Online evaluation method for data resource value

Abstract

The invention relates to the technical field of data resource management, in particular to an online evaluation method for data resource value, which comprises the following steps of: constructing a multi-dimensional evaluation index set, and acquiring metadata and market dynamic information of to-be-evaluated data in real time through a distributed interface; carrying out weight distribution on the index set based on a dynamic weighting algorithm, generating a weight coefficient set and constructing a value calculation model; performing standardized preprocessing on the collected original data, inputting the data into the model, and calculating an initial estimated value; and dynamically correcting the initial value estimation value based on the timeliness attenuation factor and the data scarcity factor, and generating a final value estimation value and an estimation report. Through multi-dimensional index construction, a dynamic weighting algorithm, standardized preprocessing and two-factor correction, automatic evaluation of the data value is realized, the problems of single dimension, static weight and neglect of timeliness scarcity of a traditional method are solved, the real-time performance and scientificity are improved, and a quantitative decision basis is provided for data capitalization.

Description

Technical Field

[0001] The present invention relates to the technical field of data resource management, and in particular to an online evaluation method for data resource value. Background Art

[0002] With the rapid development of information technology, the value of data resources as a core production factor has become increasingly prominent, and the scientific evaluation of the value of data resources has become a key link in data asset management and transaction circulation. Early evaluation methods were mostly based on a single dimension, such as focusing only on data scale or storage costs, which made it difficult to fully reflect the comprehensive value of data. As research deepened, an evaluation index system with multiple dimensions such as data quality and application scenarios was gradually formed. Some solutions introduced static weight allocation methods to construct value models, and attempted to make preliminary corrections based on timeliness factors such as data update frequency. However, existing technologies mainly remain at the static evaluation of basic data attributes, and the real-time collection and integration of market dynamic information is insufficient. In addition, the weight allocation mechanism lacks adaptability to the correlation between indicators and the dynamic changes in data characteristics, making it difficult to meet the value evaluation needs of data resources in a complex market environment.

[0003] The shortcomings of existing technologies are concentrated in three aspects: First, there are blind spots in the evaluation indicator dimensions, and the quantitative analysis of the market supply and demand characteristics and cost-effectiveness of data is not systematic enough. In particular, there is a lack of real-time tracking and modeling of external dynamic factors such as changes in policies and regulations and fluctuations in the transaction prices of similar data. Second, the weight allocation method is static and cannot dynamically adjust the importance of indicators according to changes in data characteristics, resulting in insufficient adaptability of the evaluation model to different application scenarios. Third, there are shortcomings in the data preprocessing and value correction mechanism, lacking efficient means to handle missing values ​​and outliers. In addition, there is no time-attenuation model based on data age and industry benchmark validity period, and no quantitative correction system for scarcity premium has been established. As a result, the accuracy and dynamic adaptability of the evaluation results are difficult to guarantee. In addition, the evaluation reports generated by existing solutions generally lack indicator contribution analysis and confidence interval calculation, and cannot provide comprehensive explanatory support for data trading decisions. Summary of the Invention

[0004] The purpose of the present invention is to overcome the above problems and provide an online evaluation method for the value of data resources. To achieve the above purpose, the present invention adopts the following technical solutions:

[0005] like Figure 1 As shown, an online evaluation method for data resource value includes the following steps:

[0006] Step S1: Construct a multi-dimensional evaluation indicator set and collect metadata and market dynamic information of the data to be evaluated in real time through a distributed interface;

[0007] Step S2: weight the indicator set based on a dynamic weighting algorithm, generate a weight coefficient set, and construct a value calculation model;

[0008] Step S3: Standardize and pre-process the collected raw data and input it into the model to calculate the initial value estimate;

[0009] Step S4: Dynamically revise the initial value estimate based on the timeliness attenuation factor and the data scarcity factor to generate a final value estimate and evaluation report.

