A method and system for correlation analysis of power marketing audit indicators
By using correlation analysis of electricity marketing audit indicators, the complexity of electricity marketing audit data was solved, enabling efficient data analysis and management support, and improving the level of power grid management and risk early warning capabilities.
Patent Information
- Application Number
- CN202410985617.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-23
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-07-23
AI Technical Summary
In the current electricity marketing audit work, the audit data is diverse, comes from a wide range of sources, and is large in scale, which makes the calculation of data correlation complex. Traditional manual indicator audit methods are labor-intensive, costly, and inefficient.
The correlation analysis method of electricity marketing audit indicators is adopted, including data collection, preprocessing, multi-dimensional database construction and improved limited dependent variable method, to conduct correlation analysis and output the correlation coefficient between indicators of various data types.
It improved the accuracy and reliability of power inspection data, reduced labor costs, and enhanced data analysis efficiency and management's decision support capabilities.
Smart Images

Figure CN118521346B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power marketing audit indicator correlation analysis, and in particular to a power marketing audit indicator correlation analysis method and system. Background Art
[0002] With the increasing application of modern cutting-edge technologies such as the Internet of Things and big data in the power industry, the construction of online service platforms for power grids in various regions is becoming more and more perfect. In this context, the construction of marketing audit informatization supported by big data technology has become more feasible. At present, the main problems of power marketing audit work are that the marketing audit data is diverse, has a wide range of sources, is large in scale, and has complex data correlation calculations. The traditional manual indicator audit method has a large workload, high labor costs, and low efficiency.
[0003] Therefore, correlation analysis of power marketing audit indicators based on big data technology can analyze user behavior more deeply and provide real-time data and technical support to the management of power supply companies, which is of great significance for improving the refined management level and risk warning capabilities of power grids. Summary of the invention
[0004] In view of the technical problems pointed out in the background technology, the purpose of the present invention is to provide a method and system for correlation analysis of power marketing audit indicators.
[0005] To achieve the purpose of the present invention, the technical solution provided by the present invention is specifically as follows:
[0006] First aspect
[0007] This application provides a method for analyzing the correlation of power marketing audit indicators, including the following steps:
[0008] Step S1: Collect power marketing audit data and pre-process it to obtain a pre-processed power marketing data set;
[0009] Step S2: using the pre-processed power marketing data set, constructing a multi-dimensional power marketing audit database including data time, data region and data user category;
[0010] Step S3: Based on the multi-dimensional power marketing audit database, the power marketing audit data in the multi-dimensional power marketing audit database is subjected to correlation analysis using an improved restricted dependent variable method, and the correlation coefficients between indicators of each data type are output.
[0011] Second aspect
[0012] Corresponding to the above method, the present application also provides a power marketing audit indicator correlation analysis system, comprising the following units:
[0013] Data collection preprocessing unit, multi-dimensional power marketing audit database construction unit and correlation coefficient calculation unit;
[0014] The data collection and preprocessing unit is used to collect power marketing audit data and perform preprocessing to obtain a preprocessed power marketing data set;
[0015] The multi-dimensional power marketing audit database construction unit is used to construct a multi-dimensional power marketing audit database including data time, data region and data user category using the pre-processed power marketing data set;
[0016] The correlation coefficient calculation unit is used to perform correlation analysis on the power marketing audit data in the multi-dimensional power marketing audit database based on the multi-dimensional power marketing audit database using an improved restricted dependent variable method, and output the correlation coefficients between indicators of each data type.
[0017] Compared with the prior art, the present invention has the following beneficial effects:
[0018] This application evaluates the effectiveness of the proposed power marketing audit indicator correlation analysis method and system through simulation experiments, proving its reliability in power audit. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] Figure 1 A schematic diagram of the method flow provided in an embodiment of the present application. DETAILED DESCRIPTION
[0020] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Example 1
[0021] like Figure 1 As shown, this embodiment provides a method for analyzing the correlation of power marketing audit indicators, including the following steps:
[0022] Step S1: Collect power marketing audit data and pre-process it to obtain a pre-processed power marketing data set;
[0023] The power marketing audit data includes the following types: abnormal power consumption monitoring data, user profile information, electricity fee and price data, metering data, etc. By preprocessing the marketing audit data, the accuracy and reliability of the data can be ensured and the data quality can be improved.
[0024] Step S1 specifically includes the following steps:
[0025] Step S1.1: Collect power marketing audit data;
[0026] Step S1.2: Locate the information source and classify the types of power marketing audit data;
[0027] The main information sources of marketing audit data include market information and business information of power companies. In power marketing audit, it is necessary to locate the information source of the data first, and determine the type of data based on the information source.
[0028] The process of locating the source of data information is as follows:
[0029] (1)
[0030] In the formula, f (·) represents the data information source location function; s Indicates the data information source: C Indicates constraints in the data collection process; D i Indicates the input power marketing audit data sample; i Indicates the number of data acquisition times; j Indicates the number of data integrations; A , T , P They represent the accuracy, invalidity and completeness evaluation indicators of power marketing audit data respectively; l a , l t and l p Indicates the corresponding evaluation index weight;
[0031] When the power marketing audit data sample is input, the data location information source is output.
[0032] On the basis of locating the source of data information, a parallel computing method is adopted to classify the types of power marketing audit data that have completed the information source location, and the membership degree of the power marketing audit data compared with each data type is output, and the type with the highest membership degree is selected as the type of the data. This process is shown in formula (2):
[0033] (2)
[0034] In the formula, C FCM Indicates the number of types of power marketing audit data; c i* Indicates i *Class initialization data class number center, i *=1,2,…, C FCM; x j* Represents each power marketing audit data point; j *Indicates the number of power marketing audit data; C k Indicates k Class initialization data class number center, k Indicates the number of divisions, k =1,2,…, C FCM ; u i*j* Indicates power marketing audit data points x j* For data types i *'s membership degree, d Represents the fuzzy coefficient, and its value is 1.
[0035] Step S1.3: Perform data cleaning and extraction on the power marketing audit data after information source location and type classification, and output the processed power marketing audit data set.
[0036] In order to avoid redundant data in the collected data, data cleaning and extraction are carried out on the basis of "Step S1.2: Locating the information source and classifying the types of power marketing audit data". This process is shown in the following equations (3)-(4):
[0037] (3)
[0038] (4)
[0039] In the formula, Data represents the power marketing audit data set after cleaning and extraction; Indicates the data status. >3 or <-3, =0, the rest =1; m Represents the mean value of a data set; s Represents the standard deviation of the data set. After this step, the audit data set after cleaning and extraction is output.
[0040] Step S2: using the pre-processed power marketing data set, constructing a multi-dimensional power marketing audit database including data time, data region and data user category;
[0041] It should be noted that the time dimension of the data includes year, month and day; the regional dimension of the data includes province, city, county (district); and the user category dimension of the data includes residential, commercial and industrial. Step S2 builds a multi-dimensional marketing audit database based on the technology of the power marketing data set in step S1, in order to gather a wider and more multi-dimensional power marketing data set, ensure the organized storage of power marketing data, and improve the efficiency of data analysis.
[0042] Estimate the storage requirements and query costs of the multi-dimensional power marketing audit database and set up a database concurrency control mechanism. The database concurrency control mechanism ensures the integrity and consistency of data when multiple users or applications access and modify the database at the same time. Perform data processing and analysis on the multi-dimensional power marketing audit database based on relational online analytical processing (ROLAP) technology.
[0043] Step S3: Based on the multi-dimensional power marketing audit database, the power marketing audit data in the multi-dimensional power marketing audit database is subjected to correlation analysis using an improved restricted dependent variable method, and the correlation coefficients between indicators of each data type are output.
[0044] The marketing audit data types mainly include abnormal electricity consumption monitoring data, user profile information, electricity charges and electricity price data, and metering data. Each marketing audit data type corresponds to its own audit indicator, that is, each data type represents a marketing audit data type indicator.
[0045] The details are as follows:
[0046] Step S3.1: Establish an improved restricted dependent variable model for marketing audit data type indicators;
[0047] Clarify the relationship between the dependent and independent variables of the audit indicators corresponding to each data type, and define the i The dependent variable of the power marketing audit data is F i* , the corresponding independent variable is X i* Each data F i* There are two types of values: the first type is greater than F 0, F 0 represents the non-zero minimum value of the data; the second category is F i* =0.
[0048] The first i * Power marketing audit data dependent variable F i* and independent variables X i* The relationship between them is expressed as:
[0049] (5)
[0050] (6)
[0051] (7)
[0052] In the formula, X i* for( n +1)-dimensional independent variable vector, n is a positive integer, α for( n +1)-dimensional unknown parameter vector, T Represents the transpose of a vector. g i* represents the error correction vector, with the degrees of freedom being k Chi-square distribution of * 2 ( k *); Formula (7) is the corresponding chi-square distribution density function, g express g i* is the value of a random variable in , Γ(·) represents the gamma function, and e represents the base of the natural logarithm.
[0053] Then the improved restricted dependent variable model of power marketing audit data type indicators is expressed as:
[0054] (8)
[0055] Improved restricted dependent variable model for marketing audit data type indicators, independent variables X i* is observable (i.e. X i* Take the actual observed value), the dependent variable F i* It cannot be obtained directly, but is observed in a restricted way.
[0056] Step S3.2: Based on the improved restricted dependent variable model of marketing audit data type indicators, solve the correlation coefficients between indicators of each data type.
[0057] The details include:
[0058] Step S3.2.1: Before solving the correlation coefficient between data type indicators, the data in the improved restricted dependent variable is first subjected to centroid convergence processing, the purpose of which is to cluster the discrete data in the independent variable and improve the efficiency of data indicator correlation mining.
[0059] For any discrete sample data in an independent variable, calculate the distance between the sample data and the two centroids LEF , and classify the sample data into the class where the centroid with the shortest distance is located. L EF The calculation process is as follows:
[0060] (9)
[0061] (10)
[0062] In the formula, and represents the weight coefficient, represents the Euclidean distance, represents the discretized F-distance, U and I represents the centroid of any two sample data. i t represents the number of features, j t Indicates j t Features, d (·) represents the distance metric function, m t and s t represents a parameterized function in the unit interval, d Indicates the value size of the unit interval.
[0063] Among them, the output L EF It represents the centroid distance of discrete sample data. The smaller the centroid distance is, the higher the similarity coefficient of the sample data is in certain features.
[0064] Step S3.2.2: Calculate the dependent variable F i* .
[0065] according to F i* and X i* of C FCM observations to measure α and k *Estimated value. F i* When ≤0, the observed value of the dependent variable is G 0, when F i* >0, the observed value of the dependent variable is G 1, then the observed value of the dependent variable G n = G 0+ G 1. If you ignoreF i* Observations at ≤0 G 0, then the rest G 1 observation is a complete observation ( F i* >0), for linear or nonlinear data, the least squares method and nonlinear least squares method are used to estimate α respectively.
[0066] but, F i* >0, the dependent variable F i* The conditional expectation of the observed value is:
[0067] (11)
[0068] therefore,
[0069] (12)
[0070] in, and is the probability density function and distribution function of the standard chi-square distribution. Since the least squares method does not calculate the probability density function that is not independent of X i* Item , so the least squares estimates are biased and inconsistent. ,but:
[0071] (13)
[0072] That is, the least squares estimator of α is biased. k The maximum likelihood estimator of * is a consistent estimator.
[0073] Using centroid distance L EF Correction F i* ,but:
[0074] (14)
[0075] represents the modified dependent variable.
[0076] Output It shows that if the centroid distance L EF The smaller it is, the greater the correlation coefficient between data type indicators.
[0077] Step 3.2.3: Calculate the correlation coefficient between data type indicators
[0078] The modified dependent variable Then, the Pearson correlation coefficient method was used to calculate and X i* The correlation coefficient of .
[0079] (15)
[0080] In the formula, Indicates the data type indicator and X i* The correlation coefficient between and Respectively X i* and The average value of .
[0081] This application evaluates the effectiveness of the proposed power marketing audit indicator correlation analysis method and system through simulation experiments, proving its reliability in power audit.
[0082] Example 2
[0083] This embodiment provides a power marketing audit indicator correlation analysis system, including the following units:
[0084] Data collection preprocessing unit, multi-dimensional power marketing audit database construction unit and correlation coefficient calculation unit;
[0085] The data collection and preprocessing unit is used to collect power marketing audit data and perform preprocessing to obtain a preprocessed power marketing data set;
[0086] The multi-dimensional power marketing audit database construction unit is used to construct a multi-dimensional power marketing audit database including data time, data region and data user category using the pre-processed power marketing data set;
[0087] The correlation coefficient calculation unit is used to perform correlation analysis on the power marketing audit data in the multi-dimensional power marketing audit database based on the multi-dimensional power marketing audit database using an improved restricted dependent variable method, and output the correlation coefficients between indicators of each data type.
[0088] Among them, the power marketing audit data includes the following types: abnormal power consumption monitoring data, user profile information, electricity charges and prices data, and metering data, etc.
[0089] The data collection preprocessing unit is used to specifically perform the following steps:
[0090] Step S1.1: Collect power marketing audit data;
[0091] Step S1.2: Locate the information source and classify the collected power marketing audit data;
[0092] Step S1.3: Perform data cleaning and extraction on the power marketing audit data after information source location and type classification, and output the processed power marketing audit data set.
[0093] Among them, in step S1.2, the process of locating the information source is as follows:
[0094] (1)
[0095] In the formula, f (·) represents the data information source location function; s Indicates the source of data information; C Indicates constraints in the data collection process; D i Indicates the input power marketing audit data sample; i Indicates the number of data acquisition times; j Indicates the number of data integrations; A , T , P They represent the accuracy, failure and completion evaluation indicators of power marketing audit data respectively. l a , l t and l p Indicates the corresponding evaluation index weight.
[0096] The optional implementation modes of the embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above implementation modes. Within the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical scheme of the embodiments of the present invention, and these simple modifications all belong to the protection scope of the embodiments of the present invention.
Claims
1. A method for correlation analysis of power marketing audit indicators, characterized in that: The steps include: Step S1: Collect power marketing audit data and pre-process it to obtain a pre-processed power marketing data set; Step S2: using the pre-processed power marketing data set, constructing a multi-dimensional power marketing audit database including data time, data region and data user category; Step S3: Based on the multi-dimensional power marketing audit database, the power marketing audit data in the multi-dimensional power marketing audit database are subjected to correlation analysis using an improved restricted dependent variable method, and the correlation coefficients between indicators of each data type are output; Step S1 specifically includes the following: Step S1.1: Collect power marketing audit data; Step S1.2: Locate the information source and classify the collected power marketing audit data; Step S1.3: Perform data cleaning and extraction processing on the power marketing audit data after information source location and type classification, and output the processed power marketing audit data set; In step S1.2, the process of locating the information source is as follows: In the formula, represents the data information source location function; s represents the data information source; C represents the constraint factor in the data collection process; D i represents the input power marketing audit data sample; i represents the number of data acquisition times; j represents the number of data integration times; A, T, and P represent the accuracy, failure, and completeness evaluation indicators of power marketing audit data, respectively, a , t and λ p Indicates the corresponding evaluation index weight; In step S1.2, based on the location of the information source of the data, a parallel computing method is adopted to classify the power marketing audit data that has completed the information source location, and the membership degree of the power marketing audit data compared to each data type is output, and the type with the highest membership degree is selected as the type of the data. This process is shown in formula (2): In the formula, C FCM Indicates the number of types of power marketing audit data; c i* represents the center of the i*th class of initialized data, i*=1,2,…,C FCM ;x j* represents each power marketing audit data point; j* represents the number of power marketing audit data; C k represents the center of the k-th class of initialized data, k represents the number of divisions, k = 1, 2, ..., C FCM ;u i*j* Represents power marketing audit data point x j* For the membership of data type i*, δ represents the fuzzy coefficient, which takes the value of 1; In step S1.3, the power marketing audit data after information source location and type classification is cleaned and extracted. This process is shown in the following equations (3)-(4): Data=∑y i*j* x j* ·S j* (3) In the formula, Data represents the power marketing audit data set after cleaning and extraction; S j* Indicates data status; Z j* >3 or Z j* <-3, S j* =0, the rest S j* =1; μ represents the mean value of the data set; σ represents the standard deviation of the data set.
2. The method for analyzing the correlation of power marketing audit indicators according to claim 1, characterized in that: In step S1, the power marketing audit data includes the following types: power consumption anomaly monitoring data, user profile information, electricity fee and price data, and metering data.
3. A power marketing audit indicator correlation analysis system, characterized in that: The following units are included: Data collection and preprocessing unit, multi-dimensional power marketing audit database construction unit and correlation coefficient calculation unit; The data collection and preprocessing unit is used to collect power marketing audit data and perform preprocessing to obtain a preprocessed power marketing data set; The multi-dimensional power marketing audit database construction unit is used to construct a multi-dimensional power marketing audit database including data time, data region and data user category using the pre-processed power marketing data set; The correlation coefficient calculation unit is used to perform correlation analysis on the power marketing audit data in the multi-dimensional power marketing audit database based on the multi-dimensional power marketing audit database using an improved restricted dependent variable method, and output the correlation coefficient between indicators of each data type; The data acquisition preprocessing unit is used to specifically perform the following steps: Step S1.1: Collect power marketing audit data; Step S1.2: Locate the information source and classify the collected power marketing audit data; Step S1.3: Perform data cleaning and extraction processing on the power marketing audit data after information source location and type classification, and output the processed power marketing audit data set; In step S1.2, the process of locating the information source is as follows: In the formula, represents the data information source location function; s represents the data information source; C represents the constraint factor in the data collection process; D i represents the input power marketing audit data sample; i represents the number of data acquisition times; j represents the number of data integration times; A, T, and P represent the accuracy, failure, and completeness evaluation indicators of power marketing audit data, respectively, a , t and λ p Indicates the corresponding evaluation index weight; In step S1.2, based on the location of the information source of the data, a parallel computing method is adopted to classify the power marketing audit data that has completed the information source location, and the membership degree of the power marketing audit data compared to each data type is output, and the type with the highest membership degree is selected as the type of the data. This process is shown in formula (2): In the formula, C FCM Indicates the number of types of power marketing audit data; c i* represents the center of the i*th class of initialized data, i*=1,2,…,C FCM ;x j* represents each power marketing audit data point; j* represents the number of power marketing audit data; C k represents the center of the k-th class of initialized data, k represents the number of divisions, k = 1, 2, ..., C FCM ;u i*j* Represents power marketing audit data point x j* For the membership of data type i*, δ represents the fuzzy coefficient, which takes the value of 1; In step S1.3, the power marketing audit data after information source location and type classification is cleaned and extracted. This process is shown in the following equations (3)-(4): Data = ∑ yi*j*xj*·Sj*(3) In the formula, Data represents the power marketing audit data set after cleaning and extraction; S j* Indicates data status; Z j* >3 or Z j* <-3, S j* =0, the rest S j* =1; μ represents the mean value of the data set; σ represents the standard deviation of the data set.
4. The power marketing audit indicator correlation analysis system according to claim 3 is characterized in that: The power marketing audit data includes the following types: power consumption anomaly monitoring data, user profile information, electricity fee and price data, and metering data.
Citation Information
Patent Citations
Risk inspection monitoring platform based on cloud platform under full-power-grid multi-source data
CN112434924A
Financial data intelligent auditing system and method based on big data platform system
CN113628024A
Power marketing inspection analysis system and method based on big data
CN115730962A
Source network load storage multi-source data acquisition and control means convergence method based on cloud resources
CN115864527A
KR20210037488A
