Hydrological data accurate recommendation method and system

Abstract

The application discloses a kind of hydrological data accurate recommendation method and system, method includes: the analysis of customer demand, based on hydrological database establishes hydrological calculation feature library, calculates similar river basin, similar site, similar hydrological demand and the list of data that meet;Establish customer feature library, calculate user portrait;Combining hydrological calculation feature library and customer feature library, all possible data combination scheme calculates hydrological data recommendation scheme index, selects the scheme with the highest hydrological data recommendation scheme index as final recommended scheme.The scheme is more than 60% than traditional keyword matching method recommendation accuracy rate is improved;Introduce river basin similarity judgment mechanism, further improve the correlation and applicability of hydrological data recommendation.

Description

Technical Field

[0001] This invention relates to the field of hydrological information technology, and in particular to a method and system for accurate recommendation of hydrological data. Background Technology

[0002] Hydrological data includes various hydrological elements such as precipitation, runoff, water level, and water quality, and is a crucial foundation for water resources management, water conservancy project design, flood control and drought relief decision-making, and hydrological research. With the development of hydrological monitoring technology, the amount of hydrological data has grown exponentially, forming a massive amount of multi-source heterogeneous data resources.

[0003] Currently, the recommendation of hydrological data mainly relies on manual classification and retrieval or simple keyword matching, which has the following drawbacks:

[0004] (1) Low accuracy, unable to accurately understand the user's deep needs, and poor relevance of recommendation results;

[0005] (2) Insufficient adaptability, making it difficult to handle the spatiotemporal characteristics of hydrological data and the personalized needs of users;

[0006] (3) The update is delayed and the recommendation strategy cannot be dynamically optimized based on user behavior.

[0007] The above problems result in low efficiency for users to obtain effective hydrological data, hindering the full realization of the value of hydrological data. Previous hydrological data did not incorporate actual calculation needs, leading to inaccurate cost calculations or costs that did not meet actual calculation requirements. Summary of the Invention

[0008] To address the aforementioned technical problems, this invention provides a method and system for accurate hydrological data recommendation, overcoming the shortcomings of low accuracy and poor adaptability in existing hydrological data recommendation technologies.

[0009] This invention is achieved using the following technical solution:

[0010] Firstly, a method for accurately recommending hydrological data includes the following steps:

[0011] Step S1: Analyze customer needs, establish a hydrological calculation feature library based on the hydrological database, and calculate similar watersheds, similar stations, similar hydrological requirements, and a list of data that meet the requirements.

[0012] Step S2: Establish a customer feature database and calculate user profiles;

[0013] Step S3: Combining the hydrological calculation feature library and the customer feature library, calculate the hydrological data recommendation scheme index for all possible data combination schemes, and select the scheme with the highest hydrological data recommendation scheme index as the final recommended scheme.

[0014] Specifically, the analysis of customer needs in step S1 includes: receiving the user's input demand text, and extracting demand features through natural language processing, including demand keywords, demand type, time range, and spatial range for hydrological data.

[0015] Specifically, the calculation of the hydrological data recommendation scheme index in step S3 includes:

[0016] Calculate and define each feature factor, dynamically adjust the weight coefficients of each feature factor according to the project type, and calculate the recommendation index. The calculation formula is as follows:

[0017] ;

[0018] in, ~ These are the weighting coefficients corresponding to each characteristic factor; The distance between the project site and the hydrological station; For watershed similarity; The calculation rules and formulas require, including both minimum and maximum requirements; To ensure calculation accuracy, settings are configured within a range of 0% to 100% based on customer requirements; To meet customer needs, including just meeting and fully meeting requirements; The cost calculated based on the data combination scheme; This is the adjustment coefficient for the recommendation strategy derived from user profiles; The larger the value, the better the solution matches the customer's needs.

[0019] Specifically, the distance between the project address and the hydrological station The traditional method is used to calculate the distance between the target site of the project and multiple hydrological stations. The calculation formula is as follows:

[0020] ;

[0021] in, The three-dimensional coordinates of the target site for the project are represented as follows: The three-dimensional coordinates of the hydrological station are represented as follows: … … , Indicates the first Hydrological stations and .

