Financial risk assessment method and system based on prediction index

Abstract

The invention relates to a financial risk assessment method and system based on a prediction index. The method comprises the following steps: generating the prediction index, displaying the prediction index, updating the prediction index and tracing the prediction index, wherein the prediction index comprises a basic prediction index and a calculation prediction index. A basic prediction index is generated through multiple prediction rule models, a calculation prediction index is generated by means of a prediction calculation model based on the basic prediction index, a dependency relationship among index data is established, and a prediction curve graph is drawn and dynamically updated according to real-time data. The defects of a traditional method in the aspects of prediction capability, data real-time performance and index understanding are effectively overcome, and therefore the accuracy, timeliness and operability of financial risk assessment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of financial data analysis, and particularly to a financial risk assessment method and system based on prediction indicators. Background Art

[0002] In the financial field, accurate risk assessment is crucial for the decision-making of financial institutions, enterprises, and investors. With the increasing complexity and variability of the financial market, traditional financial risk assessment methods have gradually revealed many limitations and are difficult to meet the requirements of high efficiency, accuracy, and forward-looking in risk assessment in the current market environment.

[0003] In traditional data analysis methods for financial risk assessment, historical data is mainly relied on for analysis. This approach lacks the effective prediction ability for future trends and cannot capture the potential risks brought by the dynamic changes in the market in a timely manner. For example, when there are sudden policy adjustments, economic fluctuations, or industry changes in the market, the analysis based on historical data is difficult to quickly give an accurate judgment on the future risk situation, making financial practitioners lack a strong decision-making basis when facing uncertainties.

[0004] Existing data analysis and prediction tools, such as some statistical analysis software and business intelligence (BI) platforms, although they can process and analyze data to a certain extent, generally have obvious defects. On the one hand, they often only provide basic statistical analysis functions and lack in-depth logical deduction and multi-model prediction capabilities. In actual financial scenarios, a single analysis model is difficult to comprehensively reflect complex financial phenomena, and different types of financial indicators require diversified calculation models for comprehensive analysis, while these tools cannot meet this requirement. On the other hand, the prediction updates of these tools are not real-time enough. The financial market changes rapidly, and market data may change significantly at any time. If the prediction data cannot be updated in a timely manner, the provided risk assessment results will be out of touch with the actual market situation, leading decision-makers to make wrong judgments based on outdated information and miss the best opportunity to respond to risks or seize opportunities. In addition, the user interfaces of these tools are not intuitive enough, which increases the learning cost and operation difficulty of users, prolongs the time cycle from data acquisition to decision-making, and ultimately leads to a reduction in decision-making efficiency.

[0005] In the current rapidly developing financial market environment, decision-makers urgently need tools that can quickly respond to market changes. However, existing tools often cannot meet the requirements in terms of response speed and accuracy. When an emergency or new risk factor appears in the market, existing tools cannot quickly and accurately assess risks, resulting in financial institutions and investors being difficult to quickly make effective risk response strategies and possibly suffering unnecessary economic losses.

[0006] In summary, existing financial risk assessment methods and tools have deficiencies in aspects such as prediction ability, data real-time nature, and operation experience. There is an urgent need for a financial risk assessment method and system based on prediction indicators to solve these problems, so as to improve the accuracy, timeliness, and operability of financial risk assessment, and provide more reliable decision-making support for financial market participants. Summary of the Invention

[0007] The purpose of the present invention is to provide a financial risk assessment method and system based on prediction indicators to solve the above problems.

[0008] To achieve the above purpose, on the one hand, the present invention provides a financial risk assessment method based on prediction indicators, including the following steps:

[0009] Generate prediction indicators: The prediction indicators include basic prediction indicators and calculated prediction indicators;

[0010] The steps for generating basic prediction indicators include: Select basic indicators and configure prediction time periods; Apply at least one prediction rule model within each prediction time period to generate prediction data; Integrate the prediction data of each time period to form basic prediction indicators, and establish a dependency relationship between the basic prediction indicators and the basic indicators;

[0011] The steps for generating calculated prediction indicators include: Select multiple of the basic prediction indicators, generate calculated prediction indicators through a prediction calculation model; Establish a dependency relationship between the calculated prediction indicators and the basic prediction indicators cited;

[0012] Display prediction indicators: Display the prediction indicators in the form of charts or lists;

[0013] Update prediction indicators: Trigger the update of the basic indicator data in any one of the ways of automatic system update and manual active update, so that the prediction indicators are synchronized with external real-time data;

[0014] Trace back prediction indicators: Starting from the prediction indicators, recursively trace their dependency chain until the basic indicators are found.

[0015] Further, in the financial risk assessment method based on prediction indicators, the prediction rule model includes at least one of a latest value model, a fixed value model, a year-on-year model, a year-on-year difference model, a month-on-month model, a month-on-month difference model, an N-period moving average model, an N-period linear extrapolation value model, a dynamic month-on-month difference model, an interpolation method model after a given final value, a seasonal model, a moving average year-on-year model, a year-on-year growth rate interpolation model, a unary linear fitting model, an N-year average value model, and an annual value backward deduction model.

[0016] Further, in the financial risk assessment method based on prediction indicators, the prediction calculation model includes at least one of a year-on-year ratio calculation model, a year-on-year difference calculation model, an N-value moving average calculation model, a cumulative value conversion to monthly / quarterly value calculation model, an N-value month-on-month ratio calculation model, an N-value month-on-month difference calculation model, an upsampling model, a downsampling model, a time shift model, a super-seasonal model, a fitting residual method model, an annualization model, a diffusion index model, a cumulative value calculation model, a year-to-date calculation method model, and an exponential smoothing model.

[0017] Further, in the financial risk assessment method based on prediction indicators, in the step of displaying prediction indicators, for the chart display, first query and assemble the indicator data, perform the alignment processing of the Gregorian calendar and the Spring Festival for the seasonal chart display, and then configure the options for plotting; for the list display, directly present the data.

[0018] Further, in the financial risk assessment method based on prediction indicators, in the step of updating prediction indicators, when the system is automatically updated, it is triggered according to the set time, and when manually updated, manually click the button to trigger the update of the basic indicator data, and when updating, query the associated indicators, classify and refresh and store the data.

[0019] Further, in the financial risk assessment method based on prediction indicators, in the step of tracing the source of prediction indicators, first determine the node type, recursively construct a tree structure, use the specified plug-in for plotting, and mark the relevant information of the basic indicators.

