Financial index real-time analysis method and system

Abstract

The invention relates to the technical field of financial data processing, in particular to a financial index real-time analysis method and system, and the method comprises the following steps: extracting financial fields, constructing a mapping chain, matching an archiving path, judging whether to switch, analyzing historical cycle data, establishing a trend chain, merging a state chain, and generating a real-time analysis result. According to the method, through structural acquisition of income, expenditure, asset and liability fields in the financial statement and deep-level analysis of type distribution, sequence density and crossing frequency, construction of the field chain composition density is realized, and data structure characteristics can be accurately mastered at the initial stage of index identification; and the matching deviation risk caused by heterogeneous field distribution is effectively avoided. On the basis, clustering comparison between a field chain structure and archiving path nodes is adopted, an error graph is constructed, fine measurement of path node similarity is achieved, and the reasonability of archiving path selection can be improved through a path error minimum matching mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of financial data processing, and in particular to a real-time analysis method and system for financial indicators. Background Art

[0002] The field of financial data processing technology includes the means of collecting, organizing, calculating, analyzing and managing various financial information generated by enterprises or organizations in the process of operation and management. The core content of this technical field includes accounting data collection, financial statement generation, financial risk identification, cost control, budget management and financial forecasting. Through the systematic processing of structured and unstructured financial data, this field supports decision makers in the dynamic control and analysis of corporate economic activities.

[0003] Among them, the real-time analysis method of financial indicators refers to the financial data generated by the enterprise in its business activities, through the real-time acquisition and dynamic calculation of core financial indicators such as asset-liability ratio, current ratio, net profit margin, accounts receivable turnover rate, etc., to achieve the grasp of key financial conditions. This method is usually based on the setting of indicator analysis models, combined with data collection rules and timestamp marking mechanisms, to perform structured extraction and indicator calculation processing on the original data from the financial system, and further by setting threshold conditions and logical judgment rules, the analysis results are classified and archived in real time, which is convenient for subsequent statistical comparison and periodic analysis.

[0004] The existing technology takes the static indicator model as the core, performs preset structural analysis on financial fields and sets thresholds for judgment and classification. It lacks the ability to dynamically perceive the distribution of fields, which leads to structural recognition errors and path archiving deviations when processing different financial report formats or field change scenarios. The field structure has not been deeply analyzed, and it is difficult to identify implicit structural rules in reports with high field combinations and cross-frequency, and there is a high error rate in field positioning and matching. In terms of archiving path selection, it only relies on static rules or direct comparison of current indicator values ​​with set standards, and has not established an effective structural matching mechanism, making path determination susceptible to interference from outliers, resulting in path mismatch. In the trend judgment link, the trend estimation method relies on a single period or simplified average value, lacks the investigation of the consistency of field differences in consecutive periods and the degree of trend overlap, resulting in the inability to accurately depict the true change trajectory of indicators and insufficient adaptability to periodic fluctuations. For example, when a company's financial indicators undergo short-term adjustments, the existing methods are prone to identify them as trend changes, triggering misjudgments and incorrect archiving, affecting the accuracy of subsequent statistical analysis and decision support. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a real-time analysis method and system for financial indicators.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: a real-time analysis method of financial indicators, comprising the following steps: S1: Obtain the income fields, expenditure fields, asset fields and liability fields in the financial statements, build a field mapping chain, and analyze the type distribution, sequence density and cross frequency of each field to generate field chain composition density information; S2: Based on the field chain composition density information, perform field type matching and sequence combination comparison on the current indicator field chain and the archive path node field chain structure clustered according to the field chain mapping structure of all financial indicators in the current period, and generate a field chain matching error map; S3: According to the matching error distribution of each archiving path node in the field chain matching error map, select the path with the lowest error, judge whether the index archiving path needs to be switched, and obtain the path switching judgment result; S4: Based on the nodes that have been determined as paths that need to be switched in the path switching judgment result, the numerical sequence of the current indicator in the previous continuous financial period under the path structure after switching is collected, and the consistency judgment of the period-to-period difference and the interval overlap analysis are performed on each group of fields in the sequence to establish an indicator change trend chain table; S5: Based on the indicator change trend chain table, consistent paths with overlapping intervals are screened, state chains are merged, and real-time analysis results of financial indicators are generated.

[0007] As a further solution of the present invention, the field chain composition density information includes field sequence density, field type crossover frequency, and field type occurrence position; the field chain matching error map specifically includes field overlap rate, sequence deviation level, and matching error distribution; the path switching judgment result includes path adaptation degree, structural path position, and node switching suggestion; the indicator change trend chain table specifically refers to period value difference, period direction consistency number, and change amplitude interval distribution; the financial indicator real-time analysis results include state path matching degree, period segment consistency, and state chain merge path.

[0008] As a further solution of the present invention, the step of obtaining the field chain composition density information is specifically as follows: S111: Obtain the income field, expenditure field, asset field and liability field in the financial statement, collect the field combination structure included in each financial indicator in the current period, obtain the arrangement order of each field, the number of field types and the location data of the field cross-dense area, organize the fields in order to build a field mapping chain, and obtain the field mapping chain structure sequence; S112: Call the field arrangement order in the field mapping chain structure sequence to determine whether the types of adjacent fields are consistent, using the formula: ; Calculate the Sequential continuity of group field chain structures , which is used to measure the consistency and stability of the order in which financial fields are arranged in the structure; in, Indicates The sequential continuity of the group field chain structure, Indicates The total number of fields in the group field chain, Indicates the number of fields from the first to the All pairs of fields between fields are accumulated. Indicates The first Fields and Field type consistency indicators, where the same is 1 and different is 0. Indicates The position index of the field in the field chain. Indicates the absolute value of the field index spacing. is the normalization factor used to convert the sum to the average between field pairs; S113: Based on the sequential continuity of the field mapping chain structure sequence and the field chain structure, count the field occurrence position and type distribution frequency corresponding to each field type in the field chain structure, identify the cross-appearing field combination segments, obtain the combination characteristics of the cross-type and position, and generate field chain composition density information.

