Data display processing method and device, electronic equipment and storage medium

Abstract

The invention provides a data display processing method and device, electronic equipment and a storage medium, and relates to the technical field of data processing.The method comprises the steps that attribute text processing is conducted on target data, and word frequency features corresponding to the target data are obtained; inputting the word frequency features into a pre-trained hierarchical attribute classifier to obtain a hierarchical classification result of the target data, the hierarchical classification result at least comprising a plurality of hierarchical levels and a data association relationship of hierarchical attributes and inheritance of each hierarchical level for the target data; performing clustering processing on the target data layer by layer under a plurality of hierarchical levels based on a hierarchical classification result by using a time-space constrained spectral clustering algorithm model to obtain a clustering category under each hierarchical level; and taking the plurality of hierarchical levels as zoom levels, taking the clustering category as a legend, and constructing a data map corresponding to the target data, the data map being used for displaying the target data in a page in a zoom manner. According to the method, all data are displayed in one page according to hierarchical level zooming.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data display processing method, device, electronic device and storage medium. Background Art

[0002] In data visualization technology, data presentation is particularly important because it helps users quickly identify patterns and trends in the data.

[0003] For massive amounts of data, if all the data is displayed flatly, it will appear messy. Loading a large amount of data can easily cause the system to freeze. Layered display is complicated to operate and it is impossible to view the data globally. Summary of the invention

[0004] The present invention provides a data display processing method, device, electronic device and storage medium, which can realize the scaling and display of all target data in a page according to the hierarchical level, and can avoid the problems of jumping pages back and forth during layered display, inflexible operation and cluttered pages when the full amount of data is tiled.

[0005] In a first aspect, a data display processing method is provided, comprising: Perform attribute text processing on the target data to obtain the word frequency features corresponding to the target data; Input the word frequency feature into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification result of the target data, wherein the hierarchical classification result includes at least a plurality of hierarchical levels and hierarchical attributes of the target data and inherited data association relationships for each hierarchical level; Using the spectral clustering algorithm model with time and space constraints, the target data is clustered layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain the clustering categories at each hierarchical level; Using multiple hierarchical levels as zoom levels and cluster categories as legends, a data map corresponding to the target data is constructed. The data map is used to display the target data in a zoomed manner within a page.

[0006] In a second aspect, a data display processing device is provided, comprising: A processing module is used to perform attribute text processing on the target data to obtain word frequency features corresponding to the target data; A hierarchical module, used for inputting word frequency features into a pre-trained hierarchical attribute classifier to obtain a hierarchical classification result of the target data, wherein the hierarchical classification result includes at least a plurality of hierarchical levels and hierarchical attributes of the target data and inherited data association relationships for each hierarchical level; A clustering module is used to utilize a spectral clustering algorithm model with time and space constraints to cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; A construction module is used to use multiple hierarchical levels as zoom levels and cluster categories as legends to construct a data map corresponding to the target data. The data map is used to zoom and display the target data within a page.

[0007] According to a third aspect, an electronic device is provided, comprising: a processor and a memory, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method in the first aspect or its various implementations.

[0008] According to a fourth aspect, a computer-readable storage medium is provided for storing a computer program, wherein the computer program enables a computer to execute the method according to the first aspect or its various implementations.

[0009] Through the technical solution provided by the present invention, the word frequency features corresponding to the target data can be obtained by processing the target data with attribute text; then the word frequency features are input into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification results of the target data, and the hierarchical classification results include at least multiple hierarchical levels and hierarchical attributes and inherited data association relationships of the target data at each hierarchical level; further, the spectral clustering algorithm model with time and space constraints is used to cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; finally, multiple hierarchical levels are used as zoom levels, and clustering categories are used as legends to construct a data map corresponding to the target data, and the data map is used to zoom and display the target data on a page. The present application can realize the zoom display of all target data on a page according to the hierarchical levels, which can avoid the problems of jumping back and forth pages during hierarchical display, inflexible operation, and cluttered pages when the full amount of data is tiled.

[0010] It should be understood that the above general description and the following detailed description are only exemplary and explanatory and cannot limit the present invention. Other features and advantages of the present invention will be described in detail in the following specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 An application scenario diagram provided for an embodiment of the present application; Figure 2 A flowchart of a data display processing method provided in an embodiment of the present application; Figure 3 A schematic diagram of the updating principle of a spectral clustering algorithm model provided in an embodiment of the present application; Figure 4 A schematic diagram of a process for updating the weight of a clustering evaluation index provided in an embodiment of the present application; Figure 5 A schematic diagram of an example of a constructed data map provided in an embodiment of the present application; Figure 6 A flowchart of a data display processing method provided in another embodiment of the present application; Figure 7 An example diagram of data association inheritance provided in an embodiment of the present application: Figure 8 A schematic diagram of a principle flow of data display processing provided in an embodiment of the present application; Fig. 9 A schematic diagram of the structure of a data display processing device provided in an embodiment of the present application; Fig.10 A schematic diagram of the structure of a data display processing device provided in another embodiment of the present application; Fig.11 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0014] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0015] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0016] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0017] In data visualization technology, data presentation is particularly important because it helps users quickly identify patterns and trends in the data.

[0018] For massive monitoring data, if all the data is displayed flatly, it will appear messy. Loading a large amount of data will easily cause the system to freeze. Layered display will be complicated to operate and it will be impossible to view the data globally.

[0019] In order to solve the above technical problems, the invention of the present invention is to introduce the concept of data map, use automatic stratification to determine the zoom level, and superimpose a clustering algorithm that is more in line with the characteristics of monitoring data to perform data clustering in each layer. It can realize the viewing of all data on one page, simplify the operation process, and display data information more intuitively to avoid data clutter.

