Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a data processing system suitable for accounting finance management, which solves the efficiency problem possibly encountered by the traditional data processing system in the accounting finance management field when processing a large amount of accounting finance data, and solves the problems of inaccurate analysis result, incomplete information and untimely feedback caused by incomplete or inaccurate data.
The data processing system comprises a task acquisition module, a real-time data screening module, a historical data screening module, a related data calling module, a data integration feedback module and a data storage and management module, wherein the task acquisition module is used for being in communication connection with a user end to acquire an accounting financial management demand task of the user end, the real-time data screening module is used for calling a real-time data stream stored in the data storage and management module and screening the real-time data stream according to correlation matching information between the real-time data stream and the accounting financial management demand task to obtain task primary data, the historical data screening module is used for searching historical call data of the accounting financial management demand task and screening historical call data of the accounting financial management demand task according to timeliness and usability of the historical call data to obtain task secondary data, the related data calling module is used for calling task primary data, task secondary data and task tertiary data based on the degree of correlation with the task secondary data, the data integration feedback module is used for integrating the task primary data, the task secondary data and the task tertiary data to obtain task demand data set, and further, the task primary data set is used for storing the task primary demand data and the task data and the financial management demand data, and the historical call data are provided for the task primary data and the task management demand data and the task primary data and the task data and the financial management demand data and the data.
Further, the accounting finance management demand task specifically comprises a data analysis report task, a risk assessment prediction task, a budget cost control task, a tax planning task, a business decision support task and a user credit assessment task.
Further, the associated matching information specifically comprises data field matching degree, data type format matching degree and service logic matching degree.
Further, the real-time data stream is filtered to obtain specific analysis of task level data, wherein the specific analysis comprises the steps of obtaining data fields, data type formats and business logic rules required by an accounting financial management demand task, cleaning the real-time data stream, removing invalid, repeated and abnormal data, further carrying out conversion normalization on the data formats, matching the data fields of each data in the real-time data stream with the data fields required by the accounting financial management demand task to obtain the data field matching degree of each data in the real-time data stream, screening each data based on the comparison of the data field matching threshold of the accounting financial management demand task with the data field matching degree of each data, screening data with the data type matching degree smaller than the data field matching threshold, matching the type formats of each data in the real-time data stream after the screening data field matching degree is smaller than the data field matching threshold with the data type formats required by the accounting financial management demand task to obtain the data type format matching degree of each data, screening the data type formats of each data in the real-time data stream after the screening data field matching degree is smaller than the data type matching threshold of the accounting financial management demand task with the data type format matching degree required by the business logic rules, matching the data type of each data type in the real-time data stream is smaller than the data type matching degree required by the data type matching threshold of the accounting financial management demand task, and screening the data with the service logic matching degree smaller than the service logic matching threshold based on the comparison of the service logic matching threshold of the accounting financial management demand task and the service logic matching degree of each data, and marking the real-time data stream after the data with the service logic matching degree smaller than the service logic matching threshold is finally screened as task first-level data.
Further, historical call data of the accounting financial management demand task is searched, the historical call data of the accounting financial management demand task is screened according to timeliness and usability of the historical call data, and specific analysis of task second-level data is obtained, wherein the historical call data of the accounting financial management demand task is searched based on a data storage and management module, the historical call data are sorted according to time, further historical call data falling in a time window are screened out according to a set time window, usability information is further identified according to the screened historical call data falling in the time window, the usability information specifically comprises data abnormal marking conditions, data security attack conditions and data source credibility conditions, and the historical call data which is identified as the task second-level data without data abnormal marking conditions or data security attack conditions or data source credibility conditions is screened out.
The method comprises the steps of setting a time window, setting a time window label for the time window of different sets according to the time length, dividing the time window label into different sets according to the size, setting a characteristic evaluation value label for the characteristic evaluation value of different sets according to the size, wherein the characteristic evaluation value label comprises a first characteristic evaluation value, a second characteristic evaluation value and a third characteristic evaluation value, the historical call data corresponding to the first characteristic evaluation value is matched with the first time window, the historical call data corresponding to the second characteristic evaluation value is matched with the second time window, and the third characteristic evaluation value is matched with the call history data corresponding to the third characteristic evaluation window.