[0010] Furthermore, in step S1, the multi-dimensional evaluation index set includes basic attribute indicators, market value indicators and cost efficiency indicators; the basic attribute indicators include data scale, update frequency and structural complexity;

[0011] The data model balances data size differences based on the logarithmic transformation of the total number of records, and the calculation formula is:

[0012] M=log 10 (N+1)

[0013] Among them, M is the data model, N is the total number of data records;

[0014] The update frequency maps the data timeliness to the inverse of the average update interval, and the calculation formula is:

[0015]

[0016] Where F is the update frequency, Δt avg The average update interval of the data source in days;

[0017] The structural complexity measures the degree of data structuring by a weighted combination of the number of entity associations and the number of dimensions. The calculation formula is:

[0018] C b =αE+βD(α+β=1)

[0019] Among them, C b is the structural complexity, E is the total number of relationships between entities in the data model, and D is the number of data dimensions;

[0020] The market value indicators include demand heat H and competition scarcity S;

[0021] H=γ1Q s +γ2P t (γ1+γ2=1)

[0022] Among them, Q s is the search volume growth rate, P t is the transaction price volatility, γ1 and γ2 are adjustment coefficients;

[0023]

[0024] Among them, V u is the number of unique features, N c The number of competitors.

[0025] The cost-effectiveness indicators include storage cost rate, computing cost rate and revenue conversion rate.

[0026] Furthermore, in step S1, the market dynamic information includes industry search index, policy and regulation change events, and similar data transaction price volatility; the industry search index is obtained by crawling the search volume of search engine keywords in real time, and the policy and regulation change events are obtained by extracting keyword change information in government announcements through natural language processing to trigger the update of industry benchmark validity period parameters; the similar data transaction price volatility is calculated by standard deviation through blockchain transaction records, and the formula is:

[0027]

[0028] Among them, P t is the transaction price volatility, p is the i-th transaction price, μ p is the historical average price, n is the number of transactions, when the volatility exceeds the threshold When , the recalculation of demand heat H is triggered; σ p is the standard deviation of historical prices.

[0029] Furthermore, in step S2, the dynamic weighting algorithm is an improved CRITIC algorithm, and its indicator weight calculation formula is:

[0030]

[0031] Among them, w j Represents the weight coefficient of the jth indicator σ j is the standard deviation of index j, μ j is the mean value of index j, ε=1×10 -5 A very small constant to prevent division by zero; r ji is the Pearson correlation coefficient between indicator j and indicator i.

[0032] Furthermore, in step S3, the standardization preprocessing includes missing value filling, outlier processing and normalization operations;

[0033] The missing value filling adopts the random forest regression algorithm to predict the missing fields, and its model is expressed as:

[0034]

[0035] in, is the predicted result of the i-th missing value, X observed is the observation feature matrix, f RF is a random forest regression model;

[0036] The outlier processing is based on the 3σ principle to truncate the data that exceeds the mean ± 3 times the standard deviation. The truncation rule is:

[0037]

[0038] Where x′ is the truncated data value, μ is the data mean, and σ is the standard deviation;

[0039] The normalization operation maps the numerical index to the interval [0,1] through the Min-Max algorithm, and the formula is:

[0040]

[0041] Among them, x norm is the normalized value, x min is the minimum value in the sliding window, x max is the maximum value in the sliding window.

[0042] Furthermore, in step S4, the calculation formula of the time-effect attenuation factor θ is:

[0043]

[0044] Where t is the data age, T ref is the industry benchmark validity period, and k is the decay rate coefficient.

[0045] Furthermore, in step S4, the calculation formula of the data scarcity factor ρ is:

[0046]

[0047] Among them, V u is the number of unique features of the data, V th is the top 10% percentile threshold of the industry, and η=0.1 is the adjustment coefficient.

[0048] Furthermore, in step S4, the evaluation report includes final value estimation, indicator contribution analysis and confidence interval;

[0049] The calculation formula for the terminal value valuation is:

[0050]

[0051] Among them, V final is the final valuation, V initis the initial estimate, exp is the natural exponential function, k is the decay rate coefficient, t is the data age, T ref is the industry benchmark validity period, η=0.1 is the adjustment coefficient, Vu is the number of unique data features, V th The threshold for the top 10% of the industry;

[0052] The calculation formula of the indicator contribution is:

[0053]

[0054] Among them, I j is the contribution of the jth indicator, w j is the weight coefficient of the jth indicator, x j is the original index value, x j,min and x j,max is the normalized boundary value of the jth indicator;

[0055] The calculation formula of the confidence interval is:

[0056] Confidence interval = μ v ±1.96σ v

[0057] Among them, μ v is the historical evaluation mean, σ v is the historical evaluation standard deviation, and 1.96 is the quantile of the standard normal distribution at the 95% confidence level.

