability of "bid for string and bid for co" in the Q and Z attribute indexes.
From the above, it can be seen that the set of bidding targets in the POS domain is U in the M, N, and P attribute indexes 1 Then U is 1 ＝{c 1 }; in the Q and Z attribute indexes, the set of bidding objects in the NEG domain is U 2 Then U is determined 2 ＝{c 1 ,c 2 }. Further, the set U is solved through the simplest result fusion method 1 And set U 2 Of (2) intersection U * I.e. U * ＝{c 1 H, in which the set U is collected * The bidding object in (1) represents the possibility of having larger 'cosets and serial bids' under the indexes of M, N and P attributes and the attributes of Q and Z; thus, set U * The bidding target in (1) is a company with the highest possibility of "crossing and accompanying bidding" with the bidding target A of the selected survey, namely c 1 And a large possibility of 'accompany mark and string mark' exists between the company A and the company A, so that the company A is worthy of further investigation by investigators and preferentially recommends.
And (3) giving a recommendation suggestion: since the investigator can promote the current investigation progress according to the existing information in the research of bidding "series bidding, accompanying bidding", in the method of the present invention, the investigator recommendation suggestion can also be given according to a single attribute index, as follows:
to is directed atAttribute indexes M, N and P, for bid object c classified into POS domain i Preferentially recommending so as to obtain more attributes to provide data for further decision recommendation, wherein the attribute indexes Q and Z are opposite;
bidding object c for division into BND domain i The investigation is suggested to be suspended, and if the cost such as time is sufficient, more attributes are obtained by considering the investigation to make further decision recommendation;
for attribute indexes M, N and P, for bid object c divided into NEG domain i Basically, suspicion is excluded, and no investigation is recommended, and the attribute indexes Q and Z are reversed.
A bid ' string mark, accompany mark ' object processing apparatus based on fuzzy entropy mean value shadow album, includes interconnect's processor, memory and output port, its characterized in that: the memory is used for acquiring bidding data and storing data in the processing process of the processor, and the output port is used for setting and outputting recommended survey objects to the terminal; the processor calculates the recommended survey object by adopting the 'string bid and accompanying bid' object recommendation method based on the fuzzy entropy mean negative album and sends the recommended survey object to an output port.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (10)
1. A 'series bid, accompany bid' object recommendation method based on fuzzy entropy mean shadow album is characterized by comprising the following steps:
s1, selecting a bidding target of survey, and collecting bidding data related to the bidding target;
s2, preprocessing data, and arranging a bidding condition table, a bid-winning condition table and a bid amount table according to bid-winning data;
s3, extracting relevant attribute indexes of the bidding objects, including extracting attribute indexes which are positively correlated with the 'string bid and accompany bid' and attribute indexes which are negatively correlated from a bidding condition table, a winning bid condition table and a bidding amount table, and respectively calculating characteristic values of the bidding objects under different attribute indexes;
s4, normalizing the characteristic values of the bidding objects under different attribute indexes to obtain membership values mu (x) of the bidding objects under different attribute indexes, wherein mu (x) is more than or equal to 0 and less than or equal to 1;
s5, constructing a fuzzy entropy mean shadow set model, reducing uncertainty difference between a fuzzy entropy mean shadow set and a fuzzy set A by adopting a decision division method, calculating a fuzzy entropy loss function of each bidding object according to decision division actions adopted by each bidding object, and calculating decision division threshold value pairs (alpha, beta) under different attribute indexes by minimizing a total fuzzy entropy loss function;
s6, based on the fuzzy entropy mean negative album model, membership values mu (x) of the bidding objects under different attribute indexes and decision division threshold value pairs (alpha, beta) under different attribute indexes, performing three-branch approximate division on the bidding objects under different attribute indexes, and respectively dividing the bidding objects under different attribute indexes into a POS domain, a BND domain and a NEG domain to obtain three-branch approximate division results of the bidding objects under different attribute indexes;
s7, fusing three approximate division results of the bidding objects under different attribute indexes to obtain a fused division result;
s8, outputting recommended bidding objects possibly having bidding 'string bid, accompanied bid' behaviors according to the fused division result, wherein the recommendation mode comprises the following steps: preferentially recommending bidding objects classified to the POS domain for bidding objects under attribute indexes positively correlated with the 'serial bids and accompanying bids'; for bidding targets under the attribute index having a negative correlation with the "string bid and the accompanied bid", bidding targets classified in the NEG domain are preferentially recommended.
