Market information prediction method and system based on financial consultation analysis

Abstract

The invention relates to a market information prediction method and system based on financial consultation analysis, and relates to the technical field of finance, and the method comprises the steps: collecting user browsing information in real time, and obtaining a predicted user portrait; according to the predicted user portrait and the user portrait set, determining a portrait matching degree through a matching degree formula; when the portrait matching degree is greater than a high matching degree threshold value, determining a current user portrait, and determining a pushed financial information sequence; when the portrait matching degree is smaller than a high matching degree threshold value, generating a portrait matching degree sequence, and forming a financial information sequence; when the financial information sequence exists, pushing to a browsing page of the user according to the financial information sequence; financial information is generated according to the financial information sequence or the current user portrait; and screening the financial information to obtain market prediction information, and outputting the market prediction information. The method has the effects of ensuring that the user obtains the information content related to the interest of the user and pushing the corresponding market prediction information in time.

Description

Technical Field

[0001] This invention relates to the field of financial technology, and in particular to a market information forecasting method and system based on financial consulting analysis. Background Technology

[0002] The rapid development and increasing complexity of financial markets, influenced by a complex interplay of macroeconomic factors, policies and regulations, and market sentiment, have resulted in a high degree of uncertainty. Financial information, as a "forward-looking signal source" of market demand and risk, directly impacts the quality of investment decisions in terms of its timeliness and accuracy.

[0003] Currently, with the rise of AI technology, various financial apps have built-in AI systems. Users can quickly obtain relevant financial sector information through the AI ​​system and send the latest financial news to consult the AI ​​system in order to receive market forecast information in response.

[0004] Regarding the technologies mentioned above, traditional financial information delivery methods often rely on fixed channel classifications or simple keyword matching, resulting in low efficiency in obtaining financial information. Summary of the Invention

[0005] To address the inefficiency of traditional financial information delivery methods, this invention provides a market information forecasting method and system based on financial consulting analysis.

[0006] In a first aspect, the present invention provides a market information forecasting method based on financial consulting analysis, employing the following technical solution:

[0007] A market information forecasting method based on financial consulting analysis includes:

[0008] Step 1: Respond to user inquiry signals and collect user browsing information in real time;

[0009] Step 2: Calculate the expected user profile based on the user browsing information and the preset user profile feature value formula;

[0010] Step 3: Based on the projected user profile and the preset user profile set, determine the profile matching degree using a preset matching degree formula;

[0011] Step 4: When the profile matching degree is greater than the preset high matching degree threshold, the current user profile is determined. The current user profile includes information category feature values ​​corresponding to various financial sector information categories.

[0012] Step 5: Determine the sequence of financial information to be pushed based on the current user profile;

[0013] Step 6: When the profile matching degree is less than the high matching degree threshold, generate a profile matching degree sequence based on the profile matching degree, and form a financial information sequence based on the financial information preset by the user profile corresponding to the profile matching degree sequence;

[0014] Step 7: When the financial information sequence exists, push it to the user's browsing page according to the financial information sequence;

[0015] Step 8: Generate financial information based on the financial information sequence or the current user profile;

[0016] Step 9: Filter the financial information based on the user inquiry signal to obtain market forecast information, and output the market forecast information, which includes a portion of the financial information.

[0017] By employing the aforementioned technical solution, user browsing information is collected and analyzed to generate a sequence of financial information that aligns with users' current interests and needs. This allows users to access the financial information they require in a timely manner, and also provides them with timely market forecasts. This method overcomes the limitations of traditional financial information delivery methods that rely on fixed channel categories or simple keyword matching. It achieves dynamic capture of shifting user interests, enabling users to access the latest financial information promptly and obtain accurate market forecasts, thereby effectively preventing them from missing the optimal investment window.

[0018] Optionally, the method for determining the current user profile when the profile matching degree is greater than the high matching degree threshold includes:

[0019] Step 40: Obtain user search history and determine keywords based on the user search history;

[0020] Step 41: Determine the user's weekly search frequency, total user search frequency, and target information category based on the keywords;

[0021] Step 42: Determine the keyword intent weight based on the user's weekly search frequency, the user's total search frequency, and the target information category;

[0022] Step 43: Determine the core intent based on the keyword intent weights and the preset search intent tag library;

[0023] Step 44: Update the current user profile based on the core intent and the preset weight adjustment method.

[0024] By adopting the above technical solutions and acquiring and analyzing users' search records, it is possible to more accurately grasp users' search intentions and changes in interests, thereby dynamically updating user profiles. This method ensures the timeliness and accuracy of user profiles, providing a more reliable foundation for subsequent financial information delivery and making the content more closely aligned with users' actual needs.

[0025] Optionally, it also includes a method for generating an intent-exploratory information pool when the core intent does not exist, the method comprising:

[0026] Step 45: Determine the click-through rate of exploratory information based on the preset exploratory information;

[0027] Step 46: Determine the total relevance of information categories by using the keyword intent weight and the corresponding financial information categories;

[0028] Step 47: Calculate the push weight of exploratory information based on the sum of information category feature values, click-through rate of exploratory information, and information category relevance;

[0029] Step 48: Sort the exploratory information in descending order based on the exploratory information push weight to obtain the exploratory information sequence, and push the corresponding information to the user's browsing page according to the exploratory information sequence.

[0030] By employing the aforementioned technical solution, when the user's core intent is not explicitly stated, the push weight of exploratory information is calculated by analyzing the click-through rate and relevance of information categories. This method can intelligently push exploratory information that users may be interested in, helping them discover new investment opportunities or points of interest, while avoiding information gaps caused by a lack of clear intent, thus improving user experience and information utilization efficiency.

[0031] Optionally, it also includes a method for determining the sequence of corrected information, the method comprising:

[0032] Step 10: Obtain the corresponding market prediction tags through financial information. The prediction tags include the predicted sector, predicted trend, predicted cycle, and predicted probability value.

[0033] Step 11: Obtain the user click-through rate of information corresponding to the market prediction tag;

[0034] Step 12: After the forecast period ends, obtain the actual market trend and calculate the forecast deviation value based on the forecast trend and the actual market trend;

[0035] Step 13: When the prediction deviation value exceeds the preset maximum prediction deviation value, the corrected information category feature value is calculated based on the prediction deviation value and the information category feature value using the push strategy correction formula;

[0036] Step 14: Determine the correction information sequence based on the correction information category feature value, and push the correction information sequence to the user's browsing page.