[0022] Specifically, the calculation of watershed similarity selects judgment indicators from two dimensions: natural geographical features and hydrological features. An improved cosine similarity algorithm is used to calculate the overall similarity between watersheds, as shown in the formula:

[0023] ;

[0024] in, For the first The weights of each characteristic factor indicator and , representing the weight coefficients corresponding to each characteristic factor. ~ ; For the first The cosine similarity of the characteristic factor indicators; the natural geographical features include watershed area, average elevation, average slope and vegetation coverage, and the hydrological features include multi-year average precipitation, multi-year average runoff, runoff coefficient and flood duration; at the same time, a similarity threshold for watershed similarity is set, and watersheds with similarity to the target watershed above the threshold are selected, and their relevant hydrological data are included in the recommended candidate set.

[0025] Specifically, the cost is calculated as follows:

[0026] ;

[0027] in, The service index is 0 for free and 1 for paid. This is the discount factor, with a value range of (0,1]. For the first The first site Item name, For data prices, This refers to the number of documents.

[0028] Specifically, the calculation rules and formulas require To measure the quantitative indicators of the recommended hydrological data combination and its derived calculation scheme, The value is obtained through the matching degree function. The matching degree function is calculated with a value range of [0,1], where 1 represents a perfect match. The calculation formula is expressed as:

[0029] ;

[0030] in, Indicates the matching weight of the calculation method. This indicates the matching degree of the calculation method; if it matches, then... Equals 1, if not matched Equal to 0; The weight represents the data completeness matching degree. Indicates the degree of data completeness and matching. = (Number of complete data items required in the recommendation scheme) / (Total number of data items required); This indicates the weight of the matching degree of the output results. This indicates the matching degree of the output results; a match indicates a good match. Equals 1, if not a match It equals 0.

[0031] Specifically, the customer needs The value is obtained through the sufficiency function. To evaluate, it is represented as:

[0032] ;

[0033] in, Indicates basic satisfaction level. Indicates overflow reward. This indicates a lack of punishment.

[0034] Specifically, the recommendation strategy adjustment coefficient derived from the user profile The recommendation strategy continuously optimizes the behavioral weights of user profile tags and recommendation model parameters based on user feedback on the recommendation results through a reinforcement learning mechanism; the recommendation strategy adjustment coefficient... The value is the behavioral weight of the user profile tag. The calculation formula is expressed as:

[0035] ;

[0036] in, As the basic weight of behavior, The attenuation coefficient is... The time interval between the occurrence of the behavior and the present. For duration coefficient, This represents the natural exponential function.

[0037] On the other hand, a hydrological data accurate recommendation system, used to implement the aforementioned hydrological data accurate recommendation method, includes the following modules:

[0038] Hydrological data storage module: Compiles and stores hydrological data from hydrological stations;

[0039] User Requirement Analysis Module: Receives user input of requirement text and uses a natural language processing model to extract requirement keywords, requirement type, time range, and spatial range;

[0040] Precise matching recommendation module: Uses a fusion algorithm to generate a recommendation list;

[0041] Dynamic update module: Connects the recommendation precision matching module and the user interaction module interface, updates model parameters based on user behavior; updates user demand feature weights and recommendation model parameters based on user clicks, downloads, and ratings of recommendation results;

[0042] User interaction module: Displays recommendation results and receives user feedback.