[0020] Further, in the financial risk assessment method based on prediction indicators, in the step of generating and calculating prediction indicators, the processing methods for the null values of the dates in the basic prediction indicators include: searching forward / backward for the values within the most recent 35 days, filling the null values with 0 values, and skipping the null value dates without participating in the calculation.

[0021] In addition, on the other hand of this embodiment, a financial risk assessment system based on prediction indicators is also proposed, including a prediction indicator generation module, a prediction indicator calculation module, a prediction indicator display module, a prediction indicator update module, and a prediction indicator traceability module;

[0022] The prediction indicator generation module is used to generate basic prediction indicators. The prediction indicator generation module includes a basic indicator acquisition unit, a prediction time period configuration unit, a prediction rule model selection unit, and an integrated prediction data unit; the basic indicator acquisition unit is used to acquire basic indicator data, the prediction time period configuration unit is used to configure the prediction time period for the selected basic indicators, the prediction rule model is used to select at least one from multiple prediction rule models and apply it to the basic indicator data for prediction, and the integrated prediction data unit is used to form basic prediction indicators from the prediction data of each time period and establish the dependency relationship between the basic prediction indicators and the basic indicators;

[0023] The calculation and prediction index module is used to generate calculation and prediction indexes. The calculation and prediction index module includes a prediction index data acquisition unit, a calculation model acquisition unit, and an integrated calculation result unit. The prediction index data acquisition unit is used to acquire multiple pieces of the basic prediction index data. The calculation model acquisition unit is used to acquire a prediction calculation model. The integrated calculation result unit is used to calculate the basic prediction index data through the prediction calculation model, and integrate the calculation results to form a calculation and prediction index, and establish a dependency relationship between the calculation and prediction index and the referenced basic prediction index.

[0024] The prediction index display module is used to display prediction indexes.

[0025] The prediction index update module is used to update prediction indexes.

[0026] The prediction index traceability module is used to trace prediction indexes.

[0027] Further, in the financial risk assessment system based on prediction indexes, the prediction index display module includes a query index data unit, a curve graph drawing unit, a seasonal graph drawing unit, and a list data display unit. The query index data unit is used to query and assemble index data. The curve graph drawing unit and the seasonal graph drawing unit draw graphs according to requirements respectively. The list data display unit is used to present index data.

[0028] The prediction index update module includes a trigger index update unit, a recursive query association unit, a classified refresh index unit, and an update and storage index unit. The trigger index update unit is used to trigger the update of index data. The recursive query association unit is used to query associated index data. The classified refresh index unit is used to classify and refresh index data. The update and storage index unit is used to store index data.

[0029] The prediction index traceability module includes a query and determine node unit, a recursive construct tree structure unit, a draw tree graph unit, and a mark relevant information unit. The query and determine node unit is used to determine nodes. The recursive construct tree structure unit is used to construct a tree. The draw tree graph unit is used to draw a graph. The mark relevant information unit is used to mark index information.

[0030] Further, in the financial risk assessment system based on prediction indexes, the prediction index update module refreshes indexes in the order of basic indexes, basic prediction indexes, and calculation and prediction indexes.

[0031] Compared with the prior art, the present invention has at least the following technical effects:

[0032] A financial risk assessment method and system based on prediction indicators provided by the present invention generate basic prediction indicators through a variety of prediction rule models, generate calculation prediction indicators based on the basic prediction indicators by means of a prediction calculation model, and establish the dependency relationships between index data. The present invention supports the display of prediction indicators in the form of charts or lists. In addition, the update method combines both automatic system update according to the set time and manual trigger update. When updating, query, classification refresh, and storage operations are performed. The present invention supports tracing back from the prediction indicators to the basic indicators recursively, displaying the dependency relationships in a tree diagram and annotating relevant information, effectively solving the deficiencies of traditional methods in terms of prediction ability, data real-time performance, and index understanding. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a flowchart of the financial risk assessment method based on prediction indicators in an embodiment of the present invention;

[0034] Figure 2 It is a logic diagram for generating prediction indicators in the financial risk assessment method based on prediction indicators in an embodiment of the present invention;

[0035] Figure 3 It is a flowchart of the financial risk assessment system based on prediction indicators in another embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] The following will describe in more detail a financial risk assessment method and system based on prediction indicators of the present invention with reference to the schematic diagrams, in which the preferred embodiments of the present invention are shown. It should be understood that those skilled in the art can modify the present invention described herein while still achieving the advantageous effects of the present invention. Therefore, the following description should be understood as a broad guidance for those skilled in the art and not as a limitation to the present invention.

[0037] For clarity, not all features of the actual embodiments are described. In the following description, well-known functions and structures are not described in detail because they would obscure the present invention with unnecessary details. It should be considered that in the development of any actual embodiment, a large number of implementation details must be made to achieve the specific goals of the developer, such as changing from one embodiment to another according to the limitations of the relevant system or business. In addition, it should be considered that such development work may be complex and time-consuming, but it is only routine work for those skilled in the art.

[0038] In the following paragraphs, the present invention will be described more specifically by way of example with reference to the accompanying drawings. According to the following description, the advantages and features of the present invention will be clearer. It should be noted that the drawings are all in a very simplified form and use non-precise scales, only for the purpose of facilitating and clearly assisting in explaining the objectives of the embodiments of the present invention.

[0039] On the one hand, as Figure 1As shown in the figure, this embodiment provides a financial risk assessment method based on prediction indicators, including the following steps:

[0040] S1: Generate prediction indicators

[0041] The prediction indicators include basic prediction indicators and calculated prediction indicators.

[0042] The steps for generating basic prediction indicators in S11 include:

[0043] Select basic indicators and configure the prediction period; apply at least one prediction rule model within each prediction period to generate prediction data; integrate the prediction data of each period to form basic prediction indicators, and establish a dependency relationship between the basic prediction indicators and the basic indicators.

[0044] The steps for generating calculated prediction indicators in S12 include:

[0045] Select multiple of the basic prediction indicators, generate calculated prediction indicators through a prediction calculation model; establish a dependency relationship between the calculated prediction indicators and the basic prediction indicators they reference.

[0046] S2: Display prediction indicators

[0047] Display the prediction indicators in the form of charts or lists.

[0048] S3: Update prediction indicators

[0049] Trigger the update of the basic indicator data in any one of the system automatic update and manual active update methods, so that the prediction indicators are synchronized with external real-time data.

[0050] S4: Trace back prediction indicators

[0051] Starting from the prediction indicators, recursively trace their dependency chain until the basic indicators are found.

