Index calculation system for aided analysis of user transactions and monitoring method thereof

Abstract

The invention relates to an index calculation system for assisting in analyzing user transactions and a monitoring method thereof. The index calculation system comprises the following steps: selecting field transaction data serving as current segment dynamic data and technical transaction data serving as historical complete data; according to the selected five core influence indexes, transaction data is subdivided, a classification logic calculation scheme is generated to carry out transaction value calculation on specific data, index calculation is carried out, and various analysis operations of the system are carried out based on a calculation result. The data quality is improved, the possibility that the transaction index deviates from the original data in the calculation processing process is eliminated, the workload of classification screening in subsequent processing is reduced, and the efficiency of the whole preprocessing process is indirectly improved. The method is suitable for reflecting the unique technology development direction and technology cooperation change degree of the technology market and comprehensively reflecting the dynamic development of the technology transaction market.

Description

Technical Field

[0001] The present invention relates to a data analysis technology, and in particular to an index calculation system and a monitoring method thereof for assisting in analyzing user transactions. Background Art

[0002] In the process of technology transactions, accurate evaluation of technology content requires in-depth statistical analysis of historical data and market dynamic monitoring based on this. Given that current statistical analysis methods focus on single quantitative observations, we urgently need to build comprehensive indicators to accurately map the development dynamics of the technology market. In addition, the diversity of technology types requires differentiated evaluation standards, and the ever-changing market demand and supply emphasize the flexibility and timeliness of the evaluation system to ensure the scientificity and foresight of technology transaction decisions, which brings great challenges to the evaluation guidance of the technology market. Summary of the invention

[0003] In view of the non-normal distribution of technology transaction data, data dispersion, poor stability, and difficulty in effective analysis using algorithmic models or conventional statistical methods, an index calculation system and a monitoring method for assisting in the analysis of user transactions are proposed. The system is used to appropriately reflect the unique technology development direction and degree of change of technological cooperation in the technology market, and comprehensively reflect the dynamic development of the technology transaction market.

[0004] The technical solution of the present invention is: an index calculation system for assisting in analyzing user transactions, comprising data receiving modules connected in sequence, receiving transaction data sent by a database;

[0005] The selected module adds the selected data process based on the original data, collects the data according to the filtering mechanism, and obtains stable and accurate data to be sent to the pre-processing module;

[0006] The preprocessing module adds a preprocessing data process to the selected data, classifies the selected data, calls the algorithm in the calculation module, and calculates the data transaction value according to the classification logic calculation scheme; the calculation module, various calculation algorithms, assist the preprocessing module and the operation management module in calculation;

[0007] The operation management module receives the data from the preprocessing module during the calculation process, performs index calculation, and performs system operations based on the calculation results, including storage, processing, management, and monitoring and display.

[0008] Preferably, the transaction data includes on-site transaction data and technical transaction data, and the on-site transaction data includes current segment dynamic data in formal listing, intended listing, and non-public agreement transaction modes; the technical transaction data covers complete historical data on technical transactions in a wide range of technical fields and national economic industries.

[0009] Preferably, the filtering mechanism of the on-site transaction data in the selected module is: first filter the data of "issued transaction vouchers" as valid transaction data; and set only the on-site transaction data with a single amount ranking percentile between [5%, 95%] to be retained each year.

[0010] Preferably, the filtered on-site transaction data in the pre-processing module is classified according to date and technical field, and the classification logic calculation scheme is to calculate the average value of each technical field as the transaction value-sent to the on-site operation management module.

[0011] Preferably, the filtering mechanism of the technical transaction data in the selected module includes annual data deviation filtering and daily data extreme value filtering:

[0012] Annual data deviation filtering: First, implement annual transaction amount unit ranking filtering, sort the technology transaction amount by unit, and after sorting, calculate the transaction amount difference between each unit and the previous unit. If the difference between the amount of the unit and the previous unit exceeds the set threshold, mark the unit as a "significant growth difference unit". Then, among all the units marked as "significant growth difference units", select the unit with the highest amount in the current year as the benchmark, and calculate the ranking percentage of the highest ranking unit in all units; then traverse all units, and remove the unit data with a ranking percentage higher than the ranking percentage of the highest ranking unit from the analysis set, capturing and reflecting the normal value distribution characteristics of the vast majority of ordinary units in the technology transaction market; secondly, use the percentile method to select technology transaction data between [5%, 95%] for the remaining data to eliminate the improper influence of extreme values ​​on the overall average;

[0013] Daily data extreme value filtering: After annual data deviation filtering, the 3σ de-extreme value method is used to perform refined extreme value filtering on daily data. Relying on the stable 3σ principle in statistics, 99.73% of the observations in the data set will fall within the range of plus or minus three standard deviations of the mean.