[0058] The advantages of the present invention are:

[0059] 1. This invention achieves comprehensive coverage and real-time tracking of factors influencing the value of data resources by constructing a multidimensional evaluation index set encompassing basic attributes, market value, and cost-effectiveness. It also utilizes a distributed interface to collect metadata, industry search indices, policies and regulations, transaction prices, and other market dynamics in real time. Combined with specific indicator formulas such as logarithmic transformation of data size and reciprocal calculation of update frequency, this method accurately quantifies the underlying quality of data and the state of market supply and demand, laying a multidimensional data foundation for subsequent value calculations and effectively avoiding the one-sidedness of single-dimensional evaluations.

[0060] 2. This paper employs a modified CRITIC algorithm for dynamic weighting. This method uses the standard deviation of indicators to measure data variability and the Pearson correlation coefficient to reflect the correlation between indicators. This model constructs a weight coefficient calculation model, enabling dynamic adaptation of weight assignment to the relationship between data characteristics and indicators. Compared to fixed weighting methods, this mechanism can adjust the importance of each indicator in real time based on market dynamics, making the weights more closely aligned with the value contribution of data resources in different scenarios, thereby enhancing the scientific nature and flexibility of the evaluation model.

[0061] 3. This paper performs standardized preprocessing on raw data using random forest regression to fill missing values, the 3σ principle to truncate outliers, and Min-Max normalization. It then dynamically adjusts the initial valuation using a time-sensitive decay factor and a data scarcity factor, achieving dual optimization of data quality and dynamic value adjustment. Preprocessing ensures the reliability of input data, while dynamic correction effectively reflects the value decay over time and the premium effect of data scarcity. Ultimately, an evaluation report containing indicator contribution and confidence intervals is generated, providing a precise and interpretable quantitative basis for data management and trading decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The drawings constituting a part of this application are used to provide a further understanding of this application and make other features, purposes and advantages of this application more apparent. The drawings and descriptions of the exemplary embodiments of this application are used to explain this application and do not constitute an improper limitation on this application.

[0063] In the attached figure:

[0064] Figure 1 This is a flow chart of an online evaluation method for data resource value in Example 1.

[0065] Figure 2 This is a flow chart of the data preprocessing process in an online evaluation method for data resource value in Example 1. DETAILED DESCRIPTION

[0066] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0067] The present invention is described in detail and specifically below through specific examples to provide a better understanding of the present invention. However, the following examples do not limit the scope of protection of the present invention.

[0068] Example 1

[0069] like Figure 1-2 As shown, an online evaluation method for data resource value includes the following steps:

[0070] Step S1: Construct a multi-dimensional evaluation indicator set and collect metadata and market dynamic information of the data to be evaluated in real time through a distributed interface;

[0071] Step S2: weight the indicator set based on a dynamic weighting algorithm, generate a weight coefficient set, and construct a value calculation model;

[0072] Step S3: Standardize and pre-process the collected raw data and input it into the model to calculate the initial value estimate;

[0073] Step S4: Dynamically revise the initial value estimate based on the timeliness attenuation factor and the data scarcity factor to generate a final value estimate and evaluation report.

[0074] For example, when a financial institution conducts a customer credit data valuation, it first executes step S1 to construct a multi-dimensional evaluation indicator set encompassing basic attributes, market value, and cost-effectiveness. Basic attribute indicators measure the data's quality, market value indicators reflect market supply and demand, and cost-effectiveness indicators assess the data's input-output efficiency. A distributed interface collects customer credit data metadata in real time, such as the number of transaction records and average update interval, as well as market dynamics such as industry search volume, policy documents, and transaction prices for similar data. This data provides the foundation for subsequent weight assignment and value calculation. Step S2 utilizes an improved CRITIC algorithm to dynamically assign weights based on indicator variability and correlation, constructing a value calculation model. Step S3 preprocesses the raw data by filling in missing values, addressing outliers, and normalizing the data to improve data quality. The data is then fed into the model to calculate the initial value. Step S4 dynamically adjusts the initial estimate based on factors such as data age, industry benchmark validity period, and uniqueness, ultimately generating an evaluation report containing multi-dimensional analysis. Each step is interconnected, forming a complete online valuation process.