[0037] By employing the aforementioned technical solution, and by acquiring market prediction tags and user click-through rates corresponding to financial information, combined with the deviation between the actual trend and the predicted trend after the prediction period ends, the trust weight of users for similar information can be dynamically adjusted. Furthermore, a new sequence of financial information is generated based on the corrected information push weight, ensuring that the pushed content is closer to the actual market situation.

[0038] Optionally, it also includes a method for identifying and blocking high-risk information, which includes:

[0039] Step 15: Based on the financial information, obtain the financial sector corresponding to the financial information, and obtain the market volatility, industry valuation deviation and policy risk corresponding to the financial sector;

[0040] Step 16: Determine the mean volatility of the market based on the overall market volatility, and obtain the volatility deviation through the overall market volatility and the mean volatility.

[0041] Step 17: Calculate the real-time risk level based on volatility deviation, industry valuation deviation, and policy risk;

[0042] Step 18: Determine high-risk information based on the real-time risk level, and determine the current corrected information sequence based on the high-risk information, and push the current corrected information sequence to the user's browsing page.

[0043] By employing the aforementioned technical solution, the real-time risk level of information is calculated and assessed. This method can accurately identify high-risk information and adjust and correct the information sequence accordingly, ensuring that the information pushed to users not only meets their interests and needs but also effectively avoids potential risks, thereby improving the safety and stability of users' investment decisions.

[0044] Optionally, methods for not blocking high-risk information may also be included, such as:

[0045] Step 19: Calculate the user preference dimension using information category feature values, predicted trust level, and core intent;

[0046] Step 20: Calculate the user's risk tolerance coefficient based on user preference dimensions and real-time risk level;

[0047] Step 21: When the user's risk tolerance coefficient exceeds the preset high risk tolerance threshold, a modified information sequence is pushed to the user's browsing page;

[0048] Step 22: When the user's risk resistance coefficient does not exceed the high risk resistance threshold, the information is pushed to the user's browsing page according to the current correction information sequence.

[0049] By adopting the above technical solutions, calculating a user's risk tolerance coefficient can more comprehensively assess the user's ability to withstand high-risk information. When the user has a high risk tolerance, the content in the corrected information sequence can be directly pushed to ensure that the user does not miss any potentially important information. When the user's risk tolerance is low, the corrected information sequence will be pushed to ensure the security of the user's investment decisions.

[0050] Optionally, it also includes a method for constructing a cross-market information relationship graph, which includes:

[0051] Step 23: Obtain the association frequency of corresponding keywords in different markets in real time;

[0052] Step 24: Determine the correlation coefficients for different markets based on the overall market data;

[0053] Step 25: Calculate the degree of association using association frequency and association coefficient;

[0054] Step 26: When the correlation exceeds the preset strong correlation threshold, the corresponding markets are grouped together and marked as strong correlation markets and recorded in the preset strong correlation market table;

[0055] Step 27: Determine the market preference value corresponding to the current user profile based on the core intent, user risk preference, and information category feature value;

[0056] Step 28: Determine single-market preferences and cross-market preferences based on market preference values;

[0057] Step 29: When the user has a single market preference, execute steps 1 through 9;

[0058] Step 30: When a user has cross-market preferences, determine the current preferred market based on the information category feature value;

[0059] Step 31: Determine the current associated market based on the current preference market and strong association market table, and adjust the current correction information sequence according to the current associated market and push it to the user's browsing page.

[0060] By adopting the above technical solution and constructing a cross-market information correlation graph, the system can capture the dynamic relationships between different markets in real time, providing users with more comprehensive and accurate cross-market information pushes. When users exhibit cross-market preferences, the system uses a strongly correlated market table to intelligently adjust and correct the information sequence, ensuring that the pushed information content not only matches the user's current interests but also reflects the mutual influence between markets, providing users with more forward-looking and strategic market forecast information.

[0061] Optionally, methods for pushing information to the user's browsing page based on the corrected information sequence include:

[0062] Step 210: Extract and analyze the information content corresponding to all financial information in the corrected information sequence to obtain positive and negative information;

[0063] Step 211: Adjust the feature values ​​of the information categories corresponding to the positive information according to the preset information priority adjustment formula, and form the final information sequence;

[0064] Step 212: Push the corresponding financial information to the user's browsing page according to the final information sequence;

[0065] Step 2130: When current corrected information exists, extract the information content corresponding to all financial information in the corrected information sequence and analyze it to obtain positive and negative information;

[0066] Step 2131: Adjust the information category feature value corresponding to the negative information according to the information priority adjustment formula, and form the final information sequence to execute step 212.

[0067] By employing the aforementioned technical solution, the system can perform in-depth analysis of financial information in the revised information sequence, distinguishing between positive and negative information. When no current revised information exists, it indicates the user has a strong risk tolerance, and the system will adjust the information category feature values ​​of the positive information accordingly, thereby generating the final information sequence. Similarly, if current revised information exists, it indicates the user has a weak risk tolerance, and the system will adjust the information category feature values ​​of the negative information to form the final information sequence. This process not only improves the accuracy of information delivery but also enhances the efficiency and satisfaction of users in obtaining information.

[0068] Optionally, methods for identifying positive and negative information may also be included, such as:

[0069] Step 2100: Extract text information from the news based on the built-in AI port to determine the expected positive and negative news;

[0070] Step 2101: Retrieve the historical accuracy rates corresponding to the expected positive and negative information;

[0071] Step 2102: When the historical accuracy rate reaches the preset reliable accuracy rate, obtain the same market analysis information corresponding to the expected positive information and the expected negative information;

[0072] Step 2103: Extract text information of the same market analysis information from the AI ​​port to determine the comparison information. The prediction result of the comparison information is expected to be positive or negative.

[0073] Step 2104: When the expected positive information is consistent with the prediction result of the corresponding comparative information, the positive information is determined;

[0074] Step 2105: When the predicted negative information is consistent with the prediction result of the corresponding comparative information, the negative information is determined.

[0075] By employing the aforementioned technical solution, the system can initially filter out expected positive and negative information using its built-in AI interface. By retrieving historical accuracy data for this information, it identifies those reaching a preset reliable accuracy rate. The system then compares this high-accuracy information with other information analyzed under the same market conditions, enabling it to accurately distinguish between genuine positive and negative information. This process not only improves the accuracy of information analysis but also provides a more reliable basis for subsequent information delivery, ensuring that users receive financial information that best suits their needs.