[0014] Preferably, the filtered technical transaction data in the selected module enters the preprocessing module to simultaneously perform classification standardization classification logic calculation as transaction value-classification, and single-dimension standardization classification logic calculation as transaction value-single dimension, and according to the classification standardization classification and single-dimension standardization classification logic calculation results, the average value of technical transactions per day as transaction value-transaction is calculated and sent to the operation management module; classification standardization classification logic calculation:

[0015] Step ①: Calculate the average transaction value based on the five dimensions of transaction data: date, technology field, national economic industry, subject type, and contract type;

[0016] Step ②: If a valid value cannot be obtained in step ① due to missing or insufficient data, the next best option is to calculate the average value based on only the four dimensions of date, technical field, national economic industry and entity type; Step ③: If step ② cannot be implemented, that is, the first two steps cannot produce valid results due to insufficient data, the calculation dimensions are further simplified, and the average value is calculated based on only the three dimensions of date, technical field and national economic industry;

[0017] Transaction value-classification result selection logic: If step ① can produce a valid value, the result of step ① will be used first; if step ① cannot be implemented, the result of step ② will be used instead; if step ② is also not feasible, the result of step ③ will be used in the end, and finally the transaction value-classification will be obtained. The progressive selection logic ensures the maximum utilization of data;

[0018] Single-dimensional standardized classification logic calculation:

[0019] Step ④: Calculate the average value based on the date and technology field of the transaction data;

[0020] Step ⑤: Calculate the average value based on the date of transaction data and the dual dimensions of national economic industry;

[0021] Step ⑥: Calculate the average value based on the dual dimensions of transaction data date and entity type;

[0022] Step 7: Calculate the average value based on the date and contract type of the transaction data;

[0023] Step ⑧: Calculate the average value based on the date dimension of the transaction data;

[0024] Transaction value - single dimension result calculation logic: Transaction value - single dimension = [(④+⑥+⑦) / 3+(⑤+⑥+⑦) / 3] / 2;

[0025] Average value of technology trades per day:

[0026] If transaction value-classification produces a valid value, this result shall be used first; if transaction value-classification cannot be implemented, the transaction value-single dimension result shall be used instead; if transaction value-single dimension is also not feasible, the result of step ⑧ shall be finally used.

[0027] Preferably, the operation management module receives transaction value-on-site and transaction value-transaction calculation result data,

[0028] The following logic is used to select the total average transaction value as the transaction value: when the transaction value-market is not empty, the weighted average method is used to calculate the transaction value, transaction value = (transaction value-market + transaction value-transaction) / 2, which fully reflects the true situation of the market or transaction; when the transaction value-market is empty, the constant substitution method is used, transaction value = transaction value-transaction;

[0029] The transaction value on the starting date of the transaction is used as the benchmark value, and the transaction value on subsequent dates is compared with it and enlarged or reduced proportionally to calculate the corresponding index value.

[0030] A method for monitoring an index calculation system for assisting in analyzing user transactions, utilizing the index calculation system for assisting in analyzing user transactions as described in claims 1 to 7, receiving market transaction data sent by a technology transaction database, and monitoring dynamic changes in technology transaction market data.

[0031] A method for calculating a technology transaction index comprises sending the technology field classification, national economic industry classification, subject type classification and contract type classification to which the technology transaction belongs, as well as the on-site transaction data of the technology transaction into the index calculation system for assisting in analyzing user transactions as described in claim 7, to obtain the transaction value and index of the technology transaction.

[0032] A patent value assessment method, wherein the patent number, the patent's IPC classification number, the national economic industry classification number, the application date, and the current status field are input into the index calculation system for assisting in analyzing user transactions as described in claim 7, and the transaction value obtained is the patent value valuation.

[0033] The beneficial effects of the present invention are as follows: the index calculation system and monitoring method of the present invention for assisting in analyzing user transactions improves the quality of data, eliminates the possibility of deviation of the transaction index from the original data during the calculation and processing process, reduces the workload of classification and screening in subsequent processing, and indirectly improves the efficiency of the entire preprocessing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flow chart of the index calculation and monitoring method for assisting in analyzing user transactions of the present invention;

[0035] Figure 2 This is a rating distribution diagram of an embodiment of the present invention;

[0036] Figure 3 This is a distribution diagram of the technical transaction value of the embodiment of the present invention;

[0037] Figure 4 This is a technology maturity distribution diagram of the embodiments of the present invention;

[0038] Figure 5This is a classification distribution diagram of technical topics of the embodiments of the present invention. DETAILED DESCRIPTION