[0020] Because monitoring data does not have natural location information, distance information, direction information, and subordinate information like geographic map data; in order to map monitoring data and relationships to display on the map, it is necessary to construct the three elements of the map: scale, direction, and legend; so when constructing a massive monitoring data map, it is necessary to construct the scale (zoom level), direction (location), and legend (cluster category) of the data. Therefore, the design idea of ​​this solution is to construct the zoom level and cluster category of the data map and calibrate its display position: 1. Automatic stratification based on classifiers: First, use the classifier to determine the stratification level according to the data attributes, and use the specific stratification as the zoom level of the data map (similar to the scale in the map).

[0021] 2. Cluster the data in each level based on the spectral clustering algorithm as the legend of the data map (province, city, district, etc.).

[0022] 3. Considering the dependence of monitoring data on time and call span, constrain the spectral clustering algorithm and optimize the clustering algorithm to achieve more accurate clustering (more accurate location to avoid cities in province A being classified into province B).

[0023] Accordingly, the beneficial technical effects that this solution can bring are: 1. Improved user experience: Through a map-like data display method, all data can be viewed on one page, avoiding the problem of jumping back and forth between pages and inflexible operation when displaying in layers; avoiding the problem of cluttered pages when displaying all data in a flat format.

[0024] 2. Performance improvement: Through automatic stratification and automatic clustering, the map scale is introduced to scale data, reduce the amount of displayed data, and avoid interface delays and page freezes caused by loading a large amount of data at one time.

[0025] 3. Reduced work costs: Automatic classification and clustering are performed based on models trained based on historical data, which reduces the steps of manual configuration and reduces work costs.

[0026] 4. Improved accuracy: Different algorithms are used to obtain the three elements of data map construction in the process to improve the accuracy of data display. The algorithm is optimized for monitoring data, and time constraints are introduced by using the real-time characteristics of monitoring data. Distance constraints are introduced by using the call relationship characteristics of monitoring data to optimize the spectral clustering algorithm. At the same time, the evaluation indicators are automatically converged to obtain the optimal parameter configuration to improve data clustering accuracy.

[0027] It should be understood that the technical solution of the present invention can be applied to the following scenarios, but is not limited to: In some possible implementations, Figure 1 An application scenario diagram provided by an embodiment of the present invention, such as Figure 1 As shown, the application scenario may include an electronic device 110 and a network device 120. The electronic device 110 may establish a connection with the network device 120 via a wired network or a wireless network.

[0028] Exemplarily, the electronic device 110 may be a desktop computer, a laptop computer, a tablet computer, etc., but is not limited thereto. The network device 120 may be a terminal device or a server, but is not limited thereto. In one embodiment of the present invention, the electronic device 110 may send a request message to the network device 120, and the request message may be used to request to obtain a data map corresponding to the target data. Further, the electronic device 110 may receive a response message sent by the network device 120, and the response message includes a data map corresponding to the target data.

[0029] also, Figure 1 An electronic device 110 and a network device 120 are provided as an example, but other numbers of electronic devices and network devices may be included in practice, and the present invention is not limited thereto.

[0030] In some other possible implementations, the technical solution of the present invention may also be executed by the above-mentioned electronic device 110, or the technical solution of the present invention may also be executed by the above-mentioned network device 120, and the present invention does not limit this.

[0031] After introducing the application scenarios of the embodiments of the present invention, the technical solutions of the present invention will be described in detail below: Figure 2 A flowchart of a data display processing method provided by an embodiment of the present invention, the method can be performed as follows Figure 1 The electronic device 110 shown in the figure performs, but is not limited to this. Figure 2 As shown, the method may include the following steps: Step 210: Perform attribute text processing on the target data to obtain word frequency features corresponding to the target data.

[0032] For the disclosed embodiments, when performing attribute text processing on target data, it can be first determined whether the target data is of text type. When the target data is of text type, the target data can be directly segmented to obtain multiple words, and word frequency features can be extracted based on the multiple words. When the target data is of a data type other than text type, the target data can be first converted into text data of text type, and then the text data can be segmented to obtain multiple words, and word frequency features can be extracted based on the multiple words.

[0033] When converting target data under other data types into text data under text type, exemplarily, when the other data type is image type, text extraction can be performed on target data of image type through text extraction technology to obtain text data; when the other information type is voice type, voice-to-text technology can be used to convert target data of voice type into text data of text type.

[0034] It should be noted that the target data in the present application may be data of any data type. In the following implementation steps of the present disclosure, the target data is monitoring data as an example to illustrate the technical solution in the present application, but it does not constitute a specific limitation.

[0035] Step 220: Input the word frequency feature into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification result of the target data.

[0036] Among them, the hierarchical attribute classifier is a classifier trained for hierarchical attribute classification tasks, which can be composed of a naive Bayes classification model. When training for hierarchical attribute classification tasks, the existing data set can be used to obtain different classification probabilities of new input data through supervised learning, and the highest probability is determined as the final hierarchical classification result. The hierarchical attribute classifier can be based on the word frequency characteristics of the target data, and can perform data stratification by analyzing the data attributes or the relationship between the data, confirm the preliminary aggregation attributes of the data (i.e., hierarchical attributes) and the data association relationship, and form the zoom level of the data map; thus, the hierarchical classification result includes at least multiple hierarchical levels and the hierarchical attributes and inherited data association relationships of each hierarchical level for the target data.

[0037] Step 230 : Using the spectral clustering algorithm model with time and space constraints, cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain cluster categories at each hierarchical level.

[0038] In specific application scenarios, according to the above design ideas, in the face of massive big data, after classification, the data in each layer may still be a lot, and direct display will appear messy. In this case, it is necessary to cluster the data within the layer. This article uses the optimized spectral clustering algorithm to aggregate the data within the layer. Considering that the different call spans of monitoring data are different, and the monitoring data is related to time (data and call requests at different times are different), in order to be more accurate in subsequent clustering, time and space constraints can be introduced to optimize the spectral clustering algorithm, and the optimal parameters are obtained by iterative convergence of clustering evaluation indicators, and finally the optimized spectral clustering algorithm model is obtained. Specifically, the time and space constraints of the spectral clustering algorithm can be realized by adjusting the weight values ​​of the time weight and the distance weight. Among them, in the process of adjusting the weight values ​​of the time weight and the distance weight, the sum of the time weight and the distance weight is always kept as 1.

