A method, system, and platform for evaluating the leading stocks in a theme based on continuous advancement criteria.

Abstract

This invention belongs to the technical field of theme leadership assessment, and discloses a method, system, and platform for theme leadership assessment based on continuous advancement judgment. The method includes: based on the leadership results of a target trading day, analyzing the theme advancement level of the target trading day using a leadership persistence and advancement assessment algorithm, and assessing the number of theme leadership times based on a special judgment technique for the number of hot topic leadership times; filtering the theme advancement level and the number of theme leadership times for validity; constructing a scoring model based on the filtered theme advancement level and the number of theme leadership times to assess the quality of theme leadership; and storing the theme leadership quality, the filtered theme advancement level, and the number of theme leadership times in a theme record table. This invention achieves daily and window-level dual-layer assessment of theme leadership status by constructing a multi-dimensional indicator optimization mechanism, a rolling window counting mechanism, a continuous advancement judgment mechanism, and an extended indicator completion mechanism.

Description

Technical Field

[0001] This invention relates to the field of theme-leading stock evaluation technology, and in particular to a theme-leading stock evaluation method, system and platform based on continuous advancement judgment. Background Technology

[0002] In securities trading, themes serve as a crucial vehicle for organizing the movement of individual stocks. Investors and quantitative systems often use themes to observe market risk appetite, capital flows, and short-term sentiment changes. As market pace accelerates, relying solely on abnormal fluctuations at the individual stock level is no longer sufficient to promptly grasp the market's main trend; the industry is gradually shifting towards a comprehensive assessment of themes.

[0003] Existing, relatively similar implementation schemes mainly include the following categories: First, selecting the top-performing themes by ranking their daily price changes; second, using the number of stocks hitting their daily limit, the number of consecutive limit-up days, or the strength of the limit-up moves as indicators of theme popularity; third, judging market attention by trading volume, net capital inflow, and other trading dimensions; and fourth, measuring the trend strength of a theme by its cumulative gains over a number of trading days. While these schemes can reflect theme popularity in certain dimensions, they usually operate independently, lacking a unified multi-dimensional leading framework and rarely combining real-time market data caching, historical daily databases, and rolling window progression relationships for coordinated evaluation.

[0004] Especially during trading phases with frequent sector rotation, relying solely on daily rankings can easily lead to misjudging one-off, short-lived themes as mainstream themes. Conversely, relying solely on cumulative occurrences can easily overlook the actual daily capital inflows. Therefore, existing technologies have the following shortcomings: Single-indicator judgment is distorted: relying solely on one of the following—price increase, number of limit-up days, or net capital inflow—is easily affected by extreme values, news stimuli, or disturbances from individual constituent stocks.

[0005] Lack of sustained judgment: Most solutions only output the strongest theme on a certain day, and cannot distinguish whether the theme is breaking out for the first time, advancing continuously, or has already faded away.

[0006] Inconsistent real-time and historical data: Real-time market data comes from cache or API, while historical data comes from daily charts. If these are not processed in a consistent manner, it can easily lead to incomparable assessments of the same subject on different dates.

[0007] Insufficient utilization of consensus on hot topics: The number of times a theme is repeatedly selected by the market in recent trading days often represents a higher level of capital consensus, but existing methods lack a dedicated mechanism to accumulate and classify this characteristic.

[0008] Evaluation results are difficult to retain: many implementations only output real-time rankings, without generating traceable details of leading stocks, the number of times they have led, and their ranking data, making it difficult to provide a basis for subsequent strategies, risk control, and knowledge discovery.

[0009] Therefore, how to provide a method, system, and platform for evaluating the leading stocks in a theme based on continuous advancement judgment is an urgent problem to be solved. Summary of the Invention

[0010] This invention provides a method, system, and platform for evaluating leading stocks based on continuous advancement determination, in order to solve the problems mentioned above in the prior art.

[0011] According to a first aspect of the present invention, a method for evaluating the leading stocks in a theme based on continuous advancement determination is provided.

[0012] In one embodiment, a method for evaluating the leading sectors based on continuous advancement criteria includes: Obtain real-time thematic ranking data or data already stored in the database for the target trading day, and combine it with a multi-dimensional indicator leading stock screening mechanism to determine the leading stock results for the target trading day; Based on the leading gains results of the target trading day, the theme advancement level of the target trading day is analyzed using the leading gains persistence and advancement evaluation algorithm, and the number of themes leading gains is evaluated based on the hot spot leading gains special judgment technology. The effectiveness of the theme advancement level and the number of times the theme has led the market is filtered. A scoring model is built based on the filtered theme advancement level and the number of times the theme has led the market to evaluate the quality of the theme's leading position. The quality of the theme's leading position, the filtered theme advancement level, and the number of times the theme has led the market are stored in the theme record table.

[0013] In one embodiment, obtaining real-time sector ranking data or existing data for the target trading day, and combining it with a multi-dimensional indicator-based leading stock selection mechanism to determine the leading stock results for the target trading day includes: If the target trading day is the current trading day, read the real-time theme ranking data from the remote cache. If the target trading day is a historical trading day, read the corresponding date's stored data from the historical theme ranking table. Based on real-time thematic ranking data and existing data, the leading sector for the target trading day is determined. Then, based on the leading sector, sorting rules and filtering logic are matched to output the leading sector results for the target trading day. The performance of each sector under the leading sector results is statistically analyzed to determine the leading sector results for the target trading day.

[0014] In one embodiment, based on the leading gainers of the target trading day, the algorithm for evaluating the sustainability and advancement of leading gains is used to analyze the advancement level of the theme on the target trading day, and the number of times a theme has led gains is evaluated based on a special judgment technique for the number of times a hot topic has led gains, including: Based on the leading stock analysis results, analyze the real-time or daily data of the themes for the target trading day, including the increase, number of limit-up stocks, trading volume, net amount of main funds, and closing price for the target trading day; Obtain the comparison data of the day before the target trading day, and combine it with the increase, number of limit-up stocks, turnover, net amount of main funds and closing price to calculate the incremental funds, ten-day increase and main funds increase of the target trading day, so as to complete the supplementary processing of the theme leading expansion indicator and obtain the leading result of the target trading day after the supplementation. Extract target input data from the leading stocks results of the target trading day after completion, and combine the target input data with the theme's leading stock sustainability and advancement assessment technology to determine the theme's advancement level for the target trading day; Based on the cumulative ranking of each theme within the target window and the number of times the hot topics led the market on the previous trading day, the number of hot topics in the target trading day is analyzed.