[0039] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0040] An index calculation system for assisting in analyzing user transactions, including sequentially connected

[0041] A data receiving module receives transaction data sent by the database;

[0042] The selected module adds the selected data process based on the original data, collects the data according to the filtering mechanism, and obtains stable and accurate data to be sent to the pre-processing module;

[0043] The preprocessing module adds a preprocessing data process to the selected data, classifies the selected data, calls the algorithm in the calculation module, and calculates the transaction value according to the logical calculation plan;

[0044] Calculation module, various calculation algorithms, assisting the pre-processing module and operation management module in calculation;

[0045] The operation management module receives the data from the preprocessing module during the calculation process, performs index calculation, and performs system operations based on the calculation results, including storage, processing, management, and monitoring and display.

[0046] The selected modules are first grouped according to different "subject classifications" based on the technology name field data; then they are filtered based on the "transaction amount" field data and the amount value, which solves the classification filtering and screening of technology names and subjects. However, due to the non-normal characteristics of the distribution of technology transaction data, ranging from tens of thousands to hundreds of millions of yuan, there is a huge difference between the mean and the median. Directly using the average is likely to cause data distortion. Therefore, it is necessary to introduce a more sophisticated extreme value filtering mechanism, such as the percentile-based truncation method or robust statistical method, to scientifically and rationally eliminate extreme values, ensure that the technology transaction index can more truly reflect market dynamics, and provide solid data support for technology transaction value assessment and technology transfer strategies. Therefore, it is further improved.

[0047] The preprocessing module first calculates the total index by comprehensive method according to different field classifications, and then calculates the total index by average method according to different subject classifications, which solves the classification and calculation of transaction amounts for different technology fields and subject types. However, it is difficult to fully capture the deep structure of technology transaction value by classification based on only two dimensions. Given that the value distribution of technology transactions often spans multiple complex factors, and the direct correlation with technology fields and subject types is not always significant, a more refined analysis framework is proposed to be introduced. By integrating relevant indicators such as the activity of national economic industries and the complexity of cooperation networks, a multi-dimensional evaluation system is intended to be constructed, in order to more accurately reveal the inherent logic and distribution patterns of technology transaction value, and provide decision makers with richer and deeper insights, and therefore further improvements are made.

[0048] The operation management module further improves the data processing segmentation and trend, so in addition to obtaining index calculation results, it can also obtain various data change trends for analysis reference.

[0049] When constructing a technology transaction index calculation system, we first need to select effective influencing indicators from a large amount of transaction information. We selected 52 positioning fields that are sensitive to technology transactions, used models such as random forests and decision trees, and combined with the AUC evaluation algorithm to conduct in-depth mining and fine screening of 400,000 sample data. By eliminating null values ​​and significantly irrelevant fields, such as technology names, numbers, and names of buyers and sellers, the data dimension is effectively reduced, the purity and concentration of data information are improved, the sample data volume is reduced to 200,000, and the positioning field volume is reduced to 35. After repeated testing of indicators, five core influencing indicators were finally locked in: date, technology field, national economic industry, subject type, and contract type. The AUC value of these five indicators reached 0.876, which is in the excellent range of [0.85,0.95]. It is a very effective classifier, and the selected five core influencing indicators are obtained.

[0050] like Figure 1 As shown, the technology transaction index is generally calculated and generated by two parts, one is the on-site transaction data (current dynamic data), and the other is the technology transaction data (historical complete data). According to the selected five core impact indicators, the transaction data is subdivided. The two types of transaction data include 11 categories of technology fields, 20 categories of national economic industries, 13 categories of subject types, and 5 categories of contract types.

[0051] 1. On-site transaction data processing

[0052] 1.1 Filtering of on-site transaction data

[0053] The data receiving module receives the on-site transaction data, which includes formal listing, intended listing, non-public agreement transactions and other transaction methods. In order to ensure the value accuracy and representativeness of the on-site transaction data, the data of "issued transaction vouchers" are first filtered in the selected module as valid transaction data; in addition, according to the statistics of the original data, the overall average transaction amount is 13.7225 million, while the median is only 1.5 million. The difference between the two is large, indicating that the data with higher transaction amounts has a great impact on the overall average. In order to balance the gap between the average and the median, the percentile method is used to remove the data that has a greater impact on the overall average. Generally, the percentile is used to select data points less than the 5th percentile or greater than the 95th percentile as abnormal values. Therefore, in the selected module, it is set to retain only the on-site transaction data with a single amount ranking percentile between [5%, 95%] each year. Through the filtering of the percentile method, the overall average transaction amount of the filtered data is reduced to 4.6575 million, and the median is increased to 1.7611 million. The main purpose of this step is to eliminate some technical trading data that are far below or above the average level, so as to control the overall on-site technical trading value at a stable level.

