A data processing method, apparatus, and electronic device based on a real-time database

Abstract

This application provides a data processing method, apparatus, and electronic device based on a real-time database, relating to the field of data processing. In this method, the original observation sequences of multiple target measurement points within a set time interval are extracted to form a set of time-value pairs. These pairs are then input into a kernel transformation unit, where a time-sensitive kernel function based on exponential decay is constructed to generate a kernel response sequence. A measurement point mapping function is constructed using Tikhonov regularization to map the kernel response sequence to a timestamp-independent statistical potential space, resulting in statistical potential response variables containing mean response, trend response, and offset response dimensions. The statistical potential response variables of each target measurement point are input into a statistical calculation engine to execute the corresponding statistical formula, outputting the statistical results. These results are then bound to the request path, and the kernel response function and statistical potential response variables are cached for subsequent reuse. Implementing the technical solution provided in this application facilitates improved efficiency in statistical calculations using real-time databases.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and specifically to a data processing method, apparatus and electronic device based on a real-time database. Background Technology

[0002] In scenarios such as industrial production process monitoring, energy efficiency assessment, and multi-system collaborative analysis, real-time databases are widely used as critical infrastructure to store historical observation data from various monitoring points.

[0003] Currently, to improve server storage efficiency and read / write performance, the real-time database built into the server typically employs a compressed storage strategy, recording only the observation points and their corresponding timestamps when the measured point values ​​change. This results in some measurement points having high-density records within the same time period, while others may have completely missing observation points, forming a sparse temporal structure. When statistical calculation formulas involve weighted operations, correlation analysis, or joint judgments among multiple measurement points, without a unified time benchmark, the calculation engine often needs to perform secondary processing on the original observation sequence, such as interpolation, truncation, or resampling, to achieve forced alignment of the time dimension. Therefore, this type of processing not only increases the complexity of the data preprocessing workflow but also easily introduces additional calculation errors, thereby reducing the efficiency of real-time database statistical calculations.

[0004] Therefore, there is an urgent need for a data processing method, device, and electronic device based on a real-time database. Summary of the Invention

[0005] This application provides a data processing method, apparatus, and electronic device based on a real-time database, which facilitates improving the efficiency of real-time database statistical calculations.

[0006] The first aspect of this application provides a data processing method based on a real-time database. The method includes: acquiring multiple target measurement points involved in a statistical calculation formula based on the real-time database; extracting the original observation sequence of each target measurement point within a set time interval, each original observation sequence including a timestamp sequence and a corresponding value sequence of the target measurement point, forming a time-value pair set of the target measurement point; inputting the time-value pair set of each target measurement point into a kernel transformation unit; constructing a time-sensitive kernel function with exponential decay characteristics for each target measurement point in the original timestamp sequence, generating a kernel response sequence of the target measurement point under the original time axis; and applying a TiK algorithm based on the kernel response sequence of each target measurement point. Honov regularization constraints are used to construct a measurement point mapping function, which maps the kernel response sequence to a unified statistical potential energy space to obtain the statistical potential energy response variable corresponding to the target measurement point. The statistical potential energy space is a computational domain independent of timestamps and includes mean response dimension, trend response dimension, and offset response dimension. The statistical potential energy response variables corresponding to multiple target measurement points are input into a statistical calculation engine, and the corresponding statistical calculation formulas are executed in the corresponding statistical potential energy spaces to obtain statistical calculation results. The statistical calculation results are bound to the original request path, and the kernel response function corresponding to the target measurement point and the statistical potential energy response variable are stored in a local mapping unit for reuse in subsequent similar statistical calculation requests.

[0007] Optionally, the step of obtaining multiple target measurement points involved in the statistical calculation formula based on a real-time database, and extracting the original observation sequence of each target measurement point within a set time interval, wherein each original observation sequence includes a timestamp sequence and a corresponding value sequence of the target measurement point, forming a time-value pair set of the target measurement point, specifically includes: parsing the measurement point identifier for each target measurement point in the statistical calculation formula to generate a target measurement point mapping table; constructing a data extraction request by combining the target measurement point mapping table with the set time interval, and sending the data extraction request to the real-time database interface layer; controlling the real-time database interface layer to call the compressed storage reading unit to perform decompression processing on the compressed data of each target measurement point within the set time interval to obtain the original observation sequence of the target measurement point; pairing the timestamp sequence and the corresponding value sequence included in the original observation sequence one by one to form the time-value pair set, wherein the time-value pair set includes a measurement point identifier, an observation sequence sampling density, a data source identifier, a data integrity flag, and a time interval parameter.

[0008] Optionally, the step of inputting the time-value pair set of each target measurement point into the kernel transformation unit, constructing a time-sensitive kernel function with exponential decay characteristics in the original timestamp sequence for each target measurement point, and generating a kernel response sequence of the target measurement point under the original time axis specifically includes: for each target measurement point, extracting the timestamp sequence and the corresponding value sequence from the corresponding time-value pair set, arranging them in ascending order of timestamps to construct an observation trajectory; constructing a time-sensitive kernel function on the original time axis of the observation trajectory, with each reference time point as the center; using the time-sensitive kernel function to perform a weighted summation of the historical observations near the reference time point, and using the ratio of the weighted result to the sum of the kernel function weights as the kernel response value corresponding to the reference time point, and constructing a continuous kernel response sequence point by point along the original time axis; dynamically adjusting the value of the time decay factor in the time-sensitive kernel function according to the observation density of each target measurement point to obtain the kernel response sequence of the target measurement point under the original time axis.

[0009] Optionally, the step of constructing a measurement point mapping function based on the kernel response sequence of each target measurement point using Tikhonov regularization constraints to map the kernel response sequence to a statistical potential energy space with a unified structure, thereby obtaining the statistical potential energy response variable corresponding to the target measurement point, specifically includes: inputting the kernel response sequence of each target measurement point into a mapping function construction unit to construct a measurement point mapping function model with the kernel response sequence as input and multiple statistical response dimensions corresponding to the statistical potential energy space as output; introducing the Tikhonov regularization constraints into the measurement point mapping function model, and controlling the variation amplitude of the mapping function corresponding to the target measurement point in the parameter space by adding a penalty term for the parameter magnitude of the measurement point mapping function to the loss function of the mapping function corresponding to the target measurement point, thereby generating the measurement point mapping function; and using the measurement point mapping function to map the kernel response sequence of each target measurement point to the mean response dimension, trend response dimension, and offset response dimension in the statistical potential energy space, respectively, thereby generating the statistical potential energy response variable of the target measurement point in the statistical potential energy space.

[0010] Optionally, the step of inputting the statistical potential energy response variables corresponding to each of the multiple target measurement points into the statistical calculation engine, and executing the corresponding statistical calculation formula in the corresponding statistical potential energy space to obtain the statistical calculation result specifically includes: inputting the statistical potential energy response variables corresponding to each of the multiple target measurement points into the statistical calculation engine in the form of a structure vector; the statistical calculation engine is used to perform response dimension matching verification on each statistical calculation formula to ensure that the response dimension of the statistical potential energy response variables of each target measurement point matches the input dimension of the statistical calculation formula; performing weighted operations, algebraic transformations, and logical relation operations on the statistical potential energy response variables in the statistical potential energy space according to the calculation structure of the statistical calculation formula, without relying on timestamps, to generate output results; and binding the output results with the identifier of the statistical calculation formula, the identifier of the statistical potential energy response variables participating in the statistical calculation, the weight parameters used, and the confidence index to form the structured statistical calculation result.

[0011] Optionally, the step of binding the statistical calculation results with the original request path and storing the kernel response function corresponding to the target measurement point and the statistical potential response variable in the local mapping unit specifically includes: extracting the original request path parameters of the statistical calculation task in the real-time database, wherein the original request path parameters include the statistical calculation formula identifier, the target measurement point identifier set, the set time interval parameter, the response dimension configuration, and the request initiation identifier code; establishing an index mapping between the statistical calculation results and the corresponding original request path parameters to construct a statistical calculation path index structure; determining the parameters corresponding to the kernel response function and the statistical potential response variable, and structurally binding them with the statistical calculation path index structure; uniformly encapsulating the statistical calculation path index structure, the statistical calculation results, the parameters of the kernel response function, and the statistical potential response variable into a mapping record unit, and writing the mapping record unit into the local mapping unit for direct reuse or incremental calculation of subsequent statistical calculation requests for similar paths.

