Data large-screen visual method, system, device and storage medium

Abstract

The application relates to the technical field of artificial intelligence, and specifically provides a data large-screen visualization method, system, device and storage medium, which comprises the following steps: a mixed density network is used to capture field distribution characteristics of to-be-displayed data; a fuzzy type membership degree calculation method is used to determine distribution type membership degrees corresponding to the field distribution characteristics; the contribution degrees of the distribution type membership degrees to display modes are calculated, and type weights are determined based on the contribution degrees; decision values are calculated according to the distribution type membership degrees, the contribution degrees and the type weights, and an optimal display mode is determined according to the decision values; and the display mode comprises a statistical method and a chart type. The application can automatically identify an internal mode of data and match an optimal visualization scheme, and solves the problems of poor adaptability and one-sided display effect of a traditional method.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of artificial intelligence, and particularly relates to a data large-screen visualization method, system, device and storage medium. BACKGROUND

[0002] The existing large-screen visualization technology mainly relies on artificial preset rules or simple statistical analysis to select a chart type, for example, directly mapping a column chart, a line chart or the like based on field name matching or basic statistical quantities (such as mean value and variance). Such a method has two defects: one is that the complex characteristics (such as multi-peak and skewness) of data distribution are ignored, resulting in that the chart cannot accurately reflect the nature of the data; the other is that the rule configuration is rigid and difficult to adapt to mixed data (such as fields containing normal distribution and power-law distribution at the same time). SUMMARY

[0003] In view of the above defects of the prior art, the application provides a data large-screen visualization method, system, device and storage medium to solve the above technical problems.

[0004] In a first aspect, the application provides a data large-screen visualization method, comprising:

[0005] capturing a field distribution feature of to-be-displayed data by using a mixed density network;

[0006] determining a distribution type membership degree corresponding to the field distribution feature by using a fuzzy type membership degree calculation method;

[0007] calculating a contribution degree of the distribution type membership degree to a display mode, and determining a type weight based on the contribution degree;

[0008] calculating a decision value according to the distribution type membership degree, the contribution degree and the type weight, and determining a best display mode according to the decision value;

[0009] The display mode includes a statistical method and a chart type.

[0010] In an optional implementation, the capturing of the field distribution feature of the to-be-displayed data by using the mixed density network comprises:

[0011] inputting an original data field of the to-be-displayed data into the mixed density network to obtain a Gaussian mixture distribution field distribution feature, the field distribution feature including a mean value, a standard deviation and a mixing coefficient;

[0012] The mixed density network includes 5 layers of residual networks, and each layer of residual network contains 128 GRU units.

[0013] In an optional implementation, the determining of the distribution type membership degree corresponding to the field distribution feature by using the fuzzy type membership degree calculation method comprises:

[0014] The membership degree calculation formula of the to-be-displayed data and the distribution type t is:

[0015]

[0016] K is the total number of the to-be-displayed data decomposed into Gaussian components; is the mean value of the kth Gaussian component; is the standard deviation of the kth Gaussian component; is the mixing coefficient of the kth Gaussian component; is the ideal mean value of the tth distribution type, which is a fixed value set according to expert experience; is the semantic vector of the field corresponding to the kth Gaussian component is the similarity between the semantic vector of the field corresponding to the kth Gaussian component and the standard semantic vector of the tth distribution type; is the similarity between the semantic vector of the field corresponding to the kth Gaussian component and the standard semantic vector of the tth distribution type; is the dynamic adjustment factor.

[0017] In an optional implementation, the contribution of the distribution type membership degree to the display mode is calculated, and the type weight is determined based on the contribution, including:

[0018] The membership degree of the to-be-displayed data to each distribution type is taken as a feature, and the display mode is taken as a label, and the contribution of the membership degree to the display mode is calculated;

[0019] The contribution is mapped to the type weight.

[0020] In an optional implementation, the mapping of the contribution to the type weight includes:

[0021]

[0022] wherein, r m ∈{0,1} indicates whether the display mode m is selected or not; and a indicates an adjustment intensity factor; indicates the new weight of the tth distribution type, reflecting its importance; M indicates the total number of display modes; and t,m indicates the original contribution of the tth distribution type to the mth display mode; and t,m indicates the new contribution of the tth distribution type to the mth display mode.