2. The method for recommending bidding 'string bid, accompany bid' objects based on the fuzzy entropy mean shadow album according to claim 1, wherein the characteristic values under different attribute indexes are normalized by a dispersion normalization method, and the normalization calculation method comprises:
wherein, mu (x) represents the membership value of the bidding object under different attribute indexes, and mu (x) belongs to [0,1]]Indicating the degree to which the bid target belongs or does not belong to the company of the companion or the cross bid, x is a radical of a fluorine atom i Representing the ith characteristic value in the original data set,
represents the minimum value of the characteristic values in the original data set,
representing the maximum value of the characteristic values in the original dataset.
3. The method for recommending bidding 'string bid, co-bid' objects based on the fuzzy entropy mean negative album according to claim 1, wherein the construction of the fuzzy entropy mean negative album model comprises:
wherein the content of the first and second substances,
expressing a polynomial of a fuzzy entropy mean shadow set for carrying out three-branch division on the bidding object according to decision division threshold values (alpha, beta), wherein alpha and beta are real numbers and satisfy that beta is more than or equal to 0 and less than or equal to delta * ≤α≤1，μ A (x) Representing membership values of bidding objects in the fuzzy set, A representing the fuzzy set, delta * Representing membership values of bid objects divided into shaded areas under different attribute indexes, wherein the shaded areas representAnd judging the area in which the bid object cannot be judged according to the current information.
4. The method as claimed in claim 3, wherein the membership value δ of the bid object divided into shadow areas under different attribute indexes is used as a membership value δ of the bid object * The calculation method is as follows:
s31, solving the fuzzy entropy of each bidding object in the fuzzy set A according to a fuzzy entropy solving formula, and calculating the average fuzzy entropy of all bidding objects in the fuzzy set A
S32, according to the average fuzzy entropy
Calculating membership constant value delta of bidding objects divided into shadow areas under different attribute indexes * 。
5. The method for recommending bid-inviting 'string bid, companion bid' object based on fuzzy entropy mean shadow set as claimed in claim 4, wherein the average fuzzy entropy mean shadow set
The calculation method of (c) is as follows:
for the continuum domain of discourse U,
for a discrete type of domain of discourse U,
wherein the content of the first and second substances,
represents the average fuzzy entropy of all bidding objects in the fuzzy set A, x represents a bidding object, mu A (x) A membership value representing a bid subject in the fuzzy set, and card (-) representing the number of subjects having a membership of not 0 or 1 in the bid subject set, S (. Mu.) A ) And a set of objects with membership degree not being 0 or 1 in the set representing the bidding objects.
6. The method for recommending bidding 'string bid, companion bid' objects based on fuzzy entropy mean shadow set according to claim 4, wherein the membership constant value δ of the bidding object divided into shadow areas under different attribute indexes * Having two values, i.e.
And is provided with
Satisfy the requirements of
Wherein, the first and the second end of the pipe are connected with each other,
and
all used to represent membership constant value, delta, of bidding object in shadow area * The calculation method comprises the following steps:
7. the method for recommending bidding 'string bid, companion bid' objects based on fuzzy entropy mean shadow set according to claim 1, wherein the fuzzy entropy loss function of each bidding object is calculated by:
El(a|x)＝|E e -E b |
where El (a | x) represents a fuzzy entropy loss function of a bidding object, E e Indicating that the bidding object x conducts a e Fuzzy entropy after decision partitioning, E b Representing the fuzzy entropy of the bidding target x at the beginning, a = { a = { (a) } e ，a r ，a s↑ ，a s↓ Denotes a decision partitioning operation, a e 、a r 、a s↑ 、a s↓ Representing four decision division operations aiming at bidding objects in a fuzzy entropy mean shadow set respectively;
the decision division operation comprises:
(1) If a bid object x in the set A to be investigated satisfies mu A (x)＞δ * Then, the bidding object x has three decisions of (E1), (R1) and (S1), and the division method is:
(E1) The method comprises the following steps If El (a) e |x)≤El(a r |x)∧El(a e |x)≤El(a s↓ | x), then the object takes action a e I.e. by
(R1): if El (a) r |x)≤El(a e |x)∧El(a r |x)≤El(a s↓ | x), then the object takes action a r I.e. by
(S1): if El (a) s↓ |x)≤El(a e |x)∧El(a s↓ |x)≤El(a r | x), then the object takes action a s↓ I.e. by
(2) If a bidding object x in the bidding object set A satisfies mu A (x)≤δ * Then, the bidding object x has three decisions of (E2), (R2) and (S2), and the division method is:
(E2) The method comprises the following steps If El (a) e |x)≤El(a r |x)∧El(a e |x)≤El(a s↓ | x), then the object takes action a e I.e. by
(R2): if El (a) r |x)≤El(a e |x)∧El(a r |x)≤El(a s↓ | x), then the object takes action a r I.e. by
(S2): if El (a) s↑ |x)≤El(a e |x)∧El(a s↑ |x)≤El(a r | x), then the object takes action a s↓ I.e. by
Wherein, mu A (x) Representing membership values of bidding objects in the fuzzy set, A representing the fuzzy set, delta * Representing membership values of bid objects divided into shadow areas under different attribute indexes,
expressing a polynomial of a fuzzy entropy mean value shadow set for carrying out three-branch division on an alpha and beta bidding object according to a decision division threshold value, wherein alpha and beta are two real numbers and satisfy that beta is more than or equal to 0 and less than or equal to delta * ≤α＜1。
8. The method for recommending bidding 'string bid, companion bid' objects based on fuzzy entropy mean negative album according to claim 7, wherein the calculation of the total fuzzy entropy loss of all bidding objects comprises:
where El (α, β) represents the total fuzzy entropy loss function of all bidding objects, x represents one bidding object, el (a | x) represents the fuzzy entropy loss function of one bidding object, and U represents the domain of discourse.
9. The method for recommending bidding 'string bid, co bid' object based on fuzzy entropy mean shadow album according to claim 8, wherein the decision division threshold value pair (α, β) is calculated by:
wherein, delta * And the membership constant value represents the membership constant value of the bidding object divided into the shadow areas under different attribute indexes.
10. A bid ' string mark, accompany mark ' object processing apparatus based on fuzzy entropy mean value shadow album, includes interconnect's processor, memory and output port, its characterized in that: the memory is used for acquiring bidding data, storing data in the processing process of the processor,
the output port is used for setting and outputting recommended survey objects to the terminal;
the processor calculates a recommended survey object by adopting the bidding serial bid and co-bid object recommendation method based on the fuzzy entropy mean negative album of any one of the claims 1 to 9, and sends the recommended survey object to an output port.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911384753.6A CN111144910B (en) | 2019-12-28 | 2019-12-28 | Bidding 'series bid, companion bid' object recommendation method and device based on fuzzy entropy mean shadow album |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911384753.6A CN111144910B (en) | 2019-12-28 | 2019-12-28 | Bidding 'series bid, companion bid' object recommendation method and device based on fuzzy entropy mean shadow album |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111144910A CN111144910A (en) | 2020-05-12 |
| CN111144910B true CN111144910B (en) | 2022-11-15 |
Family
ID=70521367
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911384753.6A Active CN111144910B (en) | 2019-12-28 | 2019-12-28 | Bidding 'series bid, companion bid' object recommendation method and device based on fuzzy entropy mean shadow album |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111144910B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7298897B1 (en) * | 2004-02-11 | 2007-11-20 | United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Optimal binarization of gray-scaled digital images via fuzzy reasoning |
| US8688579B1 (en) * | 2010-06-08 | 2014-04-01 | United Services Automobile Association (Usaa) | Automatic remote deposit image preparation apparatuses, methods and systems |
| CN106503929A (en) * | 2016-11-14 | 2017-03-15 | 西安交通大学 | A kind of method that intellectual analysis enclose mark and string bid behavior |
| CN109858544A (en) * | 2019-01-28 | 2019-06-07 | 重庆邮电大学 | The steel product quality detection method clustered based on section shade collection and density peaks |
| CN109871676A (en) * | 2019-03-14 | 2019-06-11 | 重庆邮电大学 | Three identity identifying methods and system based on mouse behavior |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232518B (en) * | 2019-06-11 | 2023-07-14 | 西北工业大学 | Threat assessment method based on three decisions |