[0012] Optionally, the current statistical calculation request path parameters are obtained. If it is determined that the current statistical calculation request path parameters are completely consistent with the statistical calculation path index structure stored in the local mapping unit, the statistical calculation result in the corresponding mapping record unit is directly extracted, skipping the calculation process of the statistical calculation engine, and the result is directly reused. If it is determined that the current statistical calculation request path parameters are partially similar to the statistical calculation path index structure, and there is partial overlap in the identifier set of the target measurement point, partial overlap in the time interval parameters, or partial overlap in the response dimension configuration, the parameters of the cached kernel response function and the statistical potential response variable are extracted as basic inputs. Incremental calculations are performed for the new measurement point or extended time interval, and the parameters of the new kernel response function, the new statistical potential response variable, and the basic inputs are merged to generate the current statistical calculation result.

[0013] A second aspect of this application provides a data processing apparatus based on a real-time database. The data processing apparatus includes an acquisition module and a processing module. The acquisition module is used to acquire multiple target measurement points involved in a statistical calculation formula based on the real-time database, and extract the original observation sequence of each target measurement point within a set time interval. Each original observation sequence includes a timestamp sequence and a corresponding value sequence for the target measurement point, forming a time-value pair set for the target measurement point. The processing module is used to input the time-value pair set of each target measurement point to a kernel transformation unit, and construct a time-sensitive kernel function with exponential decay characteristics for each target measurement point in the original timestamp sequence, generating a kernel response sequence of the target measurement point on the original time axis. The processing module is further used to... The kernel response sequence is used to construct a measurement point mapping function using Tikhonov regularization constraints, mapping the kernel response sequence to a unified statistical potential energy space to obtain the statistical potential energy response variable corresponding to the target measurement point. The statistical potential energy space is a computational domain independent of timestamps, and includes mean response dimension, trend response dimension, and offset response dimension. The processing module is also used to input the statistical potential energy response variables corresponding to multiple target measurement points into the statistical calculation engine, execute the corresponding statistical calculation formula in the corresponding statistical potential energy space, and obtain the statistical calculation result. The processing module is also used to bind the statistical calculation result with the original request path, and store the kernel response function corresponding to the target measurement point and the statistical potential energy response variable in the local mapping unit for reuse in subsequent similar statistical calculation requests.

[0014] A third aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, and both the user interface and the network interface are used to communicate with other devices. The processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method described above.

[0015] A fourth aspect of this application provides a computer-readable storage medium storing instructions that, when executed, perform the method described above.

[0016] In summary, one or more technical solutions provided in this application have at least the following technical effects or advantages:

[0017] 1. By maintaining the original observation sequence of each target measurement point in its original time axis structure and inputting it into the kernel transformation unit in the form of a time-value pair set, the steps of interpolation, truncation or resampling of multi-point observation sequences in traditional methods are avoided, which significantly reduces the preprocessing complexity. At the same time, the original authenticity and time distribution characteristics of the data are maintained, which facilitates the improvement of the efficiency of real-time database statistical calculation.

[0018] 2. A time-sensitive kernel function with exponential decay characteristics is used to locally weight the response of the measurement point time axis and construct a kernel response sequence. This processing method can effectively capture the local change trend and delay characteristics of the measurement point in its own time structure, enhance the dynamic adaptability to the non-uniform sampling behavior of the measurement point, and effectively improve the response stability and trend expression ability.

[0019] 3. Based on kernel response sequences, a Tikhonov regularization constraint is introduced to construct a measurement point mapping function, mapping the measurement point data to a unified statistical potential energy space. This allows statistical calculations to no longer rely on a unified timestamp dimension, but rather to be performed in an abstract response space. The construction of this potential energy space represents the dynamic behavior of the original measurement points as a structured response vector, including mean response dimension, trend response dimension, and offset response dimension, improving the operability and uniformity of cross-measurement point comparison and collaborative analysis.

[0020] 4. During the statistical calculation phase, the statistical potential energy response variables of multiple target measurement points are uniformly input into the statistical calculation engine as structural inputs. Statistical calculation formulas are directly executed in the statistical potential energy space, bypassing the accuracy errors and computational redundancy caused by traditional time alignment. This achieves efficient collaborative computation of asynchronous data from multiple measurement points. By structurally binding the statistical calculation results with the original request path and writing the calculated kernel response function and statistical potential energy response variables into the local mapping unit, subsequent statistical requests with the same or partially similar paths can directly reuse existing calculation results or perform incremental calculations. This significantly improves the overall system response efficiency, reduces the consumption of redundant computational resources, and constructs an intelligent statistical support system with high reusability and elastic scalability. Attached Figure Description

[0021] Figure 1 A flowchart illustrating a data processing method based on a real-time database provided in an embodiment of this application;

[0022] Figure 2 A schematic diagram of a data processing device based on a real-time database provided in an embodiment of this application;

[0023] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0024] Explanation of reference numerals in the attached figures: 21. Acquisition module; 22. Processing module; 31. Processor; 32. Communication bus; 33. User interface; 34. Network interface; 35. Memory. Detailed Implementation

[0025] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0026] In the description of the embodiments of this application, the words "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design that is described as "for example" or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design options. Rather, the use of the words "for example" or "for instance" is intended to present the relevant concepts in a specific manner.

[0027] In the description of the embodiments of this application, the term "multiple" means two or more. For example, multiple systems means two or more systems, and multiple screen terminals means two or more screen terminals. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. The terms "comprising," "including," "having," and variations thereof all mean "including but not limited to," unless otherwise specifically emphasized.

[0028] In applications such as industrial production process monitoring, energy efficiency assessment, and multi-system collaborative analysis, real-time databases serve as core data infrastructure and are widely used to store historical observation information from a large number of measurement points.

[0029] To improve storage utilization and data access performance, current real-time databases generally employ compressed storage strategies, recording observations and their timestamps only when measurement point values ​​change. This sampling-based mechanism leads to significant differences in the recording frequency of different measurement points within the same time interval. Some measurement points form high-density observation sequences, while others may have no records during that time period, resulting in a non-uniform sparse temporal structure. Under this structure, if statistical calculation formulas involve weighted calculations, correlation judgments, or logical combinations among multiple measurement points, the lack of a unified time reference will severely affect the execution of the calculation process. To achieve consistency in the time dimension, the system needs to perform secondary processing operations such as interpolation, truncation, or resampling on the original observation sequences of each measurement point. These operations not only significantly increase the data preprocessing burden but may also introduce unnecessary errors during time alignment, thereby reducing the overall accuracy and execution efficiency of statistical calculations.

[0030] To address the aforementioned technical problems, this application provides a data processing method based on a real-time database, referring to... Figure 1 , Figure 1 This is a flowchart illustrating a data processing method based on a real-time database, provided as an embodiment of this application. The method is applied to a server and includes steps S110 to S150, as follows:

[0031] S110. Obtain multiple target measurement points involved in the statistical calculation formula based on the real-time database, and extract the original observation sequence of each target measurement point within a set time interval. Each original observation sequence includes the timestamp sequence and the corresponding value sequence of the target measurement point, forming a time-value pair set of the target measurement point.

[0032] Specifically, a server refers to a data processing system or computing node that performs statistical calculations, possessing the ability to read data from a real-time database, execute computational logic, and manage result storage. For example, in an energy management platform, the central data processing server communicates with the real-time database through an interface to periodically summarize data from measurement points such as temperature, pressure, and flow rate. A statistical calculation formula refers to calculation rules defined on multiple target measurement points, used to derive statistical indicators, such as mean, maximum, and rate of change, from the measurement point data. This formula typically includes weighting factors, calculation operators, and a logical structure. For example, a boiler thermal efficiency evaluation formula might be: "Boiler outlet steam flow rate multiplied by calorific value divided by fuel mass flow rate," which involves multiple target measurement points.

[0033] A target measurement point refers to the specific data source unit called in the statistical calculation formula, representing a continuous observation record of a specific physical variable. In the real-time database, each target measurement point has a unique identifier, such as "Main Steam Pressure of Unit #1" or "Inlet Temperature of Cooling Pump #2". A set time interval refers to the time range specified by the statistical calculation request, used to define the boundaries for extracting target measurement point data from the real-time database. It is usually defined in the form of a start time and an end time, for example, "00:00 on June 1, 2025 to 00:00 on June 2, 2025" is a set time interval. The original observation sequence refers to all unprocessed observation records of a target measurement point within the set time interval, including the measurement point values ​​and their corresponding time information, in the format of raw data arranged in ascending order of time. For example, for the measurement point "Main Steam Pressure of Unit #1", multiple observation points may be recorded within the set time interval, such as "12.3 MPa at 08:00" and "12.4 MPa at 08:03", constituting the original observation sequence of that measurement point. A timestamp sequence refers to all the observation time nodes in the original observation sequence, representing the recording time of each observation. In the example above, "08:00" and "08:03" are the timestamp sequences of the target measurement point.