[0075] This embodiment has constructed a complete online evaluation process covering data collection, weight allocation, value calculation and dynamic correction. Through step S1, a multi-dimensional evaluation indicator set is constructed, and the metadata and market dynamic information of the data to be evaluated are collected in real time using a distributed interface to ensure the comprehensiveness and real-time nature of the evaluation data; step S2 uses a dynamic weighting algorithm to assign weights to indicators, combined with the standardized preprocessing of step S3 to improve data quality, and then the initial valuation is dynamically corrected through the timeliness attenuation factor and scarcity factor of step S4, and finally an evaluation report containing multi-dimensional analysis is generated. This process realizes fully automated processing from data collection to value output, solving the problem that traditional evaluation methods rely on manual intervention and data lag.

[0076] Furthermore, in step S1, the multi-dimensional evaluation index set includes basic attribute indicators, market value indicators and cost efficiency indicators; the basic attribute indicators include data scale, update frequency and structural complexity;

[0077] The data model balances data size differences based on the logarithmic transformation of the total number of records, and the calculation formula is:

[0078] M=log 10 (N+1)

[0079] Among them, M is the data model, N is the total number of data records;

[0080] The update frequency maps the data timeliness to the inverse of the average update interval, and the calculation formula is:

[0081]

[0082] Where F is the update frequency, Δt avg The average update interval of the data source in days;

[0083] The structural complexity measures the degree of data structuring by a weighted combination of the number of entity associations and the number of dimensions. The calculation formula is:

[0084] C b =αE+βD(α+β=1)

[0085] Among them, C b is the structural complexity, E is the total number of relationships between entities in the data model, and D is the number of data dimensions;

[0086] The market value indicators include demand heat H and competition scarcity S;

[0087] H=γ1Q s +γ2P t (γ1+γ2=1)

[0088] Among them, Q s is the search volume growth rate, P t is the transaction price volatility, γ1 and γ2 are adjustment coefficients;

[0089]

[0090] Among them, V u is the number of unique features, N c The number of competitors.

[0091] The cost-effectiveness indicators include storage cost rate, computing cost rate and revenue conversion rate.

[0092] In the multi-dimensional indicator set constructed in step S1, the basic attribute indicators include data scale, update frequency, and structural complexity. The data scale is transformed by the logarithmic transformation formula M=log 10 (N+1) balance difference, if the total number of customer credit data records N = 50000, then M≈4.698; the update frequency is calculated as the inverse of the average update interval, If the average update interval of the data source is 10 days, then F = 0.0909; the structural complexity is a weighted combination of the number of entity associations and the number of dimensions, C b=αE+βD, assuming that the entity association number E = 4, the data dimension D = 15, the weighting coefficient α = 0.5, β = 0.5, then C b =9.5. In the market value index, demand heat H = γ1Q s +γ2P t , if the search volume growth rate Q s =15%, trading price volatility P t =10%, adjustment coefficients γ1 = 0.6, γ2 = 0.4, then H = 13%; competitive scarcity If the unique feature V u =2,Number of competitors c =3, then S≈0.0687. The cost-effectiveness indicators cover storage cost rate, computing cost rate, and revenue conversion rate, which are 3%, 496%, and 12%, respectively. These specific indicators provide a clear quantitative basis for subsequent dynamic weighting and value calculation.

[0093] Furthermore, in step S1, the market dynamic information includes industry search index, policy and regulation change events, and similar data transaction price volatility; the industry search index is obtained by crawling the search volume of search engine keywords in real time, and the policy and regulation change events are obtained by extracting keyword change information in government announcements through natural language processing to trigger the update of industry benchmark validity period parameters; the similar data transaction price volatility is calculated by standard deviation through blockchain transaction records, and the formula is:

[0094]

[0095] Among them, P t is the transaction price volatility, p i is the i-th transaction price, μ p is the historical average price, n is the number of transactions, when the volatility exceeds the threshold When , the recalculation of demand heat H is triggered; σ p is the standard deviation of historical prices.

