Multi-dimensional data processing method, apparatus, device, medium and product

Abstract

The application provides a multi-dimensional data processing method, device, equipment, medium and product, and relates to the technical field of data processing. The method comprises the following steps: obtaining multi-dimensional data to be processed and a historical data set, the historical data set comprising a plurality of historical data, each historical data being associated with a historical algorithm; calculating a similarity index between the multi-dimensional data and each historical data in the historical data set; determining a target similarity index from the similarity indexes corresponding to each historical data; determining a historical algorithm associated with the historical data corresponding to the target similarity index as a target historical algorithm; and processing the multi-dimensional data according to the target historical algorithm to obtain a processing result. The method of the application improves the multi-dimensional data processing accuracy and efficiency by flexibly selecting a suitable historical algorithm for data processing, and improves the user experience.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a multi-dimensional data processing method, apparatus, equipment, medium and product. Background Technology

[0002] In modern enterprise decision support systems, the need for intelligent processing of multi-dimensional data is becoming increasingly urgent. With the deep integration of big data, artificial intelligence, and Internet of Things technologies, enterprises need to quickly extract valuable information from massive amounts of heterogeneous data to support dynamic decision-making.

[0003] In existing technologies, a three-tier architecture is typically used to implement multidimensional data processing. The data access layer aggregates data from databases, application programming interfaces (APIs), or real-time streams through an Extract-Load-Transform (ETL) process. The computation and processing layer, based on an Online Analytical Processing (OLAP) engine, processes data according to predefined metric calculation rules or fixed algorithm models. The visualization layer displays the processing results through configurable components (such as charts and dashboards) and supports basic user interactions such as filtering and drill-down.

[0004] However, existing methods use fixed algorithms when processing multi-dimensional data, which cannot be flexibly adjusted according to the characteristics of the data, resulting in low processing accuracy and efficiency, and a poor user experience. Summary of the Invention

[0005] This application provides a multi-dimensional data processing method, apparatus, device, medium, and product, which is used to obtain multi-dimensional data to be processed and historical data sets, calculate similarity indicators, determine target similarity indicators, process multi-dimensional data through corresponding historical algorithms, and generate processing results. This allows for flexible processing of multi-dimensional data, improves the accuracy and efficiency of multi-dimensional data processing, and enhances user experience.

[0006] Firstly, this application provides a multi-dimensional data processing method, which includes:

[0007] Acquire the multi-dimensional data to be processed and the historical data set. The historical data set includes multiple historical data points, and each historical data point is associated with a historical algorithm.

[0008] Calculate similarity indicators between multidimensional data and historical data in the historical dataset;

[0009] Target similarity indicators are determined from the similarity indicators corresponding to each historical data.

[0010] The historical algorithms associated with the historical data corresponding to the target similarity index are identified as the target historical algorithms;

[0011] The target historical algorithm is used to process multi-dimensional data to obtain the processing results.

[0012] In one possible implementation, the similarity index between the multi-dimensional data and each historical data point in the historical dataset is calculated, including:

[0013] Calculate the Pearson correlation coefficient between the multidimensional data and each historical data point in the historical dataset;

[0014] Each Pearson correlation coefficient is used as a similarity indicator for each historical data point.

[0015] In one possible implementation, the target similarity index is determined from the similarity indices corresponding to each historical data point, including:

[0016] From the Pearson correlation coefficients corresponding to each historical data, the Pearson correlation coefficient with the largest value is determined as the first similarity index;

[0017] If the first similarity index is greater than or equal to the first preset threshold, then the first similarity index is determined as the target similarity index.

[0018] In one possible implementation, calculating the similarity index between the multi-dimensional data and each historical data point in the historical dataset further includes:

[0019] If the first similarity index is less than the first preset threshold, then calculate the Bayesian factor between the multi-dimensional data and each historical data in the historical data set.

[0020] Each Bayes factor is used as a similarity index for each historical data point.

[0021] From the similarity indicators corresponding to various historical data, the target similarity indicators are determined, including:

[0022] From the Bayesian factors corresponding to each historical data, the Bayesian factor with the largest value is determined as the second similarity index;

[0023] If the second similarity index is greater than or equal to the second preset threshold, then the second similarity index is determined as the target similarity index.

[0024] In one possible implementation, the method further includes:

[0025] If the second similarity index is less than the second preset threshold, the marginal contribution of each historical algorithm to the historical task is calculated according to the Shapley value algorithm, and the marginal contribution is used as the weight of the corresponding historical algorithm.

[0026] The historical algorithms are weighted and fused according to their weights to generate a fused algorithm, which is then identified as the target historical algorithm.

[0027] In one possible implementation, after processing the multi-dimensional data according to the target history algorithm to obtain the processing result, the method further includes:

[0028] Based on the processing results, multiple business metrics are obtained;

[0029] A metric association network is constructed based on business metrics. The metric association network includes multiple nodes and connecting edges. Each node corresponds to a business metric, and each connecting edge indicates that there is a correlation, causal relationship, or attribution relationship between the metrics corresponding to two nodes.