[0034] A value sequence refers to the numerical value corresponding to each timestamp in the original observation sequence, representing the physical measurement result of the target measuring point at that moment. Using the example above, "12.3 MPa" and "12.4 MPa" are the numerical items in the value sequence. A time-value pair set refers to the data set formed by pairing the timestamp sequence with the value sequence one-to-one. It is an ordered pair representing the measured values ​​of the target measuring point at different time points. Taking "Main Steam Pressure of Unit 1" as an example, its time-value pair set might be {(08:00, 12.3 MPa), (08:03, 12.4 MPa), (08:06, 12.3 MPa)}. In summary, after receiving the statistical calculation request, the server parses the statistical calculation formula, determines all the target measurement points involved in the formula, and extracts the original observation sequence for each target measurement point from the real-time database according to the set time interval. Furthermore, each original observation sequence is decomposed into a timestamp sequence and a value sequence, and then reassembled into a time-value pair set to prepare for subsequent kernel response generation and statistical calculation.

[0035] In one possible implementation, multiple target measurement points involved in the statistical calculation formula based on a real-time database are obtained. The original observation sequence of each target measurement point is extracted within a set time interval. Each original observation sequence includes a timestamp sequence and a corresponding value sequence for the target measurement point, forming a time-value pair set for the target measurement point. Specifically, this includes: parsing the measurement point identifier for each target measurement point in the statistical calculation formula to generate a target measurement point mapping table; constructing a data extraction request by combining the target measurement point mapping table with the set time interval, and sending the data extraction request to the real-time database interface layer; controlling the real-time database interface layer to call the compressed storage and reading unit to decompress the compressed data of each target measurement point within the set time interval to obtain the original observation sequence of the target measurement point; pairing the timestamp sequence and corresponding value sequence included in the original observation sequence one-to-one to form a time-value pair set, which includes the measurement point identifier, observation sequence sampling density, data source identifier, data integrity flag, and time interval parameters.

[0036] Specifically, the first step is for the statistical computing engine to parse the statistical calculation formulas contained in the statistical calculation request, identifying all target measurement points involved in the calculation. For each target measurement point, its measurement point identifier in the statistical calculation formula is extracted and semantically standardized, for example, by using measurement point mapping rules or a measurement point metadata index table, and uniformly converted into a unique measurement point code in the real-time database. All target measurement point codes are then aggregated to form a target measurement point mapping table, which is used to indicate the database address and data structure definition corresponding to each measurement point in the subsequent data extraction process.

[0037] The second step involves constructing a data extraction request based on all target measurement point codes contained in the target measurement point mapping table, together with the specified time interval specified in the statistical calculation request. The specified time interval includes the start time. With end time The time parameter pairs , and measurement point code set After binding, a data retrieval request object is generated:

[0038]

[0039] The request object is submitted to the real-time database interface layer as a standardized data access interface specification at the structural layer.

[0040] Thirdly, the control logic invokes commands to the real-time database interface layer. Upon receiving the data extraction request, the interface layer accesses the corresponding data compression area through the measurement point code and calls the compression storage and reading unit to perform the decompression operation. Based on the storage structure index and compression coding rules, the compression storage and reading unit extracts all records of the target measurement point within the time interval from the compressed block and restores them to the original observation sequence. The original observation sequence is in the following form:

[0041]

[0042] in For the target measurement point The original observation sequence, To record timestamps, This is the observation value corresponding to that moment.

[0043] The fourth step involves extracting the timestamp sequence and value sequence from the decompressed original observation sequence to form a timestamp vector. AND value vector Then, pair them one by one to form a set of time-value pairs:

[0044]

[0045] This time-value pair set Indicates the target measurement point All original observation points within the set time interval.

[0046] Fifth step, for each time-value pair set This data is then structured and encapsulated to form a standard target measurement point data structure. This structure includes the following five key fields: measurement point identifier... That is, the unique code of the target measurement point; the sampling density of the observation sequence. The data source identifier indicates whether the data comes from the main real-time database system, secondary system, or archive system; the data integrity flag records whether there are gaps, breakpoints, or outliers in the observation sequence within the time interval; and the time interval parameter. This refers to the boundary conditions of the data extraction range.

[0047] Through the above five steps, the raw data preparation work for the statistical calculation request on the server side is completed, ensuring that the time-value pair set of all target measurement points within the set time interval has been completely extracted, structurally encapsulated and ready for use in subsequent kernel response construction and statistical potential mapping, and maintaining a unified structure, consistent indexing and traceability at the conceptual and data layers.

[0048] S120. Input the time-value pair set of each target measurement point into the kernel transformation unit, construct a time-sensitive kernel function with exponential decay characteristics in the original timestamp sequence for each target measurement point, and generate the kernel response sequence of the target measurement point under the original time axis.

[0049] Specifically, the kernel transformation unit refers to the computational module used to receive the set of time-value pairs and generate the structural response results. It possesses functions such as kernel function modeling, response weight allocation, and time-sensitive mapping. The internal structure of this unit is based on the kernel response principle to model the local behavior of the target measurement point under the original time structure. It is a key component connecting the original observation structure and the statistical potential energy space. The original timestamp sequence refers to the timestamp entries in the set of time-value pairs, recording the time nodes when the observed values ​​of the target measurement point actually change within the observation interval. For example, for the above set P, its original timestamp sequence is [08:00, 08:03, 08:08]. The time-sensitive kernel function refers to the function structure that performs weighted response modeling on the original timestamps. The kernel function form is based on the time difference and introduces exponential decay characteristics. The kernel response sequence refers to the set of response values ​​obtained after weighted calculation of the target measurement point at multiple reference time points based on the time-sensitive kernel function. Each reference time point t corresponds to a kernel response value.

[0050] In summary, the server submits the set of time-value pairs extracted from each target measurement point within a defined time interval to the kernel transformation unit. The kernel transformation unit, while preserving the original timestamp sequence of the target measurement points, constructs a time-sensitive kernel function controlled by exponential decay. Based on this kernel function, it performs weighted calculations on the original observations of the target measurement points, ultimately outputting a kernel response sequence. This kernel response sequence serves as the structured input for subsequent statistical potential mapping and statistical calculations, thus solving the problem of behavioral consistency modeling under timestamp misalignment conditions.

[0051] In one possible implementation, the time-value pair set of each target measurement point is input into the kernel transformation unit. For each target measurement point, a time-sensitive kernel function with exponential decay characteristics is constructed in the original timestamp sequence to generate a kernel response sequence of the target measurement point under the original time axis. Specifically, this includes: for each target measurement point, extracting the timestamp sequence and corresponding value sequence from the corresponding time-value pair set, arranging them in ascending order of timestamps to construct an observation trajectory; constructing a time-sensitive kernel function on the original time axis of the observation trajectory, centered on each reference time point; using the time-sensitive kernel function to perform a weighted summation of the historical observations near the reference time point, and using the ratio of the weighted result to the sum of the kernel function weights as the kernel response value corresponding to the reference time point, and constructing a continuous kernel response sequence point by point along the original time axis; dynamically adjusting the value of the time decay factor in the time-sensitive kernel function according to the observation density of each target measurement point to obtain the kernel response sequence of the target measurement point under the original time axis.

[0052] Specifically, the first step is for the server to extract the timestamp sequence and corresponding value sequence from the time-value pair set for each target measurement point. The timestamp sequence is an ordered set of observation times, denoted as . ,in Indicates the first There are several observation time points; the corresponding value sequence is the set of observation values ​​of the target measuring point at each time point, denoted as... ,in For the target measurement point in time The observed values. The server's timestamp sequence. Perform an ascending sort operation to maintain the temporal consistency of the observation order in the time-value pair set, and construct the observation trajectory of the target measurement point within the set time interval as the temporal basis for subsequent kernel transformation operations.

[0053] The second step involves the server selecting a continuous set of reference time points on the original timeline of the target measurement point, denoted as... This set can be dynamically generated based on observation density or an equal-interval strategy. For each reference time point... Server constructs time-sensitive kernel function The kernel function is an exponential decay function, defined as:

[0054]

[0055] in, This is the current reference time. For historical observation time points, This is the time decay factor, used to control the decay rate of the weights corresponding to different time differences. This kernel function assigns higher weights to observations closer to the current reference point, reflecting local temporal correlation.

[0056] The third step is to consider each reference time point. The server calculates its corresponding kernel response value. The response value is given by the ratio of the weighted observations to the total weight of the kernel function, and the specific calculation formula is as follows:

[0057]

[0058] Wherein, the numerator represents each historical observation value. Weighted by kernel function The weighted summation, with the denominator being the sum of the weights of all kernel function values, ensures that the kernel response value is in the form of a weighted average. This response value describes the target measurement point at time point... The smooth response to its historical observation behavior.