The method comprises the steps of acquiring correlation information between the rest data of task secondary data and each task secondary data, excluding the rest data of the task secondary data, and the co-occurrence times, wherein the correlation information comprises covariance absolute values and co-occurrence times, averaging the covariance absolute values of the rest data and each task secondary data to obtain covariance absolute value average values of the rest data and task secondary data sets, averaging the co-occurrence times of the rest data and each task secondary data to obtain co-occurrence times average values of the rest data and the task secondary data sets, acquiring covariance absolute value threshold values and co-occurrence times threshold values, comparing the covariance absolute value average values and the co-occurrence times average values of the rest data and the task secondary data sets with corresponding covariance absolute value threshold values and co-occurrence times threshold values, and screening out the rest data with the covariance absolute value average values larger than the covariance absolute value threshold values and the co-occurrence times average values larger than the co-occurrence times threshold values as task tertiary data.
The invention has the following beneficial effects:
the data processing system suitable for accounting financial management can be used for rapidly matching and screening out data meeting task requirements through a real-time data screening and historical data screening module, reducing time for manual screening and processing, improving efficiency and response speed of data processing, timely capturing and processing data according to real-time data streams, guaranteeing real-time performance and accuracy of task primary data, screening historical data according to timeliness and availability standards through the historical data screening module, guaranteeing quality and applicability of task secondary data, calling task tertiary data according to the correlation degree of the task secondary data through a related data calling module, further enriching content and diversity of task requirement data sets, improving the correlation and the comprehensiveness of the data, integrating task primary data, task secondary data and task tertiary data into the task requirement data sets through a data integrating feedback module, feeding back the task primary data, the task primary data and the task tertiary data to a user side, conducting decision support and business analysis based on comprehensive and multi-level data, improving accuracy and effect of decision, and providing rapid, accurate and comprehensive data, improving service experience, and improving user satisfaction value and application satisfaction of a system.
Detailed Description
The embodiment of the application realizes the optimization of the data screening process through the data processing system suitable for accounting finance management, improves the real-time data processing capacity and the data analysis precision, and efficiently utilizes the history call data.
The problems in the embodiment of the application have the following general ideas:
The method comprises the steps of directly communicating with a user side, obtaining a specific accounting finance management demand task, screening and processing data according to the relevance of a real-time data stream and task demands to obtain first-level data of the task, ensuring that a system can respond immediately when the data are generated, providing latest information support, utilizing existing historical call data, screening according to timeliness and availability standards of the data, generating second-level data of the task, calling data of a relevant degree based on the second-level data of the task, obtaining deeper third-level data of the task, helping the system understand and analyze the data from multiple angles, providing a more accurate and comprehensive data view for a user, integrating the obtained first-level, second-level and third-level data of the task into a complete task demand data set, feeding the complete task demand data back to the user side, helping the user to make accurate and orderly decisions, and improving management efficiency and decision quality.
Referring to fig. 1 and 2, an embodiment of the invention provides a technical scheme, which comprises a task acquisition module, a real-time data screening module, a historical data screening module, a related data calling module, a data integration feedback module and a data storage and management module, wherein the task acquisition module is used for being in communication connection with a user end to acquire an accounting financial management demand task of the user end, the real-time data screening module is used for calling a real-time data stream stored in the data storage and management module and screening the real-time data stream according to correlation matching information between the real-time data stream and the accounting financial management demand task to obtain task primary data, the historical data screening module is used for searching historical call data of the accounting financial management demand task and obtaining task secondary data according to timeliness and usability of the historical call data, the related data calling module is used for calling task primary data based on the degree of correlation with the task secondary data, the data integration feedback module is used for integrating the task primary data, the task secondary data and the task tertiary data to obtain a task demand data set, and further, and the task primary data set is used for storing and the task demand data, and the historical call data of the accounting financial management demand task primary data, and the historical call data are used for searching the task primary demand data, and the historical call data of the task primary demand data and the data.
Specifically, the accounting finance management demand task specifically comprises a data analysis report task, a risk assessment prediction task, a budget cost control task, a tax planning task, a business decision support task and a user credit assessment task.
