Data analysis method for equipment material database

Abstract

The invention relates to an energy power industry equipment material supply chain data processing technology, discloses a data analysis method for an equipment material database, and solves the problems of inaccurate cost calculation, no data support for purchasing decision and difficulty in real-time judgment of price trend in the aspect of energy power industry equipment material price analysis in the prior art. According to the scheme, firstly, the deviation ratio between the supplier quotation and the standard price is calculated, and the quotation credibility is judged; secondly, mining a quarterly / annual fluctuation rule of the price by using a time sequence algorithm; thirdly, outputting the correlation degree of the equipment price and the raw material price through the correlation analysis model; and finally, judging the next change condition of the price in combination with the first three analysis results. The method can improve the accuracy of equipment material cost measurement and calculation, provides data support for enterprise purchase decision, assists enterprises in mastering the market price trend, and enhances the toughness and efficiency of an energy power industry equipment material supply chain.

Description

Technical Field

[0001] This invention relates to data processing technology for the equipment and materials supply chain in the energy and power industry, specifically to a data analysis method for an equipment and materials database. Background Technology

[0002] In the development of the energy and power industry, equipment and materials, as a core component of the supply chain, are diverse in type and specifications, and their prices are affected by multiple factors such as market supply and demand, raw material fluctuations, and policy adjustments, leading to severe challenges in price information management. (1) Cost calculations lack scientific basis and their accuracy is difficult to guarantee: Currently, when estimating equipment and material costs, companies often rely on the past experience of their staff or quotation data from a single source, lacking a systematic mechanism for judging the credibility of supplier quotations. On the one hand, the determination of standard prices lacks a weighted logic based on the fusion of multi-source data, relying solely on some historical contract prices or scattered inquiry results, failing to objectively reflect the true market price level. On the other hand, there is a lack of effective screening methods for outliers in supplier quotations, resulting in inconsistent quotation data quality and ultimately causing significant discrepancies between cost calculation results and actual procurement costs.

[0003] (2) Procurement decisions lack data support, making it difficult to grasp market trends: The energy and power industry involves long procurement cycles and large capital investments for equipment and materials. Understanding price fluctuations directly impacts procurement costs and supply chain stability. Currently, companies lack the systematic capability to mine historical equipment and material price data, making it impossible to identify cyclical price fluctuations and miss cost optimization opportunities during price troughs. Furthermore, they fail to quantify the correlation between equipment and material prices and core raw material prices, making it impossible to predict equipment and material price trends based on raw material price changes. Procurement decisions often rely on subjective judgment, making them susceptible to market volatility.

[0004] (3) Insufficient ability to judge price trends in real time, and slow supply chain response: As the energy and power industry becomes increasingly market-oriented, equipment and material prices are more significantly affected by dynamic factors such as the macroeconomy and industry policies, necessitating real-time data updates and dynamic adjustments to analytical conclusions. However, existing technologies lack mechanisms for multi-dimensional data integration and analysis, failing to collaboratively utilize information such as price credibility, price cycle patterns, and raw material correlations, making it difficult to predict future price trends in real time. Furthermore, low data update frequency and fixed analytical model parameters cause analysis results to lag behind market changes, preventing companies from adjusting their procurement strategies in a timely manner. This can not only increase procurement costs but also lead to supply chain disruptions due to price fluctuations, weakening the company's core competitiveness.

[0005] In summary, existing technologies for analyzing equipment and material prices in the energy and power industry suffer from inaccurate cost calculations, lack of data support for procurement decisions, and difficulty in real-time price trend assessment. There is an urgent need for a systematic data analysis method based on equipment and material databases, which can mine data value through multi-dimensional algorithms, provide accurate support for enterprise cost calculations and procurement decisions, and thereby enhance the resilience and efficiency of the equipment and material supply chain. Summary of the Invention

[0006] The technical problem to be solved by this invention is to provide a data analysis method for an equipment and materials database, which solves the problems of inaccurate cost calculation, lack of data support for procurement decisions, and difficulty in real-time judgment of price trends in the existing technology for equipment and materials price analysis in the energy and power industry.