[0030] The indicator association network is displayed for users to view, and in response to the user's selection of any node, other nodes associated with the user's selected node and their corresponding association strength are dynamically highlighted.

[0031] Secondly, this application provides a multi-dimensional data processing apparatus, the apparatus comprising:

[0032] The acquisition module is used to acquire the multi-dimensional data to be processed and the historical data set. The historical data set includes multiple historical data, and each historical data is associated with a historical algorithm.

[0033] The calculation module is used to calculate the similarity index between multi-dimensional data and each historical data in the historical data set;

[0034] The calculation module is also used to determine the target similarity index from the similarity indexes corresponding to each historical data.

[0035] The algorithm determination module is used to determine the historical algorithms associated with the historical data corresponding to the target similarity index as the target historical algorithm;

[0036] The data processing module is used to process multi-dimensional data according to the target historical algorithm to obtain the processing results.

[0037] In one possible implementation, the calculation module is also used to calculate the Pearson correlation coefficient between the multidimensional data and each historical data in the historical data set;

[0038] The calculation module is also used to use each Pearson correlation coefficient as a similarity index for each historical data.

[0039] In one possible implementation, the calculation module is further configured to determine the largest Pearson correlation coefficient from the Pearson correlation coefficients corresponding to each historical data, and use it as the first similarity index.

[0040] The calculation module is also used to determine the first similarity index as the target similarity index if the first similarity index is greater than or equal to the first preset threshold.

[0041] In one possible implementation, the calculation module is further configured to calculate the Bayesian factor between the multi-dimensional data and each historical data in the historical data set if the first similarity index is less than the first preset threshold.

[0042] The calculation module is also used to use each Bayes factor as a similarity index for each historical data.

[0043] The calculation module is also used to determine the Bayes factor with the largest value from the Bayes factors corresponding to each historical data, as the second similarity index;

[0044] The calculation module is also used to determine the second similarity index as the target similarity index if the second similarity index is greater than or equal to the second preset threshold.

[0045] In one possible implementation, the calculation module is further configured to calculate the marginal contribution of each historical algorithm to the historical task according to the Shapley value algorithm if the second similarity index is less than the second preset threshold, and use the marginal contribution as the weight of the corresponding historical algorithm.

[0046] The algorithm determination module is also used to perform weighted fusion of each historical algorithm according to the weight, generate a fusion algorithm, and determine the fusion algorithm as the target historical algorithm.

[0047] In one possible implementation, the data processing module is also used to obtain multiple business indicators based on the processing results;

[0048] The data processing module is also used to construct an indicator association network based on business indicators. The indicator association network includes multiple nodes and connecting edges, where each node corresponds to a business indicator, and each connecting edge indicates that there is a correlation, causal relationship or attribution relationship between the indicators corresponding to two nodes.

[0049] The data processing module is also used to display the indicator association network for users to view, and in response to the user's selection of any node, dynamically highlight other nodes associated with the user's selected node and their corresponding association strength.

[0050] Thirdly, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor.

[0051] The memory stores the instructions that the computer executes.

[0052] The processor executes computer execution instructions stored in memory to implement a multi-dimensional data processing method according to the first aspect of the invention.

[0053] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement a multi-dimensional data processing method according to the first aspect of the invention.

[0054] Fifthly, this application provides a computer program product, including a computer program, which, when executed by a processor, is used to implement a multi-dimensional data processing method according to the first aspect of the invention.

[0055] Based on the implementation methods provided in the above aspects, this application can be further combined to provide more implementation methods.

[0056] This application provides a multi-dimensional data processing method, apparatus, device, medium, and product, comprising: first, acquiring the multi-dimensional data to be processed and a historical data set, wherein the historical data set includes multiple historical data points, each historical data point being associated with a historical algorithm; next, calculating the similarity index between the multi-dimensional data and each historical data point in the historical data set; subsequently, determining a target similarity index from the similarity indices corresponding to each historical data point; then, determining the historical algorithm associated with the historical data corresponding to the target similarity index as the target historical algorithm; finally, processing the multi-dimensional data according to the target historical algorithm to obtain the processing result. This achieves the following technical effects: based on the characteristics of the data to be processed, the most suitable historical algorithm is flexibly selected to process the multi-dimensional data, resulting in accurate and efficient processing results. By flexibly processing multi-dimensional data, the accuracy and efficiency of multi-dimensional data processing are significantly improved, while providing users with higher-quality and more personalized services, effectively enhancing the user experience. It realizes intelligent algorithm matching and reuse based on historical experience, avoiding the lack of adaptability caused by fixed algorithms in traditional multi-dimensional data processing methods, and can improve the accuracy, efficiency, and scenario generalization ability of multi-dimensional data processing. Attached Figure Description

[0057] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0058] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0059] Figure 1This is a schematic diagram illustrating an application scenario of a multi-dimensional data processing method provided in an embodiment of this application.