[0059] The fourth step involves the server moving along each reference time point on the original timeline. Repeat the kernel function construction and weighted response calculation operations described above to generate the kernel response sequence point by point.

[0060]

[0061] This kernel response sequence is a smooth response vector that preserves the original time structure. It is used to describe the trend change behavior of the target measurement point at different time points and provides a consistent input basis for statistical potential mapping.

[0062] Fifth, the server dynamically adjusts the time decay factor based on the observation density of each target measurement point. The value of is taken as follows. Observation density is defined as the number of observation points per unit time, denoted as . ,in For the number of observation points, , These are the first and last time points in the time-value pair set, respectively. If the observation density is high, a larger value should be selected. The value is chosen to enhance the local sensitivity of the kernel function; if the observation density is low, a smaller value is selected. The value is adjusted to maintain a stable response range. The dynamic adjustment rule can be expressed as:

[0063]

[0064] in The density adjustment ratio coefficient set for the system ensures that the attenuation degree is adaptively matched with the distribution characteristics of the measurement point data, thereby constructing a high-fidelity kernel response sequence that reflects the original observation trend of the measurement point.

[0065] In summary, the above five steps constitute the complete process of the server constructing the kernel response for the target measurement point. This process not only preserves the original observation time structure, but also achieves a weighted expression of local behavior through a time-sensitive kernel function, and improves the dynamic stability of the response results through density-aware parameter tuning, ensuring that a unified, continuous, and comparable response input is provided for subsequent statistical potential energy space mapping.

[0066] S130. Based on the kernel response sequence of each target measurement point, a measurement point mapping function is constructed using Tikhonov regularization constraints to map the kernel response sequence to a statistical potential energy space with a unified structure, thereby obtaining the statistical potential energy response variable corresponding to the target measurement point. The statistical potential energy space is a computational domain that does not depend on the timestamp and includes the mean response dimension, the trend response dimension, and the offset response dimension.

[0067] Specifically, Tikhonov regularization is a parameter constraint strategy used to prevent overfitting of the mapping function during modeling. Its basic idea is to introduce a penalty term into the least squares loss function to control the magnitude of the model parameters and constrain the model complexity. Its optimization objective can be expressed as:

[0068]

[0069] in The kernel response sequence of the target measurement point. Let be the parameter vector of the measurement point mapping function to be optimized. The kernel response characteristic matrix, The regularization coefficient controls the influence of parameter amplitude on the fitting process. For example, when there are abrupt fluctuations in the kernel response, Tikhonov regularization can suppress overfitting to outliers and improve the robustness of the overall response results.

[0070] A measurement point mapping function is a transformation function that takes the kernel response sequence of a target measurement point as input and outputs a statistically consistent response with a uniform structure. The mapping function takes the form of... ,in For the kernel response sequence, This is a structured response vector containing multiple response dimension values ​​from the statistical potential energy space. This function can be implemented using regularized regression, embedding models, or structural projection. The statistical potential energy space is a multidimensional structured space that does not rely on timestamps and is specifically used to represent the statistical response characteristics of a target measurement point; each dimension corresponds to a response quantity with a specific statistical meaning. The statistical potential energy space removes temporal information, retaining only the behavioral pattern of the target measurement point over a period of time. For example, this space can be three-dimensional. Its three coordinate axes are the mean response dimension, the trend response dimension, and the offset response dimension, respectively.

[0071] The mean response dimension refers to the response quantity corresponding to the overall average level of the target measurement point kernel response sequence. The calculation formula is:

[0072]

[0073] For example, if the kernel response sequence is [12.2, 12.4, 12.3, 12.1], then the mean response dimension is... .

[0074] The trend response dimension refers to the intensity and direction of the trend of the target measurement point kernel response sequence over time. It can be represented by the slope of the linear fit of the sequence or the mean of the first difference. The calculation formula is:

[0075]

[0076] For example, if the sequence is [12.2, 12.4, 12.6, 12.8], then the trend response dimension is positive, reflecting a clear upward trend.

[0077] The offset response dimension refers to the degree of deviation of the overall position of the target measurement point in the kernel response sequence relative to the reference baseline value (such as the starting value of the time interval or the global mean). It is commonly represented by the difference between the first and last points of the sequence or the average deviation relative to the starting point. The calculation formula is:

[0078]

[0079] For example, if the sequence is [12.2,12.4,12.3,12.5], the offset response dimension is positive, indicating that the measurement point rises slightly throughout the time period.

[0080] The statistical potential energy response variable refers to the structural response result generated by projecting the target measurement point into the statistical potential energy space. It is in the form of a structural vector, as follows:

[0081]

[0082] This vector, detached from the time dimension, represents the statistical behavioral characteristics of the measurement points and is used for subsequent unified structural calculations and comparisons among multiple measurement points.

[0083] In summary, the above scheme illustrates how the server constructs a measurement point mapping function based on the kernel response sequence through Tikhonov regularization constraints. This function projects the target measurement point behavior from the time-dependent observation sequence to a unified, time-independent, and structured statistical potential energy space, resulting in a statistical potential energy response variable that can be used for unified analysis across measurement points and structures. This provides a structural foundation for subsequent efficient and stable collaborative computing.

[0084] In one possible implementation, based on the kernel response sequence of each target measurement point, a measurement point mapping function is constructed using Tikhonov regularization constraints. This maps the kernel response sequence to a statistical potential energy space with a unified structure, obtaining the statistical potential energy response variable corresponding to the target measurement point. Specifically, this includes: inputting the kernel response sequence of each target measurement point into the mapping function construction unit to construct a measurement point mapping function model with the kernel response sequence as input and multiple statistical response dimensions corresponding to the statistical potential energy space as output; introducing Tikhonov regularization constraints into the measurement point mapping function model, and controlling the variation amplitude of the mapping function corresponding to the target measurement point in the parameter space by adding a penalty term for the parameter magnitude of the measurement point mapping function to the loss function of the mapping function corresponding to the target measurement point, thereby generating the measurement point mapping function; and using the measurement point mapping function to map the kernel response sequence of each target measurement point to the mean response dimension, trend response dimension, and offset response dimension in the statistical potential energy space, respectively, generating the statistical potential energy response variable of the target measurement point in the statistical potential energy space.

[0085] Specifically, in the first step, the server inputs the kernel response sequence of each target measurement point into the mapping function construction unit as the input dataset for training samples. The kernel response sequence consists of multiple continuous response values, representing the dynamic change trend of the target measurement point under the original time axis. The task of the mapping function construction unit is to establish a multi-output regression model using this kernel response sequence as the input variable. The output of the model is the three statistical response dimensions of the target measurement point in the statistical potential space, namely the mean response dimension, the trend response dimension, and the offset response dimension. This process is essentially establishing a mapping relationship model from the response sequence to the statistical structure space, forming a measurement point mapping function with the following structure: the input is a set of continuous kernel response values, and the output is a structured statistical feature vector.

[0086] The second step involves introducing Tikhonov regularization constraints during the construction of the measurement point mapping function model to control the convergence stability of the function's parameters. Specifically, a regularization term is added to the loss function used in the measurement point mapping function. This regularization term is defined as the squared magnitude of the measurement point mapping function parameters multiplied by a regularization coefficient. That is, if the core parameter of the measurement point mapping function is the weight matrix W, then the total loss function is the sum of the squares of the prediction errors plus the squared magnitude of W multiplied by the regularization coefficient. This penalty term suppresses the overfitting tendency of the mapping function at certain extreme values ​​in the kernel response sequence, thereby enabling the mapping function to maintain better generalization ability across multiple target measurement points.

[0087] Third, the server uses the constructed measurement point mapping function to input the kernel response sequence of the target measurement point into the function, and maps the function output to the three dimensions of the statistical potential energy space respectively. The mapping process is performed according to the following calculation logic: First, the overall value of the kernel response sequence is weighted and averaged, and the output is the mean response dimension; then, the average growth rate of the difference sequence between adjacent points in the kernel response sequence is calculated, and the output is the trend response dimension; finally, the average offset of each value in the kernel response sequence relative to the first value is calculated, and the output is the offset response dimension. The above three response dimensions constitute the response vector of the target measurement point in the statistical potential energy space, which is called the statistical potential energy response variable.

[0088] The implementation of this technical solution has the following effects: On the one hand, by mapping kernel response sequences of unequal length, different densities, and different time structures to a structured statistical potential energy space through a measurement point mapping function, standardized modeling of the response behavior of different target measurement points is achieved; on the other hand, the introduction of Tikhonov regularization suppresses parameter instability and model overfitting, ensuring the consistency and reliability of statistical response variables under multiple measurement points and multiple time conditions; finally, through a unified statistical potential energy response variable structure, the originally time-heterogeneous and incomparable target measurement points can be transformed into input variables that can directly participate in unified statistical calculations, fundamentally solving the problems of inconsistent calculation dimensions and unstable responses in traditional statistical engines when processing compressed sampling and asynchronous data from multiple measurement points, greatly improving the structural adaptability, computational efficiency, and robustness of statistical calculations.

