CN115170166B - Big data sensing method and system for judging monopoly behavior - Google Patents
Big data sensing method and system for judging monopoly behavior Download PDFInfo
- Publication number
- CN115170166B CN115170166B CN202211081363.3A CN202211081363A CN115170166B CN 115170166 B CN115170166 B CN 115170166B CN 202211081363 A CN202211081363 A CN 202211081363A CN 115170166 B CN115170166 B CN 115170166B
- Authority
- CN
- China
- Prior art keywords
- data
- price
- monopoly
- time
- model
- 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
- 238000000034 method Methods 0.000 title claims abstract description 72
- 230000008859 change Effects 0.000 claims abstract description 67
- 230000006399 behavior Effects 0.000 claims abstract description 55
- 238000004458 analytical method Methods 0.000 claims abstract description 35
- 230000000694 effects Effects 0.000 claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims abstract description 16
- 238000009826 distribution Methods 0.000 claims abstract description 16
- 238000004519 manufacturing process Methods 0.000 claims abstract description 16
- 238000012544 monitoring process Methods 0.000 claims abstract description 12
- 230000035772 mutation Effects 0.000 claims description 58
- 238000012360 testing method Methods 0.000 claims description 35
- 238000012545 processing Methods 0.000 claims description 23
- 230000008447 perception Effects 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 18
- 230000000007 visual effect Effects 0.000 claims description 13
- 230000002776 aggregation Effects 0.000 claims description 11
- 238000004220 aggregation Methods 0.000 claims description 11
- 238000007689 inspection Methods 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 9
- 230000011218 segmentation Effects 0.000 claims description 9
- MFYSYFVPBJMHGN-UHFFFAOYSA-N Cortisone Natural products O=C1CCC2(C)C3C(=O)CC(C)(C(CC4)(O)C(=O)CO)C4C3CCC2=C1 MFYSYFVPBJMHGN-UHFFFAOYSA-N 0.000 claims description 8
- 238000007621 cluster analysis Methods 0.000 claims description 8
- 230000007787 long-term memory Effects 0.000 claims description 7
- 230000015654 memory Effects 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 230000004931 aggregating effect Effects 0.000 claims description 2
- 238000010998 test method Methods 0.000 claims description 2
- 239000013598 vector Substances 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 24
- 238000012790 confirmation Methods 0.000 abstract description 2
- 238000001514 detection method Methods 0.000 description 16
- 238000010586 diagram Methods 0.000 description 16
- 230000001364 causal effect Effects 0.000 description 14
- 239000004568 cement Substances 0.000 description 8
- 230000000875 corresponding effect Effects 0.000 description 7
- 238000004364 calculation method Methods 0.000 description 6
- 238000013461 design Methods 0.000 description 6
- 230000002596 correlated effect Effects 0.000 description 5
- 238000012795 verification Methods 0.000 description 5
- 238000007405 data analysis Methods 0.000 description 4
- 238000000354 decomposition reaction Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000007774 longterm Effects 0.000 description 4
- 238000013488 ordinary least square regression Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 108010057186 Insulin Glargine Proteins 0.000 description 3
- COCFEDIXXNGUNL-RFKWWTKHSA-N Insulin glargine Chemical compound C([C@@H](C(=O)N[C@@H](CC(C)C)C(=O)N[C@H]1CSSC[C@H]2C(=O)N[C@H](C(=O)N[C@@H](CO)C(=O)N[C@H](C(=O)N[C@H](C(N[C@@H](CO)C(=O)N[C@@H](CC(C)C)C(=O)N[C@@H](CC=3C=CC(O)=CC=3)C(=O)N[C@@H](CCC(N)=O)C(=O)N[C@@H](CC(C)C)C(=O)N[C@@H](CCC(O)=O)C(=O)N[C@@H](CC(N)=O)C(=O)N[C@@H](CC=3C=CC(O)=CC=3)C(=O)N[C@@H](CSSC[C@H](NC(=O)[C@H](C(C)C)NC(=O)[C@H](CC(C)C)NC(=O)[C@H](CC=3C=CC(O)=CC=3)NC(=O)[C@H](CC(C)C)NC(=O)[C@H](C)NC(=O)[C@H](CCC(O)=O)NC(=O)[C@H](C(C)C)NC(=O)[C@H](CC(C)C)NC(=O)[C@H](CC=3NC=NC=3)NC(=O)[C@H](CO)NC(=O)CNC1=O)C(=O)NCC(=O)N[C@@H](CCC(O)=O)C(=O)N[C@@H](CCCNC(N)=N)C(=O)NCC(=O)N[C@@H](CC=1C=CC=CC=1)C(=O)N[C@@H](CC=1C=CC=CC=1)C(=O)N[C@@H](CC=1C=CC(O)=CC=1)C(=O)N[C@@H]([C@@H](C)O)C(=O)N1[C@@H](CCC1)C(=O)N[C@@H](CCCCN)C(=O)N[C@@H]([C@@H](C)O)C(=O)N[C@@H](CCCNC(N)=N)C(=O)N[C@@H](CCCNC(N)=N)C(O)=O)C(=O)NCC(O)=O)=O)CSSC[C@@H](C(N2)=O)NC(=O)[C@H](CCC(N)=O)NC(=O)[C@H](CCC(O)=O)NC(=O)[C@H](C(C)C)NC(=O)[C@@H](NC(=O)CN)[C@@H](C)CC)[C@@H](C)CC)[C@@H](C)O)NC(=O)[C@H](CCC(N)=O)NC(=O)[C@H](CC(N)=O)NC(=O)[C@@H](NC(=O)[C@@H](N)CC=1C=CC=CC=1)C(C)C)C1=CN=CN1 COCFEDIXXNGUNL-RFKWWTKHSA-N 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 229960002869 insulin glargine Drugs 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000012800 visualization Methods 0.000 description 3
- 238000003657 Likelihood-ratio test Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000000546 chi-square test Methods 0.000 description 2
- 238000012937 correction Methods 0.000 description 2
- 238000007728 cost analysis Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000001976 improved effect Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000008092 positive effect Effects 0.000 description 2
- 238000012827 research and development Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000000630 rising effect Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 238000009960 carding Methods 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000013075 data extraction Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000000227 grinding Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000012847 principal component analysis method Methods 0.000 description 1
- 238000004445 quantitative analysis Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000008093 supporting effect Effects 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/018—Certifying business or products
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Business, Economics & Management (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Human Resources & Organizations (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Primary Health Care (AREA)
Abstract
The invention belongs to the technical field of big data, and discloses a big data sensing method and a big data sensing system for judging monopoly behaviors, wherein the big data sensing method comprises the following steps: monitoring products in key industries, determining the geographical distribution and the market influence distribution of production enterprises, and analyzing the trend difference between the price change of the products in the key industries and the price change of the products in the province; meanwhile, the method is used for identifying the price change period when the market has the risk of abnormal price change and feeding back early warning information; and determining high correlation among enterprises with high influence on enterprise products after administrative filing, analyzing the cost and price cost of the production enterprises, and providing traceable and provable model analysis evidence. The invention can realize early warning and evidence collection of the monopoly illegal activities, carry out data expansion, risk warning and evidence obtaining and evidence fixing by using various algorithms, and realize the discovery and confirmation of the monopoly illegal activities by using an algorithm model and an index system of big data from two dimensions of time and space.