[0076] Secondly, this invention provides a market information forecasting system based on financial consulting and analysis, employing the following technical solution:

[0077] A market information forecasting system based on financial consulting analysis includes:

[0078] The acquisition module is used to acquire user browsing information and user search history;

[0079] The memory is used to store the program that implements the market information forecasting method based on financial consulting analysis as described above;

[0080] The processor loads and executes programs from memory.

[0081] By adopting the above technical solution, the system collects users' browsing information and search records through an acquisition module, enabling key functions such as financial information collection and analysis, dynamic updating of user profiles, and generation and delivery of market forecast information. This system architecture improves the accuracy and personalization of financial information delivery, providing users with timely and accurate market forecast services and effectively assisting them in making investment decisions.

[0082] In summary, the present invention has at least one of the following beneficial technical effects:

[0083] 1. The invention calculates the expected user profile and updates the user profile by analyzing the keyword intent weight in the user's search history. Compared with the traditional simple keyword matching method, it more accurately reflects the user's interests and needs.

[0084] 2. Calculate real-time risk levels based on multi-dimensional data such as market volatility, industry valuation deviations, and policy risks. Based on this, identify high-risk information and block it or adjust the push strategy according to the user's risk tolerance coefficient. Attached Figure Description

[0085] Figure 1 This is a flowchart of a market information forecasting method based on financial consulting analysis, as described in an embodiment of this application.

[0086] Figure 2 This is a flowchart of a method for determining the current user profile when the profile matching degree is greater than the high matching degree threshold in an embodiment of this application;

[0087] Figure 3 This is a flowchart of a method for generating an intent exploration information pool when no core intent exists, as described in an embodiment of this application.

[0088] Figure 4 This is a flowchart of the method for determining the correction information sequence in the embodiments of this application;

[0089] Figure 5 This is a flowchart illustrating the method for identifying and blocking high-risk information in this application embodiment. Detailed Implementation

[0090] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0091] This invention discloses a market information forecasting method based on financial consulting analysis. (Refer to...) Figure 1 A market information forecasting method based on financial consulting analysis includes:

[0092] Step 1: Respond to user inquiry signals and collect user browsing information in real time.

[0093] User inquiry signals refer to requests initiated by users to obtain specific financial information or market forecasts. These signals can be natural language queries input into an AI system, or specific actions triggered while browsing financial information, such as clicking on a financial section or searching for specific keywords. User browsing information refers to various data generated by users while browsing financial information, including the duration of time spent on the information, the depth of clicks, and the actions of adding items to favorites.

[0094] Step 2: Calculate the expected user profile based on user browsing information and the preset user profile feature value formula.

[0095] The user profile feature value formula is a mathematical expression used to quantify user interests, preferences, and behavioral characteristics. A projected user profile is a quantitative model that reflects a user's current interests and behavioral characteristics; the projected user profile corresponds to the user's level of interest in various financial sectors. User browsing information includes information dwell time, click depth, and favorites actions. The specific user profile feature value formula is as follows:

[0096] ,in Let be the user's profile feature value for the i-th type of financial information, with a value range of [0, 1]. For the duration of information viewing, This refers to the average longest dwell time for similar information on the platform. When dwell time for information is a factor, the system will calculate the average longest dwell time for similar information on the platform. For click depth, For the purpose of saving, Weighted by preset dwell time For preset click depth weight, The preset collection weights are derived from data obtained through extensive experimental analysis. It is 0.5. It is 0.3. It is 0.2.

[0097] Step 3: Determine the profile matching degree based on the expected user profile and the preset user profile set using a preset matching degree formula.

[0098] A user profile set refers to a pre-defined collection of typical user profile features, encompassing user types with varying financial interests, risk preferences, and investment experiences. The matching degree formula is a mathematical expression used to measure the similarity between a projected user profile and a pre-defined user profile set. Profile matching degree refers to the degree of match between the projected user profile and the user profile set.

[0099] The specific formula for matching degree here is: , here The profile matching degree of the i-th type of information, Let n be the standard feature value of a certain type of image in the preset image set for the i-th type of information, where n is the total number of information categories. The information categories are pre-defined by humans based on the financial sectors and types corresponding to various types of information.

[0100] Step 4: When the profile matching degree is greater than the preset high matching degree threshold, the current user profile is determined. The current user profile includes the information category feature values ​​corresponding to various financial sector information categories.

[0101] The high matching threshold is a standard value used to determine whether the similarity between the expected user profile and the preset user profile set reaches a sufficiently high level. The current user profile is a model dynamically adjusted based on the user's actual behavior and interests, accurately reflecting the user's level of attention to information from various financial sectors. When the profile matching degree exceeds this threshold, the system considers the expected user profile to highly match a certain type in the preset user profile set, thus identifying that type as the current user profile. Information category feature values ​​refer to the features mentioned above. .

[0102] Step 5: Determine the sequence of financial information to be pushed based on the current user profile.

[0103] A financial news sequence refers to a method that prioritizes financial news based on its category feature values ​​within the current user profile, ranking it accordingly to generate a sequence for push notifications. This sequence is arranged from highest to lowest based on the relevance of the news to the user's interests, ensuring that the pushed content best matches the user's current needs and preferences. Specifically, this financial news sequence is formed using a preset news push priority formula. This formula calculates the push priority for various categories of news, and its specific formula is as follows: ,in The push priority for the i-th message is... The authority of the information release entity is set (this is a pre-defined value, for example, 1 for the central bank or securities firm, 0.6 for ordinary institutions, and 0.3 for individuals). The frequency of information updates is illustrated here as follows: 1 for updates on the same day, 0.5 for updates every other day, and 0.2 for updates every 3 days or more.

[0104] Step 6: When the profile matching degree is less than the high matching degree threshold, a profile matching degree sequence is generated based on the profile matching degree, and a financial information sequence is formed based on the financial information preset by the user profile corresponding to the profile matching degree sequence.

[0105] A profile matching degree sequence refers to a sequence of user profiles arranged from low to high profile matching degree. The information category with the highest feature value for each user profile sequence is obtained from this sequence, and a financial information sequence is formed.

[0106] Step 7: When a financial information sequence exists, push it to the user's browsing page according to the financial information sequence.