In this embodiment, the accounting finance management field includes a plurality of complex and diverse tasks, and the data types, data amounts and data structures processed by different tasks have significant differences, for example, the data analysis reporting task focuses on summarizing and visualizing historical finance data, while the risk assessment prediction task needs to use complex methods such as time series analysis, monte Carlo simulation and the like to predict future trends and assess potential risks, so that proper targeted data needs to be selected for analysis processing according to the different tasks.
The association matching information specifically comprises data field matching degree, data type format matching degree and business logic matching degree.
The method comprises the steps of screening real-time data streams to obtain specific analysis of task-level data, namely obtaining data fields, data type formats and business logic rules required by accounting financial management demand tasks, cleaning the real-time data streams, removing invalid, repeated and abnormal data, converting and normalizing the data formats, matching the data fields of each data in the real-time data streams with the data fields required by the accounting financial management demand tasks to obtain the data field matching degree of each data in the real-time data streams, screening each data based on the comparison of the data field matching threshold of the accounting financial management demand tasks with the data field matching degree of each data, screening data with the screening data field matching degree smaller than the data of the data field matching threshold, matching the type formats of each data in the real-time data streams with the data type formats required by the accounting financial management demand tasks after the screening data field matching degree is smaller than the data type matching threshold, screening the data type formats of each data in the data streams with the data type matching threshold of the accounting financial management demand tasks, screening the data type formats of each data type of each data with the data type of each data type in the real-time data streams based on the data type matching threshold of the accounting financial management demand tasks, matching the business logic rules, screening the data with the type of each data with the data type of each data of data in the data type in the real-time data stream after the data field matching degree smaller than the data field matching threshold is matched with the data type matching threshold, and the business logic rules required by the data type matching of each data of data type is screened, and marking the real-time data stream after the data with the service logic matching degree smaller than the service logic matching threshold value is finally screened out as task first-level data.
In this embodiment, the data fields required for the accounting finance management demand task represent specific data items or specific indexes such as amount, date, account number; the specific data field matching degree obtaining mode analysis is that considering the situation of partial matching between data fields, a word frequency statistics or a similarity calculating method based on feature vectors is used, such as cosine similarity and Jaccard similarity, the similarity degree between the data fields is evaluated, or different weights are set for different data fields, then weighted matching degree is calculated, and is specifically obtained by a real-time data stream now comprises the following fields of transaction number, account number, transaction date, transaction amount, transaction type and transaction state, assuming that a specific data field required to be focused by accounting demand task is account number, transaction date and transaction amount, for each data record in the real-time data stream, a similarity calculating method based on feature vectors is used, such as cosine similarity, account number of the data record= "12345", transaction date= "2024-07-31", transaction amount = 100.00, similarity between the account number and demand account number = 1 (complete matching), similarity between the transaction date and the demand transaction date is calculated according to the similarity degree, and the final transaction amount is calculated according to the date, if the sum of the transaction number is greater than 20207 and the demand account number is calculated according to the similarity, and if the final transaction amount is greater than the sum is calculated by the sum of 20245 and the required to be greater than the sum is calculated by the sum of the value of the sum of the required to be equal to the sum of the required and the required to be equal to the sum of the required to be equal to the sum and the required to the sum of the required to the sum and the required to be equal to the sum and the sum. Transaction amount=100.00, then account number matching = 1 (perfect match) ×0.6 (weight), transaction date matching = 1 (perfect match) ×0.2 (weight), transaction amount matching = 1 (perfect match) ×0.2 (weight), total matching= (0.6+0.2+0.2) =1 (perfect match).