[0009] As a further solution of the present invention, the step of obtaining the field chain matching error map is specifically as follows: S211: Based on the field chain density information, the current indicator field chain is called, and the archive path node field chain structure obtained by clustering the field chain mapping structure of all financial indicators in the current period is called, the overlap of the field types in the current field chain and each archive path field chain is identified, the fields with overlapping field names are extracted, the number of fields is counted and the corresponding index position offset is compared, and a field matching difference value set is generated; S212: Based on the field matching difference value set, the ratio of the field overlap values ​​in each archive path node to the total number of fields in the current field chain is calculated to obtain the field overlap rate, the index position difference values ​​of each pair of overlapping fields are counted, and divided into multiple offset level intervals, and the field overlap rates of all path nodes are compared and summarized with the offset level results to construct a field chain matching error map.

[0010] As a further solution of the present invention, the step of obtaining the path switching judgment result is specifically: S311: According to the matching error distribution of each archive path node in the field chain matching error map, the field overlap rate and the order deviation level corresponding to each node are identified, and the node chains whose field overlap rate is lower than the field overlap rate threshold are extracted, and the node chains whose order deviation level exceeds the field order deviation threshold are extracted, and the two types of node chains are merged and marked, and the marked path nodes are screened out to obtain a path node elimination result set; S312: Based on the path node elimination result set, call the retained archive path node set, calculate the matching error corresponding to each node, select the single structure path with the lowest matching error, extract the corresponding field chain structure, call the field arrangement order and field combination content in the current indicator field chain, compare the field order structure and field combination distribution in the target node path field chain, and generate archive path matching difference information; S313: Based on the archive path matching difference information, determine the range of change in the arrangement continuity and the magnitude of the field combination difference of the current indicator field chain structure between the original path and the target path, and make a decision on the path switching in combination with the archive path node switching trigger condition, and establish a path switching judgment result representing the switching status identifier between the indicator field chain and the archive node.

[0011] As a further solution of the present invention, the steps of obtaining the indicator change trend linked list are specifically as follows: S411: Based on the node determined as the one that needs to be switched in the path switching judgment result, collect the numerical sequence of the current indicator in the first three consecutive financial periods under the path structure of the node, extract the value content of the corresponding field of the indicator in each period, and generate a continuous period sequence of the indicator; S412: Based on the continuous period sequence of the indicator, select the accounting item fields associated with the original field group structure, extract the numerical sequences corresponding to the main business income, total period expenses, increase or decrease of non-current assets and repayment of current liabilities, calculate the difference between two adjacent periods of each field, record the direction of the numerical change of each field within three periods, and generate the statistical results of the field period change; S413: Based on the statistical result of the field periodic change, the value difference of each field between two periods is judged for sign consistency, using the formula: ; Calculate the The calculation field is in The overlap of the changing trends under the period , used to measure the consistency of the change direction and magnitude of the same field in adjacent financial periods, build a field-level trend chain by combining the trend expression results of multiple period pairs, and establish an indicator change trend chain table; in, Indicates The fields in The range of fluctuations in the cycle, Indicates The fields in The range of fluctuations in the cycle, represents the intersection length of two periodic intervals, represents the total length of the union of two period intervals, Indicates The fields in The direction of the value change of each cycle, where +1 indicates an increase, -1 indicates a decrease, and 0 indicates no change. Indicates that the corresponding field is in The direction of change of the cycle, It is the direction difference, which reflects the consistency of the changing direction of adjacent cycles.

[0012] As a further solution of the present invention, the steps for obtaining the real-time analysis results of the financial indicators are specifically as follows: S511: Based on the indicator change trend chain table, call the continuous period indicator state sequence corresponding to the current indicator field chain in the path switching node structure, extract the value change trend information, change interval range and period time series index of the field corresponding to each path segment in the continuous period, and generate a period state sequence path set; S512: Call each path segment field sequence in the periodic state sequence path set, using the formula: ; Calculating path segments The total value of trend overlap matching , used to evaluate the consistency between the field change trend of the path segment in multiple cycles and the current indicator path, and filter all A set of path segments exceeding the path matching threshold is used to generate trend matching path segments; in, represents the number of cycle pairs covered by the path segment, Indicates the number of fields in the path segment, It is The calculation field is in The overlap of the changing trends in the next cycle; S513: Based on the trend matching path segment, select the path segment with the matching degree closest to the upper limit of the path matching threshold, extract the continuous indicator state sequence covered by the target path segment, merge and fuse the field structure in each state segment with the current field chain, integrate them into a continuous trend expression structure, and generate real-time analysis results of financial indicators.