[0007] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows: A data analysis method for an equipment materials database includes the following steps: S1. Equipment and Material Price Deviation Analysis: S11. Extract standard price data and corresponding supplier quotation data of the target equipment materials from the equipment and materials database. The standard price data is obtained by weighted calculation based on historical contract prices and periodic inquiry data. The supplier quotation data includes valid quotations from at least 3 qualified suppliers for the same specification of equipment materials. S12. Preprocess supplier quotation data to remove outliers; S13. Calculate the deviation rate of each supplier's quotation relative to the standard price: Deviation rate = (Supplier quotation - Standard price) / Standard price × 100%; S14. Determine the credibility level of the supplier's quotation by comparing the absolute value of the deviation rate with the corresponding thresholds preset for different credibility levels; S2. Cyclical Analysis of Equipment and Material Prices: S21. Extract at least 5 years of historical price data for the target equipment materials from the equipment and materials database, and organize them into a continuous price time series by quarter or year; S22. The ADF test is used to test the stationarity of price time series. Non-stationary series are transformed into stationary series through first-order differencing. S23. The stationary time series is analyzed using the ARIMA model. The optimal parameters of the model are selected using the AIC criterion. The optimized ARIMA model is used to fit the price change curve to explore the quarterly / annual fluctuation pattern of equipment and material prices and determine the price cycle. S24. Verify the price cycle: If the coefficient of variation of price fluctuation amplitude is ≤15% within 3 consecutive cycles, and the time interval deviation between peak and trough values ​​meets the requirements of quarterly cycle ≤1 month and annual cycle ≤3 months, then the price cycle is deemed valid. S3. Correlation analysis of equipment and material prices: S31. Extract at least three years of monthly price data for the target equipment material from the equipment material database, as well as monthly price data for raw materials directly related to the equipment material; S32. Preprocess the two types of monthly price data to remove missing and outlier values; S33. The Pearson correlation coefficient is used to calculate the linear correlation between equipment material prices and raw material prices, and the correlation level between the two is determined by the absolute value of the correlation coefficient. S34. Construct a linear regression model with raw material prices as independent variables and equipment material prices as dependent variables, and output the regression equation to quantify the correlation between the two. S4. Real-time analysis of equipment and material prices: S41. Integrate the quote credibility level results from step S1, the effective price cycle pattern from step S2, and the price correlation data from step S3; S42. Based on the integrated data, determine the next change in equipment and material prices: If the current price cycle is at its trough, raw material prices are trending downwards, and the proportion of highly reliable quotes is ≥60%, then the price is likely to decrease; if the current price cycle is at its peak, raw material prices are trending upwards, and the proportion of highly reliable quotes is ≥60%, then the price is likely to increase; if there are contradictions in the integrated data, combine the regression equation output from step S3 to predict the price change range. When the absolute value of the change range is ≤3%, the price is determined to be stabilizing. S43. Update the equipment and materials database data quarterly and repeat steps S1-S4 to optimize the analysis results.

[0008] Furthermore, in step S11, when calculating the standard price data, historical contract prices are weighted according to their time period: historical contract prices within the past year have a weight of 30%-40%, historical contract prices within the past 1-2 years have a weight of 20%-30%, and historical contract prices within the past 2 years have a weight of 10%-20%; the reliability weight of periodic inquiry data is 70%-80%.

[0009] Furthermore, in step S23, if the ARIMA model does not achieve the preset standard for fitting the stationary time series, it is replaced by the Fourier transform method to decompose the time series into sine waves of different frequencies. By analyzing the sine wave frequencies, the quarterly / annual fluctuation patterns and price cycles of equipment and material prices are determined.