[0023] In an optional implementation, the method further includes:

[0024] The dynamic adjustment factor is updated using the type weight:

[0025]

[0026] wherein, γ is a learning rate, and ε is a prevention zero constant.

[0027] In an optional embodiment, the decision value is calculated according to the distribution type membership degree, the contribution degree and the type weight, and the optimal display mode is determined according to the decision value, including:

[0028]

[0029] wherein DVm is a decision value of selecting the mth display mode, m is a decision value of selecting the mth display mode, is a membership degree of the data to be displayed and the tth distribution type, is a weight of the tth distribution type, is a contribution degree of the tth distribution type to the mth display mode; is a bias term;

[0030] The maximum decision value is screened out, and the display mode corresponding to the maximum decision value is output as the optimal display mode.

[0031] In a second aspect, the present application provides a data large-screen visualization system, including:

[0032] A feature extraction module is configured to capture field distribution features of the data to be displayed by using a hybrid density network.

[0033] A type determination module is configured to determine distribution type membership degrees corresponding to the field distribution features by using a fuzzy type membership degree calculation method.

[0034] A mode matching module is configured to calculate contribution degrees of the distribution type membership degrees to the display modes, and determine type weights based on the contribution degrees.

[0035] A mode determination module is configured to calculate decision values according to the distribution type membership degrees, the contribution degrees and the type weights, and determine an optimal display mode according to the decision values.

[0036] The display modes include statistical methods and chart types.

[0037] In a third aspect, a device is provided, including:

[0038] A memory is configured to store a data large-screen visualization program.

[0039] A processor is configured to implement the steps of the data large-screen visualization method provided in the first aspect when the data large-screen visualization program is executed.

[0040] In a fourth aspect, a computer readable storage medium is provided, and the storage medium stores a data large-screen visualization program. The data large-screen visualization program is executed by a processor to implement the steps of the data large-screen visualization method provided in the first aspect.

[0041] The data large screen visualization method, system, device and storage medium provided by the application can automatically identify the internal mode of data and match the optimal visualization scheme by dynamically capturing distribution characteristics through a mixed density network, combining fuzzy type membership degree calculation and contribution degree weighted decision, and solving the problems of poor adaptability and one-sided display effect of traditional methods.

[0042] In addition, the application has reliable design principles, simple structure and very wide application prospects. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, for those of ordinary skill in the art, other drawings can also be obtained without creative labor based on these drawings.

[0044] Figure 1 is a schematic flow chart of the method of one embodiment of the present application.

[0045] Figure 2 is a schematic block diagram of the system of one embodiment of the present application.

[0046] Figure 3 is a structural schematic diagram of a device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0047] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application.

[0049] The data large screen visualization method provided by the embodiments of the present application is executed by a computer device, and accordingly, the data large screen visualization system runs in the computer device.

[0050] Figure 1 is a schematic flow chart of the method of one embodiment of the present application. Wherein, Figure 1The execution subject can be a data large-screen visualization system. The order of steps in the flowchart can be changed according to different needs, and some steps can be omitted.

[0051] As shown in the figure, the method comprises: Figure 1

[0052] S1. Utilize the mixed density network to capture the field distribution characteristics of the data to be displayed;

[0053] S2. Determine the distribution type membership degree corresponding to the field distribution characteristics by a fuzzy type membership degree calculation method;

[0054] S3. Calculate the contribution degree of the distribution type membership degree to the display mode, and determine the type weight based on the contribution degree;

[0055] S4. Calculate the decision value according to the distribution type membership degree, the contribution degree and the type weight, and determine the best display mode according to the decision value; the display mode includes a statistical method and a chart type.

[0056] In an embodiment of the present application, based on step S1, a possible embodiment will be given below to specifically and non-limitingly illustrate the specific implementation scheme.

[0057] The original data field of the data to be displayed is input into the mixed density network to obtain the field distribution characteristics of the Gaussian mixture distribution, and the field distribution characteristics include the mean, the standard deviation and the mixing coefficient; the mixed density network includes 5 layers of residual networks, and each layer of residual network contains 128 GRU units.

[0058] The original data field is modeled by a mixed density network (MDN) to obtain the field distribution characteristics of the Gaussian mixture distribution. Specifically, the input original data field of the data to be displayed is first preprocessed, including standardization, missing value filling and the like, to ensure the standardization and availability of the data. The preprocessed data is input into the mixed density network, and the parameters of the Gaussian mixture distribution, i.e. the mean μ, the standard deviation σ and the mixing coefficient π, are finally output through layer-by-layer calculation of the internal structure of the network.