[0015] In one embodiment, extracting target input data from the leading stocks results of the completed target trading day, and combining the target input data with the theme's leading stock sustainability and advancement assessment technology to determine the theme's advancement status on the target trading day includes: Data is extracted from the top performers of the target trading day after completion, according to a preset window range, to establish a mapping between the themes of the target trading day and the cumulative number of top performers in history, and to obtain the historical mapping of each theme within the preset window. Read the number of times the theme led the rise on the previous trading day before the target trading day, establish a mapping from the theme to the number of times the theme led the rise on the previous trading day, and obtain yesterday's mapping; Read the set of leading themes for the target trading day, the set of leading themes for the previous trading day, and the set of leading themes at the beginning boundary of the window, and calculate the intersection and union of the set of leading themes for the target trading day and the set of leading themes for the previous trading day, and the set of leading themes for the target trading day and the set of leading themes at the beginning boundary of the window, respectively. Based on the intersection and union of data, themes with continuous characteristics and candidate themes for advancement are selected, and the advancement status of themes on the target trading day is analyzed by combining historical mapping and yesterday's mapping.

[0016] In one embodiment, based on the cumulative ranking value of each theme within the target window and the number of times hot topics led the gains on the previous trading day, the analysis of the number of hot topics in the target trading day includes: Within the target trading window, count the number of times each theme appears in the theme ranking table, and define the number of times the theme leads the market on the target trading day under the hot topic type; Read the number of times the hot sectors led the gains on the previous trading day before the target trading day, and compare the number of times the hot sectors led the gains on the target trading day with the number of times the hot sectors led the gains. If the number of times the hot sectors led the gains on the target trading day is greater than the number of times the hot sectors led the gains, it means that the number of times the hot sectors led the gains on the target trading day is a consensus-enhanced upgrade. If the number of times the top gainers on the target trading day is equal to the number of times the top gainers in hot sectors are top gainers, it indicates that the number of times the top gainers on the target trading day is a stable and sustained advancement. If the number of times the top gainers on the target trading day is less than the number of times the top gainers in hot sectors are top gainers, it indicates that the number of times the top gainers on the target trading day is a consensus weakening advancement. Based on the advancement results, complete the information on the theme's price increase, number of limit-up days, 10-day price increase, increase in trading volume, net amount of main funds, and increase in main funds. Based on the completed information, output the number of hot topics in the theme on the target trading day.

[0017] In one embodiment, the effectiveness of the theme advancement level and the number of times a theme has led the market is filtered, and a scoring model is constructed based on the filtered theme advancement level and the number of times a theme has led the market to evaluate the quality of the theme's leading position, including: The theme advancement level and the number of times the theme has led the rise are converted into an index queue, and the inertia moment of the index queue is analyzed. The inertia moment threshold is compared with the inertia moment threshold to remove theme advancement levels and the number of times the theme has led the rise that do not meet the validity requirements. The filtered themes' advancement levels are broken down into three structures: the length of the continuous advancement chain, the interval between each board level, and the distribution of the number of tiers. The filtered themes' leading rise times are also broken down into three substructures: the distribution of intraday leading rise periods, the length of the cross-day leading rise relay chain, and the proportion of leading rise times of each tier, in order to restore the flow path. A scoring model is constructed based on the flow path to describe the matching degree of the advancing leading structure. The synchronization coefficient between the continuous advancement chain and the leading relay chain is calculated to quantify the risk resistance of the advancing leading structure, output a comprehensive health score, and determine the quality of the leading theme.

[0018] In one embodiment, the theme advancement level and the number of times a theme has led the market are converted into an index queue, and the inertia moment of the index queue is analyzed and compared with the inertia moment threshold to remove theme advancement levels and the number of times a theme has led the market that do not meet the effectiveness requirements, including: The theme advancement level and the number of times the theme has led the rise are used as the storage carrier of the feature index queue. Then, one theme advancement level and one number of times the theme has led the rise are randomly selected from the feature index queue storage carrier as the un-benchmark sample, and the rest are used as the benchmark sample. Calculate the time interval between the unbenchmarked sample and the benchmarked sample, determine whether the leading combination of the unbenchmarked sample and the benchmarked sample is equal based on the time interval, and generate a theme leading feature index queue based on the result. Based on the probability values ​​and time intervals of the feature index queue, calculate the feature inertia moments of the theme advancement level and the number of times the theme has led the rise, and compare the feature inertia moments with the inertia moment threshold. If the characteristic moment of inertia is greater than or equal to the moment of inertia threshold, it means that the corresponding theme's advancement level and the number of times the theme has led the market are valid data. If the characteristic moment of inertia is less than the moment of inertia threshold, it means that the corresponding theme's advancement level and the number of times the theme has led the market are invalid data and need to be removed.

[0019] In one embodiment, a scoring model is constructed based on the flow path to describe the matching degree of the advancing leading structure, and the synchronization coefficient between the continuous advancing chain and the leading relay chain is calculated to quantify the risk resistance of the advancing leading structure, including: Based on the flow path, independent and dependent variables are defined, and the independent variables are divided into two groups according to the dimension to which the features belong: the continuous advancement chain feature group and the leading relay chain feature group. A binning strategy relationship table is established. Based on the predefined binning intervals corresponding to the independent variables in the binning strategy relationship table, each feature is initialized into bins, and adjacent bins with insufficient significant differences are merged through the chi-square test to obtain feature bins. The contribution of independent variables to the dependent variable is quantified based on feature binning, and the observation matrices of independent and dependent variables are constructed according to the contribution level to build the regression equation; By using regression equations to eliminate features that do not conform to business logic until all independent variable coefficients in the regression equations are positive, a structural matching degree scoring model is obtained. Based on the structural matching degree scoring model, the synchronization coefficient between the continuous advancement chain and the leading relay chain is calculated. The synchronization coefficient is combined with the comprehensive score to output a quantitative value of risk resistance capability.

[0020] According to a second aspect of the present invention, a theme-leading evaluation system based on continuous advancement determination is provided.