[0039] Among them, the spectral clustering algorithm model may include a spectral clustering composition conversion module, a matrix construction module, a graph cutting module and a clustering module. The spectral clustering composition conversion module is used to perform graph data conversion on the target data layer by layer based on the hierarchical attributes and inherited data association relationships at each hierarchical level to obtain the target spectral clustering composition of the target data at the target hierarchical level; the matrix construction module is used to construct the target Laplacian matrix corresponding to the target spectral clustering composition based on the time and space constraint information; the graph cutting module is used to cut the spectral clustering composition into multiple non-overlapping target subgraphs based on the target Laplacian matrix using the graph cutting algorithm; the clustering module is used to perform clustering processing on the sub-target data in each target subgraph respectively to obtain the clustering category at the target hierarchical level.

[0040] In a specific application scenario, before executing the steps of this embodiment, the spectral clustering algorithm model may be trained and optimized in advance. Accordingly, the training and optimization process of the spectral clustering algorithm model may include: determining first time-space constraint information, the first time-space constraint information at least including a distance matrix, an initial adjacency matrix, a timestamp, a first time weight, and a first distance weight; iteratively executing the updating process of the first time-space constraint information until a preset iteration termination condition is reached to obtain final time-space constraint information, and constructing a spectral clustering algorithm model based on the final time-space constraint information; wherein the updating process of the first time-space constraint information includes: clustering the training data based on the first time-space constraint information using the initial spectral clustering algorithm model to obtain a first clustering evaluation index; adding a random offset to the first time weight and the first distance weight to obtain a second time weight and a second distance weight; clustering the training data based on the second time-space constraint information using the initial spectral clustering algorithm model to obtain a second clustering evaluation index, the second time-space constraint information at least including a distance matrix, an initial adjacency matrix, a timestamp, a second time weight, and a second distance weight; and judging whether to update the second time-space constraint information to the first time-space constraint information based on the first clustering evaluation index and the second clustering evaluation index.

[0041] Among them, Figure 3 As shown, the training data can be clustered based on the first time-space constraint information to obtain a first clustering evaluation index DB (Davies-Bouldin). Then, the first time weight in the first time-space constraint information can be and the first distance weight Add random perturbations (positive and / or negative perturbations) to obtain the second time weight and the second distance weight . Then, the second time weight can be included and the second distance weight The second time-space constraint information is used to cluster the training data to obtain the second clustering evaluation index . Then, based on the first clustering evaluation index DB and the second clustering evaluation index , determine whether to update the second time-space constraint information to the first time-space constraint information. The lower the DB value, the more compact the cluster and the better the separation effect. The first time weight and the first distance weight is the initial default value, such as 0.5. Repeat the above iterative update process of the first time and space constraint information until the preset iteration termination condition is reached and the optimal solution is obtained. and The final time and space constraint information. The preset iteration termination condition can be reaching the set number of iterations, or using the second time and space constraint information The updated first time space constraint information DB is less than a certain threshold .

[0042] It should be noted that the first clustering evaluation index and the second clustering evaluation index can also be set as other indicators for clustering effect evaluation. In the embodiment of the present disclosure, the clustering evaluation index DB is taken as an example to illustrate the technical solution in this application, but it does not constitute a specific limitation.

[0043] Accordingly, when judging whether to update the second time-space constraint information to the first time-space constraint information based on the first cluster evaluation index and the second cluster evaluation index, the embodiment steps may include: calculating the index difference between the second cluster evaluation index and the first cluster evaluation index; when judging that the index difference is less than or equal to the second preset threshold, using the second time-space constraint information to update the first time-space constraint information, and continuing to execute the update process of the first time-space constraint information; when judging that the index difference is greater than the second preset threshold, in the update process of the first time-space constraint information, keeping the first time-space constraint information unchanged. Among them, the value of the second preset threshold can be set according to the actual application, such as being set to 0.

[0044] Among them, Figure 4 As shown, based on the first clustering evaluation index DB and the second clustering evaluation index , calculate the index difference In judging When using the second time-space constraint information Update the first time-space constraint information DB, that is, accept the new solution On the contrary, in judging When the first time-space constraint information DB is kept unchanged.

[0045] Step 240: Use multiple hierarchical levels as zoom levels and cluster categories as legends to construct a data map corresponding to the target data. The data map is used to display the target data in a zoomed manner within a page.

[0046] For the disclosed embodiment, multiple hierarchical levels can be used as zoom levels, cluster categories can be used as legends, target data can be aggregated layer by layer, and the amount of upper layer data can be gradually reduced to ensure page display efficiency. Each time you zoom in and drill down a level, you can ensure that the target data can be viewed layer by layer on a page. The data map can be freely zoomed, such as Figure 5 As shown, by gradually zooming in on the data map, a more specific (localized) clustering division of the target data at the next higher hierarchical level can be displayed on the page; or by gradually zooming out on the data map, a more global clustering division of the target data at the previous lower hierarchical level can be displayed on the page.

[0047] In summary, according to the data display processing method provided by the present invention, the word frequency features corresponding to the target data can be obtained by performing attribute text processing on the target data; then the word frequency features are input into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification results of the target data, and the hierarchical classification results include at least multiple hierarchical levels and hierarchical attributes and inherited data association relationships for the target data at each hierarchical level; further, the spectral clustering algorithm model with time and space constraints is used to cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; finally, multiple hierarchical levels are used as zoom levels, and clustering categories are used as legends to construct a data map corresponding to the target data, and the data map is used to zoom and display the target data within a page. The present application can realize the zoom display of all target data according to the hierarchical levels within a page, which can avoid the problems of jumping back and forth pages during hierarchical display, inflexible operation, and cluttered pages when the full amount of data is tiled.