[0013] A financial indicator real-time analysis system, the financial indicator real-time analysis system is used to execute the above-mentioned financial indicator real-time analysis method, the system comprises: The field density building module obtains the income field, expenditure field, asset field and liability field in the financial statement, builds a field mapping chain, and analyzes the type distribution, sequence density and cross frequency of each field to generate field chain composition density information; The field matching analysis module performs field type matching and sequence combination comparison on the current indicator field chain and the archive path node field chain structure clustered according to the field chain mapping structure of all financial indicators in the current period based on the field chain composition density information, and generates a field chain matching error map; The path switching judgment module selects the path with the lowest error according to the matching error distribution of each archiving path node in the field chain matching error map, judges whether the index archiving path needs to be switched, and obtains the path switching judgment result; The trend chain table generation module collects the numerical sequence of the current indicator in the previous continuous financial period under the path structure after the switch based on the nodes that have been determined as the paths that need to be switched in the path switch judgment result, performs consistency judgment of the period difference and interval overlap analysis on each group of fields in the sequence, and establishes the indicator change trend chain table; The indicator status merging module selects consistent paths with overlapping intervals based on the indicator change trend chain list, merges the status chains, and generates real-time analysis results of financial indicators.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, through the structural collection of income, expenditure, assets and liabilities fields in financial statements and the in-depth analysis of their type distribution, sequence density and cross frequency, the construction of field chain composition density is realized, which can accurately grasp the data structure characteristics at the early stage of indicator identification, and effectively avoid the matching deviation risk caused by the heterogeneous field distribution. On this basis, the clustering comparison between the field chain structure and the archiving path node is adopted to construct the error spectrum, realize the fine measurement of the similarity of the path node, and improve the rationality of the archiving path selection through the path error minimum matching mechanism. Further, through the collection of the historical period field value sequence and the consistency of the period difference and the trend overlap analysis, the indicator change trend chain table is established to provide sufficient trend support information for path switching. The dual trend matching of the field level change direction and amplitude can effectively avoid the misjudgment caused by abnormal fluctuations in short periods and improve the accuracy of the indicator archiving path switching. On the basis of path selection, the period path segment is screened based on the trend overlap matching value, the field structure and state information in the path segment are integrated, and the indicator state structure sequence with strong continuity and consistent trend is constructed to ensure that the analysis results have temporal logic consistency and trend continuity. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the workflow of the present invention; Figure 2 This is a flow chart of obtaining field chain composition density information in step S1 of the present invention; Figure 3 This is a flow chart of obtaining a field chain matching error map in step S2 of the present invention; Figure 4 This is a flow chart of obtaining the path switching judgment result in step S3 of the present invention; Figure 5 This is a flow chart of obtaining the indicator change trend list in step S4 of the present invention; Figure 6 This is a flow chart of obtaining the real-time analysis results of financial indicators in step S5 of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0018] See also Figure 1 The present invention provides a technical solution: a real-time analysis method of financial indicators, comprising the following steps: S1: Obtain the income field, expenditure field, asset field and liability field in the financial statement, collect the basic field combination structure of each financial indicator in the current period, build the field mapping chain before indicator generation according to the field sequence arrangement, the number of field types and the position of the field cross-dense area, calculate the continuity of the field mapping chain according to the field sequence density, and count the cross frequency and occurrence position of each field type in each group of field mapping chains to generate field chain composition density information; S2: Based on the field chain composition density information, perform field type matching and sequence combination comparison on the current indicator field chain and the archive path node field chain structure clustered according to the field chain mapping structure of all financial indicators in the current period, compare the number of overlapping fields and the sequence position of the field chain respectively, calculate the position adaptation degree of the current field chain structure in all archive path nodes, and generate a field chain matching error map according to the field overlap rate and sequence deviation level; S3: According to the matching error distribution of each archiving path node in the field chain matching error map, the node chains whose field overlap rate is lower than the field overlap rate threshold or whose sequence deviation exceeds the field sequence deviation threshold are eliminated and screened, and the structural path corresponding to the lowest error among the remaining nodes is obtained. Combined with the field structure of the target node and the field arrangement continuity and combination difference degree of the current indicator field chain, it is judged whether the current indicator archiving path should be switched to the new node, and the path switching judgment result is obtained; S4: Based on the nodes that have been determined as paths that need to be switched in the path switching judgment results, collect the numerical sequence of the current indicator in the previous continuous financial period under the path structure after switching, select the accounting detail fields that are closely related to the original field group structure, including the main business income recognition value, the total period expenses, the increase or decrease value of non-current assets and the repayment amount of current liabilities, judge the period value difference, the number of period value direction consistency and the period value change range distribution of each field in the continuous period, perform period difference consistency judgment and interval overlap analysis on each group of fields, and establish an indicator change trend chain table; S5: Based on the obtained period-to-period difference consistency results and interval overlap analysis results in the indicator change trend chain table, the continuous period indicator state sequence corresponding to the current indicator field chain in the path switching node structure is called, and according to the continuation trend of the field value change interval in the sequence, the path segments that have period segment consistency with the current field chain and satisfy the change interval overlap degree exceeding the state path matching threshold are screened, and the path sequence with the closest matching degree is selected for state chain merging processing to generate real-time analysis results of financial indicators; The field chain composition density information includes field sequence density, field type crossover frequency, and field type occurrence location. The field chain matching error map specifically includes field overlap rate, sequence deviation level, and matching error distribution. The path switching judgment results include path adaptation degree, structural path location, and node switching suggestions. The indicator change trend chain table specifically refers to the period value difference, the number of period direction consistency, and the change range distribution. The real-time analysis results of financial indicators include state path matching, period segment consistency, and state chain merge path.

[0019] See also Figure 2 , the specific steps for obtaining the field chain density information are: S111: Obtain the income field, expenditure field, asset field and liability field in the financial statement, collect the field combination structure involved in each financial indicator in the current period, obtain the arrangement order of each field, the number of field types and the location data of the field cross-dense area, organize the fields in order to build a field mapping chain, and obtain the field mapping chain structure sequence; First, locate the specific positions of various fields in the standard financial report template, combine with the company's monthly financial report or annual consolidated report, export the original field values ​​from the electronic account book or structured database, and extract the financial data in the current cycle one by one. For example, in the first quarter of 2025, collect "main business income" of 4.58 million yuan, "sales expenses" of 730,000 yuan, "original value of fixed assets" of 6.2 million yuan, "short-term loans" of 950,000 yuan and other field data, and arrange them according to the original order in which the fields appear in the report, for example, put the income field in the first place, the expenditure field in the second place, the asset field in the third place, and the liability field in the fourth place, and establish a standard Quasi-field sequence, in which the data type of each field such as amount, percentage, and ending balance is recorded, and the possible cross-dense area positions are analyzed. Cross-dense refers to the repeated appearance of fields in multiple indicators and the formation of highly overlapping structural segments. Assuming that "main business income" and "short-term loans" appear in multiple accounting items at the same time, it can be determined that they are in the cross-dense area. Then, a field mapping chain is formed in the order of fields, such as the sequence {R1, E1, A1, L1}, where R, E, A, and L represent income, expenditure, assets, and liabilities, respectively. Finally, a structurally ordered mapping chain is formed for subsequent calculations to obtain the field mapping chain structure sequence.