[0021] In one embodiment, a theme-leading assessment system based on continuous advancement determination includes: The Leading Stock Result Determination Module is used to obtain real-time theme ranking data or data already stored in the database for the target trading day, and combine it with a multi-dimensional indicator leading stock screening mechanism to determine the leading stock result for the target trading day. The continuous advancement judgment module is used to analyze the theme advancement level of the target trading day based on the leading rise results, using the leading rise persistence and advancement evaluation algorithm, and to evaluate the theme leading rise times based on the hot spot leading rise times special judgment technology. The theme leadership evaluation module is used to filter the effectiveness of theme advancement level and theme leadership frequency. Based on the filtered theme advancement level and theme leadership frequency, a scoring model is built to evaluate the quality of theme leadership. The theme leadership quality, the filtered theme advancement level, and the theme leadership frequency are stored in the theme record table.

[0022] According to a third aspect of the present invention, a computer platform is provided.

[0023] In some embodiments, the computer platform includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0024] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention achieves a two-tiered evaluation of the leading status of themes at the daily and window levels by constructing a multi-dimensional indicator selection mechanism, a rolling window counting mechanism, a continuous advancement judgment mechanism, and an extended indicator completion mechanism. It can not only identify different types of leading themes on the same day, but also judge the theme's ability to continuously advance over multiple trading days, and highlight the role of the number of hot spots / themes selected in the identification of main themes, so as to improve the stability, interpretability and traceability of the evaluation.

[0025] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0026] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0027] Figure 1 This is a flowchart illustrating a theme-leading stock evaluation method based on continuous advancement determination, according to an exemplary embodiment. Figure 2 This is a schematic diagram of a theme-leading evaluation system based on continuous advancement determination, according to an exemplary embodiment. Figure 3 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment; Figure 4 This is a flowchart illustrating the operation of a theme-leading evaluation method based on continuous advancement determination, according to an exemplary embodiment. Figure 5 This is a flowchart illustrating the continuous advancement determination of a theme-leading stock assessment method based on continuous advancement determination, according to an exemplary embodiment. Figure 6 This is a flowchart illustrating a topic-leading evaluation method based on continuous advancement determination, according to an exemplary embodiment, to determine the number of hot topics. Detailed Implementation

[0028] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some portions and features of certain embodiments may be included in or replace portions and features of other embodiments. The scope of the embodiments herein includes the entire scope of the claims and all available equivalents thereof. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0029] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0030] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0031] Existing thematic analysis methods typically rely solely on single indicators such as daily price increase, number of limit-up stocks, or trading volume to identify hot themes, which can easily lead to the following problems: First, they can only reflect the surface strength at a certain point in time and cannot depict whether the theme has the ability to lead the market continuously; second, they cannot uniformly compare the performance of themes on real-time trading days with historical trading days; third, they cannot identify the advancement, decline, and rotation of themes over multiple trading days; and fourth, they do not make sufficient use of features such as the number of hot topics, which reflect the continuous exposure and market consensus of themes, resulting in large fluctuations in the identification results of main themes.

[0032] Therefore, the technical problem to be solved in this embodiment is: how to integrate multi-dimensional indicators such as price increase, number of limit-up stocks, 10-day price increase, main force funds, incremental funds, turnover and number of hot spots under a unified data framework, to screen the leading stocks by type, and further evaluate the sustainability of the leading stocks in the time series through algorithm steps such as rolling window counting, set intersection judgment, state transition classification and hot spot number special judgment, so as to more accurately identify the evolution path of the main theme and strong theme.

[0033] Figure 1 An embodiment of the subject-leading stock evaluation method based on continuous advancement determination of the present invention is shown.

[0034] In this optional embodiment, the theme-leading assessment method based on continuous advancement determination includes: Step S101: Obtain real-time theme ranking data or data already stored in the database for the target trading day, and determine the leading stock results for the target trading day by combining the multi-dimensional indicator leading stock screening mechanism.

[0035] In one embodiment, obtaining real-time theme ranking data or existing data for the target trading day, and determining the leading stock results for the target trading day by combining a multi-dimensional indicator leading stock screening mechanism, includes: determining the target trading day; if the target trading day is the current trading day, reading real-time theme ranking data from the remote cache; if the target trading day is a historical trading day, reading the existing data for the corresponding date from the theme historical ranking day table; determining the leading stock type for the target trading day based on the real-time theme ranking data and existing data, and matching sorting rules and screening logic based on the leading stock type to output the leading stock type results for the target trading day, so as to statistically analyze the performance of each theme under the leading stock type results and determine the leading stock results for the target trading day.

[0036] Step S102: Based on the leading gains results of the target trading day, analyze the theme advancement level of the target trading day using the leading gains sustainability and advancement evaluation algorithm, and evaluate the theme leading gains times based on the hot spot leading gains times special judgment technology.

[0037] In this optional embodiment, based on the leading performance results of the target trading day, the theme advancement level of the target trading day is analyzed using the leading performance sustainability and advancement evaluation algorithm, and the theme leading performance frequency is evaluated based on the hot spot leading performance frequency special judgment technology. This includes: analyzing the real-time or daily data of the theme on the target trading day according to the leading performance results, analyzing the increase, number of limit-up stocks, turnover, net capital inflow of main funds and closing price of the target trading day; obtaining the comparison data of the previous day of the target trading day, and combining it with the increase, number of limit-up stocks, turnover, net capital inflow of main funds and closing price to calculate the incremental funds, ten-day increase and incremental capital inflow of main funds of the target trading day, so as to complete the theme leading performance expansion indicator supplementation processing and obtain the leading performance results of the target trading day after supplementation; extracting target input data from the leading performance results of the target trading day after supplementation, and combining the target input data with the theme leading performance sustainability and advancement evaluation technology to determine the theme advancement level of the target trading day; and analyzing the number of hot spots of the theme in the target trading day based on the cumulative ranking value of each theme in the target window and the number of hot spots leading performance of the previous trading day of the target trading day.

[0038] In this optional embodiment, extracting target input data from the leading stocks results of the completed target trading day and combining the target input data with the theme's leading stock sustainability and advancement assessment technology to determine the theme's advancement status on the target trading day includes: extracting data from the leading stocks results of the completed target trading day according to a preset window range to establish a mapping between the theme on the target trading day and the historical cumulative number of leading stocks, obtaining the historical mapping of each theme within the preset window; reading the theme's leading stock frequency results of the previous trading day before the target trading day, establishing a mapping between the theme and the previous trading day's leading stock frequency, obtaining yesterday's mapping; reading the leading theme set of the target trading day, the leading theme set of the previous trading day, and the leading theme set at the window's starting boundary, and calculating the intersection and union of the leading theme set of the target trading day and the previous trading day, and the leading theme set of the target trading day and the leading theme set at the window's starting boundary, respectively; filtering themes with continuity characteristics and advancement candidate themes based on the intersection and union, and analyzing the theme's advancement status on the target trading day in conjunction with the historical mapping and yesterday's mapping.