[0043] The beneficial effects of this invention are as follows: By combining and applying multiple professional algorithms, it achieves precise processing throughout the entire process from data preprocessing to recommendation matching, improving the recommendation accuracy by more than 60% compared to traditional keyword matching methods; the introduction of a watershed similarity judgment mechanism further improves the relevance and applicability of hydrological data recommendations; through incremental training and dynamic weighting mechanisms, the system can continuously adapt to updates in hydrological data and changes in user needs, maintaining stable performance in various scenarios such as scientific research and flood control; the regular updates of the watershed feature database ensure the timeliness of similarity judgment, keeping the recommendation response time for millions of hydrological data within 1 second. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of the method for accurately recommending hydrological data according to the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0047] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0048] The following is in conjunction with the appendix Figure 1 The following describes some embodiments of the present invention in detail. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0049] This invention proposes a method and system for accurate recommendation of hydrological data. In a preferred embodiment, the method is as follows: Figure 1 As shown, it includes:

[0050] Step S1: Analyze customer needs, establish a hydrological calculation feature library based on the hydrological database, and calculate similar watersheds, similar stations, similar hydrological needs, and a list of data to be satisfied; among them, the analysis of customer needs includes: receiving the user's input demand text, extracting demand features through natural language processing, including the demand keywords, demand type, time range, and spatial range of hydrological data.

[0051] Step S2: Establish a customer feature database and calculate user profiles;

[0052] Step S3: Combining the hydrological calculation feature library and the customer feature library, calculate the hydrological data recommendation scheme index for all possible data combination schemes (such as combinations of different stations and different year series), and select the scheme with the highest hydrological data recommendation scheme index as the final recommended scheme; in this scheme, the recommended scheme types include multiple data from one station and multiple data from multiple stations.

[0053] In this embodiment, the formula for calculating the exponent is expressed as:

[0054] ;

[0055] in, ~ The weighting coefficients for each characteristic factor are dynamically adjusted according to the project type. To ensure computational accuracy, in this scheme, the computational accuracy is... The settings are configured within a range of 0% to 100% based on customer needs, and the customer determines whether fuzzy search is required. The larger the value, the better the solution matches the customer's needs.

[0056] In this embodiment, the distance between the project address and the hydrological station The traditional method is used to calculate the distance between the target site of the project and multiple hydrological stations. The calculation formula is as follows:

[0057] ;

[0058] in, The three-dimensional coordinates of the target site for the project are represented as follows: The three-dimensional coordinates of the hydrological station are represented as follows: … … , Indicates the first Hydrological stations and .

[0059] The watershed similarity is obtained by calculating the similarity of each indicator and its corresponding weight. In this embodiment, the calculation of watershed similarity selects judgment indicators from two dimensions: natural geographical features and hydrological features. Natural geographical features include watershed area, average elevation, average slope, and vegetation cover. The hydrological features include multi-year average precipitation, multi-year average runoff, runoff coefficient, and flood duration. The overall similarity between watersheds is calculated using an improved cosine similarity algorithm, with the formula:

[0060] ;

[0061] in, For the first The weight of each characteristic factor indicator represents the weight coefficient corresponding to each characteristic factor. ~ In one embodiment, the weights of natural geographic features are determined to be 0.4 and the weights of hydrological features are determined to be 0.6 using the AHP low-level analysis algorithm. These weights can also be set according to customer needs. For the first The cosine similarity of each characteristic factor index is used; a similarity threshold (e.g., 0.75) is set for watershed similarity, and watersheds with similarity to the target watershed above the threshold are selected, and their relevant hydrological data are included in the recommended candidate set.

[0062] In one embodiment, the calculation of watershed similarity may include more than 20 geographical indicators, such as: calculating and evaluating the similarity of rivers with a surface area of ​​more than 10 square kilometers, and establishing a similar watershed feature database; using the knowledge graph formula, cos(A,B), to select judgment indicators from two dimensions: natural geographical features and hydrological features; natural geographical features include watershed area (F), average elevation (H), average slope (S), and vegetation coverage (C); hydrological features include multi-year average precipitation (P), multi-year average runoff (Q), runoff coefficient (η), and flood duration (T); setting a similarity threshold (e.g., 0.75), filtering out watersheds with a similarity higher than the target watershed, and including their relevant hydrological data in the recommended candidate set.