[0010] Furthermore, in step S31, the selection of raw materials directly related to equipment materials is determined based on the correspondence between the core components and raw materials in the equipment material manufacturing process, ensuring that the selected raw materials are the core factors affecting the price of equipment materials.

[0011] Furthermore, in step S33, the specific criteria for determining the correlation level based on the absolute value of the Pearson correlation coefficient are as follows: A correlation coefficient with an absolute value greater than 0.7 is considered a strong correlation; a correlation coefficient with an absolute value between 0.3 and 0.7 is considered a moderate correlation; and a correlation coefficient with an absolute value less than or equal to 0.3 is considered a weak correlation.

[0012] Furthermore, step S4 also includes: The price change curve obtained in step S2, the correlation scatter plot generated based on the price correlation data in step S3, and the real-time analysis conclusions obtained in step S42 are presented in chart form.

[0013] Furthermore, in step S12, before removing outliers, the mean and standard deviation of all supplier quotations are calculated, and quotations that deviate from the mean by more than 3 times the standard deviation are identified as outliers and removed.

[0014] Furthermore, in step S24, the coefficient of variation of the price fluctuation range is calculated in the following way: First, calculate the mean and standard deviation of the price fluctuation range for each of the three consecutive periods. Then, calculate the coefficient of variation using the formula: coefficient of variation = standard deviation / mean × 100%.

[0015] The beneficial effects of this invention are: (1) Improve the accuracy of equipment and material cost calculation: This invention analyzes the deviation of equipment and material prices. First, it calculates an objective standard price based on weighted multi-source data. Then, it quantifies the difference between the supplier's quotation and the standard price using a deviation rate formula. Finally, it determines the credibility of the quotation according to a preset threshold. This effectively eliminates low-credibility quotation data, avoids errors caused by experience estimation or a single quotation, and controls cost calculation errors within a more accurate range. This solves the problem of cost calculation relying on subjective experience and lacking accuracy in existing technologies.

[0016] (2) Provide scientific data support for procurement decisions: On the one hand, price cycle analysis uses the ARIMA model (or Fourier transform method) to uncover the quarterly / annual price fluctuation patterns, and verifies the effectiveness of the cycle through the coefficient of variation and peak / peak interval, helping companies accurately grasp the price trough period for procurement and reduce procurement costs. On the other hand, price correlation analysis uses the Pearson correlation coefficient and linear regression model to quantify the price correlation between equipment materials and core raw materials, and can predict the price trend of equipment materials by the changes in raw material prices, avoiding procurement decisions that rely on subjective judgment and enhancing the scientific nature of decision-making.

[0017] (3) Real-time control of price change trends: The real-time price analysis integrates multi-dimensional data such as quote credibility, price cycle patterns, and raw material correlation to construct a clear trend judgment logic. When data is contradictory, regression equations are used to predict the magnitude of changes. At the same time, data is updated every quarter and the process optimization results are repeatedly analyzed to solve the problem of lagging price trend judgment in existing technologies. This helps companies adjust their procurement strategies in a timely manner and avoid missing cost optimization opportunities or facing supply chain disruption risks due to information lag.

[0018] (4) Adapt to the supply chain needs of the energy and power industry, and enhance the resilience and efficiency of the supply chain: This invention is based on an equipment and materials database, which is tailored to the characteristics of the energy and power industry, which has a wide variety of equipment and materials and many factors affecting prices. It can be directly connected to the industry's smart supply chain platform, and through systematic data analysis, it can effectively utilize price information, reduce resource waste caused by price information asymmetry in the supply chain, improve the supply chain's responsiveness to market changes, and enhance the overall resilience and operational efficiency of the energy and power industry's equipment and materials supply chain. Attached Figure Description

[0019] Figure 1 This is a flowchart of the data analysis method for the equipment and materials database in this invention. Detailed Implementation

[0020] This invention aims to provide a data analysis method for equipment and materials databases, addressing the problems of inaccurate cost calculations, lack of data support for procurement decisions, and difficulty in real-time price trend assessment in existing technologies for equipment and materials price analysis in the energy and power industry. Its core idea is to use an equipment and materials database as the data foundation, and through a progressive and collaborative analysis process involving price deviation analysis, price cycle analysis, price correlation analysis, and real-time price analysis, systematically mine the value of equipment and materials price data, thereby solving the problems of inaccurate cost calculations, lack of data support for procurement decisions, and difficulty in grasping price trends in the energy and power industry.