[0048] based on Figure 2 The embodiment shown is a refinement and extension of the above embodiment. In order to fully illustrate the specific implementation process of the method of this embodiment, this embodiment provides the following Figure 6 The specific method shown. Figure 6 based on Figure 2 As shown in Figure 6, the method comprises the following steps: Step 610: Perform attribute text processing on the target data to obtain word frequency features corresponding to the target data.

[0049] Accordingly, for the embodiment of the present disclosure, performing attribute text processing on the target data in step 610 to obtain word frequency features corresponding to the target data may include the following steps: Step 610 - 1 : segment the target data into first words based on word segmentation technology to obtain a first word sequence including a plurality of first words.

[0050] For the embodiments of the present disclosure, in a specific application scenario, an existing word segmentation tool (such as a CRF word segmenter) can be used to segment the target data in text form to obtain each independent first word and a first word sequence containing multiple first words after word segmentation, wherein each first word is marked with a corresponding target part of speech. Specifically, after obtaining the target data in text form, it is necessary to perform word segmentation on the target data and generate a word sequence, wherein each first word is used as an element in the word sequence and is arranged in sequence according to the position of appearance in the input text, and further obtains a first word sequence, wherein the format corresponding to the first word sequence can be: [word 1, word 2, word 3, ... word N].

[0051] Step 610 - 2 : According to the target part of speech of the first word, determine in the first word sequence a second word that matches the preset stop part of speech, remove the second word from the first word sequence, and obtain a second word sequence.

[0052] For the disclosed embodiments, in order to improve the efficiency of feature information extraction, after obtaining the first word marked with the target part of speech using the above method, the first word can be further identified, and some stop parts of speech in the first word sequence can be eliminated to further obtain a second word sequence containing only valid first words. Among them, the preset stop parts of speech can be modal particles, adverbs, prepositions, conjunctions, etc. These preset stop parts of speech usually have no clear meaning by themselves, and only have a certain effect when they are placed in a complete sentence, such as the common "ah", "ouch", "other", "result", "let", "but", etc. Since these words rarely express information about the relevance of documents alone, in order to improve the efficiency of data display processing and save storage space, these meaningless words can be filtered out in advance. Specifically, the identification and filtering of stop words can be achieved based on the existing stop word vocabulary.

[0053] Step 610 - 3 : Calculate the word frequency feature of the first word in the second word sequence based on the TF algorithm.

[0054] For the embodiment of the present disclosure, the preset calculation formula corresponding to the TF algorithm can be used to calculate the word frequency feature of the first word in the second word sequence. The feature description of the preset calculation formula is:

[0055] Where TF ij is the frequency feature of the first word i, ni,j is the number of times the first word i appears in the second word sequence dj, is the sum of the numbers of all first words in the second word sequence dj.

[0056] Step 620: Input the word frequency feature into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification result of the target data.

[0057] The hierarchical classification result includes at least a plurality of hierarchical levels and hierarchical attributes and inherited data association relationships of each hierarchical level for target data.

[0058] In a specific application scenario, before executing the steps of this embodiment, the hierarchical attribute classifier may be pre-trained. Accordingly, when pre-training the hierarchical attribute classifier, the embodiment steps may include: determining the sample word frequency features of the sample target data and the corresponding preset hierarchical classification labels; inputting the sample word frequency features and the preset hierarchical classification labels into the hierarchical attribute classifier, and training the hierarchical attribute classifier for the classification task of the hierarchical attributes; wherein, in the classification task training of the hierarchical attributes, the sample word frequency features are used as input features, and the preset hierarchical classification labels are used as training labels, and the hierarchical attribute classifier is iteratively trained until the loss function of the hierarchical attribute classifier is less than the first preset threshold, and it is judged that the training of the hierarchical attribute classifier is completed. wherein, the first preset threshold is a value between 0 and 1, and its specific value setting can be set according to the actual application scenario, and is not specifically limited here.

[0059] For the disclosed embodiment, a pre-trained hierarchical attribute classifier can be used to determine the hierarchical level according to the data attributes, and the target data can be hierarchized step by step according to the hierarchical attributes or relationships. For example, the hierarchical attribute classifier traverses the data attributes of the target data to obtain the hierarchical attributes: business (second layer) and app_cluster (first layer). After that, it can be aggregated according to the value of the app_cluster attribute to generate N groups of data (defined as app_cluster), each of which contains M original data. Then, according to the value of the business attribute, secondary aggregation is performed to generate J groups of data, each of which contains P app_cluster data. Finally, the target data is aggregated layer by layer to reduce the amount of data in each layer. At the same time, the aggregated data inherits the calling relationship (i.e., data association relationship) of the internal data to ensure the integrity of the data relationship. For example, if there is a calling relationship between sub-data a and b in the target data, and a is aggregated into app_cluster1, and b is aggregated into app_cluster2, then app_cluster1 and app_cluster2 retain the association relationship.

[0060] For example, Figure 7As shown, the underlying data corresponding to the target data includes: data A, data B, data C, data D, data E, data F, data G and data H, and there are data relationships between data A and data H, data B and data C, data D and data E, and data F and data G. In the first layer, data A and data B are aggregated into aggregated data 1, data C and data D are aggregated into aggregated data 2, data E and data F are aggregated into aggregated data 3, and data G and data H are aggregated into aggregated data 4. Then, the data relationship between data B and data C is inherited between aggregated data 1 and aggregated data 2, the data relationship between data D and data E is inherited between aggregated data 2 and aggregated data 3, the data relationship between data F and data G is inherited between aggregated data 3 and aggregated data 4, and the data relationship between data A and data H is inherited between aggregated data 1 and aggregated data 4. In the second layer, aggregated data 1 and aggregated data 2 are aggregated into aggregated data A, and aggregated data 3 and aggregated data 4 are aggregated into aggregated data B. Then, aggregated data A and aggregated data B can inherit the data relationship between aggregated data 1 and aggregated data 4.