The data type format required by the accounting finance management demand task comprises a character string type and a data field required by a specific format, for example, an account number is expressed as a character string type, the length is a fixed digital character, a transaction date is expressed as a date type, the format is YYYY-MM-DD, a transaction amount is expressed as a numerical value type, and the precision requirement is two digits after a decimal point; the specific acquisition mode analysis of the data type format matching degree is as follows: for each data record in the real-time data stream, acquiring the actual data type format of the data record through information of a data source, comparing whether the actual data type is matched with the required data type or not for each data field, if the actual data type is completely identical with the required data type, the data type matching degree is 1, if the actual data type is partially similar with the required data type, a fuzzy matching method such as text similarity algorithm or pattern matching can be considered to evaluate the matching degree, for data fields (such as date, amount and the like) related to specific format requirements, the data type matching degree needs to be ensured to be consistent with the required format, whether the actual date format accords with the format requirements of YYYY-MM-DD or not needs to be checked, for the amount type, whether the actual value is in a format with a fixed decimal number or not needs to be checked, and the data type matching degree and the data format matching degree needs to be comprehensively considered according to specific requirements set weights to obtain comprehensive matching degree, and specific acquisition data type format matching degree is exemplified by assuming that a financial management task requires data type format such as a character string number of 6 bits, transaction data type, transaction type of YYYY-MM-DD, transaction type data type, transaction data type and transaction data type DD, two digits after the decimal point are accurate, for the data record in the real-time data stream, the data record 1 is account number= "123456", transaction date= "2024-07-31", transaction amount = 100.00, the data record 2 is account number= "987654", transaction date= "2024/08/01", transaction amount = 150.005, for the data, account number type matching degree of the data record 1 is 1 (perfect match, length meets requirements), transaction date format matching degree is 1 (perfect match, format meets yyyyy-MM-DD), transaction amount format matching degree is 1 (perfect match, two digits after the decimal point are accurate), overall matching degree is (1+1+1)/3 = 1 (perfect match), account number type matching degree of the data record 2 is 0 (length does not meet requirements), transaction date format matching degree is 0 (format does not meet requirements), transaction amount format matching degree is 0.5 (format does not meet requirements, redundant decimal digits) and overall matching degree is (0+0+0.5)/3 = 0.17.
The business logic rules required by accounting financial management demand tasks refer to logic conditions or algorithms to be followed when processing, calculating, verifying or analyzing data, for example, for transaction verification rules, checking the validity of each transaction, including whether the amount of money is within a reasonable range, whether an account exists or not, for account balance calculation rules, updating account balances according to transaction records, ensuring the correctness and consistency of the balances, for report generation rules, generating financial reports according to specific accounting data, ensuring the accuracy and integrity of the reports, analyzing a business logic matching degree concrete acquisition mode to obtain business logic rules of actual application by a business logic definition of a system for each data record in a real-time data stream, specifically converting the business logic rules defined in accounting financial management demand tasks into computable logic conditions or algorithms, determining the business logic rules of actual application for each data record in the real-time data stream, relating to the processing process of the data, the business logic definition of a system or the real-time analysis of the data stream, comparing the business logic rules of actual application of each data record with the business logic rules defined in demand tasks, acquiring the business logic matching degree concrete rules for the account balance records according to the transaction balance calculation rules, and obtaining the transaction logic rules for the transaction balance calculation rules for the account balance calculation demand data is used for verifying that the account balance records are not matched according to the specific transaction logic definition of the system, the business logic rule applied by the data record in the existing real-time data stream is that the data record 1 is applied with transaction verification rules and account balance calculation rules, the data record 2 is applied with transaction verification rules and report generation rules, the matching degree is calculated as that the data record 1 is completely matched with the transaction verification rules in the demand task and is partially matched with the account balance calculation rules in the demand task, the business logic matching degree is (1+ partial matching degree)/the total number of rules, the data record 2 is completely matched with the transaction verification rules in the demand task and is partially matched with the report generation rules in the demand task, and the business logic matching degree is (1+ partial matching degree)/the total number of rules.