[0096] The market dynamic information collected in step S1 is crucial for the real-time nature of the evaluation results. The industry search index is obtained by crawling the search engine keyword search volume in real time. For example, the weekly search volume of "personal credit report" is used to calculate the search volume growth rate of demand popularity; policy and regulatory change events are analyzed through natural language processing to analyze government announcements, such as the credit data management policy issued by the central bank. After extracting keyword change information, it triggers the update of the industry benchmark validity period parameters, affecting the calculation of the timeliness attenuation factor; the price volatility of similar data transactions is calculated through the standard deviation of blockchain transaction records.

[0097] If the transaction price is [70, 75, 80, 65, 78] yuan / piece, the calculated mean μp=73.6, the standard deviation σ p ≈5.67, when the volatility exceeds the threshold μp±2σ p When a new product is launched, the demand heat is recalculated, ensuring that the evaluation model responds promptly to market fluctuations. The collection and processing of this dynamic information enables the evaluation system to incorporate external environmental variables in real time, improving its adaptability to market changes.

[0098] By crawling search engine keyword search volumes in real time to obtain industry search indexes, natural language processing technology is used to extract keyword changes in government announcements to identify policy and regulatory changes and trigger updates to industry benchmark validity period parameters. The volatility of transaction prices for similar data is calculated based on blockchain transaction records, and when the volatility exceeds a set threshold, a recalculation of demand heat is automatically triggered. These technical approaches integrate external environmental variables into the evaluation system in real time, overcoming the limitations of traditional methods that rely solely on static data attributes. These methods enable evaluation results to be updated in real time as industry search heat, policy adjustments, and market transaction fluctuations change, significantly improving the evaluation method's responsiveness to dynamic market environments.

[0099] Furthermore, in step S2, the dynamic weighting algorithm is an improved CRITIC algorithm, and its indicator weight calculation formula is:

[0100]

[0101] Among them, w j Represents the weight coefficient of the jth indicator σ j is the standard deviation of index j, μ j is the mean value of index j, ε=1×10 -5 A very small constant to prevent division by zero; r ji is the Pearson correlation coefficient between indicator j and indicator i.

[0102] The improved CRITIC algorithm used in step S2 is the key link between the indicator set and the value calculation model. Taking data scale, update frequency, and demand heat as examples, we first calculate the standard deviation σ of each indicator. j , reflecting data variability, such as the standard deviation of data size σ M =1.2, update frequency standard deviation σ F =0.03, standard deviation of demand heat σ H =0.05; then calculate the sum of the absolute values ​​of the Pearson correlation coefficients between the indicators R j , reflects the correlation of indicators. For example, the absolute value of the correlation coefficient between data scale and demand heat is 0.8, and the absolute value of the correlation coefficient between update frequency and demand heat is 0.6. Then the R H=1.4; contrast Final weight This formula assigns higher weight to data scale due to its high variability and strong correlation with demand. The dynamic weighting mechanism adjusts the importance of indicators in real time based on data characteristics and market conditions, making weight allocation more relevant to actual scenarios and providing a scientific set of weight coefficients for the value calculation model.

[0103] Furthermore, in step S3, the standardization preprocessing includes missing value filling, outlier processing and normalization operations;

[0104] The missing value filling adopts the random forest regression algorithm to predict the missing fields, and its model is expressed as:

[0105]

[0106] in, is the predicted result of the i-th missing value, X observed is the observation feature matrix, f RF is a random forest regression model;

[0107] The outlier processing is based on the 3σ principle to truncate the data that exceeds the mean ± 3 times the standard deviation. The truncation rule is:

[0108]

[0109] Where x′ is the truncated data value, μ is the data mean, and σ is the standard deviation;

[0110] The normalization operation maps the numerical index to the interval [0,1] through the Min-Max algorithm, and the formula is:

[0111]

[0112] Among them, x norm is the normalized value, x min is the minimum value in the sliding window, x max is the maximum value in the sliding window.