[0039] In this optional embodiment, the analysis of the number of hot topics in the target trading day is based on the cumulative ranking value of each theme within the target window and the number of times the hot topics led the gains on the previous trading day. This includes: counting the number of times each theme appears in the theme ranking table within the target trading window, and defining the number of times the theme led the gains on the target trading day under the hot topic type; reading the number of times the hot topics led the gains on the previous trading day and comparing the number of times the theme led the gains on the target trading day with the number of times the hot topics led the gains. If the number of times the theme led the gains on the target trading day is greater than the number of times the hot topics led the gains, it indicates that the number of times the theme led the gains on the target trading day is a consensus-enhancing advancement; if the number of times the theme led the gains on the target trading day is equal to the number of times the hot topics led the gains, it indicates that the number of times the theme led the gains on the target trading day is a stable maintenance advancement; if the number of times the theme led the gains on the target trading day is less than the number of times the hot topics led the gains, it indicates that the number of times the theme led the gains on the target trading day is a consensus-weakening advancement; and supplementing the theme's increase, number of limit-up days, ten-day increase, increase in trading volume, net amount of main funds, and increase in main funds information according to the advancement results, and outputting the number of hot topics in the target trading day based on the supplemented results.

[0040] like Figures 4 to 6 As shown, this embodiment uses the trading day as the basic time granularity and the theme identifier as the evaluation object. For each trading day, theme performance data is extracted from the real-time cache or historical database. Leading themes are selected from dimensions such as price increase, number of limit-up stocks, number of hot topics, ten-day price increase, main force funds, incremental funds, and trading volume. The selection results are written into the leading theme table. Based on the leading records in the most recent scrolling window, the number of times each theme has led the rise, the advancement level, and related extended indicators are calculated to form a theme leading rise frequency table. For the hot topic frequency dimension, this embodiment further adopts the theme selection frequency statistics mechanism of the past ten trading days as an important evaluation factor for identifying the main theme.

[0041] This embodiment sets up multiple leading themes, each with a set of sorting rules and filtering logic, including at least: Price Increase Type: sorted by price increase / decrease from highest to lowest, selecting the top few themes; Limit-Up Type: filtering themes with a number of limit-up stocks greater than zero, then sorting by the number of limit-up stocks from highest to lowest, selecting the top few themes; Hot Topic Type: counting the number of times a theme has been included in the theme ranking within a recent preset window, sorting by the number of occurrences from highest to lowest; Trend Type: sorted by the ten-day price increase from highest to lowest; Main Force Capital Type: sorted by the net amount of main force capital from highest to lowest; Incremental Capital Type: sorted by the increase in trading volume relative to the previous trading day from highest to lowest; Trading Volume Type: sorted by the total trading volume of the theme from highest to lowest.

[0042] Unlike the conventional approach of outputting only a single list, this embodiment generates multiple leading stock results in parallel for the same trading day, and stores the performance of each theme under each type in a unified database, including price increase, number of limit-up days, 10-day price increase, incremental funds, trading volume, number of hot topics, net amount of main funds, and incremental amount of main funds, providing a structured foundation for subsequent horizontal comparison and strategy application.

[0043] To ensure the comparability of thematic evaluation results from different sources and on different dates, this embodiment designs an extended indicator completion mechanism. For the set of themes to be evaluated, real-time / daily data of themes for the current trading day or the target historical trading day is first obtained to obtain the increase, number of limit-up stocks, turnover, net amount of main funds, and closing price. Comparative data from the previous trading day and the window start trading day are also obtained. Further calculations are made: Incremental funds = turnover of the day - turnover of the previous trading day; 10-day increase = (closing price of the day - closing price of the window start trading day) / closing price of the window start trading day × 100%; Incremental main funds = net amount of main funds of the day - net amount of main funds of the previous trading day.

[0044] With the help of the above-mentioned completion mechanism, even if the subject matter is initially selected based on a single-dimensional indicator, a complete multi-dimensional profile can be obtained in the end, providing a more sufficient chain of evidence for the determination of the main subject matter.

[0045] Meanwhile, the core of this embodiment is not just to determine "who is leading the rise today", but to further determine "whether the theme continues to lead the rise and whether it has advanced" using an executable algorithm. Specifically, this embodiment calculates the number of times the theme leads the rise for each type of leading rise within the most recent preset trading window, and combines the previous day's leading rise results, the leading rise results at the window's starting boundary, and the current day's leading rise results to determine the theme's advancement status.

[0046] The algorithm defines at least the following inputs: the current target trading day T; the previous trading day T-1; the starting boundary trading day T-n of the rolling window; the current leading theme set Lt; the leading theme set Ly of the previous trading day; the leading theme set Ls at the starting boundary of the window; the mapping table Cy of the leading times of themes on the previous trading day; the cumulative leading times table Ch of themes within the historical window. The specific calculation steps are as follows: Step 1: Intercept the data within the preset window range in reverse order of dates for the leading records, and establish a mapping Ch (i.e., historical mapping) from themes to the cumulative leading times in history to obtain the basic occurrence times of each theme within the window.

[0047] Step 2: Read the leading times results of themes on the previous trading day, and establish a mapping Cy (i.e., yesterday's mapping) from themes to the leading times yesterday.

[0048] Step 3: Read the current leading theme set Lt, the leading theme set Ly of the previous trading day, and the leading theme set Ls at the starting boundary of the window.

[0049] Step 4: Calculate the set intersections Lt∩Ly and Lt∩Ls respectively to identify the themes with continuity characteristics, and denote the union of the two intersections as R. The themes in R are defined as the candidate themes for advancement.

[0050] Step 5: Initialize the mapping Ct of the leading times of themes on the current day. Among them, first write the basic values for the themes existing in the cumulative times table Ch in history; if the theme belongs to the candidate theme set R for advancement, update the leading times of this theme on the current day according to Ct(i)=Cy(i)+1; if the theme does not belong to R, retain its historical statistical value or the basic value on the current day.

[0051] Step 6: For each theme i, compare the magnitude relationship between Ct(i) and Cy(i) to obtain the theme status grade(i); when Ct(i)>Cy(i), mark it as the advancement grade 1; when Ct(i)=Cy(i), mark it as the maintenance grade 2; when Ct(i)<Cy(i), mark it as the ebb grade 3.