[0021] Specifically, the process begins by extracting multi-dimensional raw data from the database and preprocessing it for different analytical scenarios to ensure the data is objective, complete, and effective. This provides a reliable data foundation for subsequent analyses across various dimensions and avoids analytical biases caused by data quality issues. To address the lack of scientific basis for cost calculations, price deviation analysis quantifies the differences between supplier quotes and objective standards, selecting highly credible quotes to provide accurate data references for cost calculations. To address the difficulty in grasping price trends, price cycle analysis employs the ARIMA model (or Fourier transform) to uncover quarterly / annual price fluctuation patterns, verifying the validity of the cycle by combining the coefficient of variation and peak / trough intervals, thus clarifying the characteristics of price fluctuation cycles. Finally, to address the lack of quantitative support for price correlations, price correlation analysis uses the Pearson correlation coefficient to determine the strength of the correlation between equipment materials and core raw materials, and then a linear regression model outputs a regression equation for quantitative correlation, achieving a breakthrough from qualitative correlation to quantitative calculation. Finally, the analysis results of price credibility, cyclical patterns, and raw material correlation are integrated. If the data from multiple dimensions are consistent, the price trend is directly determined; if the data are contradictory, the magnitude of change is predicted through regression equations to ensure the rationality of the trend judgment. The analysis results are dynamically optimized through quarterly data updates and process repetition, ultimately providing real-time and accurate data support for enterprise cost calculation and procurement decisions, and improving the resilience and efficiency of the equipment and materials supply chain in the energy and power industry.

[0022] In terms of specific implementation, the implementation process of the data analysis method for an equipment material database provided by this invention is as follows: Figure 1 It includes the following processes: S1. Equipment and Material Price Deviation Analysis: This step involves constructing a standard price based on multi-source data, and determining the credibility of the quote through deviation rate calculation, providing an accurate reference for cost estimation. In one exemplary implementation scheme, the process is as follows: S11. Data Extraction: Standard price data is extracted from the database. The calculation of standard prices strictly follows the weighted logic of multi-source data, integrating historical contract prices and periodic inquiry data, and assigning weights according to time and reliability to ensure the objectivity of the standard prices. Supplier quotation data includes valid quotations from at least three qualified suppliers for the same specification of equipment and materials.

[0023] S12. Pretreatment: Remove outliers (such as monopolistic prices that are significantly higher than market prices) to avoid abnormal data affecting the analysis results.

[0024] S13. Deviation Rate Calculation: The deviation rate is calculated using the formula: Deviation Rate = (Supplier Quotation - Standard Price) / Standard Price × 100%.

[0025] S14. Credibility Assessment: For example, if the standard price of a transformer is 120,000 yuan, and supplier A quotes 123,000 yuan, the deviation rate is (123,000 - 120,000) / 120,000 × 100% = 2.5%, which is considered a high-reliability quote. Supplier B quotes 135,000 yuan, with a deviation rate of 12.5%, which is considered a low-reliability quote. High-reliability quotes can be directly used for cost calculation, while low-reliability quotes require further verification to avoid cost estimation errors caused by incorrect quotations and to improve the accuracy of calculations.

[0026] S2. Cyclical Analysis of Equipment and Material Prices: This step utilizes time series algorithms to uncover patterns in price fluctuations. In one exemplary implementation, the process is as follows: S21. Time Series Construction: Extract at least 5 years of historical price data, organize it by quarter / year, and form a continuous time series.