The data field matching threshold value, the data type format matching threshold value and the business logic matching threshold value are used for controlling the accuracy and the precision of data processing, the specific data field matching threshold value represents the minimum matching degree which is required by two data fields when the data fields are matched, whether the data meets the standard of a specific requirement task is determined when screening and processing data in a real-time data stream, by designating a specific numerical value, such as 0.8, the data field matching threshold value is regarded as matching to be set only when the data field matching degree reaches more than 80%, or the matching threshold value is dynamically adjusted according to the complexity of the specific requirement task and the diversity of the data, so as to ensure the adaptability and the flexibility of a system under different conditions, the data type format matching threshold value represents the allowable difference degree between actual data and data required by the requirement task when the data type format is checked, whether the data meets the specified format and the type requirement is determined in the data processing process, the data is required by designating the accurate data type format requirement, such as YYY-MM-DD format, the amount is required to be two digits after decimal points or the allowable difference is required to be accurate, such as a certain degree can be regarded as matching, the data can be completely matched with the data logic rule is defined in a certain degree when the data type is completely matches the data type of the data, the data is completely matched with the data type logic, the required by the logic is completely matches the data in a certain logic required by the data logic, the required by the data logic required by the real-state, the data is completely matches the data required by the data logic required to be completely, and the required by the data logic required to be completely meets the data requirements, and the required by the data logic required by the accuracy is completely and the data can be completely matched condition can be matched by the data and the data type can be matched according required by the data type can be met, the application of partial business logic rules may also be allowed, for example, partial mismatch of certain rules may be tolerated in certain situations to increase the flexibility and response speed of the system. The specific data field matching threshold, the data type format matching threshold and the business logic matching threshold are set and adjusted by a system administrator or a data analyst according to actual demands and business scenes, the setting mode depends on a configuration interface or a configuration file of a system, hard coding can be performed in codes of a data processing system in a programming mode, and the characteristics of data, the complexity of processing demands and the balance of system performance are comprehensively considered during setting, so that the optimal data processing effect and user experience are achieved.
The method comprises the steps of searching historical call data of an accounting financial management demand task, screening the historical call data of the accounting financial management demand task according to timeliness and usability of the historical call data, and obtaining specific analysis of task secondary data, wherein the historical call data of the accounting financial management demand task are searched based on a data storage and management module, the historical call data are sorted according to time, further historical call data falling in a time window are screened out according to a set time window, usability information is further identified according to the screened historical call data falling in the time window, the usability information specifically comprises data abnormal marking conditions, data security attack conditions and data source credibility conditions, and the historical call data which are identified to be free of data abnormal marking conditions or data security attack conditions or data source credibility conditions are screened out and marked as the task secondary data.
In this embodiment, the data anomaly marking status refers to the possible anomaly or error condition of the data itself, such as unexpected numerical value, format error and missing value, which are identified by data verification and anomaly detection algorithm, such as statistical analysis, data rule verification and anomaly detection model, the data security attack status refers to the possible unauthorized access, tampering or other security threat to the data, the possible security attack status is detected and identified by security log analysis, intrusion Detection System (IDS), access control log, encryption and digital signature, etc., the data source credibility status refers to the validity and credibility of the data source, which ensures that the data is generated or provided by authorized and credible sources, and the credibility of the data source is confirmed by digital certificate verification, data source identity authentication, data source history examination, etc.
The specific example of the task secondary data acquisition is that an accounting finance management system is assumed to process banking transaction data, the banking transaction data of historical call is searched from a data storage and management module, the historical call data is sequenced according to time sequence, then data falling in a time window is screened out according to a set time window, data quality inspection and anomaly detection are carried out on the screened historical call data, whether data anomaly marking conditions exist or not is identified, such as numerical anomaly, missing values and the like, data security logs and security monitoring information are analyzed, whether the historical call data are subjected to unauthorized access or falsification is detected, whether the data sources of the historical call data are credible is confirmed according to identity authentication, historical credible records and the like of data sources, and the historical call data passing through the data quality, security and credibility inspection are marked as data meeting the task secondary data requirements, so that the task secondary data acquisition system can be safely used for accounting finance management demand tasks.
The method comprises the steps of screening and analyzing historical call data, ensuring high data quality and good safety when processing tasks required by accounting financial management, improving reliability and stability of a system, screening the historical call data without data abnormality and safety attack conditions, ensuring accuracy and completeness of the data in the processing process, avoiding errors and abnormal conditions caused by the data quality problems, ensuring that the used historical call data is not subjected to unauthorized access or tampering by identifying the data safety attack conditions, improving protection capability of the system on data safety, confirming credibility of data sources, ensuring validity and authenticity of the used data, avoiding using data from an unreliable source, and further ensuring validity and transparency of data processing.