[0052] It should be noted that in this embodiment, the index date sequence data refers to the data value (V) of the index on a specific date (T). In addition, there are multiple sets of date sequence data for one index. For example, the value corresponding to the index "Beijing: Zhongyuan Leading Index: Citywide" on July 31, 2024 is: 772, and there is also a corresponding value on June 30, 2024: 785.

[0053] In this embodiment, as Figure 2 shown, for step S1, the prediction indicator generation model specifically includes two modules: a basic prediction indicator generation module and a calculated prediction indicator generation module.

[0054] Specifically, the steps S11 for the basic prediction index generation module to generate basic prediction indexes are as follows: This module conducts model prediction work based on the real data of the basic indexes. First, select one index from numerous indexes in the index library as the basic index, and then configure multiple prediction periods for the selected basic index according to actual requirements. In each prediction period, select at least one model from the prediction rule models and apply it to the basic index data to generate corresponding prediction data. After generating the prediction data for each period, integrate and process these prediction data, and finally form the basic prediction index, and establish a clear dependency relationship between the basic prediction index and the basic index.

[0055] Further, the prediction rule models include at least one of the latest value model, fixed value model, year-on-year model, year-on-year difference model, month-on-month model, month-on-month difference model, N-period moving average model, N-period linear extrapolation value model, dynamic month-on-month difference model, interpolation method model after given the final value, seasonal model, moving average year-on-year model, year-on-year growth rate interpolation model, unary linear fitting model, N-year average value model, and annual value backward deduction model. The following is a detailed description of the prediction rule models:

[0056] 1. Latest value model: The predicted values are all equal to the latest value of the basic index.

[0057] 2. Fixed value model: Given a constant value, the predicted values are default to be all equal to this constant value.

[0058] 3. Year-on-year model: Given a year-on-year growth rate value, multiply the value of the same period last year by this year-on-year growth rate value to get the predicted value.

[0059] 4. Year-on-year difference model: Given a year-on-year increase value, add this year-on-year increase value to the value of the same period last year to get the predicted value.

[0060] 5. Month-on-month model: Given a month-on-month growth rate value, multiply the value of the previous period by this month-on-month growth rate value to get the predicted value.

[0061] 6. Month-on-month difference model: Given a month-on-month increase value, add this month-on-month increase value to the value of the previous period to get the predicted value.

[0062] 7. N-period moving average model: Given an N value representing the number of periods, it is the average value of the past N periods.

[0063] 8. N-period linear extrapolation value model:

[0064] 8.1. Given an N value representing the number of periods, the past N periods' values of the basic index are used as the Y value list [V0, V1, V2,..VN], and the number of periods is used as the X value list [1, 2, 3,...N], and substitute them into the linear regression equation: Y = aX + b to obtain the values of coefficients a and b.

[0065] 8.2. Substitute the future period number N+i (where i represents the first predicted period, the second period, etc.) as the value of X into the equation to obtain the predicted value Y = aX + b.

[0066] 9. Dynamic loop difference:

[0067] 9.1. Select several indicators for indicator calculation. For example, select indicator A and indicator B, configure a mathematical formula such as A + B, and substitute the contemporaneous values of indicator A and indicator B into the formula to calculate the resulting value as the dynamic loop difference for this period.

[0068] 9.2. Predicted value for this period = value of last period + dynamic loop difference.

[0069] 10. Interpolation after specifying the final value:

[0070] 10.1. Given a predicted final value Vn, calculate the period difference T and data difference S between the latest actual value V1 and the predicted final value, where T = n - 1 and S = Vn - V1.

[0071] 10.2. Loop difference = S / T * data frequency, where the data frequency is related to the indicator frequency. Daily = 1, Weekly = 7, Decade = 10, Monthly = 30, Quarterly = 90, Yearly = 365; Predicted value = value of last period + loop difference.

[0072] 11. Seasonality:

[0073] 11.1. Obtain the year list in two ways:

[0074] For N consecutive years, i.e., the current year Y, Y - 1... Y - N to obtain the year list [Y0, Y1, Y2...].

[0075] Specify N years, and obtain the year list [Y0, Y1, Y2...] according to the years selected by the user.

[0076] According to the year list, calculate the average loop difference (this period - last period) for the same period in the past N years = ((Vy0 - V’y0) + (Vy1 - Vy1)) / N, where N represents the number of items in the year list, and (Vy0 - V’y0) represents the difference between the value of this period and the value of the last period for each year. Predicted value for this period = value of last period + average loop difference.

[0077] 12. Moving average year-on-year:

[0078] 12.1. Given a value N representing the number of periods, and given a year Y as the year-on-year reference year.

[0079] 12.2. Calculate the year-on-year value (this period / last period) of the average value for the past N periods:

[0080] N-period average value = (v0+v1+vN-1+...+vN) / N, where V0 represents the current period value and vN represents the value of the previous N periods.

[0081] Find the value vY of the same period of the given year Y. If it cannot be found, look for data from 35 days before and after, and find the most recent value as vY.

[0082] 12.3. The year-on-year value of the average value of N periods = the average value of N periods / vY.

[0083] 12.4. Select the year with the same period as the previous year, and the forecast value for this period = the same period value of the same year with the same period as the previous year * the year-on-year value of the average value of N periods.

[0084] 13. Interpolation of year-on-year growth rate:

[0085] 13.1. Calculate the year-on-year growth rate of the latest value of the basic indicator: (latest value - value of the same period last year) / value of the same period last year * 100%.

[0086] 13.2. Given a final value of year-on-year growth rate, according to the final value of year-on-year growth rate (last term) on the forecast deadline, the year-on-year growth rate of the latest value of the basic indicator (first term), and the number of periods n from the latest date to the deadline, the tolerance d = (last term - first term) / (n-1) is obtained according to the arithmetic progression formula.

[0087] Calculate the year-on-year growth rate of each period up to the deadline an=a1+(n-1)d, where a1 is the year-on-year growth rate of the latest value of the basic indicator.

[0088] 13.3. The forecast value for this period = the value of the same period last year * (1 + the year-on-year growth rate for this period).

[0089] 14. Univariate linear fitting:

[0090] 14.1. Select an indicator A (independent variable) and indicator B (basic indicator), select a fitting time period, and set the time (days) that indicator A leads indicator B.

[0091] 14.2. Linear regression equation: Y = aX + b, substitute the data of indicator A into the equation to get coefficients a and b. Substitute a and b into the formula to get the fitting indicator B', concatenate the actual value of indicator B and the predicted value of B' to generate the predicted indicator.

[0092] 15.N-year average: the average of the same period in the past N years.