[0061] Step 630: Determine each hierarchical level as a target hierarchical level in turn according to the order from low to high among the multiple hierarchical levels.

[0062] Exemplarily, when the hierarchical attribute classifier outputs target data including three hierarchical levels, the first level, the second level, and the third level can be determined as the target hierarchical levels in order from low to high among the multiple hierarchical levels, and subsequent embodiment steps 640 to 670 can be executed until each hierarchical level in the multiple hierarchical levels completes cluster division and obtains the corresponding cluster category.

[0063] Step 640: Input the target data and the hierarchical attributes and inherited data association relationships of the target hierarchical level into the spectral clustering graph conversion module for graph data conversion to obtain the target spectral clustering graph of the target data at the target hierarchical level.

[0064] The spectral clustering graph after graph data conversion is essentially a graph structure that connects objects in the original data set in a specific way (such as similarity or distance measurement). This graph structure is used to represent the relationship between objects in the data set and is the basis for the spectral clustering algorithm to perform clustering analysis. In the embodiment of the present disclosure, the spectral clustering graph corresponding to the target data can be defined as the target spectral clustering graph.

[0065] Specifically, the spectral clustering graph consists of vertices and edges, where each vertex in the graph represents an object in the target data. These objects can be any type of data points, such as images, texts, biological molecules, etc. The edges in the graph represent the similarity or connection relationship between vertices (i.e., data objects). The existence of an edge and the weight of the edge (if a weighted graph is considered) depend on the measurement results of the similarity or distance between vertices. In spectral clustering, a similarity matrix is ​​usually used to represent the similarity relationship between vertices.

[0066] Step 650: Use a matrix construction module to construct a target Laplacian matrix corresponding to the target spectral clustering composition based on the time-space constraint information.

[0067] Among them, the time-space constraint information at least includes the distance matrix, the initial adjacency matrix, the timestamp, the time weight and the distance weight; the Laplace matrix L is the core matrix in spectral clustering, and the elements in the matrix represent the similarity between the corresponding vertices in the graph. It is calculated by the degree matrix D and the similarity matrix W, that is, L=DW. The Laplace matrix has some important properties, such as symmetry, semi-positive definiteness, etc., which play a key role in the process of spectral clustering. For the embodiments of the present disclosure, the Laplace matrix corresponding to the target spectral clustering composition can be defined as the target Laplace matrix.

[0068] Accordingly, for the embodiment of the present disclosure, the use of the matrix construction module in step 650 to construct the target Laplacian matrix corresponding to the target spectral clustering composition based on the time-space constraint information may include the following steps: Step 650 - 1 : Generate a degree matrix corresponding to the target spectral clustering graph according to the connection edge information of each node in the target spectral clustering graph.

[0069] The degree matrix D is a diagonal matrix used to represent the degree of each vertex in the target spectral clustering graph. The degree of a vertex is defined as the sum of the weights of all edges connected to the vertex. The elements on the diagonal of the degree matrix D are the degrees of each vertex. The connection edge information can be the out-degree and in-degree of each node. The in-degree refers to the number of connection edges pointing to the node, and the out-degree refers to the number of connection edges from the node to other nodes.

[0070] Step 650 - 2 : Based on the distance matrix, timestamp, time weight and distance weight, update the initial adjacency matrix to obtain the target adjacency matrix.

[0071] The essence of clustering is to aggregate points with high similarity. Since most monitoring data (with call relationships, deployment relationships, and dependency relationships) have time dimensions and call hop counts (distance dimensions), it is easy to aggregate data at different times based on similar features of the data using traditional clustering algorithms. For example, at time A, the service aiops-admin called the MySQL cluster, and at time B, the service aiops-alarm called the MySQL cluster. Since both of these are Java services that called the same MySQL cluster and have similar names, they are easily aggregated together by traditional clustering algorithms. In fact, these two services have different functions and should belong to different categories, which requires the introduction of time weights. On the other hand, for example, if there are two data nodes that are both unknown services and are deployed on the same IP, the timestamps of the two service data are similar, but the links called are different. If clustering is only based on time constraints, then such data aggregation cannot be avoided. Therefore, the distance matrix S and distance weight can be introduced. Constrain the algorithm.

[0072] Taking the above situation into consideration, the distance matrix, time weight, timestamp and distance weight are introduced into the monitoring data to optimize the algorithm so that it can cluster the data more accurately. The main constraint is the adjacency matrix W. The optimized formula is as follows:

[0073] In the formula, is the target adjacency matrix, represents the distance weight, represents the time weight, W is the initial adjacency matrix, S is the distance matrix, Representative Node timestamp, Representative Node The timestamp of , e is the identity matrix, This is to avoid negative numbers when timestamps are subtracted.

[0074] Considering that the multiplication of weights, timestamps, and distance matrices will make the data in the result matrix too large, in order to avoid certain features having too much influence on the results, the target adjacency matrix can also be Normalize, that is, map the values ​​in the matrix to the range of [0, 1] to obtain the normalized target adjacency matrix .

[0075] Step 650 - 3 : Calculate the target Laplacian matrix corresponding to the target spectral clustering graph according to the degree matrix and the target adjacency matrix.

[0076] For the embodiment of the present disclosure, after the normalized target adjacency matrix is ​​calculated based on step 650-2 of the embodiment, the degree matrix D and the normalized target adjacency matrix can be used to obtain the target adjacency matrix. Calculate the target Laplacian matrix The calculation formula is as follows:

[0077] Step 660: In the graph cutting module, based on the target Laplacian matrix, a graph cutting algorithm is used to cut the spectral clustering graph into a plurality of non-overlapping target subgraphs.

[0078] For the embodiments of the present disclosure, for the above-mentioned spectral clustering graph, based on the target Laplacian matrix, a graph cutting algorithm can be used to cut the entire large graph into k subgraphs that are not connected to each other. The goal is to make the connection between subgraphs as small as possible and to have no intersection between subgraphs.