[0113] Missing value filling adopts the random forest regression algorithm and uses the observation feature matrix to predict missing fields. For example, it predicts missing values ​​of structural complexity through features such as data scale and update frequency, which improves the prediction accuracy of missing data compared with simple interpolation. Outlier processing is based on the 3σ principle, and data that exceeds 3 times the standard deviation of the mean is truncated to effectively eliminate noise data. Min-Max normalization maps indicators to the range of 0 to 1, eliminating dimensional differences, and can unify the scale of storage cost rate and revenue conversion rate, solving the problems of missing data, anomalies, and dimensional inconsistency, ensuring that the data input to the model is accurate and reliable, avoiding valuation deviations caused by data quality issues, and providing high-quality data input for subsequent value calculations.

[0114] Standardization preprocessing in step S3 is a key step in ensuring data quality and directly impacts the accuracy of the initial valuation. Missing values ​​are filled using the random forest regression algorithm. For example, when a customer's income field is missing, a feature matrix constructed using observed features such as age, occupation, and debt-to-equity ratio predicts the missing value as 300,000 yuan. This improves data integrity compared to simple interpolation. Outlier handling is based on the 3σ principle. If the customer's credit score has a mean of μ = 750 and a standard deviation of σ = 50, outliers exceeding μ + 3σ = 900 or falling below μ - 3σ = 600 are truncated to ensure that the data is within a reasonable range. Normalization, using the Min-Max algorithm, maps different dimensional indicators such as age and income to the [0, 1] interval. For example, if the minimum income value is 30,000 yuan and the maximum is 500,000 yuan, an income of 150,000 yuan is normalized to 0.2553, eliminating dimensional differences. This preprocessed data provides reliable input for the valuation model, preventing valuation bias caused by data quality issues.

[0115] Furthermore, in step S4, the calculation formula of the time-effect attenuation factor θ is:

[0116]

[0117] Where t is the data age, T re f is the validity period of the industry benchmark, and k is the decay rate coefficient.

[0118] The exponential relationship between data age and the validity period of industry benchmarks quantifies the decay of data value over time. For example, if the validity period of a financial data industry benchmark is 30 days and a dataset has been stored for 60 days, the decay factor decreases as the parameter value increases. The initial estimate is significantly lower after correction, reflecting the decline in the value of outdated data. This factor, for the first time, incorporates the time dimension into data value assessment, resolving the problem of traditional methods ignoring data "freshness." It enables assessment results to dynamically reflect the changes in data value throughout its lifecycle, consistent with the objective law that data timeliness influences decision-making value in real applications.

[0119] The calculation formula of the time-effect attenuation factor in step S4 is Combined with data age t and industry benchmark validity period T ref , decay rate coefficient k, if a customer's credit data was generated 90 days ago, the industry benchmark validity period is 180 days, k = 0.8, then This indicates that the value of data decays by approximately 67.03% due to timeliness. This factor incorporates the time dimension into the assessment for the first time, resolving the problem of traditional methods ignoring the "freshness" of data. This allows the assessment results to dynamically reflect the value changes of data throughout its lifecycle, consistent with the actual law in the financial sector that the value of credit data gradually decreases over time.

[0120] Furthermore, in step S4, the calculation formula of the data scarcity factor ρ is:

[0121]

[0122] Among them, V u is the number of unique features of the data, V th is the top 10% percentile threshold of the industry, and η=0.1 is the adjustment coefficient.

[0123] The premium effect of data scarcity is quantified through the logistic function relationship between the number of unique features and the industry's top 10% quantile threshold. When the number of unique features of a piece of data exceeds the industry threshold, the scarcity factor approaches 1, significantly increasing the initial valuation and reflecting the market principle that "scarcity makes things valuable." This factor transforms the market competitiveness of data into a calculable quantitative indicator, addressing the shortcomings of traditional valuations that fail to consider scarcity. This allows the evaluation results to reflect the actual supply and demand value of data in the market, providing a more market-oriented basis for data transaction pricing.