[0027] S22. Stabilization treatment: The ADF test is used to determine the stationarity of the data. Non-stationary data is processed by first-order differencing to ensure the effectiveness of subsequent analysis. For example, the price data of a certain cable from 2019 to 2023 was non-stationary, but it became a stationary series after differencing.

[0028] S23. Price Cycle Analysis: By using the ARIMA model to fit the data, determine the optimal parameters, and explore cyclical patterns, such as discovering that the price of a certain component has an annual cycle of "rising in Q1-Q2 and falling in Q3-Q4", we can obtain the pattern that "the cyclical supply and demand relationship in the market leads to price changes".

[0029] If the model fits poorly, the Fourier transform method is used to decompose the time series to determine the period. At the same time, the validity of the period is verified by the coefficient of variation and the peak / trough interval to ensure the reliability of the pattern and provide a basis for enterprises to grasp the timing of procurement.

[0030] S24. Price Verification Period: If the coefficient of variation of price fluctuation is ≤15% within three consecutive cycles, and the time interval between peak and trough values ​​meets the requirements of ≤1 month for quarterly cycles and ≤3 months for annual cycles, then the price cycle is deemed valid.

[0031] S3. Correlation analysis of equipment and material prices: This step quantifies the correlation between equipment and raw material prices. In one exemplary implementation, the process is as follows: S31. Price Data Extraction: Data is selected based on the correspondence between equipment and raw materials, such as the price of copper corresponding to transformers and the price of silicon corresponding to components, to ensure that raw materials are the core influencing factor.

[0032] S32. Data Preprocessing: Data quality is ensured by removing missing and outlier values, such as deleting abnormal peak data for silicon material prices in a certain month.

[0033] S33. Relevance Calculation: Calculate the Pearson correlation coefficient. If the correlation coefficient between transformer prices and copper prices is 0.85, it is considered a strong positive correlation.

[0034] S34. Linear Regression Calculation: A linear regression model was constructed, outputting the regression equation "transformer price = 0.8 × copper price + 5000". The correlation was quantified. When the copper price increases by 1000 yuan / ton, it can be predicted that the transformer price will increase by 800 yuan, thus obtaining the conclusion of "correlation between equipment and raw material prices", which provides support for price prediction.

[0035] S4. Real-time analysis of equipment and material prices: This step integrates the analysis results from the preceding steps to determine price changes in real time. In one exemplary implementation, the process is as follows: S41. Data Integration: Integrate the quote credibility level results from step S1, the effective price cycle pattern from step S2, and the price correlation data from step S3.

[0036] S42. Judgment of Material Price Changes: Based on the integrated data, if the current price cycle is at a trough, raw material prices are falling, and more than 60% of the quotes are highly credible, it is determined that prices are likely to fall; if the cycle is at a peak, raw material prices are rising, and the proportion of highly credible quotes is high, it is determined that prices are likely to rise.

[0037] If the data is contradictory, such as the cycle showing a trough but raw material prices rising, the magnitude of the change can be predicted using a regression equation. If the magnitude is ≤3%, the price is considered stable. For example, if a certain equipment is in a trough but raw material prices rise slightly, the regression equation predicts a 2% price increase, and the price is considered to be trending towards stability.

[0038] S43. Regular updates and optimizations: The database is updated quarterly, the analysis process is repeated, and model parameters are optimized to ensure accurate real-time analysis and provide a basis for enterprises to dynamically adjust their procurement strategies.

[0039] The following section uses the "110kV transformer," a typical piece of equipment in the energy and power industry, as an example to illustrate the implementation process of the present invention.