Description
Technical Field
The invention belongs to the technical field of big data, and particularly relates to a big data sensing method and system for judging monopoly behaviors.
Background
At present, in the existing supervision law enforcement mode, the traditional working modes of collecting clues of monopoly cases and judging monopoly illegal behaviors are still mainly depended on-site case handling, inquiry, manual data carding and the like, and are limited by objective factors such as manpower, material resources and the like, and the result efficiency and the quality need to be improved.
In the prior art, economic analysis theory and anti-monopoly business are not combined sufficiently, an algorithm model suitable for the anti-monopoly field is not formed for main illegal behaviors such as 'monopoly protocol and abuse market dominance', the intelligent level of data utilization is not sufficient, a quantitative analysis result cannot be formed according to actual data, and a universal methodology cannot be given.
Through the above analysis, the problems and defects of the prior art are as follows: the existing method for judging monopoly behaviors has low efficiency, accurate judgment result and no universality, and meanwhile, the intelligent level is not enough.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a big data perception method for judging monopoly behaviors.
The invention is realized in such a way that a big data perception method for judging monopoly behaviors comprises the following steps:
by monitoring the information of the key industry products, determining the geographical distribution and the market influence distribution of production enterprises, and acquiring trend dissimilarity analysis data of the key industry product price change and the enterprise product price change; and identifying the price change period reflected when the market has abnormal price change risk based on the analysis data, feeding back the early warning information, and visually displaying.
And determining high correlation among enterprises with high influence on enterprise products after administrative filing, analyzing the cost and price cost of the production enterprises, and providing traceable and provable model analysis evidence.
Further, the big data perception method for judging monopoly behaviors comprises the following steps:
acquiring basic information data, tax data and sales data of an enterprise; cleaning and processing the acquired data;
determining heterogeneity of product price fluctuation characteristics from two space-time dimensions based on a structural mutation theory, and identifying monopoly behaviors of enterprises on product prices;
and thirdly, setting visual parameters, and visually displaying the recognition result and downloading the whole analysis report based on the set visual parameters.
Further, the cleaning and processing of the acquired data in the first step includes:
firstly, processing missing values, standardizing data and detecting abnormal values of acquired data;
secondly, aggregating and converting multiple time dimensions of price data; and storing the processed data into a base library and a subject library which are constructed in advance.
Further, the heterogeneity of the fluctuation characteristics of the product price is determined from two dimensions of time and space based on the structural mutation theory in the second step, and the identification of the monopoly behavior of the enterprise on the product price comprises the following steps:
firstly, constructing a multivariate time sequence model of each enterprise product in a time dimension, identifying a structure variable point of product price data by combining an endogenic multiple mutation inspection model, and constructing a GARCH model to determine the heterogeneity of the product price fluctuation characteristics;
secondly, in the spatial dimension, a spatial metering model is constructed to determine spatial correlation, aggregation and heterogeneity of different enterprise product prices.
Further, the step two is to determine the heterogeneity of the fluctuation characteristics of the product price from two dimensions of time and space based on the structural mutation theory, and the step of identifying the monopoly behavior of the enterprise on the product price comprises the following steps:
(1) Performing cluster analysis on the processed data to obtain a plurality of groups, and calculating the intra-group correlation coefficient and the inter-group correlation coefficient of each group in which each enterprise is located to obtain the characteristic of price collusion and the data with monopoly characteristic;
(2) The idea of the Bai-Perron endogenous multiple structure mutation test method is as follows:
assuming that m potential mutation points exist in time series data of a certain T phase (m +1 segmentation intervals are generated), the data generation process is as follows:
wherein,
；/>
，/>
for the value of an interpretation variable at time t, the interpretation variable is determined by &>
And &>
Is composed of two parts, is->
Variable which signifies that a coefficient has not changed>
A variable representing a change in a coefficient>
And &>
For the corresponding coefficient vector, is>
Is a residual term->
Is m unknown structural mutation points, the date on which the structural mutation points are to be defined being ^ based>
；
The Bai-Perron structural mutation test is divided into three steps: first, calculating each possible segmentation point in the formula by using the Ordinary Least Squares (OLS)
And &>
And obtaining a corresponding sum of squares of residuals; and secondly, comparing the residual square sums obtained by different segmentation modes, and taking the minimum residual square sum for segmentation: the third step is toWhether structural mutation occurs in the generation process of the time sequence is subjected to significance test;
and carrying out structural mutation detection on the market price aggregation sequence of a certain product in a certain region by utilizing a structural mutation point algorithm of Bai-Perron. Observation of data (y) Using time series t ，x t ，z t ) Simulation generation of structural change points (T) 1 8230Tm), estimating the model according to the principle of least square, and testing the statistic to judge whether structural mutation occurs. At the significance level of 5%, both UDmax and WDmax statistics were greater than their critical values, respectively, indicating the presence of structural mutations in the original sequence. From the values of the F statistics, it can be seen that the F statistics of the 5 detected structural mutation points are significantly larger than their corresponding critical values, indicating that a total of 5 structural mutations occurred during the sample period. Wherein max represents the maximum allowable value of the number of structural mutations, and the two statistics UDmax and WDmax are used to test whether there is an unknown number of structural mutations given the maximum allowable value max of the number of structural mutations; the F statistic is a statistic conforming to the F distribution when the null hypothesis is established.
(3) Whether structural mutation occurs in the generation process of the time sequence is subjected to significance test to obtain a plurality of structural change points, whether fluctuation with long-term memory is generated or not is tested in a time interval divided by two adjacent structural change points, and the GARCH model is a time sequence model for describing fluctuation rate and obtaining a good effect; for a time series
Let us order
In combination with a scale>
Obeying the GARCH (m, s) model>
If it satisfies
Wherein,
representing a rate-of-return information based on the expiration time>
Is constant and is->
For an uncorrelated white noise sequence>
Is a fluctuation rate, is a conditional standard deviation of the rate of return, is based on>
An independent co-distributed white noise column that is zero mean unit variance, and>
is based on a pattern of>
M and s are parameters of the GARCH model, alpha and beta are constant terms, and when the coefficient in the modeling of the GARCH model is close to 1, a fluctuation aggregation effect is displayed;
and (3) judging whether the time interval is a price change period or not by utilizing a GARCH (1, 1) model after the structure change point of a certain product in a certain area is divided. During the period from 2017/09/18 to 2017/12/12, the GARCH (1, 1) model is fitted to find that the sum of the coefficients (alpha and beta) is almost 1, which indicates that the fluctuation rates during the period are remarkably aggregated, long-term memory exists, and the product price enters the price change period.
(4) Establishing a spatial correlation model for products of all enterprises, and adopting first-order time correlation coefficient analysis to calculate the intra-group correlation coefficient and the inter-group correlation coefficient of each enterprise group according to different groups obtained by clustering analysis;
further, the step (1) of calculating the intra-group correlation coefficient and the inter-group correlation coefficient of the group in which each enterprise is located to obtain the feature of price collusion and the data of monopoly feature includes:
1) In the time dimension, whether the two time series have similar variation trends is calculated:
wherein, X T 、Y T Two time series of finger inputs, X t 、X t+1 、Y t 、Y t+1 Respectively at time t and time t +1 T 、Y T The value of the sequence, CORT, is the value of the correlation coefficient.