[0059] ​The mixed density network used in the application is based on a deep learning architecture, the core of which is a 5-layer residual network. The introduction of the residual network effectively alleviates the problems of gradient disappearance and gradient explosion in the training process of the deep neural network, so that the network can learn more complex feature representations. Each layer of the residual network contains 128 gated recurrent units (GRU). As a special recurrent neural network (RNN) unit, GRU can selectively remember and forget information through the gating mechanism, and can more effectively handle long-term dependencies in sequence data compared to traditional RNN units. In the present model, the GRU unit can capture the potential time series features or dependencies between data in the original data field.

[0060] In the network computing process, the input data first enters the first layer residual network. In each GRU unit, the data is calculated through the reset gate r t and the update gate z t .

[0061] r t =σ(W xr x t +W hr h t−1 +b r )

[0062] z t =σ(W xz x t +W hz h t−1 +b z )

[0063] Where x t is the input data at time t, h t−1 is the hidden state at the previous time, W xr , W hr , W xz , and W hz are weight matrices, b r and b z are bias terms, and σ is a Sigmoid activation function. Through the reset gate and the update gate, the GRU unit can dynamically determine the information to be retained and updated.

[0064] The data processed by the first layer residual network is transmitted to the next layer residual network through a residual connection. The calculation formula of the residual connection is:

[0065] h t =f(x t )+x t

[0066] wherein f(x t ) is the calculation result of the remaining part of the residual network except the identity mapping, and the structure makes the network easier to optimize the objective function in the learning process, avoiding the performance degradation problem caused by too deep network.

[0067] After layer-by-layer calculation through the 5-layer residual network, the feature vector output by the network is processed through linear transformation and activation function to obtain the parameters of the Gaussian mixture distribution. Assuming that the number of mixed components is K, for each data point, the probability density function of the Gaussian mixture distribution can be expressed as:

[0068]

[0069] wherein is the mixing coefficient of the kth mixed component, is the Gaussian distribution probability density function with the mean value and the standard deviation .

[0070] The training process of the network adopts the maximum likelihood estimation method, and the network parameters are optimized by minimizing the negative log-likelihood loss function.

[0071] In an embodiment of the present application, based on step S2, a possible embodiment will be given below to illustrate the specific implementation of the embodiment.

[0072] The membership degree calculation formula of the to-be-displayed data and the distribution type t is:

[0073]

[0074] wherein K is the total number of Gaussian components into which the to-be-displayed data is decomposed; is the mean value of the kth Gaussian component; is the standard deviation of the kth Gaussian component; is the mixing coefficient of the kth Gaussian component; is the ideal mean value of the tth distribution type, which is a fixed value set according to expert experience; is the similarity between the semantic vector of the field corresponding to the kth Gaussian component and the standard semantic vector of the tth distribution type ; is a dynamic adjustment factor.

[0075] The calculation method of the semantic similarity includes: encoding the to-be-displayed data by using an encoder to obtain a semantic vector, and then calculating the cosine similarity between the semantic vector of the to-be-displayed data and the standard semantic vector.

[0076] In one embodiment of the present application, based on step S3, a possible embodiment will be given below to specifically illustrate the non-limiting embodiment.

[0077] In the data processing and display mode optimization process, in order to effectively mine the correlation between data characteristics and display modes, this paper proposes an analysis method based on membership contribution calculation and type weight mapping. This method quantifies the influence of the membership of the data to be displayed and the distribution type on the display mode, providing a scientific basis for the selection of subsequent data visualization strategies.

[0078] S301. The membership of the data to be displayed and each distribution type is taken as a feature, and the display mode is taken as a label, and the contribution of the membership to the display mode is calculated.

[0079]

[0080] The feature SHAP value is a scalar value that quantifies the contribution of a specific feature i to the model prediction result. This value is obtained by an improved causal intervention calculation method (different from the traditional conditional expectation method), which can more accurately explain the influence of the feature. The technical core lies in generating feature subset combinations using a dynamic sampling strategy, and calculating the prediction expectation value through Monte Carlo simulation, thereby improving the credibility and stability of the explanation.