[0060] Figure 2 A flowchart illustrating a multi-dimensional data processing method provided in an embodiment of this application;

[0061] Figure 3 This is a schematic diagram of the structure of a multi-dimensional data processing device provided in an embodiment of this application;

[0062] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0063] Figure label:

[0064] 101 - Display terminal; 102 - Server;

[0065] 310 - Acquisition Module; 320 - Calculation Module; 330 - Algorithm Determination Module; 340 - Data Processing Module;

[0066] 410 - Processor; 420 - Memory; 430 - Communication components; 440 - Bus. Detailed Implementation

[0067] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0068] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" do not necessarily imply difference. It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate that something is being used as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in this application should not be construed as being better or more advantageous than other embodiments or design schemes. Specifically, the use of "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner. In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more.

[0069] It should be noted that the phrase "at...time" in the embodiments of this application can refer to the instant at which a certain situation occurs, or to a period of time after the occurrence of a certain situation; the embodiments of this application do not specifically limit this. Furthermore, the multi-dimensional data processing method provided in the embodiments of this application is merely an example; a multi-dimensional data processing method may include more or less content.

[0070] In the process of modern enterprise operation and management, decision support systems play a crucial role. With the deep integration and rapid development of big data, artificial intelligence, and Internet of Things technologies, the data environment faced by enterprises has undergone tremendous changes. Massive amounts of heterogeneous data are emerging like a tide, with wide-ranging sources and diverse formats, covering all aspects of enterprise operations, such as market dynamics data, internal production and operation data, customer behavior data, and supply chain data.

[0071] Against this backdrop, enterprises' need for intelligent processing of multi-dimensional data has become increasingly urgent. The ability to quickly and accurately extract valuable information from this massive amount of heterogeneous data to support dynamic decision-making has become crucial.

[0072] Currently, most existing multi-dimensional data processing technologies adopt a three-tier architecture. In the data access layer, ETL processes are primarily used to integrate and standardize raw data scattered across different data sources, providing a unified data foundation for subsequent processing. The computational processing layer utilizes an OLAP engine to process the data according to pre-defined rules or fixed algorithm models, calculating various indicators. The visualization layer typically constructs data dashboards in the form of configurable components, displaying the processed data in intuitive charts, reports, and other formats, facilitating multi-dimensional data viewing and analysis for users.

[0073] However, this existing processing method has significant limitations. Because the data processing logic relies heavily on manual pre-setting, the algorithm model becomes fixed once deployed and cannot be dynamically adjusted based on the distribution characteristics, modal structure, or business semantics of the input multi-dimensional data. When faced with new scenarios or abnormal patterns, it is difficult to reuse historically effective analytical experience, resulting in insufficient processing accuracy and delayed response, severely restricting the level of intelligence in decision support and the user experience.

[0074] Based on this, embodiments of this application propose a multi-dimensional data processing method, apparatus, device, medium, and product, which can be used in the field of data processing technology and aims to solve the above-mentioned technical problems of the prior art. The method involves acquiring multi-dimensional data to be processed and a historical data set, wherein the historical data set contains multiple historical data points, and each historical data point is associated with a historical algorithm. These historical data points and algorithms provide rich reference data for subsequent data processing. Next, the similarity index between the multi-dimensional data to be processed and each historical data point in the historical data set is calculated. Through scientific calculation methods, the degree of similarity between the multi-dimensional data and the historical data is quantified, providing key data support for determining the target algorithm. Then, the target similarity index is accurately determined from the similarity indices corresponding to each historical data point. This step can filter out the historical data most relevant to the data to be processed, ensuring the targeting and accuracy of subsequent processing. Afterwards, the historical algorithm associated with the historical data corresponding to the target similarity index is determined as the target historical algorithm. In this way, the most suitable historical algorithm can be flexibly selected for data processing based on the characteristics of the data to be processed. Finally, the multi-dimensional data to be processed is processed according to the determined target historical algorithm, thereby obtaining accurate and efficient processing results. By flexibly processing multi-dimensional data, the accuracy and efficiency of multi-dimensional data processing are significantly improved, while providing users with higher-quality and more personalized services, effectively enhancing the user experience.

[0075] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0076] To better understand the solutions of the embodiments of this application, an application scenario involved in the embodiments of this application will be introduced below.

[0077] For specific application scenarios of this application, please refer to [link / reference needed]. Figure 1 . Figure 1 This is a schematic diagram illustrating an application scenario of a multi-dimensional data processing method provided in an embodiment of this application. It should be noted that... Figure 1 The examples shown are merely application scenarios that can be applied to the embodiments of this application, in order to help those skilled in the art understand the technical content of this application, but do not mean that the embodiments of this application cannot be used in other devices, systems, environments or scenarios.

[0078] like Figure 1 As shown, the application scenario includes: display terminal 101 and server 102.

[0079] The display terminal 101 is used to display the processing results for user viewing. The server 102 is communicatively connected to the display terminal 101 and is used to acquire the multi-dimensional data to be processed and the historical data set, and to determine the target historical algorithm to process the multi-dimensional data based on the historical data set to obtain the processing results.

[0080] Figure 2 This is a flowchart illustrating a multi-dimensional data processing method provided in an embodiment of this application. The execution entity of this embodiment can be... Figure 1 The server 102 in the illustrated embodiment can also be other computer-related devices, and this embodiment does not impose any particular limitation on it. For ease of description, this application embodiment uniformly describes the execution subject of a multi-dimensional data processing method as a server. Figure 2 As shown, the method includes:

[0081] S201. Obtain the multi-dimensional data to be processed and the historical data set.