[0089] S140. Input the statistical potential energy response variables corresponding to multiple target measurement points into the statistical calculation engine, execute the corresponding statistical calculation formula in the corresponding statistical potential energy space, and obtain the statistical calculation results.

[0090] Specifically, a statistical computing engine refers to a computational module that executes structured computational logic. It can receive multiple statistical potential energy response variables as input and perform statistical logic such as matrix operations, algebraic operations, and normalization combinations within a structural space to output index values ​​or model parameters. For example, a statistical computing engine can implement a formula structure for "outputting the weighted ratio of temperature change amplitude to pressure change." The statistical potential energy space refers to the three-dimensional structural space that carries the statistical potential energy response variables. It does not have a timestamp dimension, but only defines response dimensions, including mean response dimension, trend response dimension, and offset response dimension. This space allows data from different measurement points and different time windows to be standardized and mapped to a comparable and computable structural domain. Statistical computing results refer to the output values ​​obtained after executing statistical calculation formulas in the statistical potential energy space. These can be single indicators or multi-dimensional output structures, depending on the formula definition. For example, "average offset of the target measurement point group," "the contribution ratio of trend to stability," and "predicted over-limit risk score" are all forms of statistical computing results.

[0091] In summary, the above technical solution describes how, after the server constructs the statistical potential energy response variables for each target measurement point, it inputs these variables into a structured statistical computing engine. The engine then performs structural vector operations in the statistical potential energy space according to preset statistical calculation formulas, thereby outputting a unified statistical index independent of time structure. This mechanism solves the problem of inconsistent computational structures in traditional statistical methods when dealing with time-displaced and heterogeneous measurement point data, significantly improving computational efficiency and result comparability.

[0092] In one possible implementation, the statistical potential energy response variables corresponding to multiple target measurement points are input into a statistical calculation engine. The corresponding statistical calculation formulas are then executed in the respective statistical potential energy spaces to obtain statistical calculation results. Specifically, this includes: inputting the statistical potential energy response variables corresponding to multiple target measurement points into the statistical calculation engine in the form of a structured vector; the statistical calculation engine performs response dimension matching verification on each statistical calculation formula to ensure that the response dimension of the statistical potential energy response variables of each target measurement point matches the input dimension of the statistical calculation formula; without relying on timestamps, performing weighted operations, algebraic transformations, and logical relational operations on the statistical potential energy response variables in the statistical potential energy space according to the calculation structure of the statistical calculation formula to generate output results; and binding the output results with the identifiers of the statistical calculation formulas, the identifiers of the statistical potential energy response variables participating in the statistical calculation, the weight parameters used, and the confidence indicators to form structured statistical calculation results.

[0093] Specifically, in the first step, the server inputs the statistical potential response variables of multiple target measurement points into the statistical computing engine in the form of a structured vector. Each statistical potential response variable consists of three defined dimensions: the mean response dimension, the trend response dimension, and the offset response dimension, represented as a triplet structured vector, as follows:

[0094]

[0095] in Indicates the target measurement point The mean response dimension value, This indicates the value of its trend response dimension. This represents the offset response dimension value. The server assigns a unique identifier to each target measurement point and binds this identifier to its structure vector to form a mapping record. All structure vectors are uniformly input into the statistical computing engine, constituting the structured input set for this statistical task.

[0096] The second step involves the statistical computation engine, upon receiving the structure vector input, first performing response dimension matching verification on the statistical computation formula. The statistical computation formula has a clearly defined input dimension, for example:

[0097]

[0098] in Indicates the statistical calculation formula identifier. , These are the statistical potential energy response variables for the two target measurement points. This represents the trend response dimension of the first target measurement point. This indicates the offset response dimension of the second target measurement point. These are the corresponding weighting coefficients. The statistical calculation engine performs a consistency check on each dimension index involved in the calculation, ensuring that all input dimension names, types, and formula input parameter structures are completely matched. If there are omissions, type mismatches, or dimension redundancy, a configuration error warning is output, and the calculation is terminated.

[0099] Third, after successful matching verification, the statistical computing engine performs time-stamp-free mathematical operations based on the structure vector data and the defined statistical calculation formulas. This process is completed in the statistical potential space, and the computational logic may include weighted combinations, algebraic function transformations, nonlinear response mappings, or combinations of logical relationships. For example, consider the following statistical calculation formula:

[0100]

[0101] in For the final output, , , , For the statistical potential response variables at different target measurement points, the numerator is a weighted sum of the mean response dimension, and the denominator is a difference correction term for the two trend response dimensions. The statistical calculation engine uses this structure to perform block parsing, value selection by dimension, mathematical operations, and calculate the numerical results.

[0102] Fourth, after generating the output results, the statistical calculation engine immediately performs a result structure binding operation. This binds the calculation results... The unique identifier of the statistical calculation formula. The identifier of each statistical potential response variable involved in the calculation. The set of weight parameters used Together with confidence metrics, these results are packaged into structured statistical calculations. Confidence metrics, which can be derived through historical variance analysis, residual error assessment, or reliability scores based on response density calculations, indicate the stability and reliability of the calculation results under the current structural input conditions.

[0103] In summary, this technical solution inputs statistical potential response variables into the statistical computing engine in the form of a structure vector. Through steps such as response dimension matching verification, structure vector kernel operation, and parameter binding of calculation results, the statistical calculation process is completely freed from the constraints of the timestamp structure, achieving a precise, controllable, and verifiable calculation path within the structure space. Furthermore, this solution supports complex cross-reference calculations between multiple target measurement points, exhibiting good versatility and scalability, and significantly enhancing the unified statistical analysis capabilities under heterogeneous data structures.

[0104] S150. Bind the statistical calculation results to the original request path, and store the kernel response function and statistical potential energy response variable corresponding to the target measurement point into the local mapping unit for reuse in subsequent similar statistical calculation requests.

[0105] Specifically, the original request path refers to the complete combination of path parameters carried in initiating a statistical calculation request, including the statistical calculation formula identifier, the target measurement point identifier set, the set time interval, the response dimension requirements, the request initiation time, and contextual constraints. The original request path constitutes a unique fingerprint identifier for this calculation, used to record the traceability of the calculation process and results. For example, a statistical request path can be represented as: formula "thermal efficiency calculation", measurement point set {T101, P105, F112}, time interval "2024-06-01 00:00 to 2024-06-01 12:00". Binding refers to establishing a structured index relationship between the statistical calculation result and the original request path, so that the corresponding calculation result can be directly located through the original request path when the result is called subsequently. The binding operation includes not only the connection between the result and the path, but also the binding of meta-information such as parameter mapping and configuration records, used to support data reuse and audit traceability.

[0106] A local mapping unit refers to a high-speed cache structure deployed within the server, used to store intermediate data and results related to the statistical calculation path. The local mapping unit supports functions such as path indexing, response variable storage, result backtracking, and reuse judgment. For example, when the system receives a request with a similar path structure, it can directly extract the existing kernel response function and statistical potential response variable from the local mapping unit, skipping the data extraction and modeling process, and only performing incremental calculations on the differences, thereby improving computational efficiency. Reuse refers to determining whether the stored kernel response function, statistical potential response variable, and calculation results can be partially or fully reused when subsequent statistical calculation requests arrive, based on path structure comparison, to avoid redundant modeling and computation. For example, if a new request only expands the time interval but uses the same target measurement points and statistical calculation formulas, the local mapping unit can retain the existing structure, only supplementing the data for the new time period before performing incremental calculations.

[0107] In summary, the above technical solution describes how, after completing a statistical calculation task, the server structurally binds the results to the request path and writes the kernel response functions and statistical potential response variables of all target measurement points involved in the modeling and calculation into the local mapping unit, thereby constructing a traceable, cacheable, and reusable computational closed-loop structure. This mechanism significantly improves the system's processing efficiency for high-frequency repetitive requests and requests with similar structures, reduces resource consumption, and provides structural support for complex model calculations.