2) Determining
The properties of (A): />
Wherein->
The two time sequences have the same trend and can rise or fall simultaneously, and the rising amplitude or the falling amplitude are the same;
showing the opposite trend of rise and fall of the two time series; />
Indicates that the two time series have no correlation in monotonicity;
3) For enterprises with sudden changes of product price structures and entering price change periods, a plurality of enterprises in a group have strong system correlation, and the correlation among the groups is weak: the inter-group correlation presents the characteristics of high value-low value, the possibility of price leadership of enterprises which judge that several pieces of furniture have market dominance capacity is increased, and the characteristics of price collusion are achieved; and (4) enterprises with any mutation of product price structures and entering the price change period, wherein the inter-group correlation in the group presents the characteristics of low value-low value, and the enterprises are judged to have monopoly characteristics.
Calculating the correlation coefficient between the price sequence group and the enterprise of the enterprise entering the price change period, and defining 0.8-1.0 strong correlation according to the actual modeling condition; 0.6-0.8 strong correlation; 0.4-0.6 moderate correlation; 0.2-0.4 weakly correlated; 0.0-0.2 are very weakly or not correlated. And judging whether the monopoly suspicion exists or not by comparing the average correlation coefficient index and the maximum correlation coefficient index.
Another object of the present invention is to provide a big data perception system for judging monopoly behavior implementing the big data perception method for judging monopoly behavior, the big data perception system for judging monopoly behavior comprising:
the risk early warning module is used for monitoring products in key industries, determining the geographical distribution and the market influence distribution of production enterprises, and analyzing the trend difference between the price change of the products in the key industries and the price change of products in the province; meanwhile, the method is used for identifying the price change period when the market has the risk of abnormal price change and feeding back early warning information;
and the evidence support module is used for determining high correlation among enterprises with high influence on enterprise products after administrative filing, analyzing the cost and price cost of the production enterprises and providing traceable and provable model analysis evidence.
It is a further object of the invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of the big data awareness method for determining monopoly behavior.
Another object of the present invention is to provide a computer-readable storage medium, storing a computer program which, when executed by a processor, causes the processor to perform the steps of the big-data-aware method for determining monopoly behavior.
Another object of the present invention is to provide an information data processing terminal for implementing the big data sensing system for judging monopoly behavior.
In combination with the above technical solutions and the technical problems to be solved, please analyze the advantages and positive effects of the technical solutions to be protected in the present invention from the following aspects:
first, aiming at the technical problems existing in the prior art and the difficulty in solving the problems, the technical problems to be solved by the technical scheme of the present invention are closely combined with results, data and the like in the research and development process, and some creative technical effects are brought after the problems are solved. The specific description is as follows:
the technology of the invention is to judge the monopoly behavior of enterprises on the product price, and research the heterogeneity of the fluctuation characteristics of the product price from two dimensions of time and space based on the structural mutation theory and the space statistics. In the time dimension, a multivariate time sequence model of each enterprise product is mainly established, the structural change points of product price data are identified by means of an endogenous multiple mutation inspection model, and a GARCH model is established to explore the heterogeneity of the product price fluctuation characteristics; and in the spatial dimension, a spatial metering model is constructed to research the spatial correlation, aggregation and heterogeneity of the product prices of different enterprises. The application and the construction of the method of the inventive algorithm model adopted by the technical scheme of the invention fill the combination of the economic analysis theory and the anti-monopoly business at home and abroad, improve the judgment accuracy and strengthen the judgment support basis.
The technical scheme provided by the invention has the beneficial effects that:
1. the invention makes the anti-monopoly law enforcement specialized and refined, and greatly improves the reliability and effectiveness of the anti-monopoly work.
2. The present invention enhances and innovations market price regulation.
3. The invention is beneficial to timely finding out the problems existing in the enterprise operation process, and effectively reduces the cost for correcting the illegal behaviors.
4. The invention ensures the utilization efficiency of data, verifies the validity of anti-monopoly detection work, enriches the monitoring dimension and effectively improves the reliability of anti-monopoly monitoring results.
Secondly, considering the technical scheme as a whole or from the perspective of products, the technical effect and advantages of the technical scheme to be protected by the invention are specifically described as follows:
the invention can realize early warning and evidence collection of the monopoly illegal behaviors, and carry out data expansion, risk warning, evidence collection and evidence fixation by using various algorithms. The invention is based on the theory of structural mutation, and realizes the discovery and confirmation of the monopoly illegal behaviors by using an algorithm model and an index system of big data from two dimensions of time and space.
(1) The invention can bring more convenient judgment means to users/use units, reduce personnel investment, improve judgment accuracy and improve judgment efficiency, and provides a method and a mode taking big data as means for judging monopoly behaviors and carrying out risk early warning.
(2) According to the invention, through analysis and utilization of historical data and real-time data, the reliability of judging the monopoly behavior result is improved, and effective foundation support is provided for decision making.
(3) The invention realizes the falling of academic theory in the field of anti-monopoly and the digitization of 'transverse monopoly' supervision, and realizes the 'quasi-real-time' intelligent supervision on non-digital economy.
Drawings
FIG. 1 is a flow chart of a big data perception method for determining monopoly behavior according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a big data perception algorithm model system for determining monopoly behavior according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of data cluster analysis provided by an embodiment of the present invention;
FIG. 4 is a schematic diagram of a Bai-Perron structural mutation test provided in the embodiment of the present invention;
FIG. 5 is a diagram of the test results provided by an embodiment of the present invention;
FIG. 6 is a chart of experimental examinations provided by an embodiment of the present invention;
FIG. 7 is a Granage causal test chart provided by an embodiment of the present invention;
fig. 8 (a) is a diagram before abnormal value detection processing provided by the embodiment of the present invention;
FIG. 8 (b) is a diagram illustrating an abnormal value detection process according to an embodiment of the present invention;
FIG. 9 is a timing diagram verification provided by an embodiment of the invention;
FIG. 10 is an exploded view of a time sequence provided by an embodiment of the present invention;
FIG. 11 (a) is an autocorrelation diagram of a smoothed sequence provided by an embodiment of the present invention;
FIG. 11 (b) is a partial correlation diagram of the smoothed sequence provided by an embodiment of the present invention;
fig. 12 is a diagram of pulse verification based on a VECM model provided by an embodiment of the invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
1. Illustrative embodiments are explained. This section is an explanatory embodiment expanding on the claims so as to fully understand how the present invention is embodied by those skilled in the art.
As shown in fig. 1, a big data perception method for determining monopoly behavior according to an embodiment of the present invention includes the following steps:
s101, acquiring basic information data, tax data and sales data of an enterprise; cleaning and processing the acquired data;
s102, determining heterogeneity of product price fluctuation characteristics from two space-time dimensions based on a structural mutation theory, and identifying monopoly behaviors of enterprises on product prices;
s103, setting visualization parameters, and visually displaying the recognition result and downloading the whole analysis report based on the set visualization parameters.