[0054] 1.2 Calculating the average on-site transaction value

[0055] Based on the filtered data set in 1.1, the preprocessing module is used to classify the data by date and technical field, and then the algorithm in the calculation module is used to calculate the average value of each technical field to obtain the steps Transaction value - operation management module is provided on site.

[0056] 2. Technical transaction data processing

[0057] The technology transaction data is received through the data receiving module. When processing the technology transaction data, given the huge amount and complexity of the data, which not only covers a wide range of technology fields and national economic industries, but also involves the uneven distribution characteristics of transaction activities, especially the high-value transaction records that may be generated by some special large enterprises or higher education institutions, whose single transaction amount can easily exceed 100 million yuan. According to the original data statistics, the overall average transaction amount is 5.4301 million, while the median is only 200,000. The gap between the two is large, and the gap with the on-site transaction data is also large. If such extreme values ​​are directly included in the statistical scope, it will significantly distort the authenticity of the average value of technology transactions, and it is difficult to accurately reflect the value distribution of ordinary units in technology transaction activities.

[0058] Therefore, in order to ensure the representativeness and accuracy of technology transaction data analysis, it is necessary to filter the data as a whole in the selected module and then enter the preprocessing module for classification standardization and single-dimensional standardized data processing.

[0059] 2.1 Overall filtering of technology transaction data

[0060] The overall filtering first involves filtering for annual data deviations, which includes a dual screening mechanism: First, filtering by annual transaction volume unit ranking. According to statistics, the number of units that conduct transactions each year is not fixed, and there is a large gap between different years, ranging from 2,000 to 8,000, and the transaction amount of each unit also varies greatly. Therefore, a structured approach is used to identify and eliminate units that show significant growth differences compared to their direct predecessors: by year, the technology transaction amount is sorted by unit, and for each unit, the difference in transaction amount between it and the previous unit (sorted by amount) is calculated, and the threshold is set to 50% of the unit amount. If the difference in amount between the unit and the previous unit exceeds this threshold, the unit is marked as a "significant growth difference unit" (for example, the unit's annual transaction amount is 1 billion yuan, ranking 5th among all units, and the threshold is 10*50%=5 1 billion yuan, while the annual transaction amount of the previous unit (ranked 4) is 4 billion yuan. The difference between the two is 3 billion yuan, which exceeds the threshold, and the unit is marked as a "significant growth difference unit"); then, among all the units marked as "significant growth difference units", the unit with the highest amount in the current year is selected as the benchmark (i.e., the highest ranking unit), and the ranking percentage of the highest ranking unit in all units is calculated (for example, if the unit ranks 5th and there are 100 units in total, the ranking percentage is 5%); then all units are traversed, and the unit data with a ranking percentage higher than the ranking percentage of the highest ranking unit are removed from the analysis set (for example, the highest "significant growth difference unit" ranking percentage is 5%, and the technology transaction data of all the top 5% units are removed). This step is intended to retain those units that are closer to the top or have higher importance in the overall ranking than the top units of "significant growth differences". Secondly, the remaining data is also selected using the percentile method to select technology transaction data between [5%, 95%] to eliminate the improper influence of extreme values ​​on the overall average. This measure aims to eliminate data bias caused by a small number of high-value transaction entities, so as to more accurately capture and reflect the normal value distribution characteristics of the vast majority of ordinary units in the technology trading market.

[0061] Secondly, daily data extreme value filtering is performed. In view of the results of technical transaction data deviation filtering, further refinement is still needed to approach the normal value distribution, and a more sophisticated preprocessing strategy is adopted. In value assessment, the percentile method is more suitable for filtering the annual extreme price points in the entire data set, regardless of whether these extreme values ​​are caused by abnormal market fluctuations, data entry errors, or other atypical factors. After using the percentile method to initially filter out the most extreme observations, for the remaining data set, the daily prices basically conform to the characteristics of normal distribution or approximate normal distribution. The 3σ de-extreme method can be further applied for refined extreme value filtering, which is more suitable for identifying and filtering those price extremes that significantly deviate from the normal fluctuation range in different market scenarios. This method relies on the stable 3σ principle in statistics, that is, about 99.73% of the observations in the data set will fall within the range of plus or minus three standard deviations (μ±3σ) of the mean. Therefore, the percentile method can be used to initially filter out the most extreme observations of the year to reduce the impact on subsequent analysis; then, for the remaining data set, the 3σ de-extreme method is further applied to perform refined extreme value filtering on the daily data. Such a comprehensive application strategy can more comprehensively identify and process outliers in technology transaction price data and improve the accuracy and reliability of analysis.