[0081] S (feature subset) represents the feature combination involved in the current analysis, and its generation process is based on the power set traversal of the feature set F. The patent innovation lies in optimizing the traversal efficiency - by using a dynamic sampling strategy to preferentially select feature combinations with high probability of association, rather than exhaustively enumerating all possibilities. This strategy significantly reduces the computational complexity, especially suitable for scenarios involving joint analysis of multi-modal features (such as text, time series, and statistical features).

[0082] F (feature set) represents all input features defined in the system, including structured data field definitions, unstructured data parsing results, and multi-modal feature fusion outputs. The patent design places special emphasis on the unified processing capability of heterogeneous features (such as numerical statistics, semantic embedding vectors, and time series pattern encoding), allowing different feature types to participate in decision-making collaboratively.

[0083] Factorial operation (S!), the factorial operation is used here to calculate the combination weight coefficient of the feature subset S, and its theoretical basis comes from the Shapley value allocation principle in cooperative game theory. By the factorial ratio of |S|! and (|F|-|S|-1)!, the system can reasonably quantify the contribution weight of different feature combinations, ensuring the fairness of mathematical distribution. This mechanism is a mathematical guarantee for the traceability of explanatory AI.

[0084] f x(S) (Model Predicted Expectation Function), which outputs the model's predicted expectation value for the feature subset S. The calculation is achieved through Monte Carlo simulation. The core innovation of the patent is the replacement of traditional conditional expectation with causal intervention (do-calculus), which avoids the interference of false correlations between features. For example, when analyzing the impact of "user age" on purchasing behavior, the system will forcibly set the age value (do operation) rather than simply calculating the conditional probability of age distribution, thus more accurately capturing causal relationships.

[0085] Feature Subset Sampling Strategy:

[0086]

[0087] λ (Temperature Parameter), a positive scalar parameter, controls the concentration of the sampling distribution. The default empirical value is 0.7, which is optimized on the validation set through grid search. Its core role is to balance the relationship between exploration (extensive sampling) and utilization (focus on high probability areas). When λ increases, the sampling distribution tends to be uniform, enhancing the exploration of rare feature combinations; when λ decreases, the sampling concentrates in high probability areas, improving computational efficiency.

[0088] q(S) (Current Model Predicted Distribution), represents the predicted probability distribution of the current model for feature subset S, which is directly obtained from the Softmax value output by the model. For example, in a classification task, if the model's prediction probability for a certain subset S is 0.8, then q(S)=0.8. This parameter dynamically reflects the real-time confidence of the model on feature combinations and is the basis for calculating sampling weights.

[0089] p base Baseline Distribution, a pre-defined baseline distribution, is usually based on the frequency of feature combinations in the training set, and is processed by Laplace smoothing to avoid zero probability problems. For example, if a certain feature combination appears 100 times in the training set with a total sample size of 100,000, its baseline probability is (100+1) / (100,000+total number of feature combinations). The baseline distribution provides stability and prevents sampling bias caused by short-term fluctuations in the model.

[0090] KL Divergence (D_KL(q||p_base)), used to measure the difference between the current model prediction distribution q(S) and the baseline distribution p_base, the calculation formula is:

[0091]

[0092] A larger KL value indicates that the current model behavior deviates significantly from the historical pattern, which may indicate data distribution drift or model overfitting.

[0093] Dynamic Adjustment Mechanism:

[0094] When data distribution drift is detected (JS divergence > 0.3):

[0095]

[0096] User feedback driven adjustment:

[0097]

[0098] Sampling probability of feature subset S is calculated by:

[0099] Feature subset sampling strategy achieves the following functions:

[0100] KL penalty: The greater the difference between model prediction distribution q(S) and baseline distribution, the lower the sampling probability, suppressing abnormal fluctuations.

[0101] Temperature control: λ adjustment adjusts the sensitivity to differences, the higher λ allows greater distribution deviation.

[0102] Model prediction difference calculation:

[0103]

[0104] f x (S), represents the function output, its technical meaning is: represents the model's prediction expectation value for the current data sample given the feature subset S. Data source: calculated by Monte Carlo simulation method, that is, by randomly sampling feature values outside the feature subset S multiple times, simulating the model's prediction results under different scenarios, and taking the average value.

[0105] Causal intervention (do-calculus) is used instead of traditional conditional expectation calculation. Traditional methods only calculate conditional probability based on existing data distribution, while causal intervention eliminates false correlations between features by actively intervening in feature values (such as forcing a certain feature to be a specific value), thus more truly reflecting the causal influence of features on prediction results.