[0093] 15.1. The past N years can be continuous or discontinuous. There are two ways to get the year list:

[0094] N consecutive years, i.e. the current year Y, Y-1...YN, get the year list [Y0, Y1, Y2...].

[0095] Specify N years, and obtain the year list [Y0, Y1, Y2,...] according to the year selected by the user.

[0096] 15.2. The indicator data are all complemented to daily data by linear interpolation:

[0097] (a) Calculate the number of days between two adjacent periods of indicators as the d value.

[0098] (b) Take the two adjacent period indicator values as the Y value list [Y0, Yd], and the number of days as the X value list [1, d], and substitute them into the linear equation Y = aX + b to obtain the a value and the b value.

[0099] (c) Take the number of days of the date corresponding to the null value relative to the date of the first indicator in two adjacent periods of indicators as the X value, and substitute the null value Y into the formula Y = aX + b for calculation.

[0100] 15.3. Calculate the mean value: Predicted value = sum of the same period values of all years / number of the year list.

[0101] 16. Annual value backtracking:

[0102] 16.1. Set the annual value, Balance = Annual value - year-to-date cumulative value of the latest value (the algorithm refers to the cumulative value), and perform balance allocation.

[0103] 16.2. Ensure that the values of each period are equal during the mean method allocation:

[0104] Daily / weekly: Remaining periods = Remaining natural calendar days / Natural calendar days of the latest date of this year's indicators * Number of periods of this year's indicators to date.

[0105] Decadal / monthly / quarterly / semi-annual: Remaining periods = Number of periods in a year (36 / 12 / 4 / 2) - Natural calendar periods to date this year

[0106] Predicted value = Balance / Remaining periods.

[0107] 16.3. Ensure that the year-on-year growth rate of each period is equal during the year-on-year method:

[0108] Specify a year-on-year year.

[0109] Year-on-year growth rate = Balance / Balance of the corresponding date in the year-on-year year.

[0110] Predicted value = Value of the same period in the year-on-year year * Year-on-year growth rate.

[0111] In addition, the steps S12 for the calculation prediction index generation module to generate the calculation prediction index are specifically as follows: Based on the basic prediction index, this module comprehensively uses real data and prediction data and conducts in-depth prediction with the aid of a prediction calculation model. In specific operations, the basic prediction index can only predict the future trend direction of the current index. However, in reality, the formation of a certain view is often the combined effect of multiple different types of indexes in various model calculations. Therefore, finally, the calculated prediction index data needs to be integrated to form the calculation prediction index data. First, select multiple generated basic prediction indexes, and then call the prediction calculation model to perform arithmetic processing on the basic prediction index data. Through the accurate calculation of the prediction calculation model, the calculation prediction index is generated, and the dependency relationship between the calculation prediction index and the referenced basic prediction index is clearly established, providing more comprehensive and in-depth data support for subsequent financial risk assessment.

[0112] Furthermore, the prediction calculation model includes at least one of a year-on-year value calculation model, a year-on-year difference calculation model, an N-value moving average calculation model, a cumulative value conversion to monthly / quarterly value calculation model, an N-value month-on-month ratio calculation model, an N-value month-on-month difference calculation model, an upsampling model, a downsampling model, a time shift model, a super-seasonal model, a fitting residual method model, an annualization model, a diffusion index model, a cumulative value calculation model, a year-to-date calculation method model, and an exponential smoothing model. The following is a detailed description of the prediction calculation model:

[0113] 1. Index operation: The researcher uses the data of multiple indexes to perform mathematical operations to obtain corresponding results and generate index Z.

[0114] 1.1. The researcher first selects a series of relevant basic prediction indexes, including index A, index B, index C, etc.

[0115] 1.2. Confirm the date sequence of the to-be-generated index Z. The model supports two specified methods:

[0116] (a) Specify the time sequence of a certain index participating in the calculation.

[0117] (b) The union of the time sequences of all indexes participating in the operation.

[0118] 1.3. Determine the null value handling method:

[0119] (a) Find the nearest value within 35 days before and after the current date of the index: When a certain index has no value in the date sequence participating in the calculation, find the nearest value before / after this index as the value of the current day for calculation. The traversal allows crossing years, up to 35 days forward and up to 35 days backward.

[0120] (b) The current date of the indicator is not calculated: As long as one indicator has no value (i.e., a null value) on a certain date, it is considered that the indicator has no value on that date.

[0121] (c) Filling the previous value for the current date of the indicator: Null values are preferentially filled with the most recent previous value. When there is no previous value, they are filled with the subsequent value.

[0122] (d) Filling the subsequent value for the current date of the indicator: Null values are preferentially filled with the most recent subsequent value. When there is no previous value, they are filled with the subsequent value.

[0123] (e) The current date of the indicator equals 0: Null values are involved in the calculation with a value of 0.

[0124] 1.4. Input the calculation formula, such as: A + B + C.

[0125] 1.5. The model, based on the date sequence data of the selected indicators A, B, and C, obtains the corresponding data v of each indicator at date t under the date sequence determined in 1.2. If the v of a certain indicator is a null value, corresponding processing is carried out according to the null value processing method in 1.3. After finally obtaining the relevant data, mathematical operations are performed according to the calculation formula input in 1.4, and finally the predicted value for that date is calculated.

[0126] 2. Year-on-year ratio: For a certain indicator, perform year-on-year calculation, and the calculation formula is this year's same period / last year's same period - 1.

[0127] 2.1. Lock the current date t of the indicator.

[0128] 2.2. Match the same period of the previous year: Among the corresponding dates (t - 1 year) with values in the past year, find the day that is equal to or closest to the current value's corresponding date. If there are two dates with the same number of days as the current value's corresponding date, take the first date in the descending date order.

[0129] 2.3. Obtain the value corresponding to the matched same period of the previous year.

[0130] 2.4. Calculate the ratio of the current period's value to the value of the matched same period of the previous year.

[0131] 2.5. Fill the calculated ratio into the date corresponding to the current value to generate a new data sequence.

[0132] 2.6. Traversal allows crossing years. For daily / weekly / quarterly data, traverse up to 35 days forward and up to 35 days backward.

[0133] 3. Year-on-year difference: For a certain indicator, perform year-on-year difference calculation, and the calculation formula is this year's same period - last year's same period.

[0134] 3.1. Lock the current date t of the indicator.

[0135] 3.2. Matching the same period of the previous year: Among the corresponding dates with values in the past year (t-1 year), find the day that is equal to or closest to the current value's corresponding date. If there are two dates with the same number of days as the current value's corresponding date, take the first date in the descending order of dates.

[0136] 3.3. Obtain the value corresponding to the matched same period of the previous year.