[0124] The data scarcity factor is another important correction factor in step S4. Its formula is Quantify the market competitiveness of data. If customer credit data contains three unique features, such as no default history, a high asset rating, and overseas transaction records, and the industry top 10% threshold is 2, then ρ ≈ 0.525. Because the number of unique features exceeds the industry threshold, the scarcity factor enhances the value of the data, reflecting the market principle of "scarcity makes things valuable." This factor addresses the deficiency of traditional assessments that fail to consider scarcity, enabling assessment results to reflect the actual supply and demand value of data in the market and providing a more market-oriented basis for pricing financial data transactions.

[0125] Furthermore, in step S4, the evaluation report includes final value estimation, indicator contribution analysis and confidence interval;

[0126] The calculation formula for the terminal value valuation is:

[0127]

[0128] Among them, V final is the final valuation, V init is the initial estimate, exp is the natural exponential function, k is the decay rate coefficient, t is the data age, T ref is the industry benchmark validity period, η=0.1 is the adjustment coefficient, V u is the number of unique features of the data, V th The threshold for the top 10% of the industry;

[0129] The calculation formula of the indicator contribution is:

[0130]

[0131] Among them, I j is the contribution of the jth indicator, w j is the weight coefficient of the jth indicator, x j is the original index value, x j,min and x j,max is the normalized boundary value of the jth indicator;

[0132] The calculation formula of the confidence interval is:

[0133] Confidence interval = μ v ±1.96σ v

[0134] Among them, μ v is the historical evaluation mean, σ v is the historical evaluation standard deviation, and 1.96 is the quantile of the standard normal distribution at the 95% confidence level.

[0135] The evaluation report generated in step S4 is the final output of the entire evaluation process, including the final value valuation, indicator contribution analysis, and confidence interval. The final value valuation is calculated by multiplying the initial valuation by the timeliness attenuation factor and the scarcity factor. If the initial valuation is 800,000 yuan, θ=0.6703, ρ=0.525, then the final valuation is approximately 281,500 yuan; the indicator contribution analysis is calculated by Calculations, such as the demand heat indicator weight of 0.25, the original value of 13%, the normalized bounds of 5% to 20%, and the contribution of 13.33%, clearly define the specific contribution of each indicator to value. Confidence intervals are calculated based on historical assessment data. For example, if the historical mean is 300,000 yuan and the standard deviation is 30,000 yuan, the 95% confidence interval is [241,200, 358,800] yuan, providing a reliable range for the assessment results. This reporting system addresses the problem of traditional assessments that only output a single numerical value. It provides an interpretable and verifiable scientific basis for financial institutions' data management and transaction decision-making, making the assessment results more practical.

[0136] While the specific embodiments of the present invention have been described in detail above, they are merely exemplary, and the present invention is not equivalent to the specific embodiments described above. For those skilled in the art, any equivalent modifications and substitutions of the present invention are also within the scope of the present invention. Therefore, equivalent changes and modifications made without departing from the spirit and scope of the present invention are intended to be encompassed within the scope of the present invention.

Claims

1. An online evaluation method for data resource value, characterized by: The following steps are involved: Step S1: Construct a multi-dimensional evaluation indicator set and collect metadata and market dynamic information of the data to be evaluated in real time through a distributed interface; Step S2: weight the indicator set based on a dynamic weighting algorithm, generate a weight coefficient set, and construct a value calculation model; Step S3: Standardize and pre-process the collected raw data and input it into the model to calculate the initial value estimate; Step S4: Dynamically revise the initial value estimate based on the timeliness attenuation factor and the data scarcity factor to generate a final value estimate and evaluation report.