[0107] The user's browsing page refers to the page the user enters when viewing financial information within the application. This page dynamically displays relevant information content based on a system-generated sequence of financial information. When a financial information sequence exists, the system will push it to the user's browsing page according to the sequence; otherwise, the system will push financial information to the user's browsing page using a preset information push method.

[0108] Step 8: Generate financial information based on the financial information sequence or the current user profile.

[0109] Financial information refers to financial news content automatically generated by the system based on various types of financial information contained in the financial news sequence, or on user interests and preferences reflected in the current user profile. This financial information not only covers market dynamics and industry analysis, but also includes individual stock recommendations, investment strategies, and other aspects, aiming to provide users with comprehensive, accurate, and targeted financial information services.

[0110] Step 9: Filter financial information based on user inquiry signals to obtain market forecast information, and output the market forecast information, which includes a portion of the financial information.

[0111] Market forecast information refers to content selected from financial information that is highly relevant to user inquiry signals and has predictive value.

[0112] Reference Figure 2 Methods for determining the current user profile when the profile matching degree is greater than the high matching degree threshold include:

[0113] Step 40: Obtain user search history and determine keywords based on user search history.

[0114] User search history refers to the search keywords and related search records entered by users within an application. Keywords are words or phrases extracted from user search history that represent the user's search intent and interests.

[0115] Step 41: Determine the user's weekly search frequency, total user search frequency, and target information category through keywords.

[0116] Weekly user search frequency refers to the number of times a user searches for a specific keyword within the past week. This data reflects the user's recent focus of interest. Total user search frequency refers to the total number of times a user searches for a specific keyword from the start of using the application to the current time. Target information category refers to the category of financial information most relevant to the keyword. This category is usually determined by matching the keyword with a pre-defined information category library, which records the mapping relationship between various keywords and information categories.

[0117] Step 42: Determine the keyword intent weight by using the user's weekly search frequency, total user search frequency, and target information category.

[0118] Keyword intent weight is a metric used to quantify the strength of intent conveyed by keywords in a user's search behavior. The formula for calculating keyword intent weight is as follows:

[0119] ,in The intent weight for the i-th search keyword. Let i be the weekly search frequency of the i-th keyword. This represents the total frequency of user searches per week. This refers to the number of clicks on the target information category after searching for the i-th keyword. This represents the total number of clicks after the search. and These are all obtained by the system after searching in the search field where the keywords appear, recording the number of clicks on financial information by users.

[0120] Step 43: Determine the core intent based on keyword intent weights and a pre-defined search intent tag library.

[0121] A search intent tag library is a pre-built collection containing various common search intents and their corresponding tags. These tags are typically derived from financial expertise and statistical analysis of user search behavior, accurately reflecting the true intent behind users' search keywords.

[0122] Core intent refers to the core objective extracted from user search history and keyword intent weights that best represents the user's current financial information needs and interests. The system matches the calculated keyword intent weights with various tags in the search intent tag library, selecting the tag with the highest matching degree as the user's core intent.

[0123] Step 44: Update the current user profile based on the core intent and the preset weight adjustment method.

[0124] The weighting adjustment method refers to a technique that dynamically adjusts the weights of profile feature values ​​corresponding to various types of financial information in a user profile based on the user's core intent. Specifically, the system first analyzes the financial information category corresponding to the core intent, then increases the weight of profile feature values ​​related to that category, while correspondingly decreasing the weights of other unrelated categories. This adjustment method is relatively simple and will not be elaborated upon further. This approach ensures that the user profile more closely reflects the user's current actual needs and changing interests.

[0125] Reference Figure 3 It also includes a method for generating an intent-exploratory information pool when no core intent exists, the method comprising:

[0126] Step 45: Determine the click-through rate of exploratory information based on the preset exploratory information.

[0127] Exploratory information refers to financial information proactively pushed to users by the system, aiming to uncover their potential interests or guide them to explore new areas. Exploratory information does not directly correspond to a user's current explicit search intent or profile characteristics. Instead, it is selected based on market hotspots, emerging trends, or peripheral areas that the user may be interested in. The system analyzes the popularity of financial information currently appearing on the platform and the keyword intent weight to determine which financial information the user may be interested in, and then pushes it out. The update method for exploratory information will be introduced later, and will not be elaborated here.

[0128] Step 46: Determine the total relevance of information categories by using the keyword intent weight and the corresponding financial information categories.

[0129] The total relevance of information categories refers to the sum of the degree of association between different information categories by analyzing the correlation between the intent weight of keywords and their corresponding financial information categories. Specifically, the system traverses all keywords, extracts the intent weight of each keyword and its corresponding financial information category, and then counts the search frequency and click behavior of the same user across different information categories to construct a relevance matrix between information categories. Through this quantitative analysis, the system can identify which financial information categories are at the core of the user's interest network and which are at the periphery.

[0130] Step 47: Calculate the push weight of exploratory information based on the sum of information category feature values, click-through rate of exploratory information, and information category relevance.

[0131] The push weight for exploratory news refers to the weight value corresponding to exploratory news, which affects the priority of pushing exploratory news to users' browsing pages. The specific formula for calculating the push weight of exploratory news is as follows:

[0132] ,in The push weight for the i-th type of exploratory information, User profile feature values ​​for the i-th type of financial information This contributes to the click-through rate of similar exploratory information on the platform.

[0133] Step 48: Sort the exploratory information in descending order based on the weight of the exploratory information push to obtain the exploratory information sequence, and push the corresponding information to the user's browsing page according to the exploratory information sequence.

[0134] An exploratory information sequence refers to a sequence of exploratory information arranged from high to low according to the push weight of exploratory information, and its format is a queue in data structure.

[0135] Reference Figure 4 It also includes a method for determining the sequence of corrected information, the method comprising:

[0136] Step 10: Obtain the corresponding market prediction tags through financial information. The prediction tags include the predicted sector, predicted trend, predicted cycle, and predicted probability value.

[0137] Prediction tags refer to key information extracted from financial news used to describe future market trends. The prediction section clarifies the financial sector covered by the information, such as stocks, bonds, and foreign exchange. Predicted trend refers to the possible upward or downward trend of the market. Prediction period refers to a time frame that defines the validity of the prediction result, such as short-term, medium-term, or long-term. Prediction probability value is a numerical value presented as a percentage, quantifying the reliability of the prediction result.

[0138] Step 11: Obtain the user click-through rate of information corresponding to the market prediction tag.

[0139] Click-through rate (CTR) refers to the percentage of users who actually click on financial information containing specific market forecast tags.