[0106] f x (S) calculation method:

[0107] Monte Carlo simulation is a numerical calculation method based on random sampling. In calculating f x (S), the specific steps are as follows:

[0108] Fix feature subset S: Keep the feature values in S unchanged.

[0109] Randomly perturb non-S features: Randomly sample features that do not belong to S (i.e. F∖S), replacing them with other sample values in the training set or random values conforming to the distribution.

[0110] Multiple predictions and averaging: repeat the above perturbation process N times (e.g. N = 1000) and obtain the prediction results y of the model each time i Finally, calculate the average value.

[0111] Traditional conditional expectation E[y|X S =x S ] may be biased due to the correlation between features (e.g. confounding factors). For example, if features A and B are correlated, the traditional method cannot distinguish the direct causal effect of A on the result and the indirect effect through B. Causal intervention measures the independent influence of S by forcing X S =x S (e.g. do(X S =x S ), cutting off the dependence of S on other features.

[0112] S302. Map the contribution degree to a type weight.

[0113]

[0114] wherein r m ∈{0, 1} represents whether the display mode m is selected or not; α represents an adjustment intensity factor; represents the new weight of the tth distribution type, reflecting its importance; M represents the total number of display modes; ϕ t,m represents the original contribution degree of the tth distribution type to the mth display mode; ϕ t,m represents the new contribution degree of the tth distribution type to the mth display mode.

[0115] S303. Adjust the membership degree calculation formula using the type weight.

[0116] Update the dynamic adjustment factor using the type weight:

[0117]

[0118] wherein γ is a learning rate and ϵ is a prevention zero constant.

[0119] In an embodiment of the present application, based on step S4, a possible embodiment will be given below to non-restrictively illustrate the specific implementation scheme thereof.

[0120]

[0121] wherein DV m is a decision value for selecting the mth display mode, is the membership degree of the data to be displayed to the tth distribution type, is the weight of the tth distribution type, Contribution of the t-th distribution type to the m-th display mode; Bias term.

[0122] where type weight w t As a global decision regulator, global dimension w t The overall importance weight of type t (such as continuous type, categorical type, time series type) in the current system environment. It is used as a global regulatory parameter to determine the priority order of different types in the decision-making process. For example, when the system detects that the data has strong time series characteristics, the time series type weight wtime will significantly increase, making time-related statistical methods (such as moving average) have higher selection priority.

[0123] In decision value calculation, w t As a multiplication factor to scale the type influence, its value range is [0,1], and the higher the weight, the stronger the guiding role of the type on the final decision.

[0124] SHAP contribution value ϕ t,m As a local causal explainer, local dimension ϕ t,m Quantifies the causal contribution of type t to specific display mode m (such as the combination of sum, mean, quantile, and chart type). Its core role is to explain why a certain type of feature tends to choose a specific method under the current data distribution. For example, a high contribution value of the categorical type weight to "frequency statistics" (ϕcategory,frequency=0.9) indicates that the discrete characteristics of the categorical field naturally adapt to count analysis.

[0125] ϕt,m∈[−1,1], positive values indicate that type t has a positive causal impact on display mode m, and negative values indicate inhibition.

[0126] Type weight w t SHAP contribution value ϕ t,m Respectively from the system-level environment perception and data-level causal reasoning two dimensions to drive decision-making, and the product of the two fusion designs solves the contradiction between global rigidity and local overfitting in traditional methods.

[0127] Filter out the maximum decision value, and output the display mode corresponding to the maximum decision value as the best display mode.

[0128] In some embodiments, the data large-screen visualization system can include a plurality of functional modules composed of computer program segments. The computer programs of each program segment in the data large-screen visualization system can be stored in the memory of the computer device and executed by at least one processor to perform the functions of data large-screen visualization (see Figure 1 Description).

[0129] In this embodiment, the data large-screen visualization system can be divided into a plurality of functional modules according to the functions performed by the data large-screen visualization system, as shown in the figure. The module referred to in the present application refers to a series of computer program segments capable of being executed by at least one processor and capable of completing a fixed function, which are stored in the memory. In this embodiment, the functions of the modules will be described in detail in subsequent embodiments. Figure 2

[0130] The feature extraction module is configured to capture field distribution features of the data to be displayed by using a hybrid density network.