[0062] 2.2 Classification Standardization Preprocessing

[0063] 2.2.1 Classification and Standardization

[0064] In order to improve the accuracy and applicability of data processing, a classification standardization preprocessing process was implemented. This process subdivides the data into three classification systems for standardization based on the specific needs of subsequent analysis: first, classification by date, technical field, national economic industry, subject type and contract type; second, simplifying the classification dimensions, classifying by date, technical field, national economic industry and subject type when data is not available in the first one; third, further simplifying, classifying by date, technical field and national economic industry when data is not available in the second one. This classification strategy aims to ensure that the calculation and analysis at each level are based on a highly representative and approximately normally distributed data set, thereby optimizing the overall data processing process and laying a solid foundation for subsequent accurate calculations.

[0065] 2.2.2 Calculation of the average transaction value after classification standardization

[0066] Calculate the initial average value

[0067] Step ①: First, calculate the average transaction value under the five dimensions of transaction data, namely, date, technology field, national economic industry, subject type, and contract type. This step ensures the precision and accuracy of the calculation results because as many influencing factors as possible are taken into account.

[0068] Step ②: If a valid value cannot be obtained in step ① due to missing or insufficient data, the next best option is to calculate the average value based on only the four dimensions of date, technical field, national economic industry and entity type. This step is used as an alternative to increase the flexibility and adaptability of the calculation process.

[0069] Step ③: If step ② cannot be implemented, that is, the first two steps cannot produce valid results due to insufficient data, the calculation dimensions will be further simplified, and the average value will be calculated based on only three dimensions: date, technology field, and national economic industry. This step is the last resort, aiming to ensure that certain calculation results can still be obtained in extreme cases.

[0070] 2.2.3 Obtaining Transaction Value - Classification

[0071] After the calculation of the above average value is completed, the transaction value is classified. This step is marked as "Transaction Value-Classification". The result is selected based on the calculated average value and then used for subsequent calculations.

[0072] Step ⑨: Result selection logic: Based on the availability of the calculation results, the following logic is used to select the average transaction value of the classification: If step ① can produce a valid value, the result of step ① is preferred; if step ① cannot be implemented, the result of step ② is used instead; if step ② is also not feasible, the result of step ③ is finally used, and finally the transaction value-classification of step ⑨ is obtained. This progressive selection logic ensures the maximum utilization of data while ensuring the rationality and reliability of the results. 2.3 Single-dimensional standardization preprocessing

[0073] Since technology transaction data often changes according to the supply and demand relationship in the market, its attributes may only exist in one category. This situation cannot be ignored. Therefore, it is necessary to perform single-dimensional standardization preprocessing simultaneously and adopt another preprocessing strategy: single-dimensional standardization.

[0074] 2.3.1 Single Dimension Standardized Classification

[0075] The results of the overall filtering of technology transaction data in 2.1 are subdivided into five classification systems for single-dimensional standardization: date, technology field; date, national economic industry; date, subject type; date, contract type; date only. Since technology transactions are relatively special, the transaction volume and transaction value are closely related to market supply and demand, development trends, registration requirements, etc. The single-dimensional standardization preprocessing is intended to make up for the situation where the three classification standardization preprocessings in 2.2 cannot obtain result values. If the above classification methods still have no results, the final calculation is based on the date to try to ensure that there are result values ​​every day.

[0076] 2.3.2 Calculating the Average Transaction Value after Single Dimension Standardization

[0077] Calculate the initial average value

[0078] Step ④: Calculate the average value based on the transaction data’s date and technology field.

[0079] Step ⑤: Calculate the average value based on the dual dimensions of transaction data date and national economic industry.

[0080] Step ⑥: Calculate the average value based on the dual dimensions of transaction data: date and entity type.

[0081] Step 7: Calculate the average value based on the dual dimensions of transaction data: date and contract type.

[0082] Step ⑧: Calculate the average value based on the date dimension of the transaction data.

[0083] 2.3.3 Obtaining Transaction Value - Single Dimension

[0084] After completing the calculation of the above average value, the single dimension processing of the transaction value is performed. This step is marked as "transaction value-single dimension". The average value is obtained according to the calculation and used for subsequent calculations.

[0085] Step ⑩: Result calculation logic: transaction value - single dimension = [(④+⑥+⑦) / 3+(⑤+⑥+⑦) / 3] / 2

[0086] Since there is a large gap between the average contract values ​​of the two dimensions of subject type and contract type, which has a great impact on the calculation of the overall transaction value, in order to weaken the impact of these two dimensions and at the same time enhance the impact of the two dimensions of technology field and national economic industry on the overall price, specific data adjustments are made. The two dimensions of subject type and contract type are calculated twice, and then these values ​​are added and divided by the corresponding coefficients (3 and 2) to obtain the single-dimensional result of the average transaction value, and try to ensure that the result value is generated every day.