[0140] Step 12: After the forecast period ends, obtain the actual market trend and calculate the forecast deviation value based on the forecast trend and the actual market trend.

[0141] Actual market trend refers to the market rise or fall results obtained by collecting and analyzing real market data after the forecast period ends.

[0142] The prediction deviation value refers to the result obtained by comparing the predicted trend in the market prediction label with the actual observed market trend data after the prediction period ends. It reflects the degree of difference between the prediction result and the actual market performance. The specific calculation method adopts the directional matching method, that is, when the predicted trend is consistent with the actual trend, it is recorded as a positive deviation (the value can be set to 0, indicating no deviation), and when they are inconsistent, it is recorded as a negative deviation (a specific value is assigned according to the magnitude of the deviation between the predicted probability value and the actual value). For example, if a sector is predicted to rise in the short term with a prediction probability of 70%, but the sector actually falls, the deviation value needs to be calculated based on the difference between the magnitude of the fall and the predicted probability.

[0143] Step 13: When the prediction deviation value exceeds the preset maximum prediction deviation value, the corrected information category feature value is calculated based on the prediction deviation value and the information category feature value using the push strategy correction formula.

[0144] The maximum prediction deviation value refers to the maximum allowable difference between the predicted result and the actual market performance, preset by the system. The push strategy correction formula is a mathematical formula used to quantify users' level of trust in market prediction information and to adjust the pushed information based on this trust level. Specifically: ,in This represents the prediction deviation value for the i-th information category. The corrected information category feature value refers to the information category feature value after being corrected using the push strategy correction formula. Because when a piece of information frequently shows a large deviation between the predicted result and the actual result, it is easy to be disliked by users, so the push weight of that information should be reduced.

[0145] Step 14: Determine the correction information sequence based on the correction information category feature value, and push the correction information sequence to the user's browsing page.

[0146] A revised information sequence refers to a sequence generated by reordering financial information based on the revised information category feature values.

[0147] Reference Figure 5 It also includes methods for identifying and blocking high-risk information, including:

[0148] Step 15: Based on the financial information, obtain the corresponding financial sector, and obtain the corresponding market volatility, industry valuation deviation, and policy risk of the financial sector.

[0149] Market volatility refers to the magnitude of price fluctuations within a specific financial sector over a given period. It is calculated by statistically analyzing the standard deviation of the sector index over a predetermined statistical period, such as five minutes, one hour, twelve hours, twenty-four hours, or one week. Industry valuation deviation refers to the degree of deviation between the current market valuation of a specific financial sector and its historical average valuation or reasonable valuation range. Calculating industry valuation deviation involves a comprehensive analysis of indicators such as the price-to-earnings ratio and price-to-book ratio of companies within the sector, compared with historical data. Policy risk refers to the uncertainty arising from changes in government policies and regulations affecting a specific financial sector. Policy risk is determined by monitoring adjustments in macroeconomic policies, strengthening of industry regulatory policies, or changes in the international political and economic environment. Market volatility, industry valuation deviation, and policy risk are all common technical tools in this field and will not be elaborated upon further here.

[0150] Step 16: Determine the mean volatility of the market based on the overall market volatility, and obtain the volatility deviation through the overall market volatility and the mean volatility.

[0151] The average volatility of the broader market refers to the average volatility of a specific financial sector over a preset statistical period. This average is obtained by collecting volatility data of the sector index over multiple statistical periods and calculating their arithmetic mean. Volatility deviation refers to the degree of deviation between the current volatility of a specific financial sector and the average volatility of the broader market. It is calculated by subtracting the average volatility from the current volatility and taking the absolute value. This indicator is used to quantify abnormal price fluctuations in the sector. When the volatility deviation exceeds a preset threshold, the system will determine that the sector faces the risk of short-term price manipulation or irrational volatility.

[0152] Step 17: Calculate the real-time risk level based on volatility deviation, industry valuation deviation, and policy risk.

[0153] Real-time risk level is an indicator used to quantify the current risk level of a specific financial sector. The specific formula for calculating real-time risk level is as follows:

[0154] ,in The real-time risk level corresponding to the i-th type of information. This is a real-time market volatility index. This represents the average volatility of the broader market over the past three months. To measure the deviation of short-term market volatility, Let be the deviation between the industry price-to-earnings ratio corresponding to the i-th type of information and its historical average. This represents the maximum deviation between the price-to-earnings ratio (P / E ratio) of the industry corresponding to the i-th type of information and its historical average over the past five years. This refers to the policy risk coefficient. (The above content...) , , and This information is common in this field, compiled and published by various professional organizations. If there is a category of financial information with a corresponding financial section that does not exist... , , and If the information is not considered, then the risk level will not be calculated, and the information will not be recommended to users with poor risk tolerance in the subsequent process. Risk tolerance will be determined by the risk tolerance coefficient in the subsequent content, which will not be elaborated here.

[0155] Step 18: Determine high-risk information based on the real-time risk level, and adjust and correct the information sequence based on the high-risk information to determine the current corrected information sequence, and push it to the user's browsing page according to the current corrected information sequence.

[0156] High-risk information refers to financial information whose real-time risk level exceeds a preset safe volatility threshold. This type of information typically involves sectors with drastic price fluctuations, significant valuation deviations, or highly uncertain policy environments. The safe volatility threshold is a pre-set critical value used by the system to define the risk level of financial information. When the real-time risk level exceeds this threshold, the system will automatically identify and mark the relevant information as high-risk.

[0157] The current corrected information sequence refers to the sequence generated by the system after identifying high-risk information and dynamically adjusting the original corrected information sequence. The specific adjustment rule is: the sequence after removing high-risk information from the corrected information sequence.

[0158] This also includes methods for not blocking high-risk information, including:

[0159] Step 19: Calculate the user preference dimension using information category feature values, predicted trust level, and core intent.

[0160] The user preference dimension is a comprehensive indicator used to quantify users' preferences for different types of financial information. The specific formula for calculating the user preference dimension is as follows:

[0161] ,in For the user preference dimension corresponding to the i-th type of information, These are the profile feature values ​​corresponding to high-risk information. This assigns intent weights to keywords related to risk information. Since some users have a higher risk tolerance and can be excluded from the current revised information sequence (i.e., high-risk information is not excluded), the user preference dimension is calculated to subsequently assess a user's risk tolerance.

[0162] Step 20: Calculate the user's risk tolerance coefficient based on user preference dimensions and real-time risk level.