[0137] 3.4. Calculate the difference between the current value and the value of the matched same period of the previous year.

[0138] 3.5. Fill the calculated difference into the date corresponding to the current value to generate a new data series.

[0139] 3.6. Traversal allows crossing years. For daily / weekly / quarterly data, traverse up to 35 days forward and up to 35 days backward.

[0140] 4. N-value moving average calculation: The calculation formula is AVERAGE (sum of N values), where N is the number of values taken.

[0141] 4.1. Calculate the N-value moving average of the current value. Then trace back N values (including the current value) in time. If a date with a blank value is encountered, automatically skip the blank value and continue tracing back.

[0142] 4.2. If the original data corresponding to the current date is a null value, the N-value moving average is also a null value.

[0143] 5. Convert cumulative values to monthly / quarterly values.

[0144] 5.1. Calculation method for converting cumulative values to monthly values:

[0145] No value in January: January = February / 2. Value in January: January = January;

[0146] February = February / 2. February = February - January;

[0147] March = March - February. March = March - February;

[0148] April = April - March. April = April - March;

[0149] And so on. If there is no value in both January and February, no calculation is done for that year.

[0150] 5.2. Calculation method for converting cumulative values to quarterly values:

[0151] No value in the first quarter: The first quarter = The second quarter / 2. Value in the first quarter: The first quarter = The first quarter;

[0152] The second quarter = The second quarter / 2. The second quarter = The second quarter - The first quarter;

[0153] The third quarter = the third quarter - the second quarter; The third quarter = the third quarter - the second quarter;

[0154] The fourth quarter = the fourth quarter - the third quarter; The fourth quarter = the fourth quarter - the third quarter;

[0155] And so on. If there is no value in both the first quarter and the second quarter, no calculation will be done for that year.

[0156] 6. N value month-on-month ratio: For a certain indicator, calculate the N value month-on-month ratio. The calculation formula is (current period - previous period) / previous period.

[0157] 6.1. Select the corresponding value for the current period. If the current period value is empty, the month-on-month ratio is empty; if the current period has a value, traverse forward to query the previous period value. By default, N is equal to 1, and find the nearest previous period value; you can select the Nth nearest previous period value according to the set N value.

[0158] 6.2. Calculate the current period value and the matched previous period value according to the formula.

[0159] 6.3. Fill the calculated value into the date corresponding to the current period value to generate a new data series.

[0160] 6.4. When 0 and negative values appear in the original data, prompt that the month-on-month operation cannot be performed for this indicator.

[0161] 7. N value month-on-month difference: For a certain indicator, calculate the N value month-on-month difference. The calculation formula is current period - previous period.

[0162] 7.1. Select the corresponding value for the current period. If the current period value is empty, the month-on-month difference is empty; if the current period has a value, traverse forward to query the previous period value. By default, N is equal to 1, and find the nearest previous period value; you can select the Nth nearest previous period value according to the set N value to find the nearest previous period value.

[0163] 7.2. Calculate the current period value and the matched previous period value according to the formula.

[0164] 7.3. Fill the calculated value into the date corresponding to the current period value to generate a new data series.

[0165] 8. Upsampling: Support converting all frequency indicators to daily indicators. If converting a monthly indicator to a daily indicator, for example, if the current month is September but the September data has not been updated yet, assuming the September data is updated on September 30, then the data from September 1 to September 29 is equal to the data on August 31; the same applies to other frequency rules.

[0166] 9. Downsampling: Convert high-frequency indicators to low-frequency indicators.

[0167] 9.1. Select the frequency of conversion. For daily data, it can be downsampled to weekly, dekadal, monthly, quarterly, or annual; for weekly data, it can be downsampled to dekadal, monthly, quarterly, or annual; for dekadal data, it can be downsampled to monthly, quarterly, or annual; for monthly data, it can be downsampled to quarterly or annual; for quarterly data, it can be downsampled to annual.

[0168] 9.2. Data date after downsampling. For the downsampled data series of weekly data, the date is taken as every Friday; for dekadal data, it is taken as the 10th, 20th, and the last day of each month; for monthly data, it is taken as the last day of each month, for quarterly data, it is taken as the last day of the quarter, and for annual data, it is taken as the last day of the year.

[0169] 9.3. There are two options for the value of data points: a. End value, which takes the data value of the last date in the interval. b. Take the average value of the interval.

[0170] 9.4. Handling of the latest value: When the latest high-frequency data is exactly between the two dates of the low frequency to which it is to be downsampled. For example, if it is currently January 3rd and there is already data for January but January 31st has not arrived yet, then when downsampling to monthly frequency, the latest data date is January 31st.

[0171] 10. Time shift: Add (lead) or subtract (lag) the corresponding time to the time series of data to form a new data series.

[0172] 10.1. Lead by 10 days, that is, add 10 days to the date of each data point of the indicator. The original data point (2022-1-1, 100) will be converted to (2022-1-11, 100).

[0173] 10.2. Lag by 1 month, that is, subtract 30 days from the date. The original data point (2022-1-1, 100) will be converted to (2021-12-2, 100).

[0174] 11. Super-seasonality: For a certain indicator, perform super-seasonality calculation. The calculation formula is the current value - AVERAGE (the sum of the values in the same period in the past N years), where N is the number of data points taken.

[0175] 11.1. Calculate the sum of the values in the same period in the past N years, and then trace back N years (including the latest year) in time.

[0176] 11.2. The indicator data participating in the calculation is complemented to daily data (including weekends) through a linear equation.

[0177] 11.3. When performing lunar calculations, only calculate from November to May of the following year. If the current month is December, then take the Spring Festival of the second year as the time displacement standard point.

[0178] 11.4. The data frequency of the calculated result is consistent with the original indicator frequency.

[0179] 12. Fitting Residual: Calculate the difference between the actual value and the fitted value (B’) of an indicator (B). The fitted value B’ is obtained by linearly regressing the indicator A (independent variable) and the indicator B (dependent variable).

[0180] 12.1. Based on the indicator A (independent variable) and the indicator B (dependent variable) over a past time period (this N periods can be N periods backward from the latest value, including the latest, or it can be data within a selected time period of historical data), generate a linear regression equation Y = aX + b.

[0181] 12.2. From the indicator A (independent variable) and the coefficients a, b of the fitting equation, calculate the fitted series B’ = aA + b, and then calculate the difference between the fitted series B’ and the original series B to obtain a new data series Delta, Delta = B - B'.