[0052] Step 7: Output the result record including the theme identifier, the leading times Ct(i) on the current day, the status grade grade(i), and the extended index vector.

[0053] The above steps essentially constitute a time series evaluation algorithm based on rolling window counting and set state transition. The key does not lie in a single sorting, but in converting the repeated leading behavior of themes on different trading days into cumulative and comparable state variables, so as to describe the theme strength as a combined result of instantaneous strength and continuity strength.

[0054] Furthermore, in terms of engineering implementation, the algorithm can be implemented using mapping tables and set operations, where the theme identifier serves as the primary key, and yesterday's count, historical count, and today's count serve as values, achieving low-complexity updates. For multiple leading sectors on the same trading day, the algorithm can be executed in parallel, thereby improving evaluation efficiency.

[0055] For the hot topic frequency dimension, this embodiment uses a dedicated evaluation algorithm that differs from other dimensions. Specifically, within the most recent ten trading days, the cumulative number of times each theme appears in the theme ranking table is counted to obtain the theme hot topic frequency; then, this hot topic frequency is directly used as the benchmark for the theme's daily leading gains, and compared with the hot topic frequency of the previous trading day to complete the advancement level determination.

[0056] The algorithm can be expressed as follows: Let Q(i,T) represent the number of times theme i has entered the theme ranking list in the most recent m trading days up to trading day T, where m is preferably 10; let Q(i,T) be directly defined as the number of times theme i leads the gains in the hot topic type on the same day, i.e., Ct hot (i) = Q(i, T); Read the number of times Cy, the leading stock in the previous trading day. hot (i), compare Ct hot (i) with Cy hot (i): When Ct_hot(i) > Cy hot (i) indicates a consensus-enhanced advancement; when Ct hot t(i)=Cy hot When (i), it is determined to be stationary; when Ct hot (i) <Cy hot (i) is judged as a weakening of consensus; after completing the level calculation, the increase in the theme, the number of limit-up days, the ten-day increase, the increase in turnover, the net amount of main funds and the increase in main funds are supplemented to form a complete evaluation result driven by the number of hot spots.

[0057] This hot topic frequency judgment algorithm can transform whether a theme is continuously attracting market attention into a quantifiable, database-accessible, and comparable strength variable, complementing single-day increase indicators. It is particularly suitable for identifying themes with the potential to become the main theme but whose single-day increase may not be extremely leading.

[0058] Step S103: Perform validity filtering on the theme advancement level and the number of times the theme has led the rise. Based on the filtered theme advancement level and the number of times the theme has led the rise, construct a scoring model to evaluate the quality of the theme's leading rise. Store the quality of the theme's leading rise, the filtered theme advancement level, and the number of times the theme has led the rise in the theme record table.

[0059] In this optional embodiment, the effectiveness of the theme advancement level and the number of times the theme has led the rise is filtered. A scoring model is constructed based on the filtered theme advancement level and the number of times the theme has led the rise to evaluate the quality of the theme's leading position. This includes: converting the theme advancement level and the number of times the theme has led the rise into an index queue, analyzing the moment of inertia of the index queue, and comparing it with a moment of inertia threshold to remove theme advancement levels and the number of times the theme has led the rise that do not meet the effectiveness requirements; decomposing the filtered theme advancement level into three sub-structures: first, the length of the continuous advancement chain, i.e., the longest consecutive limit-up days for individual stocks within the theme, reflecting the vertical explosive strength of the theme; second, the gap interval between each board level, i.e., the number of missing trading days between adjacent limit-up heights, reflecting the risk of gaps in capital relay; and third, the distribution of the number of tiers, i.e., the proportion of individual stocks in the four tiers of 1 board, 2-3 board, 4-5 board, and 6+ board, reflecting the completeness of the theme's tiers; simultaneously, the filtered theme leading the rise... The frequency is broken down into three sub-structures: first, the distribution of intraday leading periods, i.e., the proportion of leading times in the first 30 minutes of trading, morning session, afternoon session, and closing session to the total number of leading times, reflecting the initiative of capital inflows; second, the length of the cross-day leading relay chain, i.e., the number of consecutive trading days in which the same leading stock leads the market, reflecting the sustainability of the leading stock's appeal; and third, the proportion of leading times in each tier, i.e., the proportion of leading times contributed by leading stocks in the four board-level tiers, reflecting the hierarchical diffusion of leading forces. The function of this step is to transform one-dimensional total indicators into three-dimensional structural features, fully restoring the flow path of capital in the intraday / cross-day time dimension and each board-level dimension. Based on the flow path, a scoring model is constructed to describe the matching degree of the advancing leading structure, calculate the synchronization coefficient between the continuous advancing chain and the leading relay chain to quantify the risk resistance of the advancing leading structure, output a comprehensive health score, and determine the quality of thematic leading stocks.

[0060] In this optional embodiment, the theme advancement level and theme leading rise frequency are converted into an index queue, and the inertia moment of the index queue is analyzed. The inertia moment threshold is compared with the index queue to eliminate theme advancement levels and theme leading rise frequency that do not meet the validity requirements. This includes: assigning a unique theme ID to active themes, with one observation sample corresponding to each trading day. The sample contains three core fields: trading day number T, theme advancement level G, and theme leading rise frequency C. The flag bits of all samples are initialized to "not used as benchmark sample". A hash table cache is built for each theme as the storage carrier of the feature index queue. The index queue is generated through a double loop: the outer loop sequentially takes out a sample P1 marked as "not used as benchmark sample". As a comparison base point, its marker is updated to indicate that it has been used as a benchmark sample. If no non-benchmark sample exists, the process jumps to the moment of inertia calculation stage. The inner loop sequentially retrieves all remaining non-benchmark samples P2, calculates the time interval between the two samples D = |T1-T2| (unit: trading days), and determines whether the advancing leading combination of the two samples is completely equal, i.e., simultaneously satisfying G1=G2 and C1=C2. If they are equal, it checks whether there is an index with key D in the hash table. If not, it creates index D and sets its value to 1. If it exists, it retrieves the corresponding count value E, updates it to E+1, and stores it back in the hash table. If the combination is not equal, it directly enters the next inner loop. After all benchmark samples have been processed, all indices D in the hash table are retrieved. i and the corresponding count value E i Convert the count value to a probability value P=E i / (E1+E2+…+E k This forms the final feature index queue (D1, P1), (D2, P2)...(D...). k ,P k According to the formula M=Σ （i=1） kP i ×D i 2 The characteristic moment of inertia of the current theme is calculated. The threshold T of the moment of inertia is determined by backtesting of nearly 10 years of historical data of A-shares. The initial benchmark value is set to 500, and it is adjusted to 600, 400 and 500 respectively according to bull market, bear market and volatile market. The moment of inertia M is compared with the threshold T. If M≤T, the advancement-leading data of the theme is determined to be valid, and the filter flag is set to 0 to retain the data. If M>T, it is determined to be invalid data (corresponding to moving hot spots, large-scale rotation, and non-sensitive miscellaneous themes, whose advancement-leading combination has too high a probability of repeated occurrence at different time intervals, which cannot reflect the true capital strength). The filter flag is set to 1 and removed from the evaluation dataset. In this way, false interference signals can be filtered out from the source, which can significantly improve the signal-to-noise ratio of the data in subsequent modeling and avoid invalid themes occupying computing resources and misleading the evaluation results.