[0163] The user risk tolerance coefficient is an indicator used to quantify a user's ability to withstand high-risk financial information. Its specific calculation formula is as follows: ,in This is the user's risk tolerance factor.

[0164] Step 21: When the user's risk tolerance coefficient exceeds the preset high risk tolerance threshold, the information will be pushed to the user's browsing page based on the corrected information sequence.

[0165] The high risk tolerance threshold refers to a pre-set threshold used by the system to determine whether a user has the ability to withstand high-risk financial information. When a user's risk tolerance coefficient is greater than or equal to this threshold, the system considers the user to have a strong risk tolerance and be able to accept and handle the uncertainty brought about by high-risk information. At this time, the system will no longer block high-risk information, but will directly push it based on the corrected information sequence.

[0166] Step 22: When the user's risk resistance coefficient does not exceed the high risk resistance threshold, push the current corrected information sequence to the user's browsing page.

[0167] When a user's risk tolerance coefficient does not exceed the high-risk threshold, it indicates that the user's risk tolerance is relatively weak, and the system determines that they may not be able to effectively withstand the uncertainty brought about by high-risk financial information. In this case, to protect the user's interests and avoid potential losses, the system will not directly push high-risk information, but will continue to push information based on the current corrected information sequence.

[0168] This also includes a method for constructing cross-market information correlation graphs, which includes:

[0169] Step 23: Obtain the frequency of association of corresponding keywords in different markets in real time.

[0170] Association frequency refers to the number of times keywords from different market financial information appear together within a preset time window, i.e., the number of pieces of information that simultaneously mention market A and market B. The system crawls news sources from multiple markets, social media discussions, and trading data to statistically analyze the co-occurrence frequency of specific keyword combinations (such as "crude oil price" and "US dollar index"). For example, when the international crude oil market fluctuates, the system monitors the association frequency of keywords such as "crude oil," "OPEC," "US dollar," and "inflation" in cross-market information to quantify the inter-market linkage effect. This data is collected in real-time via distributed web crawlers. The calculation of association frequency uses a sliding window model, with the window length dynamically adjusted according to market activity, shortened to 5 minutes during high-frequency trading periods and extended to 1 hour during low-activity periods to balance computational efficiency and data timeliness. The above-mentioned calculation of association frequency is a common existing technology in this field and will not be elaborated upon here.

[0171] Step 24: Determine the correlation coefficients for different markets based on the overall market data.

[0172] The correlation coefficient is an indicator used to quantify the strength of the correlation between different markets. Specifically, it refers to the price linkage coefficient of market B within 24 hours after the release of information in market A. The calculation method involves analyzing the covariance and variance of each market index in the overall market data to obtain the Pearson correlation coefficient, which is then adjusted by weighting factors such as market trading volume and capital flows. For example, when calculating the correlation coefficient between A-shares and Hong Kong stocks, the system first extracts the daily return series of the two market indices, calculates the product of their covariance and their respective standard deviations to obtain the basic correlation coefficient, and then introduces the capital flow data of Shanghai-Hong Kong Stock Connect and Shenzhen-Hong Kong Stock Connect as weights. If the net inflow of northbound capital accounts for more than 5% of the A-share turnover on a given day, the correlation coefficient is positively adjusted. The correlation coefficient ranges from -1 to 1, with the absolute value closer to 1 indicating stronger market linkage. This calculation method integrates statistical indicators and market microstructure data, and can more accurately reflect the true degree of correlation between markets.

[0173] Step 25: Calculate the correlation degree using correlation frequency and correlation coefficient.

[0174] Correlation is an indicator used to comprehensively measure the strength of information correlation between different markets. Its specific calculation formula is as follows: , here For relevance, To simultaneously mention the amount of information about both market A and market B, The maximum amount of information for a single market. The price correlation coefficient of market B within 24 hours after the release of information in market A. The historical average correlation coefficient is used to determine the above variables through correlation frequency and correlation coefficient.

[0175] Step 26: When the correlation exceeds the preset strong correlation threshold, the corresponding markets are grouped together and marked as strong correlation markets and recorded in the preset strong correlation market table.

[0176] The strong correlation threshold is a pre-set threshold used by the system to determine whether the correlation between different markets has reached a significant level. Strongly correlated markets are those identified by the system as having a significant linkage, exhibiting a high degree of consistency in information dissemination, price fluctuations, or capital flows. For example, when the correlation between the crude oil market and the airline stock market consistently exceeds the strong correlation threshold, the system will mark them as strongly correlated markets and record them in a pre-set strong correlation market table. The strong correlation market table is a database table used to store information on market combinations identified as strongly correlated. The table structure includes fields such as market combination identifier, associated market name, correlation score, first marking time, and last update time.

[0177] Step 27: Determine the market preference value corresponding to the current user profile based on the core intent, user risk preference, and information category feature value.

[0178] Market preference score is an indicator used to quantify the degree of attention a user currently pays to different markets. It is calculated by combining the information category feature value corresponding to the current user profile with the user's risk preference and core intent. The information category feature value identifies the financial sectors the user is most interested in, and the user's risk preference identifies the financial sectors with the highest risk tolerance. Finally, the user's core intent is analyzed. The calculation formula here is similar to the user preference dimension calculation formula mentioned above, so it will not be repeated.

[0179] Step 28: Determine single-market preferences and cross-market preferences based on market preference values.

[0180] Single-market preference refers to a user's tendency to focus more on a specific market than on other markets. Cross-market preference refers to a user's tendency to maintain high attention to multiple interconnected markets simultaneously. Specifically, it is determined by analyzing market preference values ​​against a table of strongly correlated markets. When a market preference value exceeds a preset standard market preference threshold, the user's preferred market is identified. Then, the corresponding strongly correlated market within the strongly correlated markets is located. If the strongly correlated market also exceeds the standard market preference threshold, cross-market preference is confirmed; otherwise, single-market preference is determined.

[0181] Step 29: When the user has a single market preference, execute steps 1 through 9.

[0182] When a user has a single market preference, it means that the user's attention to a specific market is significantly higher than that to other markets. In this case, the system will execute a complete information processing flow for the preferred market, so steps 1 to 9 will be executed.

[0183] Step 30: When a user has a cross-market preference, determine the current preferred market based on the information category feature value.

[0184] The current preference market refers to the market that the user is most interested in among all preference markets. It is determined by sorting the information category feature values ​​and selecting the information category with the highest information category feature value.