[0108] In one possible implementation, the statistical calculation results are bound to the original request path, and the kernel response function and statistical potential response variable corresponding to the target measurement point are stored in the local mapping unit. Specifically, this includes: extracting the original request path parameters of the statistical calculation task from the real-time database. The original request path parameters include the statistical calculation formula identifier, the target measurement point identifier set, the set time interval parameter, the response dimension configuration, and the request initiation identifier code. An index mapping is established between the statistical calculation results and the corresponding original request path parameters to construct a statistical calculation path index structure. The parameters and statistical potential response variables corresponding to the kernel response function are determined and structurally bound to the statistical calculation path index structure. The statistical calculation path index structure, the statistical calculation results, the parameters of the kernel response function, and the statistical potential response variables are uniformly encapsulated into a mapping record unit, and the mapping record unit is written into the local mapping unit for direct reuse or incremental calculation of subsequent statistical calculation requests for similar paths.

[0109] Specifically, the first step is that after the statistical calculation process is completed, the server immediately extracts the original request path parameters for this statistical calculation task. The original request path parameters are a structured set of information that uniquely identifies the statistical task, specifically including: the statistical calculation formula identifier. This indicates the statistical calculation logic model number used in this call; the target measurement point identifier set. Used to identify all target measurement points involved in this task; set time interval parameters. Define the temporal boundaries for data extraction and kernel response construction; configure response dimensions. Used to confirm the structural dimension of the statistical potential energy space; request initiation identifier. This is used to trace the source of computation calls. The server combines the above five parameters to construct a statistical computation path index structure, which serves as the primary key index in the local mapping unit.

[0110] The second step involves the server retrieving the kernel response function parameters and statistical potential response variables for all target measurement points from the memory cache module. The kernel response function for each target measurement point consists of the following parameters: time-sensitive kernel function form. ,in The time decay factor reflects the smoothness of the response; the original timestamp sequence With corresponding observation sequence Reference time point set Used to generate response sequences The statistical potential response variable is determined by a three-dimensional structure vector. It means that, among them The mean of the nuclear response, In response to the rate of change, This represents the relative starting point offset. The server structurally binds the aforementioned kernel response function parameters and statistical potential energy response variables to the statistical calculation path index structure one by one according to the target measurement point identifier, ensuring a one-to-one mapping relationship for subsequent data positioning and reuse.

[0111] Third, the server performs a unified encapsulation operation on the statistical computation path index structure, statistical computation results, kernel response function parameters, and statistical potential response variables to form a complete mapping record unit. The mapping record unit is constructed using a quintuple structure:

[0112]

[0113] in Indicates the statistical calculation formula identifier. The output results of this statistical calculation are as follows: The set of kernel response function parameters for all target measurement points. The set of statistical potential energy response variables corresponding to the target measurement point. This is a statistical calculation path index structure. The mapping record unit has a parsable, traceable, and indexable encapsulation format.

[0114] Fourth, the server writes the mapping record unit into the local mapping unit as a high-priority cache entry for subsequent statistical tasks. The local mapping unit supports a path matching mechanism. When a new statistical calculation request enters the system, the server compares the request path parameters with the index structure of the existing mapping record unit to determine whether one of the following two conditions is met: (1) The path parameters are completely matched, and the historical statistical calculation results are returned directly. (2) For partial matching path parameters, only the time interval or newly added target measurement points differ, the existing kernel response function parameters and statistical potential response variables are retrieved to perform incremental calculations. Through this mechanism, the server can avoid repeatedly executing the data extraction, kernel response generation and response variable mapping process, thereby greatly improving the overall response speed and computational efficiency of the statistical engine.

[0115] In summary, this technical solution uses a caching mechanism of structured path index and mapping record unit to uniformly manage key intermediate data and results in the statistical calculation process. This allows subsequent similar path requests to be directly reused or incrementally expanded based on the existing structure. This not only ensures the integrity of the calculation task in terms of accuracy and traceability, but also significantly improves the system's processing capacity and resource utilization efficiency in the face of high-frequency call scenarios.

[0116] In one possible implementation, the current statistical calculation request path parameters are obtained. If it is determined that the current statistical calculation request path parameters are completely consistent with the statistical calculation path index structure stored in the local mapping unit, the statistical calculation results in the corresponding mapping record unit are directly extracted, skipping the calculation process of the statistical calculation engine and realizing direct reuse of the results. If it is determined that the current statistical calculation request path parameters are partially similar to the statistical calculation path index structure, and there is partial overlap in the identifier set of the target measurement point, partial overlap in the time interval parameters, or partial overlap in the response dimension configuration, the parameters of the cached kernel response function and the statistical potential energy response variable are extracted as basic inputs. Incremental calculations are performed for the new measurement point or the extended time interval, and the parameters of the new kernel response function, the new statistical potential energy response variable, and the basic inputs are merged to generate the current statistical calculation result.

[0117] Specifically, the first step is for the server to obtain the current statistical calculation request path parameters. These parameters include the statistical calculation formula identifier, the target measurement point identifier set, the set time interval parameter, the response dimension configuration, and the request initiation identifier code. The server parses these path parameters into a structured path fingerprint, which is used to perform matching retrieval in the local mapping unit. The construction of the structured path fingerprint ensures the accuracy of subsequent matching operations in both semantic and data structure dimensions.

[0118] The second step involves the server comparing the current path fingerprint with all the statistical calculation path index structures stored in the local mapping unit. If a path index structure is found to be completely identical to the current path fingerprint in terms of statistical calculation formula identifier, target measurement point identifier set, set time interval parameters, and response dimension configuration, it is determined to be a "completely matched path". The server directly extracts the statistical calculation results from the mapping record unit corresponding to the path, bypassing the data extraction, kernel transformation, mapping function construction, and statistical calculation processes, thereby returning the complete calculation results with an access cost of O(1) time complexity.

[0119] Third, if no complete match is found in the path fingerprint comparison, the server continues to execute the "partially similar path" matching strategy. This strategy determines a path as incrementally reusable if any of the following conditions are met: (1) The target measurement point identifier sets partially overlap, for example, the historical path is {ABC} and the current request is {ABD}; (2) The set time interval parameters partially overlap, for example, the historical path is 00:00-12:00 and the current request is 08:00-16:00; (3) The response dimension configuration partially overlaps, for example, the current request adds "offset response dimension" calculation. The server identifies the above similar paths as incrementally reusable objects.

[0120] In the fourth step, for incrementally reusable paths, the server extracts the cached kernel response function parameters and statistical potential response variables from the similar path mapping record unit as basic inputs. For newly added target measurement point identifiers in the current request, the server re-extracts the original observation sequence and performs kernel response function construction within a set time interval; for extended time intervals, kernel response function extrapolation and mapping function calculation are performed within the newly added time period; for newly added response dimensions, the response variable projection transformation is re-performed using the original kernel response sequence as input. The incremental calculation process shortens the redundant calculation chain by reusing the basic input structure.

[0121] Fifth, the server will add the kernel response function parameters and statistical potential response variables corresponding to the new target measurement point, and structurally merge them with the historical basic input set to uniformly construct a complete structure vector set under the current request path. This set will then be input into the statistical calculation engine for computation. The statistical calculation engine will perform weighted, algebraic, or logical combinations of each dimension in the structure space, output the result value, and re-encapsulate the result and the current path parameters into a new mapping record unit, writing it to the local mapping unit and updating the local path cache.

[0122] The advantages of the above technical solution lie in its construction of a highly structured, traceable path caching mechanism and response variable reuse system, enabling the system to dynamically identify reuse opportunities, determine incremental requirements, and efficiently complete difference completion. This avoids redundant calculations, significantly saving computing resources and response time, and is particularly suitable for industrial real-time database scenarios with high-frequency statistical requests and dense changes in measurement point combinations. Furthermore, this solution achieves a full-process semantic closed loop through path index structuring and variable fusion strategies, providing stable, efficient, and scalable statistical computing support for large-scale real-time data processing systems.

[0123] This application also provides a data processing device based on a real-time database, as shown in the following figure. Figure 2 , Figure 2This is a schematic diagram of a data processing device based on a real-time database, provided in an embodiment of this application. The data processing device is a server, comprising an acquisition module 21 and a processing module 22. The acquisition module 21 acquires multiple target measurement points involved in the statistical calculation formula based on the real-time database, and extracts the original observation sequence of each target measurement point within a set time interval. Each original observation sequence includes a timestamp sequence and a corresponding value sequence for the target measurement point, forming a time-value pair set for the target measurement point. The processing module 22 inputs the time-value pair set of each target measurement point to a kernel transformation unit, constructs a time-sensitive kernel function with exponential decay characteristics for each target measurement point in the original timestamp sequence, and generates a kernel response sequence for the target measurement point under the original time axis. Based on the kernel response sequence of each target measurement point, the processing module 22 performs... A measurement point mapping function is constructed using Tikhonov regularization constraints to map the kernel response sequence to a statistical potential energy space with a unified structure, thereby obtaining the statistical potential energy response variables corresponding to the target measurement points. The statistical potential energy space is a computational domain independent of timestamps and includes mean response dimension, trend response dimension, and offset response dimension. Processing module 22 inputs the statistical potential energy response variables corresponding to multiple target measurement points into the statistical calculation engine, executes the corresponding statistical calculation formulas in the corresponding statistical potential energy spaces, and obtains the statistical calculation results. Processing module 22 binds the statistical calculation results with the original request path and stores the kernel response function and statistical potential energy response variables corresponding to the target measurement points in the local mapping unit for reuse in subsequent similar statistical calculation requests.