[0079] Step 670: Use a clustering module to perform clustering processing on the sub-target data in each target sub-graph to obtain cluster categories under the target hierarchical level.

[0080] Step 680: Use multiple hierarchical levels as zoom levels and cluster categories as legends to construct a data map corresponding to the target data. The data map is used to display the target data in a zoomed manner within a page.

[0081] For the embodiment of the present disclosure, the specific implementation process can refer to the relevant description in step 240 of the embodiment, which will not be repeated here.

[0082] In order to make this solution clearer, Figure 8, taking the target data as monitoring data as an example, the technical solution in this application is fully explained: after obtaining massive monitoring data, the three elements of the map can be constructed: scale, direction and legend. Specifically, the scale can be constructed by stratifying the monitoring data. When performing stratification processing, the monitoring data can first be processed by attribute text to obtain the word frequency characteristics corresponding to the monitoring data; then the hierarchical attribute classifier can be used to perform hierarchical attribute classification to obtain a hierarchical classification result. The hierarchical classification result includes multiple hierarchical levels and hierarchical attributes and inherited data association relationships for the target data at each hierarchical level. The hierarchical attributes and inherited data association relationships are used to aggregate the monitoring data separately at each hierarchical level. And, the construction of the legend can be achieved by aggregating the monitoring data separately at each hierarchical level. Specifically, the spectral clustering algorithm model with time and space constraints can be used to cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain the clustering categories at each hierarchical level to achieve calibration of its display position, and further obtain the direction elements of map construction. Furthermore, multiple hierarchical levels may be used as zoom levels and cluster categories may be used as legends to construct a data map corresponding to the target data, and the data map is used to display the target data in a zoomed manner within a page.

[0083] In summary, the technical solution in the present application can obtain the word frequency features corresponding to the target data by performing attribute text processing on the target data; then the word frequency features are input into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification results of the target data, and the hierarchical classification results include at least multiple hierarchical levels and hierarchical attributes and inherited data association relationships of the target data for each hierarchical level; further, the spectral clustering algorithm model with time and space constraints is used to cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; finally, multiple hierarchical levels are used as zoom levels, and clustering categories are used as legends to construct a data map corresponding to the target data, and the data map is used to zoom and display the target data on a page. The present application can realize the zoom display of all target data on a page according to the hierarchical levels, which can avoid the problems of jumping back and forth pages during hierarchical display, inflexible operation, and cluttered pages when the full amount of data is tiled.

[0084] Based on the above Figure 2 , Figure 3 A detailed description of the data display processing method provided, such as Fig. 9 As shown, Fig. 9 FIG. 1 is a structural block diagram of a data display processing device according to an exemplary embodiment. Fig. 9 As shown, the device comprises: The processing module 91 can be used to perform attribute text processing on the target data to obtain the word frequency features corresponding to the target data; A hierarchical module 92, which can be used to input word frequency features into a pre-trained hierarchical attribute classifier to obtain a hierarchical classification result of the target data, wherein the hierarchical classification result includes at least a plurality of hierarchical levels and hierarchical attributes of the target data and inherited data association relationships for each hierarchical level; The clustering module 93 may be used to utilize the spectral clustering algorithm model with time and space constraints to perform clustering processing on the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; The construction module 94 can be used to construct a data map corresponding to the target data using multiple hierarchical levels as zoom levels and cluster categories as legends. The data map is used to zoom and display the target data in a page.

[0085] In some embodiments of the present application, the processing module 91 can be specifically used to segment the target data into first words based on word segmentation technology to obtain a first word sequence containing multiple first words; determine a second word that matches a preset stop part of speech in the first word sequence according to the target part of speech of the first word; eliminate the second word in the first word sequence to obtain a second word sequence; and calculate the word frequency feature of the first word in the second word sequence based on the TF algorithm.

[0086] In some embodiments of the present application, Fig.10 As shown, the device further includes: a first training module 95; The first training module 95 can be used to determine the sample word frequency features of the sample target data and the corresponding preset hierarchical classification labels; input the sample word frequency features and the preset hierarchical classification labels into the hierarchical attribute classifier, and perform hierarchical attribute classification task training on the hierarchical attribute classifier; wherein, in the hierarchical attribute classification task training, the sample word frequency features are used as input features, and the preset hierarchical classification labels are used as training labels, and the hierarchical attribute classifier is iteratively trained until the loss function of the hierarchical attribute classifier is less than the first preset threshold value, and it is judged that the training of the hierarchical attribute classifier is completed.

[0087] In some embodiments of the present application, the spectral clustering algorithm model includes a spectral clustering composition conversion module, a matrix construction module, a graph cutting module and a clustering module; a hierarchical module 92, which can be specifically used to determine each hierarchical level as a target hierarchical level in order from low to high in a plurality of hierarchical levels; inputting the target data and the target hierarchical level into the spectral clustering composition conversion module for graph data conversion for hierarchical attributes and inherited data association relationships of the target data, so as to obtain a target spectral clustering composition of the target data at the target hierarchical level; using the matrix construction module, constructing a target Laplacian matrix corresponding to the target spectral clustering composition based on time and space constraint information, the time and space constraint information at least including a distance matrix, an initial adjacency matrix, a timestamp, a time weight and a distance weight; in the graph cutting module, based on the target Laplacian matrix, using the graph cutting algorithm to cut the target spectral clustering composition into a plurality of non-overlapping target subgraphs; using the clustering module to perform clustering processing on the sub-target data in each target subgraph respectively, so as to obtain a clustering category at the target hierarchical level.

[0088] In some embodiments of the present application, when the target Laplacian matrix corresponding to the target spectral clustering graph is constructed based on the time-space constraint information using the matrix construction module, the hierarchical module 82 can be specifically used to generate a degree matrix corresponding to the target spectral clustering graph according to the connection edge information of each node in the target spectral clustering graph; update the initial adjacency matrix based on the distance matrix, timestamp, time weight and distance weight to obtain the target adjacency matrix; and calculate the target Laplacian matrix corresponding to the target spectral clustering graph according to the degree matrix and the target adjacency matrix.