[0040] Data preparation before implementation: Based on the equipment and materials database, multi-dimensional data related to 110kV transformers were extracted to provide complete and effective data support for subsequent analysis steps. The specific data is as follows: Standard price data: Select the historical contract price of 120,000 yuan in 2023 (weight 35%), the historical contract price of 118,000 yuan in 2022 (weight 25%), and the regular inquiry data of 122,000 yuan in January 2024 (weight 40%). Calculate the standard price using the weighted formula: Standard price = 120,000 × 35% + 118,000 × 25% + 122,000 × 40% = 120,300 yuan.

[0041] Supplier quotation data: Valid quotations from 5 qualified suppliers for the same 110kV transformer were extracted, with prices of RMB 121,000, RMB 123,000, RMB 135,000, RMB 119,000, and RMB 122,000 respectively.

[0042] Historical price data: Quarterly average price data of 110kV transformers for 5 years from 2019 to 2023 were extracted to form a continuous time series of 16 data points for price cycle analysis.

[0043] Raw material price data: Price data of the core raw material (copper) directly related to the 110kV transformer were selected. Monthly copper price data for three years from 2021 to 2023 were extracted to form 36 data points, which were used to analyze the correlation between equipment and raw material prices.

[0044] Based on the above data preparation, the implementation process is as follows: I. Analysis of Deviation in Equipment and Material Prices: Data preprocessing: Calculate the mean and standard deviation of the quotations from the 5 suppliers. The mean quotation is (12.1 + 12.3 + 13.5 + 11.9 + 12.2) / 5 = 124,000 yuan. The standard deviation is calculated to be 5,800 yuan. Three times the standard deviation is 0.58 × 3 = 17,400 yuan. Since the deviation of all quotations from the mean is less than 17,400 yuan, there are no outliers to be removed.

[0045] Deviation rate calculation: Calculate the deviation rate for each supplier's quotation according to the formula: Deviation rate = (Supplier quotation - Standard price) / Standard price × 100%. Supplier 1 (121,000 RMB): Deviation rate = (121,000 - 120,000) / 120,000 × 100% ≈ 0.58%; Supplier 2 (123,000 yuan): Deviation rate = (123,000 - 120,000) / 120,000 × 100% ≈ 2.24%; Supplier 3 (135,000 yuan): Deviation rate = (135,000 - 120,300) / 120,300 × 100% ≈ 12.22%; Supplier 4 (119,000 yuan): Deviation rate = (119,000 - 120,000) / 120,000 × 100% ≈ -1.08%; Supplier 5 (122,000 yuan): Deviation rate = (122,000 - 120,000) / 120,000 × 100% ≈ 1.41%; Credibility assessment: According to the credibility grading standard, the absolute value of the deviation rate is ≤5% for high credibility, 5%-10% for medium credibility, and >10% for low credibility. The assessment results are as follows: Suppliers 1, 2, 4, and 5 are high credibility quotes, supplier 3 is a low credibility quote, and the high credibility quotes account for 80%. The high credibility quotes can be used for subsequent cost calculations.

[0046] II. Cyclical Analysis of Equipment and Material Prices: Time series construction and stationarization: The quarterly average price data of 110kV transformers from 2019 to 2023 were organized in annual-quarter order to form a continuous price time series; the stationarity test was performed using the ADF test method, and the calculated ADF test p-value was 0.25 (P>0.05), indicating that the time series was non-stationary; after first-order differencing, the ADF test was performed again, and the p-value was 0.02 (P<0.05), thus transforming the time series into a stationary series.

[0047] Cycle Mining: The ARIMA model was used to analyze the stationary time series. The optimal parameters (p,d,q) of the model were screened using the AIC criterion. The optimal parameters were determined to be (2,1,1). Based on these parameters, the price change curve was fitted to explore the price fluctuation pattern. It was found that the price of 110kV transformers exhibits an annual cyclical characteristic of "rising in Q2-Q3 and falling in Q4-Q1".