[0020] S112: Call the field arrangement order in the field mapping chain structure sequence to determine whether the types of adjacent fields are consistent, using the formula: ; Calculate the Sequential continuity of group field chain structures ; in, Indicates The sequential continuity of the group field chain structure, Indicates The total number of fields in the group field chain, Indicates the number of fields from the first to the All pairs of fields between fields are accumulated. Indicates The first Fields and Field type consistency indicators (1 for the same, 0 for different), Indicates The position index of the field in the field chain. Indicates the absolute value of the field index spacing. is the normalization factor used to convert the sum to the average between field pairs; Call the index of each field in the field mapping chain structure sequence to determine whether the types of adjacent fields are consistent, and construct a Boolean type consistency indicator. ,otherwise , combined with the index difference of the field, such as Calculate the inverse of its spacing, such as when the field , and , then the calculated value of this item is 0. , and , then the item is , the overall calculation uses the following formula: ; Now select a set of simulated data: number of fields , the position index is , the corresponding type consistency index is ,but: ; The result shows that the type continuity between fields is good and the distribution is compact, the coupling in the field chain is strong, and the field order continuity value is finally generated.

[0021] S113: Based on the sequence continuity of the field mapping chain structure sequence and the field chain structure, count the field occurrence position and type distribution frequency corresponding to each field type in the field chain structure, identify the cross-appearing field combination segments, obtain the combination characteristics of the cross-type and position, and generate field chain composition density information; Read the type of each field and its index position in the mapping chain one by one, build a field-position mapping table, and determine its concentrated area and frequency of occurrence by counting the distribution of each type of field. For example, if the "asset field" appears 12 times in 10 indicators, its average frequency of occurrence in each chain can be calculated to be 1.2 times. Identify the field segments that appear more than twice in the same chain as cross-segments. For example, "sales expenses" appear repeatedly in the second and third positions of multiple indicator chains, which is a typical cross-field. Combine the number of field types, the length of the repeated segment and the index where it is located to build a cross-segment feature record table, and then extract the frequency value and concentration value of the field distribution. For example, the records where the income field appears in the first 1 / 3 of the chain are obtained, forming a field-position density vector, such as [income: 3, expenditure: 2, assets: 5, liabilities: 4]. This is used as the basis for evaluating the field structure density, and finally generates the field chain composition density information.

[0022] See also Figure 3 , the specific steps for obtaining the field chain matching error map are: S211: Based on the field chain composition density information, the current indicator field chain is called, and the archive path node field chain structure obtained by clustering the field chain mapping structure of all financial indicators in the current period is identified, the overlap of the field types in the current field chain and each archive path field chain is identified, the fields with overlapping field names are extracted, the number of fields is counted and the corresponding index position offset is compared, and a field matching difference value set is generated; Clarify the field types and structural arrangement order of the current indicator field chain, then extract the corresponding archive path node field chain structure, and perform consistency matching operations around the field names. In actual operation, the field list in the current field chain can be defined as , the archive path node field chain is as follows , then the overlapping part of the field names is , the number of overlapping fields is 2. During the calculation process, the index positions of fields "B" and "C" in the current field chain are recorded as 2 and 3 respectively, and the corresponding archive path field chains are 1 and 2, and the index position difference is 1 and 1. Then all path nodes are processed to obtain the field name overlap and index difference of each path. For 10 groups of path node sample data, if the number of overlapping fields in a node field is 3, and its field index deviations are 1, 0, and 2 respectively, it is necessary to record the field matching pairs and their corresponding positions, and finally establish the field type and position comparison relationship to form a field matching difference value set, such as whether the names of the fields "income", "expenditure", and "assets" in different path nodes are consistent and the changes in index positions are shown in the table. The accuracy of subsequent judgments will be determined. For a more specific explanation, the current field chain is set to "main operating income, period expenses, non-current assets, and total liabilities" and compared with a path node field chain "period expenses, main operating income, total liabilities, and tax expenditures". The overlapping fields are "main operating income", "period expenses", and "total liabilities". The index position differences of the three fields are 1, 1, and 1, respectively. The number of overlapping fields is 3, and the field difference value set is 3 groups of difference 1s, reflecting that the path node is of medium similarity in type and order. No algorithm call is used during execution. The entire process is carried out item by item through field name comparison and index sequence comparison, and the difference results are recorded one by one. The result is the field matching difference value set.

[0023] S212: According to the field matching difference value set, the ratio of the field overlap value in each archive path node to the total number of fields in the current field chain is calculated to obtain the field overlap rate, the index position difference value of each pair of overlapping fields is counted, and the values ​​are divided into multiple offset level intervals. The field overlap rates of all path nodes are compared and summarized with the offset level results to construct a field chain matching error map; Based on the obtained field matching difference value set, it is necessary to further extract the field overlap values ​​and calculate the field overlap rate. The total number of fields is used as the denominator and the number of overlapping fields is used as the numerator to obtain the field overlap ratio of the path. For example, if there are 3 overlapping fields in a path and the total number of fields is 4, the overlap rate is , and then, according to the field index position difference value set, each difference value is graded. The specific interval is set as follows: the difference value of 0 is recorded as "level 0", indicating complete consistency, the difference value of 1 is "level 1", and the difference of 2 and above is "level 2". Suppose there are 3 overlapping fields in path node A, and their position deviation values ​​are 0, 1, and 2 respectively, then their offset level sequence is 0, 1, and 2. According to the above interval attribution, the level label conversion is performed, and the matching degree difference in each path node is expressed as two vector structures, namely, the combination of the overlap rate and the offset level, which is convenient for the subsequent archiving path screening between path nodes. The number of field overlaps in node B is 4, the field chain length is 5, the overlap rate is 0.8, the offset difference is 1, 1, 2, 0, the corresponding levels are 1, 1, 2, 0, and the overall average offset level is 1.0. The value source is obtained through the absolute quantification of the aligned field index difference in the field chain, and does not involve any direct calculation of non-numeric data. The overlap rate and offset level values ​​in the archive path nodes are extracted and sorted horizontally and arranged, which can further form a two-dimensional list structure for comparison. The current field chain structure is sorted by the matching adaptability of the path nodes, and finally a field chain matching error map is generated.