[0087] 2.4 Calculating the Average Technology Transaction Value

[0088] step Result selection logic: Based on the "Transaction Value - Classification" and "Transaction Value - Single Dimension" results obtained in 2.2.3 and 2.3.3, the following logic is used to select the average transaction value of technology transactions: If step ⑨ can produce a valid value, the result of step ⑨ is preferred; if step ⑨ cannot be implemented, the result of step ⑩ is adopted instead; if step ⑩ is also not feasible, the result of step ⑧ is finally adopted, and the average value of technology transactions per day is finally obtained, which is step Transaction Value - Transaction.

[0089] 2.5 Calculating the average transaction value

[0090] step Result selection logic: based on and The following logic is used to select the total average transaction value for the "Transaction Value - On-site" and "Transaction Value - Transaction" results obtained in step When not empty, the transaction value is calculated using the weighted average method. The purpose is to more fully reflect the true situation of the market or transaction; when the steps When it is empty, constant substitution method is used. The calculation process is simplified, avoiding the calculation obstacles caused by missing variables; when the steps and When both are empty, the transaction value is determined by averaging the value of all transactions on that day - the value of the transaction, ensuring that a result value can be obtained even in the most unfavorable circumstances as a note reference.

[0091] Preprocessing classifies the data and forms a logical calculation plan. Logical calculation is implemented by calling the algorithm in the calculation module, and the results of the calculation process are sent to the operation management module for storage.

[0092] 2.6 Generating Technical Trading Index

[0093] The calculation results in the above process are received and stored in the operation management module. The transaction value on the start date of the transaction is used as the benchmark value (i.e. the base period index is set to 100). The transaction value on subsequent dates is compared with it and enlarged or reduced in proportion to calculate the corresponding index value. This method ensures that the trend of transaction value changes can be quantified and intuitively displayed, providing an effective analysis tool for market participants.

[0094] Embodiment 1:

[0095] Calculation and verification of technology transaction index in the field of "electronic information" technology

[0096] First, we selected original on-site transaction data and technical transaction data based on the characteristic keywords in the technical field (e.g., the technical name contains relevant characters such as "information", "software", "communication", "computer", "network", "blockchain", "artificial intelligence", "cloud computing", etc.; or the relevant technical documents submitted contain the "computer software copyright" certificate; or the original data in the technical field marked as "electronic information"). At this time, we obtained 35,279 on-site transaction data related to the "electronic information" technical field and 182,638 technical transaction data.

[0097] Table 1 Sample data

[0098]

[0099] Next, the technical trading index calculation logic is used to calculate the data. The on-site trading data of "issued trading certificates" is filtered, and the outliers are removed using the percentile method. At this time, 33,871 filtered on-site trading data are obtained. Then the average on-site trading value is calculated according to the date and technical field.

[0100] Table 2 Results of on-site transaction value

[0101] date Technical Field Transaction Value - On-site 2024-06-06 Electronic Information * 2024-06-06 Advanced Manufacturing * 2024-06-06 Environmental protection and comprehensive utilization of resources *

[0102] We then implement annual transaction volume unit ranking filtering and percentile filtering. Based on the calculation of "significant growth difference units", we filter a batch of units each year. The filtering ranking percentage is basically in the range of 2%-5%. The number of filtered units is about 15 each year. After using the percentile method to eliminate outliers, we get 171,023 pieces of technology transaction data after filtering.

[0103] Then, the 3σ de-extreme value method was used to preprocess the classification standardization and single-dimensional standardization. The preprocessed technology transaction data were 167,092 and 162,903 respectively. Then, the average classification transaction value and the average single-dimensional transaction value were calculated according to the classification standardization and single-dimensional standardization.

[0104] Table 3. Results of classification and single-dimensional transaction value

[0105]

[0106] The average technology transaction value is then calculated.

[0107] Table 4 Transaction value-transaction results

[0108]

[0109] Finally, the average transaction value and index are calculated.

[0110] Table 5 Transaction value results

[0111]

[0112] Verify the source data based on the completion of the technology transaction value calculation.

[0113] We randomly extracted 10 data with the same date but in different technical fields and national economic industries for reverse verification and calculated the deviation rate. The basic results fell within the range of [-50%, 50%], which met the requirements of the deviation range.