[0082] In this embodiment of the application, the historical data set includes multiple historical data sets, and each historical data set is associated with a historical algorithm.

[0083] Specifically, multidimensional data refers to a data set where each observed object has multiple attributes or variables, such as traffic flow, power load, and customer behavior data from multiple sources and heterogeneous structures. This type of data is very common in fields such as statistics, machine learning, data mining, bioinformatics, and image processing. Multidimensional data can originate from multi-source heterogeneous sensors in industrial settings, business databases, or text semantic encoders. To facilitate calculation and analysis, multidimensional data is usually organized into a data matrix. This matrix can be an n×m matrix, where n represents the number of observation samples and m represents the feature dimensions contained in each observation sample.

[0084] Simultaneously, the server can retrieve historical datasets from the historical knowledge base. These datasets can consist of multiple processed historical data sets. Historical data refers to data matrices that have previously been used in calculations. A historical dataset is a collection of data matrices that have been used in calculations. Each piece of historical data is associated with a historical algorithm. A historical algorithm refers to the processing logic successfully applied to that historical data; that is, the method used to calculate and transform one multi-dimensional data set into another. Examples include anomaly detection models, dimensionality reduction transformation functions, indicator prediction formulas, or feature fusion strategies.

[0085] S202. Calculate the similarity index between the multi-dimensional data and each historical data in the historical data set.

[0086] Specifically, the server can calculate a similarity index between the multi-dimensional data to be processed and each historical data point in the historical dataset. This similarity index is used to quantify the degree of similarity between the currently input multi-dimensional data and the historical data in terms of distribution characteristics, structural patterns, or generation mechanisms. In a preferred embodiment, the calculation of the similarity index can adopt a phased strategy. First, the Pearson Correlation Coefficient can be calculated to analyze linear correlation. If the correlation does not reach a first preset threshold, the Bayesian factor can be further calculated to determine whether the data originates from the same data generation mechanism.

[0087] S203. Determine the target similarity index from the similarity indices corresponding to each historical data.

[0088] Specifically, the server can determine the target similarity index from all similarity indices corresponding to historical data. Specifically, it can select the similarity index with the highest value as a candidate similarity index and combine it with a preset threshold for validity assessment. For example, if the maximum value meets the first threshold condition (e.g., Pearson correlation coefficient ≥ 0.78), it is directly identified as the target similarity index; otherwise, it proceeds to the next stage of similarity analysis or triggers the historical algorithm fusion mechanism.

[0089] S204. The historical algorithms associated with the historical data corresponding to the target similarity index are determined as the target historical algorithms.

[0090] Specifically, the server can identify the historical algorithms associated with the historical data corresponding to the target similarity index as the target historical algorithm. This target historical algorithm is then regarded as the processing strategy that best suits the current input multi-dimensional data.

[0091] S205. Process the multi-dimensional data according to the target history algorithm to obtain the processing results.

[0092] Specifically, the server can invoke the target's historical algorithm to perform processing operations on the multi-dimensional data to be processed, generating corresponding processing results. These results may include, but are not limited to: anomaly scores, health status indices, predicted value sequences, dimensionality-reduced feature vectors, or fused indicator vectors. These results can be further used for visualization, decision-making alerts, or closed-loop control. The server can feed these processing results back to the control system of the observed object to adjust its operating parameters and achieve optimized control.

[0093] Through the above method, this embodiment realizes intelligent algorithm matching and reuse based on historical experience, avoiding the problem of insufficient adaptability caused by fixed algorithms in traditional multi-dimensional data processing methods, and can improve the accuracy, efficiency and scenario generalization ability of multi-dimensional data processing.

[0094] This embodiment provides a multi-dimensional data processing method. First, it acquires the multi-dimensional data to be processed and a historical data set, the historical data set including multiple historical data, each historical data being associated with a historical algorithm. Next, it calculates the similarity index between the multi-dimensional data and each historical data in the historical data set. Then, it determines the target similarity index from the similarity indices corresponding to each historical data. Then, it determines the historical algorithm associated with the historical data corresponding to the target similarity index as the target historical algorithm. Finally, it processes the multi-dimensional data according to the target historical algorithm to obtain the processing result.

[0095] The following technical effects were achieved: Based on the characteristics of the data to be processed, the most suitable historical algorithm is flexibly selected to process the multi-dimensional data, resulting in accurate and efficient processing results. By flexibly processing multi-dimensional data, the accuracy and efficiency of multi-dimensional data processing are significantly improved, while providing users with higher-quality and more personalized services, effectively enhancing the user experience. Intelligent algorithm matching and reuse based on historical experience are implemented, avoiding the lack of adaptability caused by fixed algorithms in traditional multi-dimensional data processing methods, and improving the accuracy, efficiency, and scenario generalization ability of multi-dimensional data processing.

[0096] In one possible implementation, the similarity index between the multidimensional data and each historical data in the historical data set is calculated, including: calculating the Pearson correlation coefficient between the multidimensional data and each historical data in the historical data set; and using each Pearson correlation coefficient as the similarity index corresponding to each historical data.