[0024] See also Figure 4 , the specific steps for obtaining the path switching judgment result are: S311: According to the matching error distribution of each archive path node in the field chain matching error map, the field overlap rate and the order deviation level corresponding to each node are identified, and the node chains whose field overlap rate is lower than the field overlap rate threshold and the node chains whose order deviation level exceeds the field order deviation threshold are extracted respectively, and the two types of node chains are merged and marked, and the marked path nodes are screened out to obtain the path node elimination result set; The error information corresponding to each archive path node is analyzed and processed one by one. First, two types of indicators, field overlap rate and sequence deviation level, are extracted, and the field overlap rate threshold and the sequence deviation level threshold are set, which are used as screening conditions for node exclusion operations. The field overlap rate threshold can be set to 0.65, that is, only path nodes with an overlap rate higher than 65% are retained. The sequence deviation level threshold can be divided into three level intervals, namely low deviation (01), medium deviation (23), and high deviation (4 and above), among which only low deviation and medium deviation path nodes are retained. In practical applications, the following example can be combined: If the current indicator field chain is [income, assets, expenditures, liabilities], its matching error with the field chain [income, expenditures, assets, liabilities] in path A is: the field overlap rate is 3 / 4=0.75, and the sequence deviation level is 2; compared with the field chain [assets, The matching error of [liabilities, other expenses, income] is as follows: if the field overlap rate is 2 / 4=0.5 and the order deviation level is 4, path B will be eliminated and path A will be retained. Path A can be included in the matching sequence later; the field overlap rate is obtained by calculating the number of fields with the same field name divided by the total number of fields in the current field chain, and the order deviation level is obtained by counting the index difference of each overlapping field and classifying it according to the preset level rules; for example, "income" in path A is at index 0, which is consistent with the index of "income" in the current field chain, and the difference is 0, the index difference of "assets" is 1, the difference of "expenditure" is 1, and the difference of "liabilities" is 0, then the deviations do not exceed 1, and the corresponding levels are classified into the medium deviation level range. Under multiple path nodes, after the above screening, the path nodes with a field overlap rate higher than 0.65 and an order deviation level not exceeding 3 are retained to obtain the path node elimination result set.

[0025] S312: Based on the path node elimination result set, call the retained archive path node set, calculate the matching error corresponding to each node, select the single structure path with the lowest matching error, and extract the corresponding field chain structure, call the field arrangement order and field combination content in the current indicator field chain, and compare the field order structure and field combination distribution in the target node path field chain respectively, and generate archive path matching difference information; After obtaining the set of path nodes after elimination, the field chain structure of each retained path node needs to be judged one by one. First, sort according to the field overlap rate, and extract the path node with the largest number of overlapping fields. If there are multiple nodes with the same number of overlaps, further compare the order deviation levels, and give priority to selecting the node with the smallest order deviation level as the matching path node. The following example can be used in the operation: Path C has 3 fields consistent with the current field chain, and Path D also has 3 fields consistent, then enter the second level of screening, the order deviation level of Path C is 1, and that of Path D is 2, so Path C is given priority to be retained as the target path node. If only a single path node has the largest number of fields overlapping, it is directly used as the target node. Next, call the field chain structure in the target path node and compare it with the current field chain structure in the word The similarity of the segment arrangement order is analyzed, and the continuous distribution of the fields is analyzed. For example, if the current field chain order is [income, expenditure, assets, liabilities] and the target node field chain is [income, assets, expenditure, liabilities], then the continuous matching field segment is only one group of [expenditure, assets], and the continuity is lower than the current structure. On this basis, the re-combination distribution of the fields is identified, such as how many field swaps are required to reach an agreement. The field position mapping table is used to mark the field alignment relationship, and it is determined whether there is a large range of field spans, that is, the total number of fields with a jump of more than 2 bits between the original index of the field and the target index. In the structural judgment, the length of the field continuous segment and the number of field index spans are combined to determine the arrangement and combination differences between the current indicator field chain and the target path node, and finally generate the archive path matching difference information.

[0026] S313: According to the archive path matching difference information, the arrangement continuity change range and field combination difference amplitude of the current indicator field chain structure between the original path and the target path are determined, and the path switching is determined in combination with the archive path node switching trigger condition, and a path switching judgment result representing the switching state identifier between the indicator field chain and the archive node is established; First, the number of consecutive matching groups in the field order of the target path node and the current field chain is compared. If the length of the consecutive segment in the target path field chain is greater than the corresponding structure of the current field chain, it is marked as order promotion. Then further judgment is made based on the field combination offset. The offset can be obtained by comparing the number of position jumps of the field in the two structures. For example, the field "asset" was originally located in the second position and is located in the first position in the target structure, which is a single offset. If there are three fields offset in the current field chain, and only one field offset in the target path field chain, it means that the degree of combination difference has decreased, and this type of combination structure is closer to the current indicator's self- According to the arrangement trend, the offset threshold is set to 2 in this judgment, that is, when the number of combined offsets of the target node is less than or equal to 2, and the number of continuous segments is not less than the current structure, it can be judged that the path structure has the basis for switching, and further combined with the structural rationality confirmation item, that is, whether the field types in the target path field chain all cover the field set required by the current indicator. If there are missing fields, switching is not allowed. On the basis of meeting all the above conditions, the path structure number corresponding to the current indicator field chain is recorded, and the pointing path in its archive path mapping table is updated, changing the original path node number to the target path node number, and finally generating the path switching judgment result.