[0131] The type determination module is configured to determine distribution type membership degrees corresponding to the field distribution features by using a fuzzy type membership degree calculation method.

[0132] The pattern matching module is configured to calculate contribution degrees of the distribution type membership degrees to the display modes, and determine type weights based on the contribution degrees.

[0133] The mode determination module is configured to calculate a decision value according to the distribution type membership degrees, the contribution degrees and the type weights, and determine an optimal display mode according to the decision value.

[0134] The display modes include statistical methods and chart types.

[0135] Figure 3 The data large-screen visualization method provided by the embodiments of the present application can be applied to a device. Those skilled in the art can understand that the device structure involved in the embodiments of the present application does not constitute a limitation on the device, and the device can include more or fewer components than the illustration, or combine certain components, or different component arrangements. In the embodiments of the present application, the device includes but is not limited to a laptop computer, a desktop computer, a workstation, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described herein and / or claimed.

[0136] The device 300 can include a processor 310, a memory 320, and a communication unit 330. These components communicate through one or more buses, and those skilled in the art can understand that the structure of the server shown in the figure does not constitute a limitation on the present application, and it can be a bus structure or a star structure, and can include more or fewer components than the illustration, or combine certain components, or different component arrangements.

[0137] ​The memory 320 can be used to store the execution instructions of the processor 310, and the memory 320 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. When the execution instructions in the memory 320 are executed by the processor 310, the device 300 can perform part or all of the steps in the following method embodiments.

[0138] The processor 310 is the control center of the storage device, connects various parts of the entire electronic device through various interfaces and lines, and performs various functions of the electronic device and / or processes data by running or executing software programs and / or modules stored in the memory 320 and calling data stored in the memory. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of multiple packaged ICs with the same function or different functions connected. For example, the processor 310 can only include a central processing unit (CPU). In the embodiments of the present application, the CPU can be a single operation core or can include multiple operation cores.

[0139] The communication unit 330 is used to establish a communication channel, so that the storage device can communicate with other devices. Receive user data sent by other devices or send user data to other devices.

[0140] The present application also provides a computer storage medium, wherein the computer storage medium can store a program, and the program can include part or all of the steps in the embodiments provided by the present application when executed. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0141] Those skilled in the art can clearly understand that the technical solutions in the embodiments of the present application can be implemented by means of software plus necessary universal hardware platforms. Based on such an understanding, the technical solutions in the embodiments of the present application can be embodied in the form of a software product, which can be stored in a storage medium such as a USB flash disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and the like, and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a second device, a network device, or the like) to execute all or part of the steps of the methods described in the embodiments of the present application.

[0142] In the present specification, the same or similar parts among various embodiments can be referred to each other. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

[0143] In the several embodiments provided in the present application, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiments described above are merely schematic. For example, the division of the modules is merely a logical function division. In actual implementation, another division manner can be used. For example, a plurality of modules or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed modules can be indirect coupling or communication connection through some interface, system or module, and can be electrical, mechanical or other forms.

[0144] The modules described as separate components can or can not be physically separate, and the components displayed as modules can or can not be physical modules, i.e., can be located in one place or distributed on a plurality of network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments.

[0145] In addition, each functional module in the various embodiments of the present application can be integrated in one processing module, or each module can exist physically independently, or two or more modules can be integrated in one module.

[0146] Although the present application has been described in detail with reference to the preferred embodiments, it should be understood that the application is not limited to those preferred embodiments. Various modifications and equivalents can be made by those skilled in the art without departing from the spirit and scope of the application. Any and all modifications and equivalents are intended to be included within the scope of the present application.

Claims

1. A method for data dashboard visualization, the method comprising: The method comprises the following steps: capturing field distribution characteristics of the data to be displayed by using a hybrid density network; determining distribution type membership degrees corresponding to the field distribution characteristics by using a fuzzy type membership degree calculation method; calculating contribution degrees of the distribution type membership degrees to display modes, and determining type weights based on the contribution degrees; calculating decision values according to the distribution type membership degrees, the contribution degrees and the type weights, and determining an optimal display mode according to the decision values; the display modes comprise statistical methods and chart types; determining distribution type membership degrees corresponding to the field distribution characteristics by using a fuzzy type membership degree calculation method comprises the following steps: a membership degree calculation formula of the data to be displayed and a distribution type t is as follows: Where K is the total number of Gaussian components into which the data to be displayed is decomposed; is the average value of the kth Gaussian component; is the standard deviation of the kth Gaussian component; is the mixing coefficient of the kth Gaussian component; is the ideal mean of the t-th distribution type, which is a fixed value set based on expert experience; is the semantic vector of the field corresponding to the kth Gaussian component With the standard semantic vector of the t-th distribution type similarity; is a dynamic regulatory factor; calculating contribution degrees of the distribution type membership degrees to display modes, and determining type weights based on the contribution degrees comprises the following steps: taking the membership degrees of the data to be displayed and each distribution type as features, taking the display modes as labels, and calculating contribution degrees of the membership degrees to the display modes; mapping the contribution degrees to the type weights.