[0097] Specifically, the server can first align the multi-dimensional data to be processed with each historical data point to the same feature dimension space. If the original dimensions of the two are inconsistent, dimension alignment can be performed through feature mapping, zero-padding, or common subset extraction to ensure the effectiveness of subsequent correlation calculations.

[0098] Subsequently, for each pair of aligned data vectors (i.e., the multidimensional data to be processed and the specific historical data), the Pearson correlation coefficient is calculated.

[0099] Specifically, the formula for calculating the Pearson correlation coefficient is as follows:

[0100]

[0101] The multi-dimensional data to be processed is (x1, x2, ..., x...). m The specific historical data is (y1, y2, ..., y). m ), where r is the Pearson correlation coefficient between the multidimensional data to be processed and the specific historical data. The mean of the multidimensional data to be processed. is the sample mean of specific historical data, and m is the number of feature dimensions. The Pearson correlation coefficient ranges from -1 to 1; the closer it is to 1, the more similar the two sets of data are in a linear relationship.

[0102] Finally, the server can use the calculated Pearson correlation coefficients as similarity indicators for the corresponding historical data, which can then be used for matching decisions in subsequent target historical algorithms. For example, if the Pearson correlation coefficient for a specific historical data is high (e.g., ≥0.78), it indicates that the current input multi-dimensional data is highly consistent with the overall trend of that historical data, and the associated historical algorithm is likely to be applicable to the current scenario.

[0103] In one possible implementation, the target similarity index is determined from the similarity indices corresponding to each historical data, including: determining the Pearson correlation coefficient with the largest value from the Pearson correlation coefficients corresponding to each historical data, as the first similarity index; if the first similarity index is greater than or equal to a first preset threshold, then the first similarity index is determined as the target similarity index.

[0104] Specifically, the server can iterate through all the Pearson correlation coefficients (i.e., the similarity index calculated above) corresponding to all historical data in the historical dataset, select the Pearson correlation coefficient with the highest value, and record it as the first similarity index. This first similarity index reflects the degree of linear correlation between the multi-dimensional data to be processed and the most similar historical data in the historical dataset.

[0105] Subsequently, the server can compare the first similarity index with a preset first threshold. The first preset threshold can be configured according to the actual business scenario and data characteristics. In a preferred embodiment, the first preset threshold can be set to 0.78. This value is based on extensive experimental verification. When the Pearson correlation coefficient is not lower than 0.78, the two sets of multidimensional data have a high degree of consistency in distribution and trend, and their underlying generation mechanisms or business semantics are usually highly similar. Therefore, the corresponding historical algorithms can be considered to have good transfer applicability.

[0106] If the first similarity index is greater than or equal to the first preset threshold, the server can determine that the current input multi-dimensional data is close enough to the most similar historical data and meets the conditions for algorithm reuse. Then, the first similarity index can be formally determined as the target similarity index, and the historical algorithm bound to the associated historical data can be used as the target historical algorithm for subsequent multi-dimensional data processing.

[0107] Conversely, if the first similarity index is less than the first preset threshold, it indicates that there is not enough similar historical data available for direct matching. The server can proceed to the next stage of similarity analysis (e.g., judgment based on the generation mechanism of Bayesian factors) or trigger a multi-algorithm fusion mechanism to ensure the stability and adaptability of the processing strategy.

[0108] In one possible implementation, calculating the similarity index between the multi-dimensional data and each historical data in the historical data set further includes: if the first similarity index is less than a first preset threshold, calculating the Bayesian factor between the multi-dimensional data and each historical data in the historical data set; and using each Bayesian factor as the similarity index corresponding to each historical data.

[0109] The process of determining the target similarity index from the similarity indices corresponding to each historical data also includes: determining the Bayes factor with the largest value from the Bayes factors corresponding to each historical data as the second similarity index; if the second similarity index is greater than or equal to the second preset threshold, then the second similarity index is determined as the target similarity index.

[0110] Specifically, if the aforementioned first similarity index (i.e., the maximum Pearson correlation coefficient) is less than a first preset threshold (e.g., 0.78), it indicates that the currently input multi-dimensional data lacks sufficient similarity to any historical data in terms of linear correlation, and the historical algorithms associated with it cannot be directly reused. In this case, the server can further use a Bayes factor based on a probabilistic generative model to analyze the deep mechanism similarity between the data.

[0111] Specifically, the server can construct two hypothesis models, a null hypothesis H0 and an alternative hypothesis H1, for the multi-dimensional data to be processed and for each historical data point. The null hypothesis H0 states that the two sets of data come from different data generation mechanisms, while the alternative hypothesis H1 states that the two sets of data come from the same data generation mechanism.

[0112] The Bayesian factor is calculated based on the joint observation sample composed of the multi-dimensional data to be processed and specific historical data.

[0113] Specifically, the formula for calculating the Bayes factor is as follows:

[0114]

[0115] in, The Bayesian factor is the relationship between the multidimensional data to be processed and the specific historical data. It is the marginal likelihood between the observed multidimensional data to be processed and the specific historical data under the null hypothesis H0. It is the marginal likelihood of the multidimensional data to be processed observed under the alternative hypothesis H1 and the specific historical data.