[0027] See also Figure 5 , the specific steps for obtaining the indicator change trend list are: S411: Based on the node determined as the one that needs to be switched in the path switching judgment result, the numerical sequence of the current indicator in the first three consecutive financial periods under the path structure of the node is collected, and the value content of the corresponding field of the indicator in each period is extracted respectively to generate a continuous period sequence of the indicator; First, call the path structure to collect the accounting value sequence of the current indicator in the first three consecutive financial periods under the path structure in turn. For each period, the field position mapped by the indicator at a specific node should be extracted first. For example, if the main business income field is 1.254 million yuan in Q1, 1.329 million yuan in Q2, and 1.297 million yuan in Q3 in 2022, then establish a time series of this field such as [125.4, 132.9, 129.7]. This step requires clear marking of the corresponding period to ensure that the sequence position is consistent with the time sequence in the subsequent calculation process, and then synchronously perform the same operation on the total period expense field, the non-current asset increase or decrease value field, and the current liability repayment amount field. For example, when the period expenses are 453,000 yuan, 447,000 yuan, and 471,000 yuan in the above three periods, they are recorded as [45.3, 44.7, 47.1]. ​​After establishing a complete time series set of four independent fields, bind them to the field dimension under the current node structure path to provide a basic data structure for the next stage of judgment and calculation, and finally obtain the indicator continuous period sequence.

[0028] S412: Based on the continuous period sequence of the indicator, select the accounting item fields associated with the original field group structure, extract the numerical sequences corresponding to the main business income, total period expenses, increase or decrease of non-current assets and repayment of current liabilities, calculate the difference between two adjacent periods of each field, record the numerical change direction of each field within three periods, and generate the statistical results of field period changes; Based on the collected continuous cycle sequence value group of indicators, the accounting fields with a strong coupling relationship with the original field group are screened out, namely, the main business income recognition value, total period expenses, non-current asset increase and decrease value and current liability repayment amount fields, and the difference judgment is performed on their values ​​in three consecutive cycles. For each field, two pairs of cycle groups are formed from the 1st cycle to the 2nd cycle and the 2nd cycle to the 3rd cycle. The difference values ​​are calculated respectively, and the direction is judged whether it is consistent. If the direction of the difference of a field is positive, positive, or negative, negative, the consistency of the direction of the field is 2 times. If it is positive, negative or negative, positive, the consistency number is 0. In the example, the main business income difference is , , the direction is inconsistent, the direction consistency times is 0, for the period expenses are -0.6, +2.4, the direction is also inconsistent, the direction consistency is 0, and if the increase or decrease value of non-current assets is 6.2, 6.9, 7.1, the difference is 0.7, 0.2, and the direction consistency is 2. In addition, the field values ​​of each pair of periods are taken to calculate the range of their changes. The calculation method is that the upper and lower limit differences constitute the interval. For example, the change range of period expenses between Q1 and Q2 is 44.7, 45.3; the interval set of all fields is stored as the basis for judging the subsequent trend overlap, and finally the field period change statistical value set is obtained.

[0029] S413: Based on the statistical results of field periodic changes, the value difference of each field between two periods is judged for sign consistency, using the formula: ; Calculate the The calculation field is in The overlap of the changing trends under the period , combine the trend expression results of multiple period pairs to build a field-level trend chain and establish an indicator change trend chain table; in, Indicates The fields in The range of fluctuations in the cycle, Indicates The fields in The range of fluctuations in the cycle, represents the intersection length of two periodic intervals, represents the total length of the union of two period intervals, Indicates The fields in The direction of the value change in each cycle (+1 indicates an increase, -1 indicates a decrease, and 0 indicates no change). Indicates that the corresponding field is in The direction of change of the cycle, It is the direction difference, which reflects the consistency of the changing direction of adjacent cycles.

[0030] The corresponding change range interval of each field in the period pair composed of two consecutive periods is called in turn, the end values ​​of the interval intersection and union are extracted for length judgment, and the absolute value of the direction difference is used as the calculation element for combined operation, using the formula: ; Taking the main business income as an example, if , , then the intersection is [129.7, 132.9], with a length of 3.2, and the union is [125.4, 132.9], with a length of 7.5. The directions are +1, -1, and the absolute value of the difference is 2. Substituting into the formula, we get: ; Taking the non-current assets field as an example, if the values ​​are 6.2, 6.9, and 7.1, the intervals are [6.2, 6.9] and [6.9, 7.1], the intersection is 0 (the endpoints do not overlap), the direction is rising, that is, +1, -1, and the difference is 0, then the formula is: ; Through the trend overlap of each field, the periodic stability and fluctuation law of the field after switching the path can be judged. The benefit of the formula is that it introduces the interval intersection ratio and the direction consistency to construct a numerical expression, breaking through the limitation of judging only by the trend slope, adapting to nonlinear trend change scenarios, and finally establishing an indicator change trend list for subsequent state path mapping.

[0031] See also Figure 6 , the specific steps for obtaining the real-time analysis results of financial indicators are as follows: S511: Based on the indicator change trend chain table, call the continuous period indicator state sequence corresponding to the current indicator field chain in the path switching node structure, extract the value change trend information, change interval range and period time series index of the field corresponding to each path segment in the continuous period, and generate a period state sequence path set; First, the direction of field value change in each period needs to be parsed from the historical records, and whether the direction of change is continuous and consistent. For example, if the values ​​of a field in periods 1, 2, and 3 are 120, 126, and 132 respectively, then its direction remains positive and consistent, and the number of directional consistency of the field is 2. At the same time, the value range of the field in each period is obtained. For example, the range of the above field can be expressed as [119, 121], [125, 127], and [131, 133] respectively. The minimum and maximum values ​​are taken to form a closed interval, and then the path switch node containing the value corresponding to the current field chain is called. The corresponding node path segment data structure, assuming that the values ​​of the corresponding fields of the path segment in the past three cycles are 118, 124, and 130, then the consistency number of change directions is also 2, and the change intervals are [117, 119], [123, 125], and [129, 131]. The data is matched with the current field data and used for subsequent path screening and judgment. After completing the above extraction, a cycle state sequence path set is established, which specifically records the path nodes, field IDs, cycle time indexes, field value intervals and directions, etc., for subsequent trend comparison and matching evaluation.