[0089] In some embodiments of the present application, Fig.10 As shown, the device further includes: a second training module 96; The second training module 96 can be used to determine first time-space constraint information when training the spectral clustering algorithm model, where the first time-space constraint information at least includes a distance matrix, an initial adjacency matrix, a timestamp, a first time weight, and a first distance weight; iteratively execute an update process of the first time-space constraint information until a preset iteration termination condition is reached to obtain final time-space constraint information, and construct a spectral clustering algorithm model based on the final time-space constraint information; The updating process of the first time-space constraint information includes: using an initial spectral clustering algorithm model to perform clustering processing on the training data based on the first time-space constraint information to obtain a first clustering evaluation index; adding a random offset to the first time weight and the first distance weight to obtain a second time weight and a second distance weight; using the initial spectral clustering algorithm model to perform clustering processing on the training data based on the second time-space constraint information to obtain a second clustering evaluation index, the second time-space constraint information at least includes a distance matrix, an initial adjacency matrix, a timestamp, a second time weight and a second distance weight; based on the first clustering evaluation index and the second clustering evaluation index, judging whether to update the second time-space constraint information to the first time-space constraint information.

[0090] In some embodiments of the present application, when judging whether to update the second time and space constraint information to the first time and space constraint information based on the first clustering evaluation index and the second clustering evaluation index, the second training module 96 can be specifically used to: calculate the indicator difference between the second clustering evaluation index and the first clustering evaluation index; when judging that the indicator difference is less than or equal to the second preset threshold, use the second time and space constraint information to update the first time and space constraint information, and continue to execute the update process of the first time and space constraint information; when judging that the indicator difference is greater than the second preset threshold, keep the first time and space constraint information unchanged during the update process of the first time and space constraint information.

[0091] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0092] In the embodiments of the present application, the word frequency features corresponding to the target data can be obtained by performing attribute text processing on the target data; then the word frequency features are input into the pre-trained hierarchical attribute classifier to obtain the hierarchical classification results of the target data, and the hierarchical classification results include at least multiple hierarchical levels and hierarchical attributes and inherited data association relationships of the target data for each hierarchical level; further, the spectral clustering algorithm model with time and space constraints is used to cluster the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; finally, multiple hierarchical levels are used as zoom levels, and clustering categories are used as legends to construct a data map corresponding to the target data, and the data map is used to zoom and display the target data on a page. The present application can realize the zoom display of all target data on a page according to the hierarchical levels, which can avoid the problems of jumping back and forth pages during hierarchical display, inflexible operation, and cluttered pages when the full amount of data is tiled.

[0093] The data display processing device of the embodiment of the present invention is described above from the perspective of the functional module in conjunction with the accompanying drawings. It should be understood that the functional module can be implemented in the form of hardware, can be implemented by instructions in the form of software, or can be implemented by a combination of hardware and software modules. Specifically, the steps of the data display processing method embodiment in the embodiment of the present invention can be completed by the hardware integrated logic circuit and / or software instructions in the processor, and the steps of the data display processing method applied in conjunction with the embodiment of the present invention can be directly embodied as a hardware decoding processor to execute, or a combination of hardware and software modules in the decoding processor to execute. Optionally, the software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory, and completes the steps in the above-mentioned data display processing method embodiment in conjunction with its hardware.

[0094] Fig.11 is a schematic block diagram of an electronic device 1100 according to an embodiment of the present invention.

[0095] like Fig.11 As shown, the electronic device 1100 may include: The memory 1110 and the processor 1120, the memory 1110 is used to store the computer program and transmit the program code to the processor 1120. In other words, the processor 1120 can call and run the computer program from the memory 1110 to implement the method in the embodiment of the present invention.

[0096] For example, the processor 1120 may be configured to execute the above method embodiments according to instructions in the computer program.

[0097] In some embodiments of the present invention, the processor 1120 may include but is not limited to: General-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0098] In some embodiments of the present invention, the memory 1110 includes but is not limited to: Volatile memory and / or non-volatile memory. Among them, the non-volatile memory can be read-only memory (ROM), programmable ROM (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM) or flash memory. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct memory bus random access memory (Direct Rambus RAM, DR RAM).

[0099] In some embodiments of the present invention, the computer program may be divided into one or more modules, which are stored in the memory 1110 and executed by the processor 1120 to complete the method provided by the present invention. The one or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the controller.

[0100] like Fig.11 As shown, the electronic device 1100 may further include: The transceiver 1130 may be connected to the processor 1120 or the memory 1110 .

[0101] The processor 1120 may control the transceiver 1130 to communicate with other devices, specifically, to send data to other devices, or to receive data sent by other devices. The transceiver 1130 may include a transmitter and a receiver. The transceiver 1130 may further include an antenna, and the number of antennas may be one or more.

[0102] It should be understood that the various components in the electronic device are connected via a bus system, wherein the bus system includes not only a data bus but also a power bus, a control bus and a status signal bus.

[0103] The present invention also provides a computer storage medium having a computer program stored thereon, which, when executed by a computer, enables the computer to perform the method of the above method embodiment. In other words, an embodiment of the present invention also provides a computer program product containing instructions, which, when executed by a computer, enables the computer to perform the method of the above method embodiment.

[0104] When software is used for implementation, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integration. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disc (Digital Video Disc, DVD)), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

[0105] Those skilled in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments applied herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0106] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the module is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0107] The modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. For example, each functional module in each embodiment of the present application may be integrated into a processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0108] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in this application, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A data display processing method, characterized in that: include: Performing attribute text processing on the target data to obtain word frequency features corresponding to the target data; Inputting the word frequency feature into a pre-trained hierarchical attribute classifier to obtain a hierarchical classification result of the target data, wherein the hierarchical classification result includes at least a plurality of hierarchical levels and hierarchical attributes and inherited data association relationships of each hierarchical level with respect to the target data; Using a spectral clustering algorithm model with time and space constraints, clustering the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level; The multiple hierarchical levels are used as zoom levels and the cluster categories are used as legends to construct a data map corresponding to the target data, and the data map is used to display the target data in a zoomed manner within a page.