[0185] Step 31: Determine the current associated market based on the current preference market and strong association market table, and adjust the current correction information sequence according to the current associated market and push it to the user's browsing page.

[0186] The current associated market refers to the set of markets that have a strong association with the current preferred market. The strong association is the associated market record that exists in the current preferred market in the strong associated market table.

[0187] The methods for pushing information to users' browsing pages based on the corrected information sequence include:

[0188] Step 210: Extract and analyze the information content corresponding to all financial information in the corrected information sequence to obtain positive and negative information.

[0189] Positive news refers to information that has a positive impact on a specific financial sector or individual stock and may drive up prices. Negative news refers to information that has a negative impact on a specific financial sector or individual stock and may cause prices to fall. The methods for determining positive and negative news will be introduced later and will not be elaborated here. When corrective information is available, it indicates that the user has a strong risk tolerance, so the priority for pushing positive news should be increased.

[0190] Step 211: Adjust the feature values ​​of the information category corresponding to the positive information according to the preset information priority adjustment formula, and form the final information sequence.

[0191] The information priority adjustment formula refers to the formula for adjusting the feature values ​​of information categories. Specifically, it increases or decreases the feature values ​​of corresponding information categories using a preset incrementing or decrementing function, which is manually preset. The final information sequence refers to the sorted list of information categories generated after adjustment by the information priority adjustment formula, containing the updated feature values ​​of the information categories.

[0192] Step 212: Push the corresponding financial information to the user's browsing page according to the final information sequence.

[0193] Step 2130: When current corrected information exists, extract the information content corresponding to all financial information in the corrected information sequence and analyze it to obtain positive and negative information.

[0194] The existence of current correction information indicates that the user's risk tolerance is low, so the priority of pushing negative information should be increased.

[0195] Step 2131: Adjust the information category feature value corresponding to the negative information according to the information priority adjustment formula, and form the final information sequence to execute step 212.

[0196] The specific process is consistent with the adjustment of the consultation category feature values ​​corresponding to the positive information mentioned above, and will not be elaborated further here.

[0197] It also includes methods for identifying positive and negative information, including:

[0198] Step 2100: Extract text information from the news based on the built-in AI port to determine the expected positive and negative news.

[0199] The AI ​​portal refers to the natural language processing module built into the system. This module uses deep learning algorithms to perform semantic analysis on financial information text, enabling it to identify the implied market impact direction within the information content. "Expected positive information" refers to information that, based on the AI ​​portal's preliminary assessment, may have a positive impact on the financial market. "Expected negative information" refers to information that, based on the AI ​​portal's preliminary assessment, may have a negative impact on the financial market.

[0200] Step 2101: Retrieve the historical accuracy rates corresponding to the expected positive and negative information.

[0201] Historical accuracy refers to the proportion of times that the actual market trend matches the predicted direction after an account or platform publishes similar information (such as predicted positive or negative news). The system maintains a historical database recording the positive / negative news published by each information source (such as specific financial media or analyst accounts) and its subsequent market reaction. Accuracy is calculated in layers based on information category (such as policy-related or industry data-related) and time window (such as the past 3 months or the past year). For example, if a financial account published positive news such as "central bank reserve requirement ratio cut" 10 times in the past 6 months, and 8 of those predictions led to a rise in related sectors, its historical accuracy is 80%. By retrieving this historical data, the system can assess the reliability of predictions from different information sources, providing a reference for subsequent steps.

[0202] Step 2102: When the historical accuracy rate reaches the preset reliable accuracy rate, obtain the same market analysis information corresponding to the expected positive information and the expected negative information.

[0203] Reliability accuracy refers to a pre-set threshold used by the system to determine whether the reliability of an information source's predictions reaches a trustworthy level. Similar market analysis information refers to other information that analyzes the same market or related financial sector as the predicted positive or negative information. For example, when the AI ​​identifies a predicted positive news article about "favorable policies for new energy vehicles," the system will retrieve other analytical information about the new energy vehicle industry published by that information source or other high-accuracy information sources, avoiding misjudgments caused by the bias of a single piece of information.

[0204] When the historical accuracy rate of anticipated positive or negative information reaches or exceeds this threshold, the system considers the prediction from that information source to have high credibility.

[0205] Step 2103: Extract text information of the same market analysis information from the AI ​​port to determine the comparison information. The prediction result of the comparison information is either expected to be positive or expected to be negative.

[0206] Comparative information refers to information targeting the same market as expected positive or negative information. Here, comparative information refers to the prediction results obtained by analyzing the information through an AI portal.

[0207] Step 2104: When the expected positive information is consistent with the prediction result of the corresponding reference information, the positive information is determined.

[0208] When the predicted positive information matches the prediction result of the corresponding comparative information, it indicates that the predicted positive information has a high degree of credibility, and the system will identify it as positive information.

[0209] Step 2105: When the predicted negative information is consistent with the prediction result of the corresponding reference information, the negative information is determined.

[0210] When the predicted negative information matches the prediction result of the corresponding comparative information, it indicates that the predicted negative information has a high degree of credibility, and the system will identify it as negative information.

[0211] Based on the same inventive concept, embodiments of the present invention provide a market information forecasting system based on financial consulting analysis.

[0212] One of them, a market information forecasting system based on financial consulting analysis, includes:

[0213] The acquisition module is used to acquire user browsing information and user search history;

[0214] The memory is used to store the program for implementing a control method that uses a market information forecasting method based on financial consulting analysis;

[0215] The processor loads and executes programs from memory.

[0216] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A market information forecasting method based on financial consulting analysis, characterized in that, include: Step 1: Respond to user inquiry signals and collect user browsing information in real time; Step 2: Calculate the expected user profile based on the user browsing information and the preset user profile feature value formula; Step 3: Based on the projected user profile and the preset user profile set, determine the profile matching degree using a preset matching degree formula; Step 4: When the profile matching degree is greater than the preset high matching degree threshold, the current user profile is determined. The current user profile includes information category feature values ​​corresponding to various financial sector information categories. Step 5: Determine the sequence of financial information to be pushed based on the current user profile; Step 6: When the profile matching degree is less than the high matching degree threshold, a profile matching degree sequence is generated based on the profile matching degree, and a financial information sequence is formed based on the financial information preset by the user profile corresponding to the profile matching degree sequence. Step 7: When the financial information sequence exists, push it to the user's browsing page according to the financial information sequence; Step 8: Generate financial information based on the financial information sequence or the current user profile; Step 9: Filter the financial information based on the user inquiry signal to obtain market forecast information, and output the market forecast information, which includes a portion of the financial information.