[0032] S512: Call each path segment field sequence in the cycle state sequence path set, using the formula: ; Calculating path segments The total value of trend overlap matching , filter all A set of path segments exceeding the path matching threshold is used to generate trend matching path segments; in, represents the number of cycle pairs covered by the path segment, Indicates the number of fields in the path segment, It is The calculation field is in The overlap of the changing trends in the next cycle; The trend matching degree of the path segment is evaluated by the trend coincidence degree calculated in the trend chain table. The calculation premise is that The corresponding field and period can be called from the trend list according to the path segment field order and period segment index matching, assuming that the current path segment contains the field number , the number of period segments , the corresponding coincidence value of each field in two period pairs is as follows: , , , , , , and the calculation is as follows: ; The calculation results show that the path segment The trend matching degree is 0.755. If the state path matching threshold is set to 0.70, the path segment meets the screening requirements and is retained in the result set. The benefit of the formula is that by integrating the directional consistency of the period segment and the interval overlap into a unified evaluation dimension, the path segment screening has the ability to evaluate the continuity of the structural trend. This calculation method avoids repeated intersection calculations, improves execution efficiency, and takes into account the continuity of the change trend.

[0033] S513: Based on the trend matching path segment, select the path segment with the matching degree closest to the upper limit of the path matching threshold, extract the continuous indicator state sequence covered by the target path segment, merge and fuse the field structure in each state segment with the current field chain, integrate them into a continuous trend expression structure, and generate real-time analysis results of financial indicators; Matching value calculated for each path segment in the trend matching path segment value group , filter out the path segments that are close to the upper limit of the state path matching threshold. If multiple path segments meet the threshold requirements, for example, path segment 1 , path segment 2 is 0.781, path segment 3 is 0.794, and the matching upper limit is set to 0.80, then path segment 3 is closest to the upper bound, and it is used as the optimal path segment to call all continuous period state structures under its structural path index, and the field value change trend and arrangement structure in the path segment are spliced ​​and fused with the continuous trend structure of the current indicator field chain to build a unified trend chain covering the start and end period indexes of the path segment, and finally integrated into the real-time trend expression of the current state of the financial indicator under the target archive path node. The result is named the real-time analysis result of the financial indicator. This structure is used to reflect the phased behavior characteristics and short-term trend composition relationship of the indicator based on the historical path extension.

[0034] A financial indicator real-time analysis system, which is used to execute the above-mentioned financial indicator real-time analysis method, and the system comprises: The field density building module obtains the income field, expenditure field, asset field and liability field in the financial statement, builds a field mapping chain, and analyzes the type distribution, sequence density and cross frequency of each field to generate field chain composition density information; The field matching analysis module performs field type matching and sequence combination comparison on the current indicator field chain and the archive path node field chain structure clustered according to the field chain mapping structure of all financial indicators in the current period based on the field chain composition density information, and generates a field chain matching error map; The path switching judgment module selects the path with the lowest error according to the matching error distribution of each archiving path node in the field chain matching error map, judges whether the index archiving path needs to be switched, and obtains the path switching judgment result; The trend chain table generation module collects the numerical sequence of the current indicator in the previous consecutive financial period under the path structure after the switch based on the nodes that have been determined as the path that needs to be switched in the path switching judgment result, and performs consistency judgment of the period difference and interval overlap analysis on each group of fields in the sequence to establish the indicator change trend chain table; The indicator status merging module selects consistent paths with overlapping intervals based on the indicator change trend chain list, merges the status chain, and generates real-time analysis results of financial indicators.

[0035] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A real-time analysis method for financial indicators, characterized in that: The following steps are involved: S1: Obtain the income fields, expenditure fields, asset fields and liability fields in the financial statements, build a field mapping chain, and analyze the type distribution, sequence density and cross frequency of each field to generate field chain composition density information; S2: Based on the field chain composition density information, perform field type matching and sequence combination comparison on the current indicator field chain and the archive path node field chain structure clustered according to the field chain mapping structure of all financial indicators in the current period, and generate a field chain matching error map; S3: According to the matching error distribution of each archiving path node in the field chain matching error map, select the path with the lowest error, judge whether the index archiving path needs to be switched, and obtain the path switching judgment result; S4: Based on the nodes that have been determined as paths that need to be switched in the path switching judgment result, the numerical sequence of the current indicator in the previous continuous financial period under the path structure after switching is collected, and the consistency judgment of the period-to-period difference and the interval overlap analysis are performed on each group of fields in the sequence to establish an indicator change trend chain table; S5: Based on the indicator change trend chain table, consistent paths with overlapping intervals are screened, state chains are merged, and real-time analysis results of financial indicators are generated.

2. The real-time analysis method of financial indicators according to claim 1, characterized in that: The field chain composition density information includes field sequence density, field type crossover frequency, and field type occurrence position. The field chain matching error map specifically includes field overlap rate, sequence deviation level, and matching error distribution. The path switching judgment result includes path adaptation degree, structural path position, and node switching suggestion. The indicator change trend chain table specifically refers to the period value difference, the number of period direction consistency, and the change amplitude interval distribution. The financial indicator real-time analysis results include state path matching degree, period segment consistency, and state chain merge path.

3. The real-time analysis method of financial indicators according to claim 1, characterized in that: The steps for obtaining the field chain composition density information are specifically as follows: S111: Obtain the income field, expenditure field, asset field and liability field in the financial statement, collect the field combination structure included in each financial indicator in the current period, obtain the arrangement order of each field, the number of field types and the location data of the field cross-dense area, organize the fields in order to build a field mapping chain, and obtain the field mapping chain structure sequence; S112: Call the field arrangement order in the field mapping chain structure sequence to determine whether the types of adjacent fields are consistent, using the formula: ； Calculate the Sequential continuity of group field chain structures , which is used to measure the consistency and stability of the order in which financial fields are arranged in the structure; in, Indicates The sequential continuity of the group field chain structure, Indicates The total number of fields in the group field chain, Indicates the number of fields from the first to the All pairs of fields between fields are accumulated. Indicates The first Fields and Field type consistency indicators, where the same is 1 and different is 0. Indicates The position index of the field in the field chain. Indicates the absolute value of the field index spacing. is the normalization factor used to convert the sum to the average between field pairs; S113: Based on the sequential continuity of the field mapping chain structure sequence and the field chain structure, count the field occurrence position and type distribution frequency corresponding to each field type in the field chain structure, identify the cross-appearing field combination segments, obtain the combination characteristics of the cross-type and position, and generate field chain composition density information.