[0124] In one possible implementation, the acquisition module 21 acquires multiple target measurement points involved in the statistical calculation formula based on a real-time database, and extracts the original observation sequence of each target measurement point within a set time interval. Each original observation sequence includes the timestamp sequence and corresponding value sequence of the target measurement point, forming a time-value pair set of the target measurement point. Specifically, the process module 22 performs measurement point identification parsing on each target measurement point in the statistical calculation formula to generate a target measurement point mapping table; the process module 22 combines the target measurement point mapping table with the set time interval to construct a data extraction request and sends the data extraction request to the real-time database interface layer; the process module 22 controls the real-time database interface layer to call the compressed storage reading unit to perform decompression processing on the compressed data of each target measurement point within the set time interval to obtain the original observation sequence of the target measurement point; the process module 22 pairs the timestamp sequence and corresponding value sequence included in the original observation sequence one by one to form a time-value pair set. The time-value pair set includes the measurement point identifier, the observation sequence sampling density, the data source identifier, the data integrity flag, and the time interval parameter.

[0125] In one possible implementation, the processing module 22 inputs the time-value pair set of each target measurement point to the kernel transformation unit, constructs a time-sensitive kernel function with exponential decay characteristics in the original timestamp sequence for each target measurement point, and generates a kernel response sequence of the target measurement point under the original time axis. Specifically, the processing module 22 extracts the timestamp sequence and the corresponding value sequence from the corresponding time-value pair set for each target measurement point, and constructs the observation trajectory by arranging them in ascending order of timestamps; the processing module 22 constructs a time-sensitive kernel function on the original time axis of the observation trajectory, with each reference time point as the center; the processing module 22 uses the time-sensitive kernel function to perform a weighted summation of the historical observation values ​​near the reference time point, and uses the ratio of the weighted result to the sum of the weights of the kernel function as the kernel response value corresponding to the reference time point, and constructs a continuous kernel response sequence point by point along the original time axis; the processing module 22 dynamically adjusts the value of the time decay factor in the time-sensitive kernel function according to the observation density of each target measurement point to obtain the kernel response sequence of the target measurement point under the original time axis.

[0126] In one possible implementation, processing module 22 constructs a measurement point mapping function based on the kernel response sequence of each target measurement point using Tikhonov regularization constraints. This maps the kernel response sequence to a statistical potential energy space with a unified structure, obtaining the statistical potential energy response variable corresponding to the target measurement point. Specifically, processing module 22 inputs the kernel response sequence of each target measurement point into the mapping function construction unit to construct a measurement point mapping function model with the kernel response sequence as input and multiple statistical response dimensions corresponding to the statistical potential energy space as outputs. Processing module 22 introduces Tikhonov regularization constraints into the measurement point mapping function model. By adding a penalty term for the parameter magnitude of the measurement point mapping function to the loss function of the mapping function corresponding to the target measurement point, the variation amplitude of the mapping function corresponding to the target measurement point in the parameter space is controlled, generating the measurement point mapping function. Processing module 22 uses the measurement point mapping function to map the kernel response sequence of each target measurement point to the mean response dimension, trend response dimension, and offset response dimension in the statistical potential energy space, respectively, generating the statistical potential energy response variable of the target measurement point in the statistical potential energy space.

[0127] In one possible implementation, the processing module 22 inputs the statistical potential energy response variables corresponding to multiple target measurement points into the statistical calculation engine, executes the corresponding statistical calculation formulas in the corresponding statistical potential energy spaces, and obtains statistical calculation results. Specifically, the processing module 22 inputs the statistical potential energy response variables corresponding to multiple target measurement points into the statistical calculation engine in the form of a structure vector. The statistical calculation engine is used to perform response dimension matching verification on each statistical calculation formula to ensure that the response dimension of the statistical potential energy response variables of each target measurement point matches the input dimension of the statistical calculation formula. Without relying on timestamps, the processing module 22 performs weighted operations, algebraic transformations, and logical relation operations on the statistical potential energy response variables in the statistical potential energy space according to the calculation structure of the statistical calculation formula to generate output results. The processing module 22 binds the output results with the identifier of the statistical calculation formula, the identifier of the statistical potential energy response variables participating in the statistical calculation, the weight parameters used, and the confidence index to form a structured statistical calculation result.

[0128] In one possible implementation, the processing module 22 binds the statistical calculation results to the original request path and stores the kernel response function and statistical potential response variable corresponding to the target measurement point in the local mapping unit. Specifically, the processing module 22 extracts the original request path parameters of the statistical calculation task from the real-time database. The original request path parameters include the statistical calculation formula identifier, the target measurement point identifier set, the set time interval parameter, the response dimension configuration, and the request initiation identifier code. The processing module 22 establishes an index mapping between the statistical calculation results and the corresponding original request path parameters to construct a statistical calculation path index structure. The processing module 22 determines the parameters and statistical potential response variables corresponding to the kernel response function and structurally binds them to the statistical calculation path index structure. The processing module 22 encapsulates the statistical calculation path index structure, the statistical calculation results, the parameters of the kernel response function, and the statistical potential response variables into a unified mapping record unit and writes the mapping record unit into the local mapping unit for direct reuse or incremental calculation of subsequent statistical calculation requests for similar paths.

[0129] In one possible implementation, the acquisition module 21 acquires the current statistical calculation request path parameters; if the processing module 22 determines that the current statistical calculation request path parameters are completely consistent with the statistical calculation path index structure stored in the local mapping unit, it directly extracts the statistical calculation results from the corresponding mapping record unit, skips the calculation process of the statistical calculation engine, and realizes direct reuse of the results; if the processing module 22 determines that the current statistical calculation request path parameters and the statistical calculation path index structure are partially similar, and there is partial overlap in the identifier set of the target measurement point, partial overlap in the time interval parameters, or partial overlap in the response dimension configuration, it extracts the parameters of the cached kernel response function and the statistical potential energy response variable as the basic input, performs incremental calculation for the new measurement point or extended time interval, and merges the parameters of the new kernel response function, the new statistical potential energy response variable, and the basic input to generate the current statistical calculation result.

[0130] It should be noted that the above embodiments of the apparatus are only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0131] This application also provides an electronic device, with reference to... Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: at least one processor 31, at least one network interface 34, a user interface 33, a memory 35, and at least one communication bus 32.

[0132] The communication bus 32 is used to enable communication between these components.

[0133] The user interface 33 may include a display screen and a camera. Optionally, the user interface 33 may also include a standard wired interface and a wireless interface.

[0134] The network interface 34 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0135] The processor 31 may include one or more processing cores. The processor 31 connects to various parts of the server via various interfaces and lines, executing instructions, programs, code sets, or instruction sets stored in the memory 35, and calling data stored in the memory 35 to perform various server functions and process data. Optionally, the processor 31 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 31 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 31 and may be implemented as a separate chip.

[0136] The memory 35 may include random access memory (RAM) or read-only memory. Optionally, the memory 35 may include a non-transitory computer-readable storage medium. The memory 35 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 35 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 35 may also be at least one storage device located remotely from the aforementioned processor 31. Figure 3 As shown, the memory 35, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program based on a real-time database data processing method.

[0137] exist Figure 3In the electronic device shown, the user interface 33 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 31 can be used to call an application program stored in the memory 35 that is a data processing method based on a real-time database. When executed by one or more processors, the electronic device executes one or more methods as described in the above embodiments.

[0138] It should be noted that, for the sake of simplicity, the foregoing method embodiments 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, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0139] This application also provides a computer-readable storage medium storing instructions. When executed by one or more processors, these instructions cause an electronic device to perform one or more of the methods described in the above embodiments.