[0063] The calculation rules and formulas require, including minimum and maximum requirements; in this embodiment, To measure the quantitative indicators of the recommended hydrological data combination and its derived calculation scheme, The value is obtained through the matching degree function. The matching degree function is calculated with a value range of [0,1], where 1 represents a perfect match. The calculation formula is expressed as:

[0064] ;

[0065] in, Indicates the matching weight of the calculation method. This indicates the degree of matching in calculation methods. For example: If a user needs to calculate the design flood using the inference formula method, the recommended watershed data must support this method (including rainfall parameters), not just regional empirical formulas. If a user needs to conduct water resource assessments, the data must include long-term runoff data, not just water level data; a good match indicates... Equals 1, if not matched It equals 0.

[0066] The weight represents the data completeness matching degree. This indicates the completeness and matching degree of the data, whether the recommended data items meet all the input variables required by the calculation formula. For example: the calculation formula requires the input of the maximum peak flow Qm and the total flood volume W → the recommended solution must include both of these data points; neither can be omitted. = (Number of complete required data items in the recommendation scheme) / (Total number of required data items).

[0067] This indicates the weight of the matching degree of the output results. This indicates the output matching degree: whether the recommended solution can directly output the type of output requested by the user. For example: the user requests a 10-year flood peak flow rate → the recommended solution should provide at least a 30-year annual maximum flow rate series to meet the requirements of frequency analysis. The user requests a daily average flow rate process line → the recommended solution should provide daily-scale flow rate data, not monthly-scale data; a match is indicated by... Equals 1, if not matched It equals 0.

[0068] The weights of the three matching degrees mentioned above are usually... (Calculation Method) has the highest weight due to its decisive role. The weight can be preset by the system based on experience or adjusted through machine learning. Example: If the user requirement is "to use the hydrological analogy method to estimate the design flood in areas with no data," then:

[0069] The recommended approach must provide design flood results or flood parameters for the reference watershed; otherwise, the method cannot be applied. =0;

[0070] Hydrological analogy methods typically require watershed area as a parameter; therefore, the recommended scheme must include the area data of the reference watershed.

[0071] Ultimately, it must be able to output the design peak flow or flood process curve of the target watershed.

[0072] In one embodiment, It is a quantitative indicator used to measure the degree of matching between the recommended hydrological data combination and its derived calculation scheme and the specific calculation rules, industry standard formulas, and output types required by the user project. It ensures that the recommended data can not only be used, but also directly meet the user's final calculation objectives.

[0073] Customer needs, including just meeting and fully meeting requirements, is a metric that measures how much a recommended solution exceeds the user's stated basic needs. It introduces the concepts of just meeting and fully meeting to distinguish the best solutions from multiple feasible options, minimizing subsequent workload and ensuring customer satisfaction. A higher value indicates a more comprehensive and reliable recommended solution, but may also be more costly. In this embodiment, customer needs... The value is obtained through the sufficiency function. To evaluate, it is represented as:

[0074] ;

[0075] in, This indicates the basic satisfaction level. When the solution just meets the customer's minimum requirements, it is recorded as 1.0, which is the baseline.

[0076] This represents the overflow bonus, measuring the additional value of a solution that exceeds customer expectations. For example:

[0077] Data Length Bonus: Clients require at least 20 years of data; the recommended plan provides 30 years. Bonus = k_years * (30 - 20) (where k_years is the bonus coefficient).

[0078] Data accuracy bonus: The client did not explicitly require accuracy, but recommended higher-precision data (e.g., compiled data > raw data);

[0079] Site density bonus: The client requested one reference site, but two highly similar reference sites were recommended, forming cross-validation and greatly increasing reliability;

[0080] Additional data bonus: We recommended relevant data that the client had not mentioned but which was of great benefit to their calculations (such as providing an additional rainstorm cause analysis report when extrapolating floods).