[0182] 13. Annualization: For a certain indicator, perform annualization calculation. The calculation formula is Annualized Value = S / a (S represents the indicator value, a represents the average annual proportion).

[0183] 13.1. Read the date T corresponding to the latest value of the indicator and the indicator value S.

[0184] 13.2. Calculate the ratio of the value corresponding to the date T of the indicator in the past three years to the value corresponding to the last date of the current year, that is, the proportion of the cumulative value up to date T in the annual value.

[0185] 13.3. Calculate the average proportion of the three years, that is, the average annual proportion a in the past three years.

[0186] 13.4. If the historical data is less than three years, calculate at least for two years. If it is less than two years, no annualized value is generated.

[0187] 13.5. If there is no value for a certain year at date T, calculate the value at date T by linear interpolation [(Y - Y1) / (X - X1) = (Y2 - Y1) / (X2 - X1)] using the two values before and after date T.

[0188] 14. Diffusion Index: For a certain indicator, perform diffusion index calculation. The calculation formula is AVERAGE (sum of the selected indicator's period-on-period difference indices).

[0189] 14.1. Select multiple indicators and set the union of the diffusion index dates.

[0190] 14.2. Fill in the missing values of the selected indicators within the union of dates with the previous period values.

[0191] 14.3. Calculate the period-on-period difference index (if the period-on-period difference > 0, take 1; if the period-on-period difference = 0, take 0.5; if the period-on-period difference < 0, take 0) for each period of the selected indicators within the union of dates, and calculate the average of the period-on-period difference indices.

[0192] 15. Cumulative value:

[0193] 15.1. Daily to weekly conversion: Select Friday for the date, calculate the sum of the daily values from last Saturday to this Friday, and the latest date is the Friday corresponding to the latest value.

[0194] 15.2. Daily to monthly conversion: Select the last day of each month for the date, calculate the sum of all daily values in the month, and the latest date is the last day of the month corresponding to the latest value.

[0195] 15.3. Daily to quarterly / annual conversion: The method is similar to the monthly conversion.

[0196] 15.4. Weekly to monthly / quarterly / annual conversion: Convert the weekly values to daily values, interpolate the null values using the interpolation method, calculate the sum of all values in the month / quarter / year, and then divide by 7.

[0197] 15.5. Monthly to quarterly / annual conversion: Add up the monthly values in the quarter / year.

[0198] 15.6. And so on. Note: Decadal indicators can be converted to lower-frequency indicators, and higher-frequency indicators cannot be converted to decadal.

[0199] 16. Year-to-date calculation method

[0200] 16.1. Year-to-date for daily data: The date is the same as the original daily data. Sum up the daily values from January 1st (inclusive) of each year to the date of the daily data (inclusive).

[0201] 16.2. Year-to-date for weekly data: The date is the same as the original weekly data. Convert the weekly values to daily frequency, interpolate the null values using the interpolation method, then the algorithm is the same as the daily year-to-date, and then divide by 7.

[0202] 16.3. Year-to-date for monthly / quarterly data: The date is the same as the original monthly / quarterly data. Sum up the monthly / quarterly values from January 1st (inclusive) of each year to the date of the monthly data (inclusive).

[0203] 17. Exponential smoothing:

[0204] 17.1. Set the initial value of the exponential smoothing value sequence = the initial value of the original time series.

[0205] 17.2. Select the smoothing coefficient alpha value: Between 0 and 1, open interval.

[0206] 17.3. The current exponential smoothing value = alpha * the current actual value + (1 - alpha) * the previous exponential smoothing value.

[0207] In summary, by constructing a comprehensive and flexible prediction configuration process, a rich variety of prediction rules are provided. Users can freely select basic indicators and set prediction intervals according to actual needs. These prediction rule models at different levels cooperate with each other, can fully adapt to various complex and changeable scenarios in the financial market, and effectively improve the accuracy and applicability of prediction results. Whether it is short-term market volatility prediction or long-term trend analysis, reliable data support can be provided for users. For example, when predicting the stock price trend, the seasonal adjustment model can be combined to consider the impact of industry seasonal factors on the stock price, and at the same time, the linear fitting model can be used to analyze the relationship between the company's financial indicators and the stock price, so as to more accurately predict the future trend of the stock price.

[0208] Furthermore, for the display of prediction indicators in step S2: Support two ways of displaying prediction indicators, namely charts or lists.

[0209] It should be noted that after configuring the prediction model of prediction indicators on the left side of the page, the generated prediction indicator data will be displayed in the form of a chart on the right side of the page. The form of the chart can not only help users better understand the data distribution, trends and changes, but also quickly identify key information and potential problems in the data.

[0210] Specifically, there are two chart type display methods for the indicator data: line chart and seasonal chart. The following are the steps for drawing line charts and seasonal charts:

[0211] 1. Query indicator data: Query indicator data according to conditions such as indicator ID or indicator code and time range. Each period of indicator data contains a date and a data value. Assemble the queried indicator data into an array.

[0212] 2. Process the indicator data of the seasonal chart:

[0213] 2.1 Gregorian calendar seasonality: Organize the indicator data into multiple arrays according to years and render them on the chart. Multiple arrays will render multiple curves.

[0214] 2.2 Spring Festival alignment seasonality:

[0215] (a) Determine the Gregorian calendar corresponding to the latest alignment of the indicator and denote it as date: If the latest date of the indicator is less than November, obtain the Gregorian calendar corresponding to this year's Spring Festival; otherwise, obtain the Gregorian calendar corresponding to next year's Spring Festival.

[0216] (b) Obtain the Gregorian calendar corresponding to the Spring Festival four years ago and denote it as date4. Calculate the number of days between date and date4 and denote it as days. Add the number of days difference days to all the indicator data dates to get new indicator dates and put them into the array corresponding to four years ago.

[0217] (c) And so on, calculate the data of three, two, and one year ago respectively, and finally form multiple arrays.

[0218] 3. Draw line charts and seasonal charts: Configure Highcharts options and set the chart type to line chart or seasonal chart. Use the dates in the array as the X-axis display and the data values as the Y-axis display.

[0219] Furthermore, for the update of the prediction indicators in step S3: The system will update the prediction indicators regularly according to the configured update time.

[0220] Specifically, according to different triggering methods, it is mainly divided into two types: automatic system update and user-initiated update. The following is the basic process of describing the update of prediction indicators:

[0221] (a) Trigger mechanism: Trigger the update according to the time set by the system (such as 19:00 every day) or when the user clicks the update button of the indicator.

[0222] (b) Query indicators: First, query the information of the indicators to be updated from the database.