[0114] Table 6 Verification results

[0115] Transaction ID Transaction Price Predicting Prices Deviation rate 2024****** 157.63 175.29 11.20％ 2024****** 99.43 85.62 -13.89％ 2024****** 92.06 113.43 23.21％

[0116] Example 2: Patent value assessment application

[0117] First, based on the calculation of the technical trading index, the obtained data is presented in tabular form.

[0118] Table 7 Example of technical trading index

[0119]

[0120] Next, based on the obtained patent number, obtain the patent's IPC classification number, national economic industry classification number, application date, current status and other fields.

[0121] Table 8 Patent Examples

[0122]

[0123]

[0124] Then, according to the comparison table, convert the IPC classification number and the national economic industry classification number into the corresponding technical field or national economic industry. If there is a complete one-to-one correspondence, match directly; if a completely consistent result cannot be matched, remove the content after the " / " in the IPC classification number and match again (for example, if the IPC number is "B43L7 / 14" and there is no matching result, select "B43L7" to match again); if there is still no matching result, only the first three digits are matched, and the first three digits must have a result (for example, if "B43L7" still has no matching result, select "B43" to match).

[0125] Table 9 IPC classification code conversion example

[0126] IPC Classification Number IPC Name Corresponding national economic sectors A01G16 Rice cultivation (A01G9 / 00 priority)〔3〕 Agriculture, forestry, animal husbandry and fishery C12N15 / 05 Plant cells〔5〕 Agriculture, forestry, animal husbandry and fishery B02C4 / 40 Remover, e.g. scraper manufacturing B02C7 / 16 Driving mechanism manufacturing

[0127] Table 10 Example of conversion of national economic industry classification code

[0128]

[0129]

[0130] Then, technical keywords are extracted based on the patent name or abstract, and the technical keywords are converted into technical field references.

[0131] Table 11 Technical keyword conversion examples

[0132] Technical Keywords Corresponding technical field medicine Biology, medicine and medical devices nuclear Nuclear Applications plastic New materials and their applications software Electronic Information

[0133] After comparing the IPC classification number, national economic industry classification number, and technical keywords, the patent application date is matched with the technology transaction index date, and the technical field and national economic industry are matched respectively. The average value after all the matches is calculated to obtain the patent value assessment. Finally, the value score and rating are formed based on the result values ​​of similar patents (patents in the same corresponding technical field or national economic industry), and the prospect description and other contents are generated.

[0134] Table 12 Patent value assessment results

[0135]

[0136]

[0137] Finally, the statistical display of all patents is realized. Figure 2 Rating distribution shown, Figure 3 The distribution of technology transaction values ​​shown, Figure 4 The technology maturity distribution shown and Figure 5 The classification distribution of technical topics shown.

[0138] This technical trading index calculation method achieves accurate calculation of transaction value under different circumstances through flexible conditional judgment and precise calculation formula, providing strong support for decision-making in financial markets, commodity trading and other fields.

[0139] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. An index calculation system for assisting in analyzing user transactions, characterized in that: It includes data receiving modules connected in sequence to receive transaction data sent by the database; The selected module adds the selected data process based on the original data, collects the data according to the filtering mechanism, and obtains stable and accurate data to send to the pre-processing module; The preprocessing module adds a preprocessing data process to the selected data, classifies the selected data, calls the algorithm in the calculation module, and calculates the data transaction value according to the classification logic calculation scheme; Calculation module, various calculation algorithms, assisting the pre-processing module and operation management module in calculation; The operation management module receives the data from the preprocessing module during the calculation process, performs index calculation, and performs system operations based on the calculation results, including storage, processing, management, and monitoring and display.

2. The index calculation system for assisting in analyzing user transactions according to claim 1, characterized in that: The transaction data includes on-site transaction data and technical transaction data. The on-site transaction data contains the current dynamic data of formal listing, intended listing, and non-public agreement transaction methods; the technical transaction data covers a wide range of technical fields and complete historical data of technical transactions in national economic industries.

3. The index calculation system for assisting in analyzing user transactions according to claim 2, characterized in that: The filtering mechanism of the on-site transaction data in the selected module is: first filter the data of "issued transaction vouchers" as valid transaction data; and set only the on-site transaction data with a single amount ranking percentile between [5%, 95%] to be retained each year.

4. The index calculation system for assisting in analyzing user transactions according to claim 3, characterized in that: The filtered on-site transaction data in the pre-processing module is classified according to date and technical field. The classification logic calculation scheme is to calculate the average value of each technical field as the transaction value - sent to the on-site operation management module.