2. The method of claim 1, wherein, capturing field distribution characteristics of the data to be displayed by using a hybrid density network comprises the following steps: inputting original data fields of the data to be displayed into the hybrid density network to obtain field distribution characteristics of a Gaussian mixture distribution, wherein the field distribution characteristics comprise a mean value, a standard deviation and a mixture coefficient; the hybrid density network comprises five layers of residual networks, and each layer of residual network comprises 128 GRU units.

3. The method of claim 1, wherein, mapping the contribution degrees to the type weights comprises the following steps: Among them, r m ∈{0,1}, indicates whether the display mode m is selected; α indicates the adjustment intensity factor; represents the new weight of the t-th distribution type, reflecting its importance; M represents the total number of display modes; Indicates the contribution of the t-th distribution type to the m-th display mode; Indicates the The contribution of the distribution type to the mth display mode.

4. The method of claim 3, wherein, The method further comprises the following steps: updating the dynamic adjustment factor by using the type weights: where γ is a learning rate, to prevent a zero constant.

5. The method of claim 1, wherein, calculating decision values according to the distribution type membership degrees, the contribution degrees and the type weights, and determining an optimal display mode according to the decision values comprises the following steps: wherein DVm is a decision value for selecting the mth display mode, m wherein DVm is a decision value for selecting the mth display mode, wherein μt is the membership of the data to be displayed to the tth distribution type, wherein wt is the weight of the tth distribution type, wherein ctm is the contribution of the tth distribution type to the mth display mode; wherein b is a bias term. filtering out a maximum decision value, and outputting a display mode corresponding to the maximum decision value as the optimal display mode.

6. A data wall visualization system, characterized by, The method comprises the following steps: a feature extraction module is configured to capture field distribution characteristics of the data to be displayed by using a hybrid density network; a type determination module is configured to determine distribution type membership degrees corresponding to the field distribution characteristics by using a fuzzy type membership degree calculation method; a mode matching module is configured to calculate contribution degrees of the distribution type membership degrees to display modes, and determine type weights based on the contribution degrees; a mode determination module is configured to calculate decision values according to the distribution type membership degrees, the contribution degrees and the type weights, and determine an optimal display mode according to the decision values; the display modes comprise statistical methods and chart types; determining distribution type membership degrees corresponding to the field distribution characteristics by using a fuzzy type membership degree calculation method comprises the following steps: a membership degree calculation formula of the data to be displayed and a distribution type t is as follows: Where K is the total number of Gaussian components into which the data to be displayed is decomposed; is the average value of the kth Gaussian component; is the standard deviation of the kth Gaussian component; is the mixing coefficient of the kth Gaussian component; is the ideal mean of the t-th distribution type, which is a fixed value set based on expert experience; is the semantic vector of the field corresponding to the kth Gaussian component With the standard semantic vector of the t-th distribution type similarity; is a dynamic regulatory factor; calculating contribution degrees of the distribution type membership degrees to display modes, and determining type weights based on the contribution degrees comprises the following steps: taking the membership degrees of the data to be displayed and each distribution type as features, taking the display modes as labels, and calculating contribution degrees of the membership degrees to the display modes; mapping the contribution degrees to the type weights.

7. A data wall visualisation device, characterised in that, The method comprises the following steps: a memory is configured to store a data large-screen visualization program; a processor is configured to implement the steps of the data large-screen visualization method according to any one of claims 1-5 when executing the data large-screen visualization program.

8. A computer readable storage medium storing a computer program, characterized in that, The readable storage medium has stored thereon a data big screen visualization program, the data big screen visualization program, when executed by the processor, implements the steps of the data big screen visualization method according to any one of claims 1-5.

Images