[0048] Cycle verification: Select three consecutive annual cycles (2021-2023), calculate the coefficient of variation of price fluctuation amplitude = 12% (≤15%), and the time interval between the peaks of each cycle is 1 month (≤3 months, which meets the annual cycle requirement). The price cycle of that year is determined to be valid and can be used to judge the subsequent price trend.

[0049] III. Correlation Analysis of Equipment and Material Prices: Data preprocessing: Monthly price data of 110kV transformers and monthly copper prices from 2021 to 2023 were preprocessed. An abnormal peak in copper prices in August 2022 (deviating from the monthly average by more than 3 standard deviations) was found. This outlier was removed to ensure the integrity and validity of the two types of data.

[0050] Correlation calculation: The Pearson correlation coefficient was used to calculate the degree of linear correlation between the two. Substituting the data, the correlation coefficient r was calculated to be 0.83 (>0.7), indicating that the price of 110kV transformers and the price of copper are strongly positively correlated.

[0051] Regression modeling: Using copper price as the independent variable x and 110kV transformer price as the dependent variable y, the regression equation is output through data fitting: y=0.78x+5200. This equation quantifies the correlation between the two, that is, for every 1000 yuan / ton increase in copper price, the 110kV transformer price is expected to increase by 780 yuan, providing a quantitative basis for price prediction.

[0052] IV. Real-time Analysis of Equipment and Material Prices: Data integration includes: 80% of high-credibility quotes obtained from price deviation analysis; "Q4-Q1 is the price trough period" obtained from price cycle analysis (currently Q4 2024, at the trough of the cycle); "transformer prices are strongly positively correlated with copper prices" obtained from price correlation analysis; and current data showing a 5% month-on-month decrease in copper prices.

[0053] Trend assessment: We are currently in the trough of the price cycle, with raw material (copper) prices trending downwards. Moreover, the proportion of highly reliable quotes is ≥60% (80% or more), and the data from multiple dimensions is consistent. Therefore, it is determined that the price of 110kV transformers is likely to decline in the next step.

[0054] Regular updates and optimizations: Regularly update the 110kV transformer price data, supplier quotation data, and copper price data in the equipment and materials database. Repeat steps one to four in this embodiment to optimize the ARIMA model parameters and linear regression equations, ensuring the accuracy of real-time price analysis and providing continuous data support for enterprises to dynamically adjust their procurement strategies.

[0055] Although embodiments of the present invention have been described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present invention, and all such changes and alterations shall not depart from the protection scope of the present invention.