Example 1
The embodiment of the invention provides a big data perception system for judging monopoly behaviors, which comprises:
and the risk early warning module is used for realizing daily monitoring on the key industry products, so that law enforcement personnel can clearly and efficiently check the geographical distribution and market influence distribution of production enterprises, and can analyze the trend difference between the key industry product price change and the province product price change according to the requirements of the law enforcement personnel. And if the market has the risk of abnormal price change, identifying the price change period and feeding back early warning information.
And the evidence support module is used for finding high correlation among enterprises with high influence aiming at enterprise products after administrative filing, and providing powerful, traceable and testable model analysis evidence for administrative punishment of monopoly behaviors after acquiring the cost of production enterprises and carrying out price cost analysis.
Example 2
The embodiment of the invention provides a big data perception method for judging monopoly behaviors, which comprises the following steps:
the method mainly relates to data acquisition, data processing, database storage, algorithm model construction and 5 parts of front-end visual display.
1. Data acquisition: the method mainly relates to basic information data, tax data and sales data of enterprises, and the acquisition method is mainly provided by the inside of a market bureau and related outside hall bureaus.
2. Data processing: after data acquisition, data cleaning and processing are carried out, wherein the data cleaning comprises processing of missing values, data standardization and detection processing of abnormal values, and the data processing comprises aggregation and conversion of multiple time dimensions of price data. And finally, performing data warehousing to form a basic library and a theme library. The data in the data lake enters a corresponding theme library, a special theme library and the like after being processed by data extraction, data cleaning, data conversion and the like, and supports algorithm modeling and data application services.
3. And (3) database storage: based on the business scene of the anti-monopoly data analysis, the MySQL database is selected for the consideration of the volume storage of the related industry product data. In the process of structuring and abstracting the original data, the concept structure design, the logic structure design and the physical structure design of the database are formed. The model algorithm part is arranged as an interface, can be embedded and is used for performing timing off-line calculation on the newly-put data.
4. An algorithm model: in order to judge the monopoly behavior of enterprises on product prices, the heterogeneity of product price fluctuation characteristics is researched from two dimensions of time and space on the basis of a structural mutation theory. In the time dimension, a multivariate time sequence model of each enterprise product is mainly established, the structural change points of product price data are identified by means of an endogenic multiple mutation inspection model, and a GARCH model is established to explore the heterogeneity of the product price fluctuation characteristics; and in the spatial dimension, a spatial metering model is constructed to research the spatial correlation, aggregation and heterogeneity of the product prices of different enterprises.
5. Visual display: the visual application of the invention is characterized by strong interaction, and ensures that law enforcement personnel can adjust the cases individually, such as selecting sub-interfaces to be displayed, selecting enterprises displayed in visual images, and downloading visual images and overall analysis reports.
In the embodiment of the invention, the model construction mode is as follows:
firstly, the model of the invention is a time series model based on the fluctuation rate, and has good model prediction effect. For a time series
Make->
Wherein->
Representing a rate-of-return information based on the expiration time>
Is constant and is->
For an unrelated white noise sequence, as shown in fig. 2, an anti-monopoly big data algorithm model system integrating cluster analysis, a multivariate time sequence model, a spatial correlation model and a price cost resonance model starting from product prices is built and realized.
Cluster analysis is a technique for statistical data analysis. The cluster analysis can reasonably carry out automatic classification according to respective characteristics, particularly carry out cluster classification under the condition of no prior knowledge; clustering is to divide similar objects into different groups or more subsets (subset) by means of static classification, so that all the object members in the same subset have similar attributes, which are usually included in a shorter spatial distance in a coordinate system.
PCA is a principal component analysis method, reduces characteristic dimensionality by mapping data from high dimension to low dimension, simultaneously reserves more characteristic information as much as possible, and then adopts a K-means clustering algorithm to group objects according to multi-characteristic variables of the objects;
the K-means algorithm is a clustering algorithm that divides a given data set into K clusters, and is also referred to as K-means because it finds K different clusters, and the center of each cluster is calculated using the mean of the data contained in the cluster, and the number of clusters, K, is user-specified, and each cluster is described by its centroid (centroid), i.e., the center of all points in the cluster. The biggest difference between clustering and classification algorithms is that the classified target class is known and the clustered target class is unknown, i.e. clustering is an unsupervised machine learning process.
In the embodiment of the invention, the K-means learns the target objects in an unsupervised mode, the number of the learned objects is uncertain, the K value of the number in a given cluster needs to be manually selected in advance, and the K value selection is very difficult to estimate in practice because different cluster centers can cause completely different clustering results. That is, the result is not necessarily global optimum, but only local optimum is guaranteed; the method is not suitable for finding clusters with non-convex shapes or clusters with large size differences, and the sample has a mean value;
the internal logic and the calculation method of the K-means are as follows:
the method according to the K-means calculation has the following three distinct features:
(1) The method is suitable for classification without past experience. For strange things, the classification of the things is unknown, no method is available, no idea is provided, forced classification is very subjective, and the things have personal colors. By collecting data dimensions and indexes and using a cluster analysis method for modeling, automatic classification can be obtained, and the method is scientific and reasonable;
(2) Multiple dimensions may be classified. Customer classification is often divided into large and small customers by the amount of consumption, but such classification knowledge mirrors one dimension: the amount of the charge is consumed. However, such classification is very simple and does not guide sales activities well, and it is desirable to combine the consumer's ability, preference characteristics, etc. for classification in addition to the amount of consumption. Therefore, the conventional method cannot be classified, but K-means can do the classification;
(3) The clustering analysis method is an exploratory self-learning method, and belongs to one of statistical learning methods. Clustering analysis is a common technology in data mining machine learning, and can group objects according to the intrinsic characteristics and rules and the principle of characteristic and index distance through indexes and dimensions, namely the similarity principle.
In the embodiment of the invention, in the actual commercial activity, it is important to research consumers and classify the consumers appropriately. The present invention often classifies consumers in three ways:
(1) Traditional experience, which is classified by customer management personnel according to familiar characteristics;
(2) Traditional data description, which is to count according to client data and classify by setting boundaries subjectively;
(3) Machine learning methods, i.e., classification using data mining and artificial intelligence methods, such as K-means.
The combination of the latter two methods is desirable for the present invention, and is also the area where clustering methods are good. The clustering method can also seek analysis variables from different angles, and provides reference for the decision of a certain theme.
Example 3
As shown in fig. 3, in embodiment 2, the embodiment of the present invention performs cluster analysis on the processed data to obtain a plurality of groups, and calculates the intra-group correlation coefficient and the inter-group correlation coefficient of each group in which an enterprise is located to obtain a feature with price collusion and data with monopoly feature;
preferably, a K-means method is adopted for clustering analysis, and the implementation steps are as follows:
1. randomly taking K elements from a sample as respective centroids of K categories;
2. respectively calculating the distances from other elements to the centroids of the K categories, and respectively dividing the element with the minimum distance from the K categories into the K categories;
3. calculating the average value of the elements in each current class as a new class center;
and repeating the steps 2 and 3 until the clustering center is not changed any more, stopping the algorithm and outputting the result, otherwise, continuing to execute.