2. The method according to claim 1, characterized in that The performing attribute text processing on the target data to obtain the word frequency features corresponding to the target data includes: Based on the word segmentation technology, the target data is divided into first words to obtain a first word sequence including a plurality of the first words; Determining, in the first word sequence, a second word that matches a preset stop part of speech according to the target part of speech of the first word; Eliminate the second word from the first word sequence to obtain a second word sequence; The word frequency feature of the first word in the second word sequence is calculated based on the TF algorithm.

3. The method according to claim 1, characterized in that The method also includes a training method for the hierarchical attribute classifier: Determine sample word frequency features of sample target data and corresponding preset hierarchical classification labels; Inputting the sample word frequency feature and the preset hierarchical classification label into a hierarchical attribute classifier, and performing hierarchical attribute classification task training on the hierarchical attribute classifier; Among them, in the classification task training of the hierarchical attribute, the sample word frequency feature is used as the input feature, and the preset hierarchical classification label is used as the training label, and the hierarchical attribute classifier is iteratively trained until the loss function of the hierarchical attribute classifier is less than the first preset threshold, and it is judged that the training of the hierarchical attribute classifier is completed.

4. The method according to claim 1, characterized in that: The spectral clustering algorithm model includes a spectral clustering composition conversion module, a matrix construction module, a graph cutting module and a clustering module; The spectral clustering algorithm model using time and space constraints performs clustering processing on the target data layer by layer at multiple hierarchical levels based on the hierarchical classification results to obtain clustering categories at each hierarchical level, including: According to the order from low to high among the plurality of hierarchical levels, each of the hierarchical levels is determined as a target hierarchical level in turn; Inputting the target data, and the hierarchical attributes and inherited data association relationships of the target hierarchical level with respect to the target data into the spectral clustering composition conversion module for graph data conversion to obtain a target spectral clustering composition of the target data at the target hierarchical level; Using the matrix construction module, constructing a target Laplacian matrix corresponding to the target spectral clustering composition based on time-space constraint information, wherein the time-space constraint information at least includes a distance matrix, an initial adjacency matrix, a timestamp, a time weight, and a distance weight; In the graph cutting module, based on the target Laplacian matrix, a graph cutting algorithm is used to cut the target spectral clustering graph into a plurality of non-overlapping target subgraphs; The clustering module is used to perform clustering processing on the sub-target data in each of the target sub-graphs to obtain cluster categories under the target hierarchical level.

5. The method according to claim 4, characterized in that The using the matrix construction module to construct a target Laplacian matrix corresponding to the target spectral clustering composition based on time-space constraint information includes: Generate a degree matrix corresponding to the target spectral clustering graph according to the connection edge information of each node in the target spectral clustering graph; Based on the distance matrix, the timestamp, the time weight and the distance weight, the initial adjacency matrix is ​​updated to obtain a target adjacency matrix; A target Laplacian matrix corresponding to the target spectral clustering graph is calculated according to the degree matrix and the target adjacency matrix.

6. The method according to claim 4, characterized in that The method also includes a training method for the spectral clustering algorithm model: Determine first time-space constraint information, wherein the first time-space constraint information at least includes a distance matrix, an initial adjacency matrix, a timestamp, a first time weight, and a first distance weight; Iteratively executing the updating process of the first time-space constraint information until a preset iteration termination condition is reached, obtaining the final time-space constraint information, and constructing the spectral clustering algorithm model based on the final time-space constraint information; The updating process of the first time-space constraint information includes: Using an initial spectral clustering algorithm model to cluster the training data based on the first time-space constraint information to obtain a first clustering evaluation index; Adding a random offset to the first time weight and the first distance weight to obtain a second time weight and a second distance weight; Using the initial spectral clustering algorithm model to cluster the training data based on second time-space constraint information to obtain a second clustering evaluation index, wherein the second time-space constraint information at least includes the distance matrix, the initial adjacency matrix, the timestamp, the second time weight, and the second distance weight; Based on the first clustering evaluation index and the second clustering evaluation index, it is determined whether to update the second time-space constraint information to the first time-space constraint information.

7. The method according to claim 6, characterized in that The determining, based on the first clustering evaluation index and the second clustering evaluation index, whether to update the second time-space constraint information to the first time-space constraint information includes: Calculating an index difference between the second clustering evaluation index and the first clustering evaluation index; When it is determined that the indicator difference is less than or equal to a second preset threshold, the first time-space constraint information is updated using the second time-space constraint information, and the updating process of the first time-space constraint information is continued; When it is determined that the indicator difference is greater than the second preset threshold, the first time and space constraint information is kept unchanged during the updating process of the first time and space constraint information.

8. A data display processing device, characterized in that: include: A processing module, used for performing attribute text processing on the target data to obtain word frequency features corresponding to the target data; A hierarchical module, used for inputting the word frequency feature into a pre-trained hierarchical attribute classifier to obtain a hierarchical classification result of the target data, wherein the hierarchical classification result includes at least a plurality of hierarchical levels and hierarchical attributes and inherited data association relationships of each hierarchical level with respect to the target data; A clustering module, used for performing clustering processing on the target data layer by layer at multiple hierarchical levels based on the hierarchical classification result by using a spectral clustering algorithm model with time and space constraints, and obtaining clustering categories at each hierarchical level; A construction module is used to use the multiple hierarchical levels as zoom levels and the cluster categories as legends to construct a data map corresponding to the target data, and the data map is used to display the target data in a zoomed manner within a page.

9. An electronic device, characterized in that: include: A processor and a memory, the memory being used to store a computer program, and the processor being used to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store a computer program, wherein the computer program enables a computer to execute the method according to any one of claims 1 to 7.

Images