2. The online data resource value evaluation method according to claim 1, characterized in that: In step S1, the multi-dimensional evaluation index set includes basic attribute indicators, market value indicators and cost efficiency indicators; the basic attribute indicators include data scale, update frequency and structural complexity; The data model balances data size differences based on the logarithmic transformation of the total number of records, and the calculation formula is: M=log 10 (N+1) Among them, M is the data model, N is the total number of data records; The update frequency maps the data timeliness to the inverse of the average update interval, and the calculation formula is: Where F is the update frequency, Δt avg The average update interval of the data source in days; The structural complexity measures the degree of data structuring by a weighted combination of the number of entity associations and the number of dimensions. The calculation formula is: C b =αE+βD (α+β=1) Among them, C b is the structural complexity, E is the total number of relationships between entities in the data model, and D is the number of data dimensions; The market value indicators include demand heat H and competition scarcity S; H=γ1Q s +γ2P t (γ1+γ2=1) Among them, Q s is the search volume growth rate, P t is the transaction price volatility, γ1 and γ2 are adjustment coefficients; Among them, V u is the number of unique features, N c The number of competitors. The cost-effectiveness indicators include storage cost rate, computing cost rate and revenue conversion rate.

3. The online data resource value evaluation method according to claim 2, characterized in that: In step S1, the market dynamic information includes industry search index, policy and regulation change events, and similar data transaction price volatility; the industry search index is obtained by crawling the search volume of search engine keywords in real time, and the policy and regulation change events are extracted through natural language processing to extract keyword change information in government announcements, triggering the update of industry benchmark validity period parameters; the similar data transaction price volatility is calculated by standard deviation through blockchain transaction records, and the formula is: Among them, P t is the transaction price volatility, p i is the i-th transaction price, μ p is the historical average price, n is the number of transactions, when the volatility exceeds the threshold When , the recalculation of demand heat H is triggered; σ p is the standard deviation of historical prices.

4. The online data resource value evaluation method according to claim 3 is characterized by: In step S2, the dynamic weighting algorithm is the improved CRITIC algorithm, and its indicator weight calculation formula is: Among them, w j Represents the weight coefficient of the jth indicator σ j is the standard deviation of index j, μ j is the mean value of index j, ε=1×10 -5 A very small constant to prevent division by zero; r ji is the Pearson correlation coefficient between indicator j and indicator i.

5. The online data resource value evaluation method according to claim 4, characterized in that: In step S3, the standardization preprocessing includes missing value filling, outlier processing and normalization operations; The missing value filling adopts the random forest regression algorithm to predict the missing fields, and its model is expressed as: in, is the predicted result of the i-th missing value, X observed is the observation feature matrix, f RF is a random forest regression model; The outlier processing is based on the 3σ principle to truncate the data that exceeds the mean ± 3 times the standard deviation. The truncation rule is: Where x′ is the truncated data value, μ is the data mean, and σ is the standard deviation; The normalization operation maps the numerical index to the interval [0,1] through the Min-Max algorithm, and the formula is: Among them, x norm is the normalized value, x min is the minimum value in the sliding window, x max is the maximum value in the sliding window.

6. The online data resource value evaluation method according to claim 5, characterized in that: In step S4, the calculation formula of the time-effect attenuation factor is: Among them, θ is the timeliness attenuation factor, t is the data age, T ref is the industry benchmark validity period, and k is the decay rate coefficient.

7. The online data resource value evaluation method according to claim 6, characterized in that: In step S4, the calculation formula of the data scarcity factor is: Where ρ is the data scarcity factor, V u is the number of unique features of the data, V th is the top 10% percentile threshold of the industry, and η=0.1 is the adjustment coefficient.

8. The online data resource value evaluation method according to claim 7, characterized in that: In step S4, the evaluation report includes the final value estimation, indicator contribution analysis and confidence interval; The calculation formula for the terminal value valuation is: Among them, V final is the final valuation, V init is the initial estimate, exp is the natural exponential function, k is the decay rate coefficient, t is the data age, T ref is the industry benchmark validity period, η=0.1 is the adjustment coefficient, V u is the number of unique features of the data, V th The threshold for the top 10% of the industry; The calculation formula of the indicator contribution is: Among them, I j is the contribution of the jth indicator, w j is the weight coefficient of the jth indicator, x j is the original index value, x j,min and x j,max is the normalized boundary value of the jth indicator; The calculation formula of the confidence interval is: Confidence interval = μ v ±1.96σ v Among them, μ v is the historical evaluation mean, σ v is the historical evaluation standard deviation, and 1.96 is the quantile of the standard normal distribution at the 95% confidence level.

Images