The method comprises the steps of setting a time window, setting a time window label for the time window of different sets according to the time length, dividing the time window label into different sets according to the size, setting a characteristic evaluation value label for the characteristic evaluation value of different sets according to the size, wherein the characteristic evaluation value label comprises a first characteristic evaluation value, a second characteristic evaluation value and a third characteristic evaluation value, the historical call data corresponding to the first characteristic evaluation value is matched with the first time window, the historical call data corresponding to the second characteristic evaluation value is matched with the second time window, and the third characteristic evaluation value is matched with the call history data corresponding to the third characteristic evaluation window.
In this embodiment, the data update period is specifically obtained by calculating the average value, standard deviation, etc. of the time interval or statistical time interval between each data point, the service period is specifically obtained by identifying the main service periodic feature by using fourier transform or other frequency analysis methods, the data stability is specifically obtained by calculating the fluctuation of the data, such as standard deviation, root mean square, etc., the feature evaluation value is specifically obtained by normalizing or normalizing each feature to ensure that different features have the same weight in the evaluation value calculation, and the data update period, the service period and the data stability are summed and averaged to obtain the feature evaluation value according to the logic relationship that the data update period is in direct proportion to the feature evaluation value, the service period is in direct proportion to the feature evaluation value, and the data stability is in direct proportion to the feature evaluation value, so as to obtain the feature evaluation value, wherein the specific feature evaluation value calculation formula is as follows: Wherein T represents a characteristic evaluation value, ε1 represents a data update period, ε2 represents a service period, ε3 represents data stability, and if the data update period is longer, the service period is longer, and the data is more stable, the characteristic evaluation value is higher, and the time window is divided into different sets such as short term, medium term and long term according to the characteristic evaluation value so as to reflect different time spans and data update frequencies.
Examples of specific time window settings are assuming that historical call data of an accounting financial management demand task comprises transaction data, financial statement data and market data, characteristic information of the data is obtained, the characteristic information comprises a data update period, a service period and data stability, the update period of the specific transaction data is usually day-to-day or even minute-level update, the update period of the financial statement data is usually quarter update, the update period of the market data is usually day-to-day index, the service period of the transaction data is highly related to day transaction and market opening time, the service period of the financial statement data is particularly related to quarter report time and annual audit time, the service period of the market data is particularly related to market fluctuation period and economic report release period, the data stability of the transaction data is stable, the fluctuation of the financial statement data may be severe during major economic events, the data stability of the financial statement data is stable, the data stability of the market data is generally stable before and after the financial statement release period is large, and the fluctuation of the market data is remarkable during market fluctuation or policy change. Each characteristic is quantized, wherein the data updating period, the service period and the data stability are respectively taken, the transaction data updating period is 1 day, the service period is 1 day, the data stability is 0.7 (higher stability), the financial report data updating period is 90 days, the service period is 90 days, the data stability is 0.9 (very high stability), the market data updating period is 1 day, the service period is 30 days, the data stability is 0.8 (high stability), the characteristic evaluation values are respectively 0.9, 60.3 and 10.6, the characteristic evaluation value is divided into a short-term (1-7 days), a second-stage time window is a middle-stage (7-30 days) and a third-stage time window is a long-term (30 days), the characteristic evaluation value is divided into a middle-stage (10-50) and a third-stage characteristic evaluation value according to the size, the market data characteristic evaluation value is further matched with the time window, the transaction data characteristic evaluation value is 0.9 and corresponds to the first-stage time window (1-7 days), the second-stage time window is a middle-stage (30 days), the characteristic evaluation value is divided into a third-stage time window is a long-stage (30 days), and the characteristic evaluation value is corresponding to the market evaluation value of the third-stage data evaluation value is a long-stage data evaluation window (60.6).
According to the characteristic information of the historical data, different kinds of data can be accurately distributed to a proper time window, the effect of a promotion strategy is improved, manual intervention is reduced through automatic processing of the historical data and the characteristic evaluation value, decision efficiency is improved, a personalized promotion strategy is customized based on different characteristics of commodities, market competitiveness is enhanced, the characteristic information of historical call data is effectively utilized, scientific basis can be provided for setting of the time window, and therefore business operation and marketing strategies are optimized.