2. The market information forecasting method based on financial consulting analysis according to claim 1, characterized in that, The method for determining the current user profile when the profile matching degree is greater than the high matching degree threshold includes: Step 40: Obtain user search history and determine keywords based on the user search history; Step 41: Determine the user's weekly search frequency, total user search frequency, and target information category based on the keywords; Step 42: Determine the keyword intent weight based on the user's weekly search frequency, the user's total search frequency, and the target information category; Step 43: Determine the core intent based on the keyword intent weights and the preset search intent tag library; Step 44: Update the current user profile based on the core intent and the preset weight adjustment method.

3. The market information forecasting method based on financial consulting analysis according to claim 2, characterized in that, It also includes a method for generating an intent-exploratory information pool when the core intent does not exist, the method comprising: Step 45: Determine the click-through rate of exploratory information based on the preset exploratory information; Step 46: Determine the total relevance of information categories by using the keyword intent weight and the corresponding financial information categories; Step 47: Calculate the push weight of exploratory information based on the sum of information category feature values, click-through rate of exploratory information, and information category relevance; Step 48: Sort the exploratory information in descending order based on the exploratory information push weight to obtain the exploratory information sequence, and push the corresponding information to the user's browsing page according to the exploratory information sequence.

4. The market information forecasting method based on financial consulting analysis according to claim 3, characterized in that, It also includes a method for determining the sequence of corrected information, the method comprising: Step 10: Obtain the corresponding market prediction tags through financial information. The prediction tags include the predicted sector, predicted trend, predicted cycle, and predicted probability value. Step 11: Obtain the user click-through rate of information corresponding to the market prediction tag; Step 12: After the forecast period ends, obtain the actual market trend and calculate the forecast deviation value based on the forecast trend and the actual market trend; Step 13: When the prediction deviation value exceeds the preset maximum prediction deviation value, the corrected information category feature value is calculated based on the prediction deviation value and the information category feature value using the push strategy correction formula; Step 14: Determine the correction information sequence based on the correction information category feature value, and push the correction information sequence to the user's browsing page.

5. The market information forecasting method based on financial consulting analysis according to claim 4, characterized in that, It also includes methods for identifying and blocking high-risk information, including: Step 15: Based on the financial information, obtain the financial sector corresponding to the financial information, and obtain the market volatility, industry valuation deviation and policy risk corresponding to the financial sector; Step 16: Determine the mean volatility of the market based on the overall market volatility, and obtain the volatility deviation through the overall market volatility and the mean volatility. Step 17: Calculate the real-time risk level based on volatility deviation, industry valuation deviation, and policy risk; Step 18: Determine high-risk information based on the real-time risk level, and determine the current corrected information sequence based on the high-risk information, and push the current corrected information sequence to the user's browsing page.

6. The market information forecasting method based on financial consulting analysis according to claim 5, characterized in that, This also includes methods for not blocking high-risk information, including: Step 19: Calculate the user preference dimension using information category feature values, predicted trust level, and core intent; Step 20: Calculate the user's risk tolerance coefficient based on user preference dimensions and real-time risk level; Step 21: When the user's risk tolerance coefficient exceeds the preset high risk tolerance threshold, push the corrected information sequence to the user's browsing page; Step 22: When the user's risk resistance coefficient does not exceed the high risk resistance threshold, the information is pushed to the user's browsing page according to the current correction information sequence.

7. The market information forecasting method based on financial consulting analysis according to claim 6, characterized in that, It also includes a method for constructing cross-market information relationship graphs, which includes: Step 23: Obtain the association frequency of corresponding keywords in different markets in real time; Step 24: Determine the correlation coefficients for different markets based on the overall market data; Step 25: Calculate the degree of association using association frequency and association coefficient; Step 26: When the correlation exceeds the preset strong correlation threshold, the corresponding markets are grouped together and marked as strong correlation markets and recorded in the preset strong correlation market table; Step 27: Determine the market preference value corresponding to the current user profile based on the core intent, user risk preference, and information category feature value; Step 28: Determine single-market preferences and cross-market preferences based on market preference values; Step 29: When the user has a single market preference, execute steps 1 through 9; Step 30: When a user has cross-market preferences, determine the current preferred market based on the information category feature value; Step 31: Determine the current associated market based on the current preference market and strong association market table, and adjust the current correction information sequence according to the current associated market and push it to the user's browsing page.

8. The market information forecasting method based on financial consulting analysis according to claim 6, characterized in that, Methods for pushing information to users' browsing pages based on corrected information sequences include: Step 210: Extract and analyze the information content corresponding to all financial information in the corrected information sequence to obtain positive and negative information; Step 211: Adjust the feature values ​​of the information categories corresponding to the positive information according to the preset information priority adjustment formula, and form the final information sequence; Step 212: Push the corresponding financial information to the user's browsing page according to the final information sequence; Step 2130: When current correction information exists, extract the information content corresponding to all financial information in the correction information sequence and analyze it to obtain positive and negative information; Step 2131: Adjust the information category feature value corresponding to the negative information according to the information priority adjustment formula, and form the final information sequence to execute step 212.

9. A market information forecasting method based on financial consulting analysis according to claim 8, characterized in that, It also includes methods for identifying positive and negative information, including: Step 2100: Extract text information from the news based on the built-in AI port to determine the expected positive and negative news; Step 2101: Retrieve the historical accuracy rates corresponding to the expected positive and negative information; Step 2102: When the historical accuracy rate reaches the preset reliable accuracy rate, obtain the same market analysis information corresponding to the expected positive information and the expected negative information; Step 2103: Extract text information of the same market analysis information from the AI ​​port to determine the comparison information. The prediction result of the comparison information is expected to be positive or negative. Step 2104: When the expected positive information is consistent with the prediction result of the corresponding comparative information, the positive information is determined; Step 2105: When the predicted negative information is consistent with the prediction result of the corresponding comparative information, the negative information is determined.

10. A market information forecasting system based on financial consulting analysis, characterized in that, include: The acquisition module is used to acquire user browsing information and user search history; A memory for storing the program of a control method for a market information forecasting method based on financial consulting analysis as described in any one of claims 1 to 9; The processor loads and executes programs from memory.

Images