[0061] In this optional embodiment, a scoring model is constructed based on the flow path to describe the matching degree of the advancing leading structure. The synchronization coefficient between the continuous advancing chain and the leading relay chain is calculated to quantify the risk resistance of the advancing leading structure. This includes: defining the independent variables as the above six sub-structure features, and defining the dependent variables as three quantitative indicators of structural risk resistance: the probability of survival of the theme in the next 3 days, the maximum drawdown in the next 3 days, and the rebound strength in the next 3 days after adjustment. Each theme corresponds to one training sample, forming a wide table of theme data and completing preprocessing such as missing value filling and outlier truncation; dividing the independent variables into two groups according to the dimension to which the features belong: the continuous advancing chain feature group and the leading relay chain feature group, and establishing a binning strategy relationship table. Define binning intervals for each variable type. For example, the length of a continuous promotion chain is divided into four initial bins: 1-2 boards, 3-5 boards, 6-8 boards, and ≥9 boards. The frequency of leading stock rotation is divided into four initial bins: ≤1 day, 1-2 days, 2-3 days, and ≥3 days. After initializing binning for each feature, adjacent bins with insufficient statistical significance are merged using a chi-square test: the chi-square value is calculated for every two bins in the binning order, and the two bins corresponding to the smallest chi-square value are merged. This process is repeated until the number of remaining bins is 5-10 or the p-value corresponding to the smallest chi-square value exceeds 0.95, ensuring that there is a statistically significant difference in the risk resistance of each bin. Subsequently, the weight of evidence (WOE) for each bin is calculated. i =ln[(Percentage of bad samples) / (Percentage of good samples)], where bad samples are defined as themes that have a pullback of ≥15% or have disappeared directly in the following 3 days, and good samples are defined as themes that have an increase or pullback of ≤5% in the following 3 days.

[0062] The informative value (IV) of each feature is calculated based on the Word of Entity (WOE) value of each bin: IV = Σ[(Percentage of bad samples - Percentage of good samples) × WOE] i The contribution of quantitative features to the structural resilience is assessed. Then, for each group of independent variables, features are selected for model inclusion: first, features with IV ≥ 0.1 are selected to form a target set; if the target set is empty, the first 3 features are selected in descending order of IV value; if the target set is not empty, the first 5 features are selected in descending order of IV value, ensuring that both feature groups have variables entering the model to guarantee the integrity of the feature dimensions. The original values ​​of the included features are converted to the corresponding binning WOE values, and an N x M column independent variable observation matrix E0 is constructed (N is the number of samples, M is the number of included features). After standardizing the three dependent variables to the [0,1] interval, an N x 3 column dependent variable observation matrix F0 is constructed. The number of introduced components is initialized to h = 1, and the symmetric matrix is ​​solved. ×F0× The unit eigenvector corresponding to the largest eigenvalue of ×E0 determines the first axis C1 of the independent variable. The first component t = E0 × C1 is calculated, and the symmetric matrix is ​​solved similarly. ×E0× The unit eigenvector corresponding to the largest eigenvalue of ×F0 determines the second axis D1 of the dependent variable, and the second component u = F0 × D1 is calculated; based on the first component, the second component, and the observation matrix, the formula e = t / ||t²|| Calculate the first load vector of the independent variable using the formula f= The second load vector of the dependent variable is calculated using the formula r = u / ||u²||. t / ||t²|| calculates the regression coefficient between the independent and dependent variables, and then uses the residual matrix E1=E0-t× F1 = F0 - t × Construct the initial regression equation F0=E0×C1× +F1.

[0063] The leave-one-out cross-validation method was used to verify the effectiveness of the next component to be introduced: after successively removing each sample, the first candidate regression equation was obtained by fitting h components, and the second candidate regression equation was obtained by fitting all samples with h-1 components. The sum of squared prediction errors of the two equations was calculated, and then the contribution index Q was obtained. 2 If there exists a Q of any dependent variable 2 If the value is ≥0.0975, the verification is successful. The observation matrix is ​​updated to the residual matrix, and the number of introduced components is incremented by 1. The component extraction process is repeated. If the Q of all dependent variables... 2If all values ​​are less than 0.0975, component extraction is stopped. Based on the current regression equation, business logic is validated, and features with negative coefficients that do not conform to business logic are removed (e.g., the coefficient for the length of the continuous advancement chain should be positive; the larger the value, the stronger the risk resistance). The remaining features are used as new input variables, and the modeling process is repeated until all independent variable coefficients in the regression equation are positive, resulting in the final structure matching score model. Finally, based on the regression coefficients and variable projection importance (VIP) values ​​of the final model, the synchronization coefficient between the continuous advancement chain and the leading relay chain is calculated, where α = Σ continuous advancement chain feature VIP value / total VIP value, β = Σ leading relay chain feature VIP value / total VIP value, and the synchronization coefficient = α × Σ (advancement chain feature coefficient × feature standardized value) + β × Σ (leading ... The model combines the synchronization coefficient (60% weight) with the comprehensive risk resistance score predicted by the model (40% weight) to map a comprehensive health score of the leading sector quality from 0 to 100. The higher the score, the better the matching degree between the advancement and leading structure, the healthier the capital flow path, and the stronger the ability to resist market fluctuations. The core function of this step is to quantify the degree of synergy between the advancement structure and the leading structure, and output interpretable and quantifiable leading quality assessment results. Finally, the filtered original thematic advancement level, the number of thematic leading times, the six substructure feature values, the synchronization coefficient of the continuous advancement chain and the leading relay chain, the comprehensive health score, the quantitative value of risk resistance, and the filter flag are uniformly stored in the thematic record table. Each record is associated with a unique thematic ID and the trading day timestamp.