Example 4
As shown in fig. 4, the big data perception method for determining monopoly behavior provided by the embodiment of the present invention specifically includes the following steps, taking price change recognition, namely, bai-Perron structure mutation test as an example:
step one, aiming at each possible division point in the formula, calculating an estimated value of a constant coefficient by using an ordinary least square method (OLS), and obtaining a corresponding residual square sum;
secondly, comparing the residual square sums obtained by different segmentation modes, and segmenting the minimum residual square sum;
and thirdly, carrying out significance test on whether structural mutation occurs in the generation process of the time sequence.
And carrying out structural mutation detection on the market price aggregation sequence of a certain product in a certain region by utilizing a structural mutation point algorithm of Bai-Perron. Observation of data (y) Using time series t ，x t ，z t ) Simulation generation of structural change points (T) 1 ,…T m ) And estimating the model according to the least square principle, and checking the statistic to judge whether the structural mutation occurs. At the significance level of 5%, both UDmax and WDmax statistics were greater than their critical values, respectively, indicating the presence of structural mutations in the original sequence. From the values of the F statistics, it can be seen that the F statistics of the 5 detected structural mutation points are significantly larger than their corresponding critical values, indicating that a total of 5 structural mutations occurred during the sample period. Wherein max represents the maximum allowable value of the number of structural mutations, and the two statistics UDmax and WDmax are used to test whether there is an unknown number of structural mutations given the maximum allowable value max of the number of structural mutations; the F statistic is a statistic that matches the F distribution when the null hypothesis holds.
After obtaining a plurality of structure change points, checking whether the time interval divided by two adjacent structure change points generates fluctuation with long-term memory, and then constructing a fluctuation rate model.
Price change recognition-GARCH model
GARCH model characteristics:
the GARCH model can characterize the clustering of market fluctuations, i.e., a larger fluctuation amplitude followed by a larger fluctuation amplitude, and a smaller fluctuation amplitude followed by a smaller fluctuation amplitude. The CARCH model also has a good prediction effect, and can estimate the established model parameters more accurately, so that the prediction effect is good. The conditional variance of the GARCH model is a function of not only the lag residual squared, but also the lag conditional variance. Estimation of the GARCH model parameters is simpler and more convenient than the ARCH model because a lower-order GARCH model can be used instead of a higher-order ARCH model. The GARCH model is more convenient and accurate to use than the ARCH model in terms of identification and estimation of the model.
Principle of the GARCH model
The GARCH model is used for describing the fluctuation rate to obtain good effect, but a higher order may be required in actual modeling, for example, an ARCH model of 11 orders is used in modeling of the Euro exchange rate fluctuation rate. GARCH model for a logarithmic rate of return sequence
Make->
For its information sequence, is called { -H {>
Obey the GARCH (m, s) model if
Satisfy +>
，/>
Wherein->
Representing a rate-of-return information based on the expiration time>
Is constant and is->
Is an uncorrelated white noise sequence, is asserted>
Is the fluctuation ratio, is the conditional standard deviation of the yield,
independently identically distributed white noise column in variance of zero mean unit>
Is based on a pattern of>
M and s are parameters of the GARCH model, alpha and beta are constant terms>
This last condition is used to ensure that the unconditional variance of the model is satisfied is finite and constant, whereas the conditional variance may vary with time t.
When the coefficient in the GARCH model is close to 1, the fluctuation aggregation effect is shown, and the product price has long-term memory, which shows that the price change is a long-term behavior, and the product price enters a price change period.
Example (c): and (3) judging whether the time interval is a price change period or not by utilizing a GARCH (1, 1) model after the structure change point of a certain product in a certain area is divided. During the period from 2017/09/18 to 2017/12/12, the GARCH (1, 1) model is fitted to find that the sum of the coefficients (alpha, beta) is almost 1, which indicates that the fluctuation rate during the period is remarkably gathered, long-term memory exists, and the product price enters the price change period.
As shown in fig. 5, is an example of a verification effect graph:
and establishing a spatial correlation model for products of all enterprises, calculating the intra-group correlation coefficient and the inter-group correlation coefficient of each enterprise group according to different groups obtained by clustering analysis, wherein the calculation of the correlation coefficient adopts first-order time correlation coefficient analysis.
First-order time correlation coefficient calculation mode: in the time dimension, whether two time series have similar variation trend is calculated
1、
，
Wherein, X T 、Y T Two time series of finger inputs, X t 、X t+1 、Y t 、Y t+1 Respectively at time t and time t +1 T 、Y T The value of the sequence, CORT is the value of the correlation coefficient;
2、
3、
indicating that the two time series hold similar trends, they will rise or fall simultaneously, and the rise or fall is also similar.
4、
Showing that the rising and falling trends of the two time series are just opposite.
Indicating that the two time series are not correlated in monotonicity.
For an enterprise with a product price structure mutation and entering a price change period, a plurality of enterprises in a group have strong system correlation, and the intergroup correlation is weak, namely the intergroup correlation in the group has the characteristic of high value-low value, so that the possibility of price leadership of several enterprises with market dominance capability is preliminarily considered to be increased, and the enterprise price collusion has the characteristic of price collusion. If a product price structure mutation occurs and enters a business in a price change period, the inter-group correlation in the group presents a characteristic of low value-low value, and the business is preliminarily considered to have a characteristic of monopoly and possibly manipulate the price for a business alone.
Example (c): calculating the correlation coefficient between the price sequence group and the enterprise of the enterprise entering the price change period, and defining 0.8-1.0 strong correlation according to the actual modeling condition; 0.6-0.8 strong correlation; 0.4-0.6 moderate correlation; 0.2-0.4 weakly correlated; 0.0-0.2 are very weakly or not correlated. And judging whether the monopoly suspicion exists or not by comparing the average correlation coefficient index and the maximum correlation coefficient index.
Shown in fig. 6 is an experimental verification chart:
price cost analysis-price-cost resonance model:
and establishing a product price and cost price coordination model and an error correction model by using methods such as nonlinear Langery causal test, impulse response, variance decomposition and the like, inspecting whether price variation and cost variation trends are consistent, and judging whether price mutation is caused by cost variation promotion. If the product price and the cost price have the grand causal effect and the co-integration relation is established, the product price and the cost price have the tendency of resonance, the rise of the product price is probably caused by the rise of the cost price, otherwise, the possibility of price manipulation of the enterprise is suspected.
Considering two price sequences, t and pz, t-pz is considered the cause of granger, if past and present values, provide information that predicts pz, t +1 at time t, proving that the information is not relevant to true causal relationships, but rather to temporal priorities. Using the grand causal price test for market definition has been summarized based on a concept that if a price rule is established in both regions, price fluctuations in one region will necessarily translate into price fluctuations in the other region; if two products or regions share a common market, the price of one region or product should be driven by granger to the price of another region or product.
The glargine cause and effect is a market-defined price test that faces three significant challenges. First, he imposes a one-way logic on the market definition. The granger causal test is essentially an assessment of the direction of association, and those in the industry claim that the emphasis is not on the direction itself. However, two-way discovery should be interpreted as more favorable evidence of market integration than one-way discovery.
Second, the granger causal test generally employs a method that emphasizes statistical significance at the expense of economic significance. For the market definition test, the question of concern is not only whether there is a statistically significant relationship between two price sequences, but also whether two price sequences are meaningfully related to their extent of passing the SSNIP test. While the glargine causal test is intended to overcome the use of any critical correlation level in assessing price related statistics, its application typically involves abandoning the scale and using statistical fitting, as shown in figure 7, the glargine causal test results indicate that cost is not a significant causal relationship to price.
2. Application examples. In order to prove the creativity and the technical value of the technical scheme of the invention, the part is the application example of the technical scheme of the claims on specific products or related technologies.
The big data perception method for judging monopoly behaviors provided by the embodiment of the invention is applied to supervision and monitoring of Shandong province markets, can be used for establishing a corresponding algorithm model by combining data analysis and anti-monopoly economic analysis, building an intelligent anti-monopoly perception platform by taking the algorithm model as a foundation, integrating enterprise behavior data of different industries, different fields and different supervision departments, determining dynamic identification indexes of enterprise monopoly behaviors, improving the efficiency of anti-monopoly work and the accuracy of data analysis, finally obtaining reliable industrial anti-monopoly analysis results and realizing anti-monopoly 'pre-prediction'.
Step S101, acquiring basic information data, tax data and sales data of an enterprise, and cleaning and processing the acquired data, wherein the specific process is as follows:
taking the data related to cement in this embodiment as an example, the registration information and the production permission information of the cement-certified enterprise are integrated to generate a cement-certified enterprise information table, and the scale and the number of the grinding production line, the clinker production line and the respective production lines are extracted from the product detail field.
And (3) creating a cement goods name dictionary table, and mapping the cement goods names which are inconsistent in different enterprise invoices into the same cement goods name, so as to solve the problem of entity identification. The HWMC column is used for distinguishing bulk bags and bagged bags, and the HWGG column is not used for distinguishing bulk bags and bagged bags.
And a screening target column is associated with a cement evidence obtaining enterprise information table, the name of a screening seller is the enterprise of a cement company, the data of a waste invoice is deleted, the data with a negative value is deleted, the purchase invoice and the sales invoice are distinguished, and the generation of a cement enterprise tax sales data information table after processing can be seen.
Outliers appear, sigma is the standard deviation according to the '3 sigma' principle, and extreme outlier data exceeding the error of + -3 sigma are deleted. Before the abnormal value detection processing in fig. 8 (a) and after the abnormal value detection processing in fig. 8 (b); the overall trend is more fit with the normal trend after the abnormal data is deleted, and the purpose of reducing the interference of the abnormal data on the result is achieved.
And after the data processing is finished, performing data warehousing to form a basic library and a theme library. In the embodiment, based on a service scene, a MySQL database is selected, and in the process of structuring and abstracting the original data, the concept structure design, the logic structure design and the physical structure design of the database are formed. The model algorithm part is arranged as an interface, can be embedded and is used for performing timing off-line calculation on the newly-put data.
Step S102, determining heterogeneity of product price fluctuation characteristics from two dimensions of time and space based on a structural mutation theory, and identifying monopoly behavior of enterprises on product prices, wherein the specific process is as follows:
in this embodiment, all enterprises in the whole province that produce a certain product are clustered according to indexes such as capacity, total sales, market share, and the like, and the enterprises are divided into multiple groups such as "have strong market influence", "have certain market influence", and "do not have market influence", and the enterprises in the group "have strong market influence" are observed and detected in a key way.
In the embodiment, irrational increase of the price of one or more enterprises can be observed, and the price trend is subjected to jump risk detection, namely, price structure mutation detection and nonparametric detection on a time sequence. Adding structure change points into the model, investigating possible mutation points one by one, selecting the minimum statistical value from the test results to compare with a critical value, finding that the original hypothesis is rejected, and considering that the structural mutation of the price occurs.
A Garch model based on a price time sequence is established for aggregation analysis, in the embodiment, the conclusion can be drawn on the data which is substituted into the Garch model and shows the fluctuation aggregation effect, the product price has long-term memory, and the price change is a long-term behavior, and the product price enters a price change period.
And then, establishing a spatial correlation model, and calculating the intra-group correlation coefficient and the inter-group correlation coefficient of each enterprise group according to different groups obtained by clustering analysis. And (4) carrying out causal test after the model is established, and carrying out spatial correlation test on enterprise products entering the price change period. In this embodiment, it is verified that a plurality of enterprises in a group in which an enterprise with a sudden change in product price structure enters a "price change period" is in strong system correlation, and the inter-group correlation is weak, that is, the inter-group correlation in the group exhibits a "high-value-low-value" characteristic, so that the probability of "price lead" of several enterprises with market dominance capability is increased, and a conclusion that the enterprises have a price "collusion" characteristic is preliminarily drawn.
And then establishing a product price and cost price coordination model and an error correction model for inspecting whether the price variation and the cost variation trend are consistent or not and judging whether the price mutation is caused by the cost variation promotion or not. In the embodiment, the product price and the cost price are found to have the granger causal effect, and meanwhile, the synergistic relationship is established, which shows that the product price and the cost price have the tendency of resonance, and the rise of the product price is probably caused by the rise of the cost price. The result shows that the method can complete the function of judging and analyzing whether the monopoly behavior exists, and can accurately and quickly draw a conclusion by constructing a model.
S103, visual parameter setting is carried out, visual display of the identification result and downloading of the whole analysis report are carried out based on the set visual parameters, and the specific process is as follows:
in this embodiment, based on the price cost data of a 42.5 model cement product, mySQL is used for data management, and a Django frame based on Python is used in cooperation with components such as echarts and the like as an application platform, so that achievement display of data processing, enterprise clustering, time sequence modeling and space sequence modeling is realized.
In the embodiment, the enterprises are clustered into two groups of high market influence and low market influence through a K-means algorithm according to the kiln productivity, the mill productivity and the annual average sales share, and the visualization of the enterprise clustering is displayed in a bubble map mode; the product price change monitoring relates to analysis results of time sequence modeling and space sequence modeling, including identification of structure change points, detection of price change periods, comparison of space correlation coefficients and the like, and is selectively displayed in a time sequence line graph mode; the result of the spatial modeling is visually output in the form of thermodynamic diagram; the price cost analysis part relates to stationarity test, a coordination model, a Glanker causal test after VAR modeling, impulse response and variance decomposition. The monitoring conclusion section will summarize each phase modeling result and inform the user of the enterprise that has a highly suspected operating market.
3. Evidence of the relevant effects of the examples. The embodiment of the invention achieves some positive effects in the process of research and development or use, and has great advantages compared with the prior art, and the following contents are described by combining data, diagrams and the like in the test process.
As shown in the timing diagram inspection of FIG. 9, the stationarity inspection in the timing analysis can be performed by timing diagram inspection, sequence decomposition, autocorrelation graph inspection, and the like. Compared with the related technology, the invention provides more modes and dimensions to finish stability inspection, thereby ensuring the reliability and accuracy of the result and supporting the judgment and analysis for judging whether the monopoly behavior exists in the next step.
The timing chart test is carried out according to the property that the mean value and the variance of the stable time sequence are constants, the timing chart of the stable time sequence in the embodiment shows that the sequence value always fluctuates randomly around a constant, and the fluctuation range is bounded; if there is a significant trend or periodicity, it is usually a non-stationary sequence. The sequence presents a slow and non-monotonous ascending trend, and has transient fluctuation and instability in local parts.
The way of performing the sequence decomposition diagram as shown in fig. 10 is to separate the time series data into different components, and this embodiment decomposes the time series data into: long-term Trend, seasonal search availability, and random residual residuals residual.
Stationary sequences have short-term correlations. The autocorrelation map 11 (a) and partial correlation map of the smoothed sequence are either smeared or truncated as in fig. 11 (b). Truncation means that after a certain order, the coefficients are all 0 or tend to be 0. Tailing is a tendency to decay, but not all 0. The autocorrelation map is neither trailing nor truncated. The partial auto-correlation map smears. The autocorrelation coefficient is longer than 0, which indicates that there is strong long-term correlation between sequences, which are not stable.
The granger causal test script based on the VECM model in the example is as follows:
the meaning is:
number of lags (no zero) 1: detection when lags is 1
Result ssr based F test: residual sum of squares F test
ssr based chi2 test: residual sum of squares chi-squared test
likelihood ratio test: likelihood ratio test results
parr F testomete: test results of parameter F
Reference value: mainly considering the p value, p is less than 0.05 to prove that b is effective on a. The P value is a parameter for determining the result of the hypothesis test, and refers to the probability that the statistical summary (e.g., the average difference between two sets of samples) is the same as the actual observed data or even larger in a probabilistic model. a. b refers to the predictor variable.
Fig. 12 is a pulse verification diagram based on the VECM model. As can be seen from fig. 12, in the embodiment, a required result can be obtained according to the established model, and the next anti-monopoly detection operation can be supported according to the inspection result, so that the anti-monopoly monitoring operation can be digitally implemented through the anti-monopoly detection big data algorithm model and the anti-monopoly detection big data algorithm system.
It should be noted that the embodiments of the present invention can be realized by hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or specially designed hardware. Those skilled in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, such code being provided on a carrier medium such as a disk, CD-or DVD-ROM, programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier, for example. The apparatus and its modules of the present invention may be implemented by hardware circuits such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of hardware circuits and software, e.g., firmware.
The above description is only for the purpose of illustrating the present invention and the appended claims are not to be construed as limiting the scope of the invention, which is intended to cover all modifications, equivalents and improvements that are within the spirit and scope of the invention as defined by the appended claims.
Claims (8)
1. A big data perception method for judging monopoly behaviors, which is characterized by comprising the following steps:
monitoring the information of the key industry products, determining the geographical distribution and the market influence distribution of a production enterprise, checking the geographical distribution and the market influence distribution of the production enterprise, analyzing the price change of the key industry products and the price change of the province products, and obtaining trend dissimilarity analysis data of the price change of the key industry products and the price change of the province products; identifying the price change period reflected when the market has abnormal price change risk based on the analysis data, and if the market has the abnormal price change risk, identifying the price change period, feeding back the early warning information, and visually displaying the early warning information;
the big data perception method for judging monopoly behaviors comprises the following steps:
acquiring basic information data, tax data and sales data of an enterprise; cleaning and processing the acquired data;
determining heterogeneity of product price fluctuation characteristics from two space-time dimensions based on a structural mutation theory, and identifying monopoly behaviors of enterprises on product prices;
setting visual parameters, and carrying out visual display of the identification result and downloading of an overall analysis report based on the set visual parameters;
determining heterogeneity of product price fluctuation characteristics from two dimensions of time and space based on a structural mutation theory in the second step, and identifying monopoly behaviors of enterprises to product prices comprises the following steps:
(1) Performing cluster analysis on the processed data to obtain a plurality of groups, and calculating the intra-group correlation coefficient and the inter-group correlation coefficient of each group in which each enterprise is positioned to obtain the characteristic of price collusion and the data with monopoly characteristic;
(2) The idea of the Bai-Perron endogenous structure mutation test method is as follows:
assuming that k potential mutation points exist in time series data of a certain T phase, k +1 segmentation intervals are generated, and the generation process of the data is as follows:
y t ＝x′ t β+z′ t γ j +u t
wherein T is the time of the mutation point, T = T 1 ,...,T n ,n＝1,...,k+1，y t To be an explained variable, the explained variable is by x' t And z' t Two parts of x' t A variable representing the coefficient unchanged, z' t Variables representing changes in the coefficients, beta and gamma j For corresponding coefficient vectors, u t Is a residual term;
the Bai-Perron structural mutation test is divided into three steps: first of allStep, calculating beta and gamma by using a common least square method aiming at each possible segmentation point in the formula j And obtaining a corresponding sum of squares of residuals; and secondly, comparing the residual square sums obtained by different segmentation modes, and taking the minimum residual square sum for segmentation: thirdly, carrying out significance test on whether structural mutation occurs in the generation process of the time sequence;
(3) Whether structural mutation occurs in the generation process of the time sequence is subjected to significance test to obtain a plurality of structural change points, whether fluctuation with long-term memory is generated or not is tested in a time interval divided by two adjacent structural change points, and the GARCH model is a time sequence model for describing fluctuation rate and obtaining a good effect; for a time series r q Let a q ＝r q -μ q ＝r q -E(r q |F q-1 ) Term of { a q Subject to the GARCH (m, s) model, { a q If it satisfies
Wherein q is the time of the time series, F q-1 Indicates the rate of return information, μ, by time q-1 q Is a constant number, a q For uncorrelated white noise sequences, σ q Is the rate of the wave motion,
is the conditional standard deviation of profitability, ε q An independent co-distributed white noise column that is zero mean unit variance, and>
is a model of q M, s are parameters of GARCH model, alpha 0 、β j Is a constant term;
(4) And establishing a spatial correlation model for products of all enterprises, and adopting first-order time correlation coefficient analysis to calculate the intra-group correlation coefficient and the inter-group correlation coefficient of each enterprise group according to different groups obtained by clustering analysis.
2. The big-data perception method for judging monopoly behavior according to claim 1, wherein the step one of cleaning and processing the acquired data comprises:
firstly, processing missing values, standardizing data and detecting abnormal values of acquired data;
secondly, aggregating and converting multiple time dimensions of price data; and storing the processed data into a base library and a subject library which are constructed in advance.
3. The big-data-aware method for determining monopolistic behavior as defined in claim 1, wherein the heterogeneity of the fluctuation features of the product price is determined from two dimensions of space and time based on the theory of structural mutation in the second step, and the identifying monopolistic behavior of the enterprise on the product price comprises:
on the time dimension, a multivariate time sequence model of each enterprise product is constructed, structural change points of product price data are identified by combining an endogenous multiple mutation inspection model, and a GARCH model is constructed to determine the heterogeneity of the product price fluctuation characteristics;
in the spatial dimension, a spatial metering model is constructed to determine spatial correlation, aggregation and heterogeneity of different enterprise product prices.
4. The big data perception method for judging monopoly behavior as claimed in claim 1, wherein the step (1) of calculating the intra-group correlation coefficient and the inter-group correlation coefficient of the group in which each enterprise is located to obtain the feature with price collusion and the data with monopoly feature comprises:
1) In the time dimension, whether two time series have similar variation trends is calculated:
wherein, X T 、Y T Two times of finger inputM sequence, X t 、X t+1 、Y t 、Y t+1 Respectively at time t and time t +1 T 、Y T The value of the sequence, CORT is the value of the correlation coefficient;
2) Determining CORT (X) T ,Y T ) The properties of (A): -1. Ltoreq. CORT (X) T ,Y T ) 1 or less, wherein CORT (X) T ,Y T ) =1 indicates that the two time series have the same trend, and will rise or fall at the same time, and the rise or fall is the same; CORT (X) T ,Y T ) = -1 denotes the opposite trend of the rise and fall of the two time series; CORT (X) T ,Y T ) =0 indicates that the two time series have no correlation in monotonicity;
3) For enterprises with sudden changes of product price structures and entering price change periods, a plurality of enterprises in a group have strong system correlation, and the correlation among the groups is weak: the inter-group correlation presents the characteristics of high value-low value, the possibility of price leadership of enterprises which judge that several pieces of furniture have market dominance capacity is increased, and the characteristics of price collusion are achieved; and (4) enterprises with any mutation of product price structures and entering the price change period, wherein the inter-group correlation in the group presents the characteristics of low value-low value, and the enterprises are judged to have monopoly characteristics.
5. A big-data-aware system for monopoly behavior, implementing the big-data-aware method for monopoly behavior determination as defined in any one of claims 1 to 4, wherein the big-data-aware system for monopoly behavior determination comprises:
the risk early warning module is used for monitoring products in key industries, determining the geographical distribution and the market influence distribution of production enterprises, and analyzing the trend difference between the price change of the products in the key industries and the price change of products in the province; meanwhile, the method is used for identifying the price change period when the market has the risk of abnormal price change and feeding back early warning information;
and the evidence support module is used for determining high correlation among enterprises with high influence on enterprise products after administrative filing, analyzing the cost and price cost of the production enterprises and providing traceable and provable model analysis evidence.
6. A computer arrangement, characterized in that the computer arrangement comprises a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to carry out the steps of the big data perception method for judging monopoly behavior according to any one of claims 1-4.
7. A computer-readable storage medium, storing a computer program which, when executed by a processor, causes the processor to perform the steps of the big-data aware method for determining monopoly behavior according to any one of claims 1 to 4.
8. An information data processing terminal, characterized in that the information data processing terminal is configured to implement the big data perception system for determining monopoly behavior according to claim 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211081363.3A CN115170166B (en) | 2022-09-06 | 2022-09-06 | Big data sensing method and system for judging monopoly behavior |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211081363.3A CN115170166B (en) | 2022-09-06 | 2022-09-06 | Big data sensing method and system for judging monopoly behavior |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115170166A CN115170166A (en) | 2022-10-11 |
| CN115170166B true CN115170166B (en) | 2023-04-11 |
Family
ID=83480512
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211081363.3A Active CN115170166B (en) | 2022-09-06 | 2022-09-06 | Big data sensing method and system for judging monopoly behavior |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115170166B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111028079A (en) * | 2019-12-08 | 2020-04-17 | 怀化学院 | Stock market systemic disc collapse time point monitoring method based on data driving |
| CN114493662A (en) * | 2021-12-27 | 2022-05-13 | 瑞格人工智能科技有限公司 | Commodity price data analysis method, platform, system and storage medium |
| CN114723287A (en) * | 2022-04-08 | 2022-07-08 | 浙江万里学院 | Quantitative statistical method for risk formation based on enterprise characteristics and operation behaviors |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5802502A (en) * | 1993-05-24 | 1998-09-01 | British Telecommunications Public Limited Company | System for selective communication connection based on transaction pricing signals |
| AU2009327490A1 (en) * | 2008-12-16 | 2011-07-28 | Genomatica, Inc. | Microorganisms and methods for conversion of syngas and other carbon sources to useful products |
| WO2015137970A1 (en) * | 2014-03-14 | 2015-09-17 | Mun Johnathan | Qualitative and quantitative modeling of enterprise risk management and risk registers |
| CN107798482A (en) * | 2017-11-16 | 2018-03-13 | 中国农业科学院农业信息研究所 | A kind of market for farm products unusual fluctuations risk monitoring method and system |
| CN109858708A (en) * | 2019-02-25 | 2019-06-07 | 天津工业大学 | A kind of user's selection incentive mechanism based on linkage of bidding |
-
2022
- 2022-09-06 CN CN202211081363.3A patent/CN115170166B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111028079A (en) * | 2019-12-08 | 2020-04-17 | 怀化学院 | Stock market systemic disc collapse time point monitoring method based on data driving |
| CN114493662A (en) * | 2021-12-27 | 2022-05-13 | 瑞格人工智能科技有限公司 | Commodity price data analysis method, platform, system and storage medium |
| CN114723287A (en) * | 2022-04-08 | 2022-07-08 | 浙江万里学院 | Quantitative statistical method for risk formation based on enterprise characteristics and operation behaviors |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115170166A (en) | 2022-10-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11250449B1 (en) | Methods for self-adaptive time series forecasting, and related systems and apparatus | |
| US11048729B2 (en) | Cluster evaluation in unsupervised learning of continuous data | |
| US9600773B2 (en) | Detecting electricity theft via meter tampering using statistical methods | |
| US8631040B2 (en) | Computer-implemented systems and methods for flexible definition of time intervals | |
| CN110865929A (en) | Abnormity detection early warning method and system | |
| CN104516808B (en) | Data prediction device and method | |
| CN107679734A (en) | It is a kind of to be used for the method and system without label data classification prediction | |
| WO2003096237A2 (en) | Electronic data processing system and method of using an electronic data processing system for automatically determining a risk indicator value | |
| CN112084229A (en) | Method and device for identifying abnormal gas consumption behaviors of town gas users | |
| CN111796957B (en) | Transaction abnormal root cause analysis method and system based on application log | |
| CN115641162A (en) | Prediction data analysis system and method based on construction project cost | |
| CN111340086A (en) | Method, system, medium and terminal for processing label-free data | |
| CN112365070A (en) | Power load prediction method, device, equipment and readable storage medium | |
| CN112132210A (en) | Electricity stealing probability early warning analysis method based on customer electricity consumption behavior | |
| CN112101471A (en) | Electricity stealing probability early warning analysis method | |
| CN116882804A (en) | Intelligent power monitoring method and system | |
| CN116383645A (en) | Intelligent system health degree monitoring and evaluating method based on anomaly detection | |
| CN115983900A (en) | Method, apparatus, device, medium, and program product for constructing user marketing strategy | |
| CN110458581B (en) | Method and device for identifying business turnover abnormality of commercial tenant | |
| CN115170166B (en) | Big data sensing method and system for judging monopoly behavior | |
| Peters et al. | Statistical features of persistence and long memory in mortality data | |
| CN109784524A (en) | Part amount prediction technique and device, equipment and storage medium | |
| Éltető et al. | Towards non-stationary grid models | |
| Emmanuel et al. | Singular spectrum analysis: an application to Kenya’s industrial inputs price index | |
| CN118071490B (en) | Credit risk trend prediction system based on moving average 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 |