[0081] This indicates a lack of penalty. A penalty is imposed if the solution fails to fully meet customer requirements in some aspects. For example, if there are missing years in the data year sequence that cannot be imputed, points are deducted based on the degree of missing data. The penalty value is usually negative. Example:

[0082] Client requirements: A maximum annual flow sequence of at least 20 years is needed for frequency analysis;

[0083] Option A (just meets the requirements): Provides a maximum annual flow sequence for exactly 20 years, with no missing data. = 1.0;

[0084] Option B (fully satisfied): Provide a 30-year annual maximum flow series, including relevant pre-survey data for rainstorm and flood investigation. = 1.0 + (k_years * 10) + (incentive during the survey period) > 1.0;

[0085] Option C (not fully satisfied): Provides 25 years of data, but 3 years of data are missing. = 1.0 - (Missing test penalty) <1.0.

[0086] In the optimization function f(x), Demand serves as a positive indicator; a higher value indicates a better solution. Ultimately, the system will balance between high Demand (high-quality solution) and high C (low-cost solution), depending on... Find the best balance between user profile preferences.

[0087] The cost is calculated based on the data combination scheme; the cost calculation is expressed as follows:

[0088] ;

[0089] in, The service index is 0 for free and 1 for paid. This is the discount factor, with a value range of (0,1]. For the first The first site Item name, For data prices, This refers to the number of documents.

[0090] The recommendation strategy adjustment coefficient is derived from the user profile. In this embodiment, a reinforcement learning mechanism is used to continuously optimize the behavioral weights of the user profile tags and the parameters of the recommendation model based on user feedback (clicks, downloads, favorites, negative reviews). The value is the behavioral weight of the user profile tag. The calculation formula is expressed as:

[0091] ;

[0092] in, As the basic weight of behavior, The attenuation coefficient is... The time interval between the occurrence of the behavior and the present. For duration coefficient, This represents the natural exponential function.

[0093] On the other hand, the present invention also provides a precise hydrological data recommendation system, comprising the following modules:

[0094] Hydrological data storage module: Compiles and stores hydrological data from hydrological stations;

[0095] User Requirement Analysis Module: Receives user input of requirement text and uses a natural language processing model to extract requirement keywords, requirement type, time range, and spatial range;

[0096] Precise matching recommendation module: Uses a fusion algorithm to generate a recommendation list;

[0097] Dynamic update module: Connects the recommendation precision matching module and the user interaction module interface, updates model parameters based on user behavior; updates user demand feature weights and recommendation model parameters based on user clicks, downloads, and ratings of recommendation results;

[0098] User interaction module: Displays recommendation results and receives user feedback.

[0099] For the foregoing embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to this application.

[0100] The above embodiments describe the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Modifications and variations made by those skilled in the art without departing from the spirit and scope of the invention should be within the protection scope of the appended claims.

Claims

1. A method for accurate recommendation of hydrological data, characterized in that, Includes the following steps: Step S1: Analyze customer needs, establish a hydrological calculation feature library based on the hydrological database, and calculate similar watersheds, similar stations, similar hydrological requirements, and a list of data that meet the requirements. Step S2: Establish a customer feature database and calculate user profiles; Step S3: Combining the hydrological calculation feature library and the customer feature library, calculate the hydrological data recommendation scheme index for all possible data combination schemes, and select the scheme with the highest hydrological data recommendation scheme index as the final recommended scheme.

2. The method for accurate recommendation of hydrological data as described in claim 1, characterized in that, The analysis of customer needs in step S2 includes: receiving the user's input demand text, and extracting demand features through natural language processing, including demand keywords, demand type, time range, and spatial range for hydrological data.

3. The method for accurate recommendation of hydrological data as described in claim 1, characterized in that, The calculation of the hydrological data recommendation scheme index in step S3 includes: Calculate and define each feature factor, dynamically adjust the weight coefficients of each feature factor according to the project type, and calculate the recommendation index. The calculation formula is as follows: ； in, ~ These are the weighting coefficients corresponding to each characteristic factor; The distance between the project site and the hydrological station; For watershed similarity; The calculation rules and formulas require, including both minimum and maximum requirements; The calculation accuracy is set within a variation range of 0% to 100% according to customer needs; To meet customer needs, including just meeting and fully meeting requirements; The cost calculated based on the data combination scheme; This is the adjustment coefficient for the recommendation strategy derived from user profiles; The larger the value, the better the solution matches the customer's needs.