4. The real-time analysis method of financial indicators according to claim 3 is characterized in that: The steps for obtaining the field chain matching error map are specifically as follows: S211: Based on the field chain density information, the current indicator field chain is called, and the archive path node field chain structure obtained by clustering the field chain mapping structure of all financial indicators in the current period is called, the overlap of the field types in the current field chain and each archive path field chain is identified, the fields with overlapping field names are extracted, the number of fields is counted and the corresponding index position offset is compared, and a field matching difference value set is generated; S212: Based on the field matching difference value set, the ratio of the field overlap values ​​in each archive path node to the total number of fields in the current field chain is calculated to obtain the field overlap rate, the index position difference values ​​of each pair of overlapping fields are counted, and divided into multiple offset level intervals, and the field overlap rates of all path nodes are compared and summarized with the offset level results to construct a field chain matching error map.

5. The real-time analysis method of financial indicators according to claim 4, characterized in that: The steps of obtaining the path switching judgment result are specifically as follows: S311: According to the matching error distribution of each archive path node in the field chain matching error map, the field overlap rate and the order deviation level corresponding to each node are identified, and the node chains whose field overlap rate is lower than the field overlap rate threshold are extracted, and the node chains whose order deviation level exceeds the field order deviation threshold are extracted, and the two types of node chains are merged and marked, and the marked path nodes are screened out to obtain a path node elimination result set; S312: Based on the path node elimination result set, call the retained archive path node set, calculate the matching error corresponding to each node, select the single structure path with the lowest matching error, extract the corresponding field chain structure, call the field arrangement order and field combination content in the current indicator field chain, compare the field order structure and field combination distribution in the target node path field chain, and generate archive path matching difference information; S313: Based on the archive path matching difference information, determine the range of change in the arrangement continuity and the magnitude of the field combination difference of the current indicator field chain structure between the original path and the target path, and make a decision on the path switching in combination with the archive path node switching trigger condition, and establish a path switching judgment result representing the switching status identifier between the indicator field chain and the archive node.

6. The real-time analysis method of financial indicators according to claim 5, characterized in that: The steps for obtaining the indicator change trend chain table are specifically as follows: S411: Based on the node determined as the one that needs to be switched in the path switching judgment result, collect the numerical sequence of the current indicator in the first three consecutive financial periods under the path structure of the node, extract the value content of the corresponding field of the indicator in each period, and generate a continuous period sequence of the indicator; S412: Based on the continuous period sequence of the indicator, select the accounting item fields associated with the original field group structure, extract the numerical sequences corresponding to the main business income, total period expenses, increase or decrease of non-current assets and repayment of current liabilities, calculate the difference between two adjacent periods of each field, record the direction of the numerical change of each field within three periods, and generate the statistical results of the field period change; S413: Based on the statistical result of the field periodic change, the value difference of each field between two periods is judged for sign consistency, using the formula: ； Calculate the The calculation field is in The overlap of the changing trends under the period , used to measure the consistency of the change direction and magnitude of the same field in adjacent financial periods, build a field-level trend chain by combining the trend expression results of multiple period pairs, and establish an indicator change trend chain table; in, Indicates The fields in The range of fluctuations in the cycle, Indicates The fields in The range of fluctuations in the cycle, represents the intersection length of two periodic intervals, represents the total length of the union of two period intervals, Indicates The fields in The direction of the value change of each cycle, where +1 indicates an increase, -1 indicates a decrease, and 0 indicates no change. Indicates that the corresponding field is in The direction of change of the cycle, It is the direction difference, which reflects the consistency of the changing direction of adjacent cycles.

7. The real-time analysis method of financial indicators according to claim 6, characterized in that: The steps for obtaining the real-time analysis results of the financial indicators are specifically as follows: S511: Based on the indicator change trend chain table, call the continuous period indicator state sequence corresponding to the current indicator field chain in the path switching node structure, extract the value change trend information, change interval range and period time series index of the field corresponding to each path segment in the continuous period, and generate a period state sequence path set; S512: Call each path segment field sequence in the periodic state sequence path set, using the formula: ； Calculating path segments The total value of trend overlap matching , used to evaluate the consistency between the field change trend of the path segment in multiple cycles and the current indicator path, and filter all A set of path segments exceeding the path matching threshold is used to generate trend matching path segments; in, represents the number of cycle pairs covered by the path segment, Indicates the number of fields in the path segment, It is The calculation field is in The overlap of the changing trends in the next cycle; S513: Based on the trend matching path segment, select the path segment with the matching degree closest to the upper limit of the path matching threshold, extract the continuous indicator state sequence covered by the target path segment, merge and fuse the field structure in each state segment with the current field chain, integrate them into a continuous trend expression structure, and generate real-time analysis results of financial indicators.

8. A real-time analysis system for financial indicators, characterized in that: According to the real-time analysis method of financial indicators according to any one of claims 1 to 7, the system comprises: The field density building module obtains the income field, expenditure field, asset field and liability field in the financial statement, builds a field mapping chain, and analyzes the type distribution, sequence density and cross frequency of each field to generate field chain composition density information; The field matching analysis module performs field type matching and sequence combination comparison on the current indicator field chain and the archive path node field chain structure clustered according to the field chain mapping structure of all financial indicators in the current period based on the field chain composition density information, and generates a field chain matching error map; The path switching judgment module selects the path with the lowest error according to the matching error distribution of each archiving path node in the field chain matching error map, judges whether the index archiving path needs to be switched, and obtains the path switching judgment result; The trend chain table generation module collects the numerical sequence of the current indicator in the previous continuous financial period under the path structure after the switch based on the nodes that have been determined as the paths that need to be switched in the path switch judgment result, performs consistency judgment of the period difference and interval overlap analysis on each group of fields in the sequence, and establishes the indicator change trend chain table; The indicator status merging module selects consistent paths with overlapping intervals based on the indicator change trend chain list, merges the status chains, and generates real-time analysis results of financial indicators.

Images