-
2019
- 2019-12-28 CN CN201911384753.6A patent/CN111144910B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7298897B1 (en) * | 2004-02-11 | 2007-11-20 | United States Of America As Represented By The Administrator Of The National Aeronautics And Space Administration | Optimal binarization of gray-scaled digital images via fuzzy reasoning |
| US8688579B1 (en) * | 2010-06-08 | 2014-04-01 | United Services Automobile Association (Usaa) | Automatic remote deposit image preparation apparatuses, methods and systems |
| CN106503929A (en) * | 2016-11-14 | 2017-03-15 | 西安交通大学 | A kind of method that intellectual analysis enclose mark and string bid behavior |
| CN109858544A (en) * | 2019-01-28 | 2019-06-07 | 重庆邮电大学 | The steel product quality detection method clustered based on section shade collection and density peaks |
| CN109871676A (en) * | 2019-03-14 | 2019-06-11 | 重庆邮电大学 | Three identity identifying methods and system based on mouse behavior |
Non-Patent Citations (7)
| Title |
|---|
| A proposed shadowed intuitionistic fuzzy numbers;M. A. El-Hawy 等;《 2015 Tenth International Conference on Computer Engineering & Systems (ICCES)》;20160128;153-160 * |
| Fuzzy Entropy: A More Comprehensible Perspective for Interval Shadowed Sets of Fuzzy Sets;Qinghua Zhang;《 IEEE Transactions on Fuzzy Systems》;20191014;第28卷(第11期);3008-3022 * |
| Hierarchical Three-Way Decisions With Intuitionistic Fuzzy Numbers in Multi-Granularity Spaces;C. Yang 等;《IEEE Access》;20190221;第7卷;24362-24375 * |
| 几类模糊多属性决策方法及其应用研究;张惠民;《中国优秀博硕士学位论文全文数据库(博士)基础科学辑》;20131115(第11期);A002-10 * |
| 基于不确定性分析的阴影集模型研究;高满;《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》;20210215(第02期);I138-658 * |
| 基于视频图像序列的交通对象检测与跟踪算法研究;吕成超;《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》;20130115(第01期);I138-1494 * |
| 直觉模糊多核聚类算法及其在乙烯原料属性聚类中的应用;崔兴华 等;《化工学报》;20170104;第68卷(第021期);739-745 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111144910A (en) | 2020-05-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lampe et al. | Trajectories of efficiency measurement: A bibliometric analysis of DEA and SFA | |
| CN107633265B (en) | Data processing method and device for optimizing credit evaluation model | |
| CN108960833B (en) | Abnormal transaction identification method, equipment and storage medium based on heterogeneous financial characteristics | |
| Daraio et al. | Introducing environmental variables in nonparametric frontier models: a probabilistic approach | |
| CN108898479B (en) | Credit evaluation model construction method and device | |
| WO2017133492A1 (en) | Risk assessment method and system | |
| CN107169628B (en) | Power distribution network reliability assessment method based on big data mutual information attribute reduction | |
| CN108665159A (en) | A kind of methods of risk assessment, device, terminal device and storage medium | |
| WO2007106786A2 (en) | Methods and systems for multi-credit reporting agency data modeling | |
| CN108898476A (en) | A kind of loan customer credit-graded approach and device | |
| Lozano | Process efficiency of two-stage systems with fuzzy data | |
| CN112700319A (en) | Enterprise credit line determination method and device based on government affair data | |
| CN112396211B (en) | Data prediction method, device, equipment and computer storage medium | |
| CN111061709A (en) | Data cleaning automation method and system for similar repeated records | |
| CN113298373A (en) | Financial risk assessment method, device, storage medium and equipment | |
| CN112884590A (en) | Power grid enterprise financing decision method based on machine learning algorithm | |
| Frey et al. | Most powerful rank tests for perfect rankings | |
| CN107133722A (en) | Power distribution network difference characteristic diagnostic analysis method based on electric power enterprise big data technology | |
| CN105447117A (en) | User clustering method and apparatus | |
| CN111144910B (en) | Bidding 'series bid, companion bid' object recommendation method and device based on fuzzy entropy mean shadow album | |
| CN111221915B (en) | Online learning resource quality analysis method based on CWK-means | |
| CN105824785A (en) | Rapid abnormal point detection method based on penalized regression | |
| Tanamal et al. | House price prediction model using random forest in surabaya city | |
| CN114626940A (en) | Data analysis method and device and electronic equipment | |
| CN112581188A (en) | Construction method, prediction method and model of engineering project bid quotation prediction model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |