CN-122020473-A - Intelligent electric power material price monitoring method and system based on same-quality identification

CN122020473ACN 122020473 ACN122020473 ACN 122020473ACN-122020473-A

Abstract

The application provides an intelligent monitoring method and system for price of electric power supplies based on identification of same products, which relate to the technical field of electric power supply management, and the method comprises the steps of obtaining technical specification data, historical settlement data and actual bidding quotation of nonstandard electric power supplies to be purchased; the method comprises the steps of encoding technical indexes in technical specification data and historical settlement data, constructing a feature vector set, iteratively screening parameter masks of the technical indexes representing core value attributes from the feature vector set by utilizing a genetic algorithm, performing space reconstruction on the feature vector set according to the parameter masks to generate a similarity sequence, obtaining a theoretical estimated price of nonstandard electric materials to be purchased based on the similarity sequence, and calculating the difference ratio of the theoretical estimated price and actual bidding price so as to realize intelligent monitoring of the electric material price, thereby solving the problem that the historical price is incomparable due to fine adjustment of the parameters of the nonstandard electric materials.

Inventors

ZHANG CHAOYANG
LI LIN
Hou Naiming
ZONG YANAN
YU YUFENG

Assignees

国家能源集团物资有限公司数据科技分公司

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. An intelligent monitoring method for electric power material price based on identification of same products is characterized by comprising the following steps: Acquiring technical specification data, historical settlement data and actual bidding quotation of nonstandard electric power materials to be purchased, wherein the historical settlement data comprises a plurality of technical indexes of historical settlement records and settlement price information; encoding technical indexes in the technical specification data and the historical settlement data to construct a feature vector set, and iteratively screening parameter masks of the technical indexes representing core value attributes from the feature vector set by utilizing a genetic algorithm, wherein the feature vector set comprises a material vector to be purchased and a plurality of historical purchasing material vectors; Performing space reconstruction on the feature vector set according to the parameter mask, calculating Euclidean distance between a reconstructed material vector to be purchased in the reconstructed feature vector set and each reconstructed historical purchasing material vector, and generating a similarity sequence; based on the similarity sequence, selecting N historical purchasing material vectors with highest similarity as target samples, and carrying out weighted average calculation on the settlement unit price of a historical settlement record corresponding to the target samples by using a weighted K nearest neighbor algorithm to obtain the theoretical estimated price of the nonstandard electric power material to be purchased; and calculating the difference ratio of the theoretical estimated price and the actual bidding price, and generating price anomaly information under the condition that the difference ratio is larger than a preset ratio threshold value so as to realize intelligent monitoring of the price of the electric power materials.
2. The method of claim 1, wherein performing weighted average calculation on the settlement unit price of the historical settlement record corresponding to the target sample by using a weighted K nearest neighbor algorithm to obtain a theoretical estimated price of the nonstandard electric power material to be purchased comprises: Acquiring a transaction time stamp of a historical settlement record corresponding to each target sample; taking the reciprocal of the sum of the Euclidean distance of each target sample and a preset smoothing factor as the spatial similarity weight of each target sample; Calculating the time attenuation weight of each target sample by using an exponential decay function according to the time difference value of each transaction time stamp and the current time; performing product operation on the spatial similarity weight and the time attenuation weight of each target sample to obtain the target weight of each target sample; and according to the target weight, carrying out weighted average calculation on the settlement unit price of the historical settlement record corresponding to each target sample to obtain the theoretical estimated price of the nonstandard electric power material to be purchased.
3. The method according to claim 1, wherein the method further comprises: calculating standard deviation of settlement unit prices of all target samples, and taking the standard deviation of preset multiples as a price fluctuation tolerance value; constructing a dynamic confidence interval according to the theoretical estimated price and the price fluctuation tolerance value; And generating price anomaly information under the condition that the actual bid offer is smaller than the lower limit of the dynamic confidence interval or larger than the upper limit of the dynamic confidence interval so as to realize intelligent monitoring of the price of the electric power material.
4. The method of claim 1, wherein constructing a feature vector set by encoding technical metrics in the technical specification data and the historical settlement data comprises: Extracting a text description field with a technical index from the technical specification data as first text information, and extracting a text description field with a technical index from each historical settlement record in the historical settlement data as second text information; Performing word segmentation processing on the first text information and each second text information by using a preset word segmentation rule to obtain a first keyword set and a plurality of second keyword sets; de-overlapping the first keyword set and all the second keyword sets, and constructing a feature index table; Constructing a material vector to be purchased according to the index position and the appearance state of each keyword in the first keyword set in the characteristic index table; constructing a plurality of historical purchasing material vectors according to the index position and the appearance state of each keyword in each second keyword set in the characteristic index table; And combining the material vector to be purchased with all the historical material vectors to be purchased to obtain a feature vector set.
5. The method of claim 1, wherein iteratively screening parameter masks representing technical metrics of core value attributes from the feature vector set using a genetic algorithm comprises: randomly generating an initial population comprising a plurality of binary strings by taking the number of feature dimensions of the feature vector set as the bit length of the binary strings; Performing preliminary reconstruction on the feature vector set according to each binary string, and selecting a plurality of temporary historical samples with the shortest Euclidean distance with the preliminarily reconstructed material vector to be purchased according to the preliminarily reconstructed feature vector set; Calculating the dispersion of the settlement price information set of the temporary history sample, and taking the dispersion as an adaptability index of a corresponding binary string, wherein the adaptability index is used for measuring whether a technical index represents a core value attribute; According to the fitness index, the initial population is arranged in an ascending order, and the binary strings with the preset quantity in the arrangement result are used as parent binary strings; Performing cross operation and mutation operation on the parent binary string to generate a child binary string; And updating the initial population according to the child binary string until the updated population meets a preset iteration convergence condition, and taking the binary string corresponding to the lowest fitness index in the updated population as a parameter mask.
6. The method of claim 5, wherein performing a crossover operation and a mutation operation on the parent binary string to generate child binary strings comprises: performing pairwise random pairing on the parent binary strings to obtain a plurality of parent pairing groups; Generating a first random integer within the bit length range of the parent binary string, and taking the first random integer as a cross point index of each parent pairing group; Exchanging binary bits of two parent binary strings in each parent pairing group after the cross point indexing to obtain a first binary string and a second binary string; generating a second random integer in the bit length range of each first binary string and each second binary string, and taking the second random integer as a variant bit index of each parent pairing group; And performing overturn operation on binary values of the variant bit indexes on each first binary string and each second binary string to obtain child binary strings.
7. The method of claim 1, wherein spatially reconstructing the set of feature vectors from the parameter mask and calculating euclidean distances between a reconstructed to-be-purchased material vector in the set of reconstructed feature vectors and each reconstructed historical purchased material vector, generating a similarity sequence comprises: Determining a binary bit in the parameter mask, which is marked as a reserved state, as a target position; removing feature dimensions of the feature vector set, which are not located at the target position, to obtain a reconstructed feature vector set; calculating Euclidean distance between the reconstructed material vector to be purchased and each reconstructed historical material vector in the reconstructed feature vector set; And arranging the history purchasing material vectors corresponding to the reconstructed history material vectors according to the ascending order of the Euclidean distance to generate a similarity sequence.
8. Electric power material price intelligent monitoring system based on same article discernment, its characterized in that includes: The system comprises an acquisition module, a storage module and a storage module, wherein the acquisition module is used for acquiring technical specification data, historical settlement data and actual bid quotation of nonstandard electric power materials to be purchased, and the historical settlement data comprises technical indexes and settlement price information of a plurality of historical settlement records; the encoding module is used for constructing a feature vector set by encoding the technical specification data and the technical indexes in the historical settlement data, and iteratively screening out a parameter mask of the technical indexes representing the core value attribute from the feature vector set by utilizing a genetic algorithm, wherein the feature vector set comprises a material vector to be purchased and a plurality of historical purchasing material vectors; The reconstruction module is used for carrying out space reconstruction on the feature vector set according to the parameter mask, calculating Euclidean distance between a reconstructed material vector to be purchased in the reconstructed feature vector set and each reconstructed historical purchasing material vector, and generating a similarity sequence; the calculation module is used for selecting N historical purchasing material vectors with highest similarity as target samples based on the similarity sequence, and carrying out weighted average calculation on the settlement unit prices of the historical settlement records corresponding to the target samples by using a weighted K nearest neighbor algorithm to obtain the theoretical estimated price of the nonstandard electric power material to be purchased; and the calculation module is also used for calculating the difference ratio of the theoretical estimated price and the actual bidding price and generating price anomaly information under the condition that the difference ratio is larger than a preset ratio threshold value so as to realize intelligent monitoring of the price of the electric power material.
9. An electronic device, comprising: A memory for storing a computer program; a processor for implementing the steps of the intelligent monitoring method for electric power material price based on identification of same as in any one of claims 1 to 7 when executing the computer program.
10. A computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and when the computer program is executed by a processor, the computer program can implement an electric power material price intelligent monitoring method based on identification of same according to any one of claims 1 to 7.

Description

Intelligent electric power material price monitoring method and system based on same-quality identification Technical Field The application relates to the technical field of electric power material management, in particular to an electric power material price intelligent monitoring method and system based on identification of same products. Background In core business plates such as coal exploitation, thermal power generation, railway transportation and the like, large-scale energy enterprises are involved in purchasing massive nonstandard equipment, accessories and engineering services each year, and the materials are various in variety, complex in technical parameters and often have differences due to project customization requirements. Reasonable management and risk management and control of material purchase price have important significance for reducing comprehensive cost of supply chains, guaranteeing stable production operation and improving purchase benefit for groups. With the deep advancement of enterprise digital transformation, the intelligent means is utilized to monitor, analyze and early warn the purchase price, and the method has become a key trend for improving purchase compliance, economy and strategic insight. In the prior art, when price monitoring is performed on enterprise purchased materials, strict code matching is usually performed on standard material main data, or price comparison is performed by referring to the transaction records of historical materials with the same type through manual experience. However, in the professional fields of coal mines, thermal power, railway transportation and the like, a large number of nonstandard and customized materials lack of unified standard codes or models exist. Because of the numerous technical parameters and the fine tuning often existing, it is difficult to find completely consistent references in the historic purchasing library, resulting in failure of the traditional price comparison mode based on accurate matching. Therefore, in the prior art, non-standard materials of the cross-service plate are difficult to accurately match with effective historical samples, so that the technical problems of low price monitoring accuracy, low efficiency, delayed risk discovery and the like are caused. Disclosure of Invention The application aims to provide an intelligent monitoring method and system for electric power material prices based on identification of same products, which are used for solving the technical problems that in the prior art, nonstandard materials of cross-service plates are difficult to accurately match with effective historical samples, so that price monitoring accuracy is low, efficiency is low, risk discovery is delayed and the like. In a first aspect, the application provides an intelligent monitoring method for electric power material prices based on identification of same products, comprising the following steps: Acquiring technical specification data, historical settlement data and actual bidding quotation of nonstandard electric power materials to be purchased, wherein the historical settlement data comprises technical indexes and settlement price information of a plurality of historical settlement records; Encoding technical indexes in technical specification data and historical settlement data to construct a feature vector set, and iteratively screening out a parameter mask of the technical indexes representing core value attributes from the feature vector set by utilizing a genetic algorithm, wherein the feature vector set comprises a material vector to be purchased and a plurality of historical purchasing material vectors; performing space reconstruction on the feature vector set according to the parameter mask, calculating Euclidean distance between a reconstructed material vector to be purchased in the reconstructed feature vector set and each reconstructed historical purchasing material vector, and generating a similarity sequence; Based on the similarity sequence, selecting N historical purchasing material vectors with highest similarity as target samples, and carrying out weighted average calculation on the settlement unit prices of the historical settlement records corresponding to the target samples by using a weighted K nearest neighbor algorithm to obtain theoretical estimated prices of nonstandard electric power materials to be purchased; Calculating the difference ratio of the theoretical estimated price and the actual bidding price, and generating price anomaly information under the condition that the difference ratio is larger than a preset ratio threshold value so as to realize intelligent monitoring of the price of the electric power material. Optionally, performing weighted average calculation on the settlement unit price of the historical settlement record corresponding to the target sample by using a weighted K nearest neighbor algorithm to obtain a theoretical estimated price of the nonstandard e