[0140] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0141] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A data processing method based on a real-time database, characterized in that, The method includes: Multiple target measurement points involved in the statistical calculation formula based on a real-time database are obtained. The original observation sequence of each target measurement point is extracted within a set time interval. Each original observation sequence includes the timestamp sequence and the corresponding value sequence of the target measurement point, forming a time-value pair set of the target measurement point. The time-value pair set of each target measurement point is input into the kernel transformation unit. For each target measurement point, a time-sensitive kernel function with exponential decay characteristics is constructed in the original timestamp sequence to generate the kernel response sequence of the target measurement point under the original time axis. Based on the kernel response sequence of each target measurement point, a measurement point mapping function is constructed using Tikhonov regularization constraints to map the kernel response sequence to a statistical potential energy space with a unified structure, thereby obtaining the statistical potential energy response variable corresponding to the target measurement point. The statistical potential energy space is a computational domain that does not depend on the time stamp, and the statistical potential energy space includes the mean response dimension, the trend response dimension, and the offset response dimension. The statistical potential energy response variables corresponding to each of the multiple target measurement points are input into the statistical calculation engine, and the corresponding statistical calculation formulas are executed in the corresponding statistical potential energy space to obtain the statistical calculation results. The statistical calculation results are bound to the original request path, and the kernel response function corresponding to the target measurement point and the statistical potential energy response variable are stored in the local mapping unit for reuse in subsequent similar statistical calculation requests.

2. The data processing method based on a real-time database according to claim 1, characterized in that, The process involves obtaining multiple target measurement points involved in the statistical calculation formula based on a real-time database, extracting the original observation sequence of each target measurement point within a set time interval, and each original observation sequence including the timestamp sequence and corresponding value sequence of the target measurement point, forming a time-value pair set of the target measurement point, specifically including: For each target measurement point in the statistical calculation formula, the measurement point identifier is parsed to generate a target measurement point mapping table; A data extraction request is constructed by combining the target measurement point mapping table with the set time interval, and the data extraction request is sent to the real-time database interface layer; The real-time database interface layer is controlled to call the compressed storage and reading unit to perform decompression processing on the compressed data of each target measuring point within the set time interval, so as to obtain the original observation sequence of the target measuring point. The original observation sequence is paired one-to-one with the timestamp sequence and the corresponding value sequence to form the time-value pair set. The time-value pair set includes the measurement point identifier, the observation sequence sampling density, the data source identifier, the data integrity flag, and the time interval parameter.

3. The data processing method based on a real-time database according to claim 1, characterized in that, The step of inputting the time-value pair set of each target measurement point into the kernel transformation unit, constructing a time-sensitive kernel function with exponential decay characteristics in the original timestamp sequence for each target measurement point, and generating the kernel response sequence of the target measurement point under the original time axis specifically includes: For each target measurement point, the timestamp sequence and the corresponding value sequence are extracted from the corresponding time-value pair set, and the observation trajectory is constructed by arranging them in ascending order of timestamps. On the original time axis of the observed trajectory, a time-sensitive kernel function is constructed with each reference time point as the center; The time-sensitive kernel function is used to perform a weighted summation of historical observations near the reference time point, and the ratio of the weighted result to the sum of the kernel function weights is used as the kernel response value corresponding to the reference time point. A continuous kernel response sequence is constructed point by point along the original time axis. The value of the time decay factor in the time-sensitive kernel function is dynamically adjusted according to the observation density of each target measuring point to obtain the kernel response sequence of the target measuring point under the original time axis.

4. The data processing method based on a real-time database according to claim 1, characterized in that, The kernel response sequence based on each target measurement point is used to construct a measurement point mapping function using Tikhonov regularization constraints, mapping the kernel response sequence to a statistical potential energy space with a unified structure, to obtain the statistical potential energy response variable corresponding to the target measurement point, specifically including: The kernel response sequence of each target measurement point is input into the mapping function construction unit to construct a measurement point mapping function model with the kernel response sequence as input and multiple statistical response dimensions corresponding to the statistical potential space as output. The Tikhonov regularization constraint is introduced into the measurement point mapping function model. By adding a penalty term for the parameter magnitude of the measurement point mapping function to the loss function of the mapping function corresponding to the target measurement point, the variation range of the mapping function corresponding to the target measurement point in the parameter space is controlled, and the measurement point mapping function is generated. The kernel response sequence of each target measurement point is mapped to the mean response dimension, trend response dimension, and offset response dimension in the statistical potential energy space using the measurement point mapping function, thereby generating the statistical potential energy response variable of the target measurement point in the statistical potential energy space.

5. The data processing method based on a real-time database according to claim 1, characterized in that, The step of inputting the statistical potential energy response variables corresponding to each of the multiple target measurement points into the statistical calculation engine, executing the corresponding statistical calculation formula in the corresponding statistical potential energy space, and obtaining the statistical calculation results specifically includes: The statistical potential energy response variables corresponding to each of the multiple target measurement points are input into the statistical calculation engine in the form of a structure vector. The statistical calculation engine is used to perform response dimension matching verification on each of the statistical calculation formulas to ensure that the response dimension of the statistical potential energy response variables of each target measurement point matches the input dimension of the statistical calculation formula. Without relying on timestamps, according to the calculation structure of the statistical calculation formula, the statistical potential energy response variable is subjected to weighted operations, algebraic transformations, and logical relational operations in the statistical potential energy space to generate output results. The output results are bound to the identifiers of the statistical calculation formula, the identifiers of the statistical potential response variables involved in the statistical calculation, the weight parameters used, and the confidence index to form the structured statistical calculation results.

6. The data processing method based on a real-time database according to claim 1, characterized in that, The step of binding the statistical calculation results with the original request path and storing the kernel response function corresponding to the target measurement point and the statistical potential response variable in the local mapping unit specifically includes: Extract the original request path parameters of the statistical calculation task of the real-time database. The original request path parameters include the statistical calculation formula identifier, the target measurement point identifier set, the set time interval parameter, the response dimension configuration, and the request initiation identifier code. Then, establish an index mapping between the statistical calculation results and the corresponding original request path parameters to construct a statistical calculation path index structure. Determine the parameters corresponding to the kernel response function and the statistical potential response variable, and structurally bind them to the statistical calculation path index structure; The statistical calculation path index structure, the statistical calculation results, the parameters of the kernel response function, and the statistical potential response variable are uniformly encapsulated into a mapping record unit, and the mapping record unit is written into the local mapping unit so that subsequent statistical calculation requests for similar paths can be directly reused or incrementally calculated.

7. The data processing method based on a real-time database according to claim 6, characterized in that, The method further includes: Get the current statistical calculation request path parameters; If it is determined that the current statistical calculation request path parameters are completely consistent with the statistical calculation path index structure stored in the local mapping unit, then the statistical calculation results in the corresponding mapping record unit are directly extracted, skipping the calculation process of the statistical calculation engine, and the results are directly reused. If it is determined that the current statistical calculation request path parameters are partially similar to the statistical calculation path index structure, and that there is partial overlap in the identifier set of the target measurement point, the time interval parameters, or the response dimension configuration, then the cached parameters of the kernel response function and the statistical potential response variable are extracted as basic inputs. Incremental calculations are performed for the newly added measurement point or the extended time interval, and the parameters of the newly added kernel response function, the newly added statistical potential response variable, and the basic inputs are merged to generate the current statistical calculation result.

8. A data processing device based on a real-time database, characterized in that, The data processing device includes an acquisition module (21) and a processing module (22), wherein, The acquisition module (21) is used to acquire multiple target measurement points involved in the statistical calculation formula based on the real-time database, and extract the original observation sequence of each target measurement point within a set time interval. Each original observation sequence includes the timestamp sequence and the corresponding value sequence of the target measurement point, which constitutes the time-value pair set of the target measurement point. The processing module (22) is used to input the time-value pair set of each target measurement point to the kernel transformation unit, construct a time-sensitive kernel function with exponential decay characteristics in the original timestamp sequence for each target measurement point, and generate the kernel response sequence of the target measurement point under the original time axis. The processing module (22) is further configured to construct a measurement point mapping function based on the kernel response sequence of each target measurement point using Tikhonov regularization constraints, map the kernel response sequence to a statistical potential energy space with a unified structure, and obtain the statistical potential energy response variable corresponding to the target measurement point. The statistical potential energy space is a computational domain that does not depend on the time stamp. The statistical potential energy space includes the mean response dimension, the trend response dimension, and the offset response dimension. The processing module (22) is also used to input the statistical potential energy response variables corresponding to each of the multiple target measurement points into the statistical calculation engine, execute the corresponding statistical calculation formula in the corresponding statistical potential energy space, and obtain the statistical calculation result; The processing module (22) is also used to bind the statistical calculation results with the original request path, and store the kernel response function corresponding to the target measurement point and the statistical potential energy response variable in the local mapping unit for reuse in subsequent similar statistical calculation requests.

9. An electronic device, characterized in that, The electronic device includes a processor (31), a memory (35), a user interface (33), and a network interface (34). The memory (35) is used to store instructions. The user interface (33) and the network interface (34) are both used to communicate with other devices. The processor (31) is used to execute the instructions stored in the memory (35) to cause the electronic device to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, perform the method as described in any one of claims 1 to 7.

Images