Claims

1. A data analysis method for an equipment materials database, characterized in that, Includes the following steps: S1. Equipment and Material Price Deviation Analysis: S11. Extract standard price data and corresponding supplier quotation data of the target equipment materials from the equipment and materials database. The standard price data is obtained by weighted calculation based on historical contract prices and periodic inquiry data. The supplier quotation data includes valid quotations from at least 3 qualified suppliers for the same specification of equipment materials. S12. Preprocess supplier quotation data to remove outliers; S13. Calculate the deviation rate of each supplier's quotation relative to the standard price: Deviation rate = (Supplier quotation - Standard price) / Standard price × 100%; S14. Determine the credibility level of the supplier's quotation by comparing the absolute value of the deviation rate with the corresponding thresholds preset for different credibility levels; S2. Cyclical Analysis of Equipment and Material Prices: S21. Extract at least 5 years of historical price data for the target equipment materials from the equipment and materials database, and organize them into a continuous price time series by quarter or year; S22. The ADF test is used to test the stationarity of price time series. Non-stationary series are transformed into stationary series through first-order differencing. S23. The stationary time series is analyzed using the ARIMA model. The optimal parameters of the model are selected using the AIC criterion. The optimized ARIMA model is used to fit the price change curve to explore the quarterly / annual fluctuation pattern of equipment and material prices and determine the price cycle. S24. Verify the price cycle: If the coefficient of variation of price fluctuation amplitude is ≤15% within 3 consecutive cycles, and the time interval deviation between peak and trough values ​​meets the requirements of quarterly cycle ≤1 month and annual cycle ≤3 months, then the price cycle is deemed valid. S3. Correlation analysis of equipment and material prices: S31. Extract at least three years of monthly price data for the target equipment material from the equipment material database, as well as monthly price data for raw materials directly related to the equipment material; S32. Preprocess the two types of monthly price data to remove missing and outlier values; S33. The Pearson correlation coefficient is used to calculate the linear correlation between equipment material prices and raw material prices, and the correlation level between the two is determined by the absolute value of the correlation coefficient. S34. Construct a linear regression model with raw material prices as independent variables and equipment material prices as dependent variables, and output the regression equation to quantify the correlation between the two. S4. Real-time analysis of equipment and material prices: S41. Integrate the quote credibility level results from step S1, the effective price cycle pattern from step S2, and the price correlation data from step S3; S42. Based on the integrated data, determine the next change in equipment and material prices: If the current price cycle is at its trough, raw material prices are trending downwards, and the proportion of highly reliable quotes is ≥60%, then the price is likely to decrease; if the current price cycle is at its peak, raw material prices are trending upwards, and the proportion of highly reliable quotes is ≥60%, then the price is likely to increase; if there are contradictions in the integrated data, combine the regression equation output from step S3 to predict the price change range. When the absolute value of the change range is ≤3%, the price is determined to be stabilizing. S43. Update the equipment and materials database data quarterly and repeat steps S1-S4 to optimize the analysis results.

2. The data analysis method for an equipment material database as described in claim 1, characterized in that, In step S11, when calculating the standard price data, historical contract prices are weighted according to their time period: the weight of historical contract prices in the past year is 30%-40%, the weight of historical contract prices in the past 1-2 years is 20%-30%, and the weight of historical contract prices in the past 2 years or more is 10%-20%; the weight of the reliability of periodic inquiry data is 70%-80%.

3. The data analysis method for an equipment material database as described in claim 1, characterized in that, In step S23, if the ARIMA model does not fit the stationary time series to the preset standard, it is replaced by the Fourier transform method to decompose the time series into sine waves of different frequencies. By analyzing the sine wave frequencies, the quarterly / annual fluctuation patterns and price cycles of equipment and material prices are determined.

4. The data analysis method for an equipment material database as described in claim 1, characterized in that, In step S31, the selection of raw materials directly related to equipment materials is determined based on the correspondence between the core components and raw materials in the equipment material manufacturing process, ensuring that the selected raw materials are the core factors affecting the price of equipment materials.

5. The data analysis method for an equipment material database as described in claim 1, characterized in that, In step S33, the specific criteria for determining the correlation level based on the absolute value of the Pearson correlation coefficient are as follows: A correlation coefficient with an absolute value greater than 0.7 is considered a strong correlation; a correlation coefficient with an absolute value between 0.3 and 0.7 is considered a moderate correlation; and a correlation coefficient with an absolute value less than or equal to 0.3 is considered a weak correlation.

6. The data analysis method for an equipment material database as described in claim 1, characterized in that, Step S4 also includes: The price change curve obtained in step S2, the correlation scatter plot generated based on the price correlation data in step S3, and the real-time analysis conclusions obtained in step S42 are presented in chart form.

7. The data analysis method for an equipment material database as described in claim 1, characterized in that, In step S12, before removing outliers, the mean and standard deviation of all supplier quotations are calculated, and quotations that deviate from the mean by more than 3 times the standard deviation are identified as outliers and removed.

8. A data analysis method for an equipment material database as described in any one of claims 1-7, characterized in that, In step S24, the coefficient of variation of price fluctuation is calculated as follows: First, calculate the mean and standard deviation of the price fluctuation range for each of the three consecutive periods. Then, calculate the coefficient of variation using the formula: coefficient of variation = standard deviation / mean × 100%.

Images