5. The index calculation system for assisting in analyzing user transactions according to claim 4, characterized in that: The filtering mechanism of the technical transaction data in the selected module includes annual data deviation filtering and daily data extreme value filtering: Annual data deviation filtering: First, implement annual transaction amount unit ranking filtering, sort the technology transaction amount by unit, and after sorting, calculate the transaction amount difference between each unit and the previous unit. If the amount difference between the unit and the previous unit exceeds the set threshold, the unit is marked as a "significant growth difference unit". Then, among all the units marked as "significant growth difference units", select the unit with the highest amount in the current year as the benchmark, and calculate the ranking percentage of the highest-ranked unit in all units; Then, we traverse all units and remove the unit data whose ranking percentage is higher than the highest ranking unit from the analysis set, so as to capture and reflect the normal value distribution characteristics of the vast majority of ordinary units in the technology trading market; secondly, we use the percentile method to select technology trading data between [5%, 95%] for the remaining data to eliminate the improper influence of extreme values ​​on the overall average; Daily data extreme value filtering: After annual data deviation filtering, the 3σ de-extreme value method is used to perform refined extreme value filtering on daily data. Relying on the stable 3σ principle in statistics, 99.73% of the observations in the data set will fall within the range of plus or minus three standard deviations of the mean.

6. The index calculation system for assisting in analyzing user transactions according to claim 5, characterized in that: The filtered technical transaction data in the selected module enters the preprocessing module to simultaneously perform classification standardization classification logic calculation as transaction value-classification, and single-dimension standardization classification logic calculation as transaction value-single dimension. According to the classification standardization classification and single-dimension standardization classification logic calculation results, the average value of technical transactions per day as transaction value-transaction is calculated and sent to the operation management module; Classification standardization classification logic calculation: Step ①: Calculate the average transaction value based on the five dimensions of transaction data: date, technology field, national economic industry, subject type, and contract type; Step ②: If a valid value cannot be obtained in step ① due to missing or insufficient data, the next best option is to calculate the average value based on only the four dimensions of date, technical field, national economic industry and entity type; Step ③: If step ② cannot be implemented, that is, the first two steps cannot produce valid results due to insufficient data, the calculation dimensions are further simplified, and the average value is calculated based on only the three dimensions of date, technical field and national economic industry; Transaction value-classification result selection logic: If step ① can produce a valid value, the result of step ① will be used first; if step ① cannot be implemented, the result of step ② will be used instead; if step ② is also not feasible, the result of step ③ will be used in the end, and finally the transaction value-classification will be obtained. The progressive selection logic ensures the maximum utilization of data; Single-dimensional standardized classification logic calculation: Step ④: Calculate the average value based on the date and technology field of the transaction data; Step ⑤: Calculate the average value based on the date of transaction data and the dual dimensions of national economic industry; Step ⑥: Calculate the average value based on the dual dimensions of transaction data date and entity type; Step 7: Calculate the average value based on the date and contract type of the transaction data; Step ⑧: Calculate the average value based on the date dimension of the transaction data; Transaction value - single dimension result calculation logic: Transaction value - single dimension = [(④+⑥+⑦) / 3+(⑤+⑥+⑦) / 3] / 2; Average value of technology trades per day: If transaction value-classification produces a valid value, this result shall be used first; if transaction value-classification cannot be implemented, the transaction value-single dimension result shall be used instead; if transaction value-single dimension is also not feasible, the result of step ⑧ shall be finally used.

7. The index calculation system for assisting in analyzing user transactions according to claim 6, characterized in that: The operation management module receives the transaction value-on-site and transaction value-transaction calculation result data, The following logic is used to select the total average transaction value as the transaction value: when the transaction value-market is not empty, the weighted average method is used to calculate the transaction value, transaction value = (transaction value-market + transaction value-transaction) / 2, which fully reflects the true situation of the market or transaction; when the transaction value-market is empty, the constant substitution method is used, transaction value = transaction value-transaction; The transaction value on the starting date of the transaction is used as the benchmark value, and the transaction value on subsequent dates is compared with it and enlarged or reduced proportionally to calculate the corresponding index value.

8. A method for monitoring an index calculation system for assisting in analyzing user transactions, characterized in that: The index calculation system for assisting in analyzing user transactions as described in claims 1 to 7 is used to receive market transaction data sent by a technology transaction database and monitor dynamic changes in technology transaction market data.

9. A method for calculating a technical trading index, characterized in that: The technical field classification, national economic industry classification, subject type classification and contract type classification to which the technology transaction belongs, as well as the on-site transaction data of the technology transaction are sent to the index calculation system for assisting in analyzing user transactions as described in claim 7 to obtain the transaction value and index of the technology transaction.

10. A patent value assessment method, characterized in that: The patent number, IPC classification number of the patent, national economic industry classification number, application date, and current status fields are input into the index calculation system for assisting in analyzing user transactions as described in claim 7, and the transaction value obtained is the patent value valuation.

Images