[0064] After obtaining the number of times a theme has led the market and its advancement level, this embodiment further filters the results for validity. It prioritizes retaining thematic records with valid trading volume, valid main force funds, or valid stage gains, while removing data that is severely missing or cannot form a valid judgment. The filtered results are written into the thematic leading table and the thematic leading number table, respectively, forming a multi-level historical data accumulation data that is daily, window-oriented, and type-oriented, providing underlying support for subsequent trading strategies, hot spot backtesting, and risk warning.

[0065] In summary, this embodiment constructs a multi-dimensional indicator-based leading stock selection framework for specific themes, enabling factors such as price increase, number of limit-up days, number of hot topics, 10-day price increase, main force funds, incremental funds, and trading volume to participate in the theme leading stock evaluation in parallel. It also constructs a theme leading stock frequency calculation algorithm based on rolling window counting, accumulating and statistically analyzing theme leading stock records within historical windows to generate a basic frequency mapping. Furthermore, it constructs a continuous advancement judgment algorithm based on the intersection of the current day's set, yesterday's set, and window boundary set to identify the theme's sustained leading stock characteristics. A level allocation algorithm based on state transition comparison outputs three states: advancement, maintenance, and decline, by comparing the current day's leading stock frequency with yesterday's leading stock frequency. Finally, it utilizes a theme hot topic frequency special judgment algorithm, using the number of times a theme was selected within the most recent preset trading days as a theme consensus variable, directly used for main theme identification and level determination. Based on the structured sedimentation mechanism of theme leading stock results, the theme leading stock table and theme leading stock frequency table can support subsequent strategy backtesting, historical review, and main line tracking.

[0066] Figure 2 An embodiment of the theme-leading evaluation system based on continuous advancement determination of the present invention is shown.

[0067] In this optional embodiment, the theme-leading assessment system based on continuous advancement determination includes: The Leading Stock Result Determination Module 201 is used to obtain real-time theme ranking data or data already stored in the database for the target trading day, and combine it with a multi-dimensional indicator leading stock screening mechanism to determine the leading stock result for the target trading day. The continuous advancement judgment module 202 is used to analyze the theme advancement level of the target trading day based on the leading rise results of the target trading day, using the leading rise persistence and advancement evaluation algorithm, and to evaluate the theme leading rise times based on the hot spot leading rise times special judgment technology. The theme leadership evaluation module 203 is used to filter the theme advancement level and the number of times the theme has led the rise in effectiveness. Based on the filtered theme advancement level and the number of times the theme has led the rise, a scoring model is constructed to evaluate the quality of the theme leadership. The quality of the theme leadership, the filtered theme advancement level, and the number of times the theme has led the rise are stored in the theme record table.

[0068] In addition, the present invention also provides a computer platform, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0069] In one embodiment, the computer platform may be a server, and its internal structure diagram may be as follows: Figure 3As shown, the computer platform includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0070] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer platform on which the present invention is applied. A specific computer platform may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0071] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0072] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0073] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A method for evaluating the leading stocks in a theme based on continuous advancement judgment, characterized in that, The method includes: Obtain real-time thematic ranking data or data already stored in the database for the target trading day, and combine it with a multi-dimensional indicator leading stock screening mechanism to determine the leading stock results for the target trading day; Based on the leading gains results of the target trading day, the theme advancement level of the target trading day is analyzed using the leading gains persistence and advancement evaluation algorithm, and the number of themes leading gains is evaluated based on the hot spot leading gains special judgment technology. The effectiveness of the theme advancement level and the number of times the theme has led the market is filtered. A scoring model is built based on the filtered theme advancement level and the number of times the theme has led the market to evaluate the quality of the theme's leading position. The quality of the theme's leading position, the filtered theme advancement level, and the number of times the theme has led the market are stored in the theme record table.

2. The theme-leading evaluation method based on continuous advancement judgment according to claim 1, characterized in that, The process of obtaining real-time sector ranking data or existing data for the target trading day, and combining it with a multi-dimensional indicator-based leading stock selection mechanism to determine the leading stock results for the target trading day, includes: If the target trading day is the current trading day, read the real-time theme ranking data from the remote cache. If the target trading day is a historical trading day, read the corresponding date's stored data from the historical theme ranking table. Based on real-time theme ranking data and existing data, the leading type of the target trading day is determined. Then, based on the leading type, sorting rules and filtering logic are matched to output the leading type results of the target trading day. The performance of each theme under the leading type results is statistically analyzed to determine the leading results of the target trading day. The leading types include those based on price increase, price limit, hot topics, trend, main force funds, incremental funds, and trading volume.

3. The theme-leading evaluation method based on continuous advancement judgment according to claim 1, characterized in that, Based on the leading gainers of the target trading day, the algorithm for analyzing the theme's advancement level on the target trading day is used to assess the theme's leading gainers frequency using a leading gainers persistence and advancement evaluation algorithm. The evaluation of the theme's leading gainers frequency based on a special judgment technique includes: Based on the leading stock analysis results, analyze the real-time or daily data of the themes for the target trading day, including the increase, number of limit-up stocks, trading volume, net amount of main funds, and closing price for the target trading day; Obtain the comparison data of the day before the target trading day, and combine it with the increase, number of limit-up stocks, turnover, net amount of main funds and closing price to calculate the incremental funds, ten-day increase and main funds increase of the target trading day, so as to complete the supplementary processing of the theme leading expansion indicator and obtain the leading result of the target trading day after the supplementation. Extract target input data from the leading stocks results of the target trading day after completion, and combine the target input data with the theme's leading stock sustainability and advancement assessment technology to determine the theme's advancement level for the target trading day; Based on the cumulative ranking of each theme within the target window and the number of times the hot topics led the market on the previous trading day, the number of hot topics in the target trading day is analyzed.

4. The theme-leading evaluation method based on continuous advancement judgment according to claim 3, characterized in that, The step of extracting target input data from the leading stocks results of the completed target trading day, and combining the target input data with the theme's leading stock sustainability and advancement assessment technology to determine the theme's advancement status on the target trading day includes: Data is extracted from the top performers of the target trading day after completion, according to a preset window range, to establish a mapping between the themes of the target trading day and the cumulative number of top performers in history, and to obtain the historical mapping of each theme within the preset window. Read the number of times the theme led the rise on the previous trading day before the target trading day, establish a mapping from the theme to the number of times the theme led the rise on the previous trading day, and obtain yesterday's mapping; Read the set of leading themes for the target trading day, the set of leading themes for the previous trading day, and the set of leading themes at the beginning boundary of the window, and calculate the intersection and union of the set of leading themes for the target trading day and the set of leading themes for the previous trading day, and the set of leading themes for the target trading day and the set of leading themes at the beginning boundary of the window, respectively. Based on the intersection and union of data, themes with continuous characteristics and candidate themes for advancement are selected, and the advancement status of themes on the target trading day is analyzed by combining historical mapping and yesterday's mapping.