[0116] The server can determine the strength of evidence for H1 relative to H0 based on the Bayesian factors. To facilitate a unified comparison, the Bayesian factors can be mapped to a standardized interval, for example, through hyperbolic tangent (tanh) transformation or normalization, ultimately obtaining a Bayesian similarity score with a value range of [-1, 1], where positive values ​​indicate support for homology and negative values ​​indicate support for heterology.

[0117] Subsequently, the server can use each calculated Bayesian factor (or its standardized form) as a similarity index for the corresponding historical data.

[0118] Accordingly, the target similarity index is determined from the similarity indices corresponding to each historical data, and also includes: selecting the Bayes factor with the largest value from the Bayes factors corresponding to all historical data as the second similarity index.

[0119] If the second similarity index is greater than or equal to a second preset threshold (e.g., 0.55), the server can determine that the currently input multi-dimensional data and the historical data corresponding to the Bayesian factor have significant consistency in their generation mechanisms, thus meeting the conditions for algorithm transfer. Therefore, the second similarity index can be determined as the target similarity index, and its associated historical algorithm can be used as the target historical algorithm.

[0120] Conversely, if the second similarity index is lower than the second preset threshold, the server can determine that no single historical algorithm can be directly applied, and can then proceed to the next stage, which involves analyzing the contribution of multiple historical algorithms and weighting and fusing them based on the Shapley value to generate a composite processing strategy that adapts to the current input multi-dimensional data.

[0121] In one possible implementation, the method further includes: if the second similarity index is less than a second preset threshold, calculating the marginal contribution of each historical algorithm to the historical task according to the Shapley value algorithm, and using the marginal contribution as the weight of the corresponding historical algorithm; weighting and fusing each historical algorithm according to the weight to generate a fusion algorithm, and determining the fusion algorithm as the target historical algorithm.

[0122] Specifically, if the aforementioned second similarity index is less than the second preset threshold (e.g., 0.55), it indicates that the currently input multi-dimensional data and any historical data lack sufficient consistency at the generation mechanism level, and a single historical algorithm cannot be directly reused. In this case, the server can no longer rely on the matching strategy of a single historical algorithm, but can switch to a multi-historical algorithm collaborative fusion mode.

[0123] Specifically, the server can call the historical algorithm set consisting of all historical algorithms. And based on the Shapley Value algorithm, each historical algorithm A is quantified. i Marginal contribution to historical tasks. Historical tasks refer to objectives such as metric prediction, anomaly detection, or feature transformation that were completed in similar business scenarios in the past. Their performance can be measured by evaluation metrics such as accuracy, F1 score, or mean squared error.

[0124] The Shapley value can be calculated based on a cooperative game theory framework, treating each historical algorithm as a participant and the joint performance of the combination of historical algorithms as the coalition value. For historical algorithm A... i Its Shapeli value It can be defined as the average marginal gain brought by the historical algorithm across all possible subset unions, and the calculation formula is as follows:

[0125]

[0126] Where v(S) represents the performance score (i.e., value function) of the historical algorithms in subset S when they are jointly executed on the historical task. That is, historical algorithm A i Marginal contribution after joining Alliance S.

[0127] The server can calculate the various Shapley values. After normalization, it becomes the corresponding historical algorithm A. i weight w i The sum of the weights corresponding to all historical algorithms is 1 and w i ≥0.

[0128] Subsequently, the server can perform weighted fusion of the historical algorithms according to their weights to generate a fusion algorithm. In one implementation, the fusion method may include, but is not limited to: weighted averaging of the output results of the historical algorithms; constructing a weighted ensemble model (such as weighted voting, stacked generalization); and using the weights as prior information to guide the initialization or regularization of the new model.

[0129] Ultimately, the server can identify the fusion algorithm as the target historical algorithm and use it to process the current multi-dimensional data, thereby achieving stable and adaptive data processing through the historical data set even in the absence of highly similar historical cases.

[0130] In one possible implementation, after processing multi-dimensional data according to the target historical algorithm to obtain the processing result, the method further includes: obtaining multiple business indicators based on the processing result; constructing an indicator association network based on the business indicators, the indicator association network including multiple nodes and connecting edges, wherein each node corresponds to a business indicator, and each connecting edge indicates that there is a correlation, causal relationship or attribution relationship between the indicators corresponding to two nodes; displaying the indicator association network for users to view, and dynamically highlighting other nodes associated with the user-selected node and their corresponding association strength in response to the user's selection operation of any node.

[0131] Specifically, the server can parse and semantically map the processing results to extract multiple business metrics with business implications. These business metrics may include, but are not limited to: equipment health index, energy efficiency ratio, fault score, capacity achievement rate, thermal anomaly degree, and supply chain turnover days. Their values ​​can be calculated and generated by the target historical algorithm based on multi-dimensional data.

[0132] Subsequently, the server can construct a metric association network based on multiple business metrics. This metric association network can be formally represented as a weighted directed graph G = (V, E, P). Here, each node v in the node set V... i Corresponding to a business metric; each edge e in the edge set E ij Indicates business metrics v i With business metrics v j There are statistical or logical dependencies between them.