[0223] (c) Recursive query: Subsequently, recursively query all other indicators associated with these indicators to ensure the integrity of the dependency chain.

[0224] (d) Classification processing: Classify all the queried indicators into three categories: basic indicators, basic prediction indicators, and calculated prediction indicators. Among them, basic indicators refer to the original data indicators directly obtained from the data source; basic prediction indicators refer to the indicators obtained through the prediction model based on the basic indicators; calculated prediction indicators: prediction indicators obtained through complex algorithms based on multiple basic prediction indicators.

[0225] (e) Refresh in sequence: According to the classification order, first refresh the basic indicators, then refresh the basic prediction indicators, and finally refresh the calculated prediction indicators.

[0226] (f) Update storage: Store the updated indicator data in the database for subsequent query and analysis.

[0227] Therefore, steps S2 and S3 are to ensure that the predicted values can respond to the update of new data in real time. When the market data changes, the system automatically adjusts the predicted values based on the latest data, so that the prediction results always closely fit the actual market situation. This mechanism can effectively avoid the prediction deviation problem caused by data lag in traditional prediction methods, and greatly improve the timeliness and reliability of prediction. Taking exchange rate prediction as an example, in the case of the rapidly changing international political and economic situation, once new macroeconomic data is released, the system can quickly incorporate it into the calculation and correct the exchange rate prediction value, providing a more timely and accurate basis for foreign exchange trading decisions.

[0228] Further, for the traceability prediction index in step S4: Starting from the prediction index, recursively trace its dependency chain until the basic index is found.

[0229] Specifically, index traceability refers to the process of tracking and tracing the source of a specific index. By recursively tracing the dependency chain of a prediction index, one can start from the prediction index and query each dependent basic index level by level. In this process, each step delves deeper into the next level of dependent indexes until the basic indexes that can no longer be decomposed are reached. Subsequently, these hierarchical dependency relationships are organized into an intuitive tree diagram, which should clearly display the complete path and hierarchical structure from the top-level prediction index to the bottom-level basic index, facilitating the understanding of the interdependency relationships between the indexes.

[0230] The following are the steps and key points on how to display this dependency chain using a tree diagram:

[0231] 1. Determine the nodes

[0232] (1) Top-level node: The prediction index, which is the starting point you want to trace.

[0233] (2) Intermediate nodes: Indexes that are dependent on the top-level node or other intermediate nodes.

[0234] (3) Basic nodes (leaf nodes): Have no dependencies and are located at the bottom layer of the tree.

[0235] 2. Construct the tree structure

[0236] (1) Take the index to be queried as a node data, query the list of indexes directly dependent on this index, and add the query results as the child nodes of this node to the tree.

[0237] (2) Traverse each index in this index list and recursively perform this process until all basic nodes are added to the tree.

[0238] 3. Draw the tree diagram

[0239] (1) Use the 'antv / x6' graphics plugin to draw the tree diagram.

[0240] (2) Place the top-level node at the top of the tree, and then add child nodes layer by layer downward.

[0241] (3) Use lines to connect the parent nodes and child nodes to represent their dependency relationships.

[0242] 4. Annotate information

[0243] Annotate the name of the index next to each node. Add other relevant information inside or next to the node, such as calculation formulas, data sources, etc.

[0244] In another embodiment, as Figure 2 shown, a financial risk assessment system based on prediction indicators is also proposed, including a prediction indicator generation module, a prediction indicator calculation module, a prediction indicator display module, a prediction indicator update module, and a prediction indicator traceability module;

[0245] The prediction indicator generation module is used to generate basic prediction indicators. The prediction indicator generation module includes a basic indicator acquisition unit, a prediction time period configuration unit, a prediction rule model selection unit, and an integrated prediction data unit. The basic indicator acquisition unit is used to acquire basic indicator data. The prediction time period configuration unit is used to configure a prediction time period for the selected basic indicators. The prediction rule model is used to select at least one from multiple prediction rule models and apply it to the basic indicator data for prediction. The integrated prediction data unit is used to form basic prediction indicators from the prediction data of each time period and establish the dependency relationship between the basic prediction indicators and the basic indicators.

[0246] The prediction indicator calculation module is used to generate calculated prediction indicators. The prediction indicator calculation module includes a prediction indicator data acquisition unit, a calculation model acquisition unit, and an integrated calculation result unit. The prediction indicator data acquisition unit is used to acquire multiple pieces of the basic prediction indicator data. The calculation model acquisition unit is used to acquire a prediction calculation model. The integrated calculation result unit is used to calculate the basic prediction indicator data through the prediction calculation model and integrate the calculation results to form calculated prediction indicators, and establish the dependency relationship between the calculated prediction indicators and the referenced basic prediction indicators.

[0247] The prediction indicator display module is used to display prediction indicators. The prediction indicator display module includes a query indicator data unit, a curve graph drawing unit, a seasonal graph drawing unit, and a list data display unit. The query indicator data unit is used to query and assemble indicator data. The curve graph drawing unit and the seasonal graph drawing unit draw graphs according to requirements respectively. The list data display unit is used to present indicator data.

[0248] The prediction indicator update module is used to update prediction indicators. The prediction indicator update module includes a trigger indicator update unit, a recursive query association unit, a classified refresh indicator unit, and an update storage indicator unit. The trigger indicator update unit is used to trigger the update of indicator data. The recursive query association unit is used to query associated indicator data. The classified refresh indicator unit is used to classify and refresh indicator data. The update storage indicator unit is used to store indicator data.

[0249] The prediction index traceability module is used to trace the prediction index. The prediction index traceability module includes a query and determination node unit, a recursive tree structure construction unit, a tree diagram drawing unit, and a relevant information annotation unit; the query and determination node unit is used to determine nodes, the recursive tree structure construction unit is used to construct a tree, the tree diagram drawing unit is used to draw a diagram, and the relevant information annotation unit is used to annotate index information.

[0250] Further, in the financial risk assessment system based on prediction indexes, the prediction index update module refreshes the indexes in the order of basic indexes, basic prediction indexes, and calculated prediction indexes.

[0251] Further, in the financial risk assessment system based on prediction indexes, the prediction index traceability module displays the dependency relationship from the top-level prediction index to the bottom-level basic index through a tree diagram, and annotates the calculation formula and data source of each node.