5. The theme-leading evaluation method based on continuous advancement judgment according to claim 4, characterized in that, The analysis of the number of hot topics in a target trading day, based on the cumulative ranking value of each theme within the target window and the number of times hot topics led the gains on the previous trading day, includes: Within the target trading window, count the number of times each theme appears in the theme ranking table, and define the number of times the theme leads the market on the target trading day under the hot topic type; Read the number of times the hot sectors led the gains on the previous trading day before the target trading day, and compare the number of times the hot sectors led the gains on the target trading day with the number of times the hot sectors led the gains. If the number of times the hot sectors led the gains on the target trading day is greater than the number of times the hot sectors led the gains, it means that the number of times the hot sectors led the gains on the target trading day is a consensus-enhanced upgrade. If the number of times the top gainers on the target trading day is equal to the number of times the top gainers in hot sectors are top gainers, it indicates that the number of times the top gainers on the target trading day is a stable and sustained advancement. If the number of times the top gainers on the target trading day is less than the number of times the top gainers in hot sectors are top gainers, it indicates that the number of times the top gainers on the target trading day is a consensus weakening advancement. Based on the advancement results, complete the information on the theme's price increase, number of limit-up days, 10-day price increase, increase in trading volume, net amount of main funds, and increase in main funds. Based on the completed information, output the number of hot topics in the theme on the target trading day.

6. The theme-leading evaluation method based on continuous advancement judgment according to claim 1, characterized in that, The process of filtering the theme's advancement level and the number of times it has led the market, and then constructing a scoring model based on the filtered theme advancement level and the number of times it has led the market, to evaluate the quality of the theme's leading position includes: The theme advancement level and the number of times the theme has led the rise are converted into an index queue, and the inertia moment of the index queue is analyzed. The inertia moment threshold is compared with the inertia moment threshold to remove theme advancement levels and the number of times the theme has led the rise that do not meet the validity requirements. The filtered themes' advancement levels are broken down into three structures: the length of the continuous advancement chain, the interval between each board level, and the distribution of the number of tiers. The filtered themes' leading rise times are also broken down into three substructures: the distribution of intraday leading rise periods, the length of the cross-day leading rise relay chain, and the proportion of leading rise times of each tier, in order to restore the flow path. A scoring model is constructed based on the flow path to describe the matching degree of the advancing leading structure. The synchronization coefficient between the continuous advancement chain and the leading relay chain is calculated to quantify the risk resistance of the advancing leading structure, output a comprehensive health score, and determine the quality of the leading theme.

7. The theme-leading evaluation method based on continuous advancement judgment according to claim 6, characterized in that, The process of converting the theme advancement level and the number of times a theme has led the market into an index queue, analyzing the moment of inertia of the index queue, and comparing it with the moment of inertia threshold to remove theme advancement levels and the number of times a theme has led the market that do not meet the effectiveness requirements includes: The theme advancement level and the number of times the theme has led the rise are used as the storage carrier of the feature index queue. Then, one theme advancement level and one number of times the theme has led the rise are randomly selected from the feature index queue storage carrier as the un-benchmark sample, and the rest are used as the benchmark sample. Calculate the time interval between the unbenchmarked sample and the benchmarked sample, determine whether the leading combination of the unbenchmarked sample and the benchmarked sample is equal based on the time interval, and generate a theme leading feature index queue based on the result. Based on the probability values ​​and time intervals of the feature index queue, calculate the feature inertia moments of the theme advancement level and the number of times the theme has led the rise, and compare the feature inertia moments with the inertia moment threshold. If the characteristic moment of inertia is greater than or equal to the moment of inertia threshold, it means that the corresponding theme's advancement level and the number of times the theme has led the market are valid data. If the characteristic moment of inertia is less than the moment of inertia threshold, it means that the corresponding theme's advancement level and the number of times the theme has led the market are invalid data and need to be removed.

8. The theme-leading evaluation method based on continuous advancement judgment according to claim 7, characterized in that, The scoring model constructed based on the flow path to describe the matching degree of the advancing leading structure, and the calculation of the synchronization coefficient between the continuous advancing chain and the leading relay chain to quantify the risk resistance of the advancing leading structure, includes: Based on the flow path, independent and dependent variables are defined, and the independent variables are divided into two groups according to the dimension to which the features belong: the continuous advancement chain feature group and the leading relay chain feature group. A binning strategy relationship table is established. Based on the predefined binning intervals corresponding to the independent variables in the binning strategy relationship table, each feature is initialized into bins, and adjacent bins with insufficient significant differences are merged through the chi-square test to obtain feature bins. The contribution of independent variables to the dependent variable is quantified based on feature binning, and the observation matrices of independent and dependent variables are constructed according to the contribution level to build the regression equation; By using regression equations to eliminate features that do not conform to business logic until all independent variable coefficients in the regression equations are positive, a structural matching degree scoring model is obtained. Based on the structural matching degree scoring model, the synchronization coefficient between the continuous advancement chain and the leading relay chain is calculated. The synchronization coefficient is combined with the comprehensive score to output a quantitative value of risk resistance capability.

9. A theme-leading stock evaluation system based on continuous advancement judgment, characterized in that, include: The Leading Stock Result Determination Module is used to obtain real-time theme ranking data or data already stored in the database for the target trading day, and combine it with a multi-dimensional indicator leading stock screening mechanism to determine the leading stock result for the target trading day. The continuous advancement judgment module is used to analyze the theme advancement level of the target trading day based on the leading rise results, using the leading rise persistence and advancement evaluation algorithm, and to evaluate the theme leading rise times based on the hot spot leading rise times special judgment technology. The theme leadership evaluation module is used to filter the effectiveness of theme advancement level and theme leadership frequency. Based on the filtered theme advancement level and theme leadership frequency, a scoring model is built to evaluate the quality of theme leadership. The theme leadership quality, the filtered theme advancement level, and the theme leadership frequency are stored in the theme record table.

10. A computer platform comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images