[0133] The association weight P represents the type and strength of the association, which can specifically include correlation, causation or attribution.

[0134] Correlation can be quantified using Pearson correlation coefficient, Spearman rank correlation, or the Maximum Information Coefficient (MIC) to determine the statistical dependence between business indicators. Causality can be inferred from Granger causality tests or Bayesian network structure learning to determine the temporal or structural causal direction. Attribution can be calculated using Shapley values ​​or Shapley Additive exPlanations (SHAP) to determine the contribution of one business indicator to the change of another.

[0135] Furthermore, the server can visualize the network of indicators, rendering it into an interactive graphical interface (such as a force-directed graph, Sankey diagram, or matrix heatmap), and deploy it on World Wide Web clients or multi-screen decision dashboards for users to view.

[0136] At the user interaction level, the server can monitor user selection actions on any node in real time (such as clicking, hovering, or voice commands). In response to this selection action, the server can dynamically highlight all adjacent nodes directly connected to the selected node and simultaneously present the association type (such as causal, positive correlation, principal attribution) and association strength values ​​(such as correlation coefficient 0.82, SHAP contribution value +15.3%) corresponding to each connection edge. In addition, users can drill down to view detailed analysis reports of the association relationships, including time series evolution, confidence intervals, or historical case support.

[0137] Through the above methods, this embodiment not only achieves intelligent processing of multi-dimensional data, but also transforms the processing results into an interpretable, explorable, and interactive decision knowledge graph, which can significantly improve enterprise users' cognitive efficiency and decision-making accuracy regarding complex business systems.

[0138] This application embodiment can divide an electronic device or main control device into functional modules according to the above method examples. For example, each function can be divided into its own functional modules, or two or more functions can be integrated into one processing unit. The integrated unit can be implemented in hardware or as a software functional module. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division; in actual implementation, there may be other division methods.

[0139] Figure 3 This is a schematic diagram of the structure of a multi-dimensional data processing device provided in an embodiment of this application. Figure 3 As shown, the device includes: an acquisition module 310, a calculation module 320, an algorithm determination module 330, and a data processing module 340.

[0140] The acquisition module 310 is used to acquire the multi-dimensional data to be processed and the historical data set. The historical data set includes multiple historical data, and each historical data is associated with a historical algorithm.

[0141] The calculation module 320 is used to calculate the similarity index between the multi-dimensional data and each historical data in the historical data set.

[0142] The calculation module 320 is also used to determine the target similarity index from the similarity indexes corresponding to each historical data.

[0143] The algorithm determination module 330 is used to determine the historical algorithms associated with the historical data corresponding to the target similarity index as the target historical algorithm.

[0144] The data processing module 340 is used to process multi-dimensional data according to the target historical algorithm to obtain the processing results.

[0145] In one possible implementation, the calculation module 320 is also used to calculate the Pearson correlation coefficient between the multidimensional data and each historical data in the historical data set.

[0146] The calculation module 320 is also used to use each Pearson correlation coefficient as a similarity index for each historical data.

[0147] In one possible implementation, the calculation module 320 is further configured to determine the largest Pearson correlation coefficient from the Pearson correlation coefficients corresponding to each historical data, and use it as the first similarity index.

[0148] The calculation module 320 is also used to determine the first similarity index as the target similarity index if the first similarity index is greater than or equal to the first preset threshold.

[0149] In one possible implementation, the calculation module 320 is further configured to calculate the Bayesian factor between the multi-dimensional data and each historical data in the historical data set if the first similarity index is less than the first preset threshold.

[0150] The calculation module 320 is also used to use each Bayes factor as a similarity index corresponding to each historical data.

[0151] The calculation module 320 is also used to determine the Bayes factor with the largest value from the Bayes factors corresponding to each historical data, as the second similarity index.

[0152] The calculation module 320 is also used to determine the second similarity index as the target similarity index if the second similarity index is greater than or equal to the second preset threshold.

[0153] In one possible implementation, the calculation module 320 is further configured to calculate the marginal contribution of each historical algorithm to the historical task according to the Shapley value algorithm if the second similarity index is less than the second preset threshold, and use the marginal contribution as the weight of the corresponding historical algorithm.

[0154] The algorithm determination module 330 is also used to perform weighted fusion of each historical algorithm according to the weight, generate a fusion algorithm, and determine the fusion algorithm as the target historical algorithm.

[0155] In one possible implementation, the data processing module 340 is also used to obtain multiple business indicators based on the processing results.

[0156] The data processing module 340 is also used to construct an indicator association network based on business indicators. The indicator association network includes multiple nodes and connecting edges, where each node corresponds to a business indicator, and each connecting edge indicates that there is a correlation, causal relationship, or attribution relationship between the indicators corresponding to two nodes.

[0157] The data processing module 340 is also used to display the indicator association network for users to view, and in response to the user's selection operation of any node, dynamically highlight other nodes associated with the user's selected node and their corresponding association strength.