4. The method for accurate recommendation of hydrological data as described in claim 3, characterized in that, The distance between the project site and the hydrological station The traditional method is used to calculate the distance between the target site of the project and multiple hydrological stations. The calculation formula is as follows: ； in, The three-dimensional coordinates of the target site for the project are represented as follows: The three-dimensional coordinates of the hydrological station are represented as follows: … … , Indicates the first Hydrological stations and .

5. The method for accurate recommendation of hydrological data as described in claim 4, characterized in that, The calculation of watershed similarity selects judgment indicators from two dimensions: natural geographical features and hydrological features. An improved cosine similarity algorithm is used to calculate the overall similarity between watersheds. The formula is as follows: ； in, For the first The weights of each characteristic factor indicator and , representing the weight coefficients corresponding to each characteristic factor. ~ ; For the first The cosine similarity of the characteristic factor indicators; the natural geographical features include watershed area, average elevation, average slope and vegetation coverage, and the hydrological features include multi-year average precipitation, multi-year average runoff, runoff coefficient and flood duration; at the same time, a similarity threshold for watershed similarity is set, and watersheds with similarity to the target watershed above the threshold are selected, and their relevant hydrological data are included in the recommended candidate set.

6. The method for accurate recommendation of hydrological data as described in claim 5, characterized in that, The cost is calculated as follows: ； in, The service index is 0 for free and 1 for paid. This is the discount factor, with a value range of (0,1]. For the first The first site Item name, For data prices, This refers to the number of documents.

7. The method for accurate recommendation of hydrological data as described in claim 6, characterized in that, The calculation rules and formulas require To measure the quantitative indicators of the recommended hydrological data combination and its derived calculation scheme, The value is obtained through the matching degree function. The matching degree function is calculated with a value range of [0,1], where 1 represents a perfect match. The calculation formula is expressed as: ； in, Indicates the matching weight of the calculation method. This indicates the matching degree of the calculation method; if it matches, then... Equals 1, if not matched Equal to 0; The weight represents the data completeness matching degree. Indicates the degree of data completeness and matching. = (Number of complete data items required in the recommendation scheme) / (Total number of data items required); This indicates the weight of the matching degree of the output results. This indicates the matching degree of the output results; a match indicates a good match. Equals 1, if not matched It equals 0.

8. The method for accurate recommendation of hydrological data as described in claim 7, characterized in that, The customer needs The value is obtained through the sufficiency function. To evaluate, it is represented as: ； in, Indicates basic satisfaction level. Indicates overflow reward. This indicates a lack of punishment.

9. The method for accurate recommendation of hydrological data as described in claim 8, characterized in that, The recommendation strategy adjustment coefficient derived from the user profile The behavioral weights of user profile tags and recommendation model parameters are continuously optimized based on user feedback on the recommendation results through reinforcement learning mechanisms. Recommended strategy adjustment coefficient The value is the behavioral weight of the user profile tag. The calculation formula is expressed as: ； in, As the basic weight of behavior, The attenuation coefficient is... The time interval between the occurrence of the behavior and the present. For duration coefficient, This represents the natural exponential function.

10. A hydrological data accurate recommendation system, used to implement the hydrological data accurate recommendation method according to any one of claims 1 to 9, characterized in that, Includes the following modules: Hydrological data storage module: Compiles and stores hydrological data from hydrological stations; User Requirement Analysis Module: Receives user input of requirement text and uses a natural language processing model to extract requirement keywords, requirement type, time range, and spatial range; Precise matching recommendation module: Uses a fusion algorithm to generate a recommendation list; Dynamic update module: Connects the recommendation precision matching module and the user interaction module interface, updates model parameters based on user behavior; updates user demand feature weights and recommendation model parameters based on user clicks, downloads, and ratings of recommendation results; User interaction module: Displays recommendation results and receives user feedback.

Images