[0252] In summary, in a financial risk assessment method and system provided by an embodiment of the present invention, basic prediction indexes are generated through multiple prediction rule models, calculated prediction indexes are generated based on the basic prediction indexes by means of a prediction calculation model, and a dependency relationship between index data is established. The present invention supports displaying prediction indexes in the form of charts or lists. In addition, the update method has both automatic system update according to the set time and manual trigger update. When updating, query, classification refresh, and storage operations are performed. The present invention can also recursively trace from the prediction index to the basic index, display the dependency relationship in the form of a tree diagram, and annotate relevant information, effectively solving the deficiencies of traditional methods in terms of prediction ability, data real-time performance, and index understanding.

[0253] The above is only a preferred embodiment of the present invention and does not impose any limitation on the present invention. Any person skilled in the art within the technical field of the present invention, without departing from the technical solution of the present invention, makes any form of equivalent replacement or modification and other changes to the technical solution and technical content disclosed by the present invention, all of which belong to the content without departing from the technical solution of the present invention and still fall within the protection scope of the present invention.

Claims

1. A financial risk assessment method based on prediction indicators, characterized in that: The following steps are involved: Generate prediction indicators: prediction indicators include basic prediction indicators and calculated prediction indicators; The steps of generating basic prediction indicators include: selecting basic indicators and configuring prediction periods; applying at least one prediction rule model to generate prediction data in each prediction period; integrating the prediction data of each period to form basic prediction indicators, and establishing a dependency relationship between the basic prediction indicators and the basic indicators; The step of generating a calculation prediction index comprises: selecting a plurality of the basic prediction indexes, generating a calculation prediction index through a prediction calculation model; establishing a dependency relationship between the calculation prediction index and the referenced basic prediction index; Display forecast indicators: Display the forecast indicators in the form of charts or lists; Update the prediction index: trigger the update of the basic index data by any one of automatic system update and manual active update, so that the prediction index is synchronized with the external real-time data; Tracing the predicted indicator: Starting from the predicted indicator, recursively trace its dependency chain until the basic indicator is found.

2. The financial risk assessment method based on prediction indicators according to claim 1 is characterized in that: The prediction rule model includes at least one of the latest value model, fixed value model, year-on-year model, same difference model, month-on-month model, ring difference model, N-period moving average model, N-period linear extrapolation value model, dynamic ring difference model, given terminal value post-interpolation model, seasonal model, moving average year-on-year model, year-on-year growth rate interpolation model, univariate linear fitting model, N-year average model and annual value reverse model.

3. The financial risk assessment method based on prediction indicators according to claim 1 is characterized in that: The prediction calculation model includes at least one of a year-on-year value calculation model, a same difference value calculation model, an N-value moving average calculation model, a cumulative value to monthly / quarterly value calculation model, an N-value month-on-month value calculation model, an N-value month-on-month value calculation model, an N-value month-on-month difference value calculation model, an up-frequency model, a down-frequency model, a time shift model, a super seasonal model, a fitting residual method model, an annualized model, a diffusion index model, a cumulative value calculation model, a year-to-date calculation method model and an index smoothing model.

4. The financial risk assessment method based on prediction indicators according to claim 1, characterized in that: In the step of displaying forecast indicators, the chart display first queries and assembles the indicator data, the seasonal chart display aligns the Gregorian calendar and the Spring Festival, and then configures the options for drawing; the list display directly presents the data.

5. The financial risk assessment method based on prediction indicators according to claim 1, characterized in that: In the step of updating the prediction index, the system automatically updates at a set time, and manually updates by clicking a button to trigger the update of the basic index data. When updating, the related indexes are queried, the classification is refreshed, and the data is stored.

6. The financial risk assessment method based on prediction indicators according to claim 1, characterized in that: In the step of tracing and predicting indicators, the node type is first determined, the tree structure is recursively constructed, the drawing is done using the specified plug-in, and the relevant information of the basic indicators is annotated.

7. The financial risk assessment method based on prediction indicators according to claim 1, characterized in that: In the step of generating and calculating the forecast index, the processing method for the date in the basic forecast index being a null value includes: searching forward / backward for values ​​within the last 35 days, filling null values ​​with 0, and skipping null value dates from participating in the calculation.

8. A financial risk assessment system based on prediction indicators, characterized in that: It includes a prediction indicator generation module, a prediction indicator calculation module, a prediction indicator display module, a prediction indicator update module and a prediction indicator tracing module; The prediction indicator generation module is used to generate basic prediction indicators. The prediction indicator generation module includes a basic indicator acquisition unit, a prediction time period configuration unit, a prediction rule model selection unit, and a prediction data integration unit; the basic indicator acquisition unit is used to acquire basic indicator data, the prediction time period configuration unit is used to configure a prediction time period for a selected basic indicator, the prediction rule model is used to select at least one from a plurality of prediction rule models to be applied to basic indicator data for prediction, and the prediction data integration unit is used to form basic prediction indicators from prediction data of each time period, and establish a dependency relationship between the basic prediction indicators and the basic indicators; The calculation prediction index module is used to generate the calculation prediction index, and the calculation prediction index module includes a prediction index data acquisition unit, a calculation model acquisition unit and a calculation result integration unit; the prediction index data acquisition unit is used to acquire a plurality of the basic prediction index data, the calculation model acquisition unit is used to acquire the prediction calculation model, and the calculation result integration unit is used to calculate the basic prediction index data through the prediction calculation model, integrate the calculation results to form the calculation prediction index, and establish a dependency relationship between the calculation prediction index and the referenced basic prediction index; The prediction indicator display module is used to display the prediction indicator; The prediction index updating module is used to update the prediction index; The prediction indicator tracing module is used to trace the prediction indicator.

9. The financial risk assessment system based on prediction indicators according to claim 8, characterized in that: The forecast indicator display module includes a query indicator data unit, a curve graph drawing unit, a seasonal graph drawing unit and a list data display unit; the query indicator data unit is used to query the assembly indicator data, the curve graph drawing unit and the seasonal graph drawing unit are used to draw as required, and the list data display unit is used to present the indicator data; The prediction index update module includes a trigger index update unit, a recursive query association unit, a classification refresh index unit and an update storage index unit; the trigger index update unit is used to trigger the update index data, the recursive query association unit is used to query the associated index data, the classification refresh index unit is used to classify and refresh the index data, and the update storage index unit is used to store the index data; The prediction indicator tracing module includes a query and determination node unit, a recursive tree structure unit, a tree diagram drawing unit and a related information labeling unit; the query and determination node unit is used to determine the node, the recursive tree structure unit is used to build a tree, the tree diagram drawing unit is used to draw, and the related information labeling unit is used to label indicator information.

10. The financial risk assessment system based on prediction indicators according to claim 8, characterized in that: The prediction index updating module refreshes the indexes in the order of basic index, basic prediction index and calculated prediction index.

Images