[0158] This embodiment provides a multi-dimensional data processing device that can execute a multi-dimensional data processing method of the above embodiment. Its implementation principle and technical effect are similar, and will not be described again here.

[0159] In the specific implementation of the aforementioned multi-dimensional data processing device, each module can be implemented as a processor, and the processor can execute computer execution instructions stored in the memory, so that the processor executes the aforementioned multi-dimensional data processing method.

[0160] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4 As shown, the electronic device includes at least one processor 410 and a memory 420. The electronic device also includes a communication component 430. The processor 410, memory 420, and communication component 430 are connected via a bus 440.

[0161] In the specific implementation process, at least one processor 410 executes computer execution instructions stored in memory 420, causing at least one processor 410 to execute a multi-dimensional data processing method as executed on the electronic device side.

[0162] The specific implementation process of processor 410 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0163] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0164] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage.

[0165] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0166] The above description of the functions implemented by electronic devices and main control devices has introduced the solutions provided by the embodiments of the present invention. It is understood that, in order to implement the above functions, the electronic device or main control device includes hardware structures and / or software modules corresponding to the execution of each function. By combining the units and algorithm steps of the various examples described in the embodiments of the present invention, the embodiments of the present invention can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the technical solutions of the embodiments of the present invention.

[0167] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the multi-dimensional data processing method described above.

[0168] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0169] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in an electronic device or a host device.

[0170] This application also provides a computer program product, which includes a computer program stored in a readable storage medium. At least one processor of an electronic device can read the computer program from the readable storage medium, and the at least one processor executes the computer program to cause the electronic device to perform the solution provided in the above embodiments.

[0171] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0172] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it is readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. The above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A multi-dimensional data processing method, characterized in that, include: Acquire multi-dimensional data to be processed and a historical data set, wherein the historical data set includes multiple historical data sets, and each historical data set is associated with a historical algorithm; Calculate the similarity index between the multi-dimensional data and each historical data in the historical data set; From the similarity indices corresponding to each of the historical data, the target similarity index is determined; The historical algorithms associated with the historical data corresponding to the target similarity index are determined as the target historical algorithms; The multi-dimensional data is processed according to the target history algorithm to obtain the processing result.

2. The method according to claim 1, characterized in that, The calculation of similarity indicators between the multi-dimensional data and each historical data in the historical data set includes: Calculate the Pearson correlation coefficient between the multidimensional data and each historical data in the historical data set; Each of the Pearson correlation coefficients is used as the similarity index corresponding to each of the historical data.

3. The method according to claim 2, characterized in that, The step of determining the target similarity index from the similarity indices corresponding to each of the historical data includes: From the Pearson correlation coefficients corresponding to each of the historical data, the Pearson correlation coefficient with the largest value is determined as the first similarity index; If the first similarity index is greater than or equal to the first preset threshold, then the first similarity index is determined as the target similarity index.

4. The method according to claim 3, characterized in that, The calculation of the similarity index between the multi-dimensional data and each historical data in the historical data set further includes: If the first similarity index is less than the first preset threshold, then calculate the Bayesian factor between the multi-dimensional data and each historical data in the historical data set; Each of the Bayesian factors is used as the similarity index corresponding to each of the historical data. The step of determining the target similarity index from the similarity indices corresponding to each of the historical data further includes: From the Bayesian factors corresponding to each of the historical data, the Bayesian factor with the largest value is determined as the second similarity index; If the second similarity index is greater than or equal to the second preset threshold, then the second similarity index is determined as the target similarity index.

5. The method according to claim 4, characterized in that, Also includes: If the second similarity index is less than the second preset threshold, the marginal contribution of each historical algorithm to the historical task is calculated according to the Shapley value algorithm, and the marginal contribution is used as the weight of the corresponding historical algorithm. The historical algorithms are weighted and fused according to the weights to generate a fused algorithm, and the fused algorithm is determined as the target historical algorithm.

6. The method according to any one of claims 1 to 5, characterized in that, After processing the multi-dimensional data according to the target history algorithm to obtain the processing result, the method further includes: Based on the processing results, multiple business metrics are obtained; A metric association network is constructed based on the business metrics. The metric association network includes multiple nodes and connecting edges, where each node corresponds to a business metric, and each connecting edge indicates that there is a correlation, causal relationship, or attribution relationship between the metrics corresponding to two nodes. The network of the aforementioned indicators is displayed for users to view, and in response to the user's selection of any node, other nodes associated with the user's selected node and their corresponding association strengths are dynamically highlighted.

7. A multi-dimensional data processing device, characterized in that, include: The acquisition module is used to acquire multi-dimensional data to be processed and a historical data set, wherein the historical data set includes multiple historical data, and each historical data is associated with a historical algorithm. The calculation module is used to calculate the similarity index between the multi-dimensional data and each historical data in the historical data set; The calculation module is further configured to determine the target similarity index from the similarity indices corresponding to each of the historical data. The algorithm determination module is used to determine the historical algorithms associated with the historical data corresponding to the target similarity index as the target historical algorithm; The data processing module is used to process the multi-dimensional data according to the target history algorithm to obtain the processing result.

8. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1 to 6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method as described in any one of claims 1 to 6.

Images