The method comprises the steps of obtaining correlation information between the rest data of the task secondary data and each task secondary data, excluding the task secondary data, in a data storage and management module, wherein the correlation information comprises covariance absolute values and co-occurrence times, averaging the covariance absolute values between the rest data and each task secondary data to obtain covariance absolute value average values of the rest data and a task secondary data set, averaging the co-occurrence times between the rest data and each task secondary data to obtain co-occurrence times average values of the rest data and the task secondary data set, obtaining a covariance absolute value threshold and a co-occurrence times threshold which are called by accounting finance management data, comparing the covariance absolute value average values and the co-occurrence times average values of the rest data and the task secondary data set with corresponding covariance absolute value threshold and co-occurrence times threshold, and screening out that the covariance absolute value average values are larger than the covariance absolute value threshold and the co-occurrence times average values are larger than the co-occurrence times threshold.
In this embodiment, the covariance measures the overall error of two random variables, the absolute value of the covariance represents the strength and direction of the linear relationship between the two variables, specifically, the covariance between the two variables is obtained by performing mathematical calculation on the historical data, and then the absolute value is taken, the co-occurrence number represents the number of times that the two variables occur simultaneously under the same condition, and for each data in the dataset, the number of times that they occur simultaneously in the same time window or event can be counted, thereby obtaining the co-occurrence number.
The covariance absolute value threshold value of the accounting financial management data call is considered to have a stronger linear relation between two data when the covariance absolute value average value is higher than the threshold value, the co-occurrence frequency threshold value is considered to have a higher relevance between the two data when the co-occurrence frequency average value is higher than the threshold value, the setting of the covariance absolute value threshold value and the co-occurrence frequency threshold value of the accounting financial management data call is adjusted and optimized according to specific business and data characteristics, and the optimal threshold value setting can be determined through experiments and statistical analysis.
The specific acquisition example of the task tertiary data is that the task secondary data is assumed to be daily transaction amount data of a certain bank account, covariance calculation and co-occurrence count are carried out on all other data except the task secondary data and the daily transaction amount data of the task secondary data, the set covariance absolute value threshold is assumed to be 0.7, the co-occurrence count threshold is 50, data, of which the covariance absolute value mean value is larger than 0.7 and the co-occurrence count mean value is larger than 50, in all the data are found to be the task tertiary data, for example, the fact that higher covariance absolute value and co-occurrence count mean value exist between monthly deposit data and daily transaction amount data of a customer is found through the steps, therefore, the deposit data of the customer are marked as the task tertiary data, the fact that the task tertiary data and the task secondary data have higher correlation is ensured, and the accuracy and efficiency of data processing are improved.
The task three-level data highly related to the task two-level data is screened out by analyzing the degree of correlation with the task two-level data, so that the accuracy and efficiency of data processing are improved, the task three-level data and the task two-level data are ensured to have higher correlation, the subsequent data analysis and processing are more accurate and effective, the data with lower correlation with the task two-level data are eliminated, the system resources and the processing time are reduced, and the data processing flow is optimized.
In summary, the present application has at least the following effects:
The system comprises a task acquisition module, a real-time data screening module, a historical data screening module, a data integration feedback module, a historical data processing module, a data integration feedback module and a data processing module, wherein the task acquisition module can communicate with a user side and acquire specific accounting financial management demand tasks, ensure that data can be processed according to actual demands rather than all data in a one-step way, the real-time data screening module can screen and process data according to the associated matching information of real-time data flow and task demands so as to acquire primary data of the tasks, can guarantee that a system can respond immediately when needed and provide timely updated data, the historical data screening module can screen and process the historical data according to timeliness and availability standards of historical call data so as to acquire secondary data of the tasks, can utilize existing data resources to improve the reuse rate and efficiency of the data, and the related data call module can help the system to analyze the data deeply and provide more accurate and comprehensive data support, and the data integration feedback module integrates the primary data, the secondary data and the tertiary data of the tasks into a complete task demand data set and feeds back the complete data set to the user side so as to help the user to make more accurate and decision-making quality and more effective improvement.
Those skilled in the art will appreciate that embodiments of the present invention may be provided as a system. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to block diagrams of systems according to embodiments of the present invention. It will be understood that each structure of the block diagrams can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the block diagram each structure.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in each structure of the block diagram.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in each structure in the block diagram.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.