A state analysis method and system based on Internet

Abstract

The present invention provides a state analysis method and system based on the Internet, which relates to the field of data processing technology. In the present invention, for each Internet device to be analyzed, the predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed is predicted; the predicted device state description data corresponding to each of the multiple Internet devices to be analyzed is subjected to feature mapping processing to output the target description data mapping feature representation corresponding to each Internet device to be analyzed; based on the target description data mapping feature representation corresponding to each Internet device to be analyzed, the multiple Internet devices to be analyzed are subjected to device state analysis processing to output the target device state information corresponding to each Internet device to be analyzed. Based on the above method, the accuracy of device state analysis can be improved to a certain extent.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a state analysis method and system based on the Internet. Background Art

[0002] With the improvement of Internet technology and data processing technology, there are more application scenarios of Internet technology and data processing technology. Among them, Internet technology will involve more Internet devices in application, and the status of Internet devices directly affects the operation of the Internet system. Therefore, it is necessary to analyze the status of Internet devices. However, in the existing technology, there is a problem of poor accuracy of device status analysis. Summary of the invention

[0003] In view of this, an object of the present invention is to provide an Internet-based status analysis method and system to improve the accuracy of device status analysis to a certain extent.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0005] A state analysis method based on the Internet, comprising:

[0006] For each Internet device to be analyzed, predict the predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the predicted device state description data belongs to the predicted device state description data of the device state data cluster to be analyzed in the first data dimension, and there are multiple Internet devices to be analyzed, and the device state data cluster to be analyzed includes multiple pieces of device state data to be analyzed;

[0007] Performing feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed, respectively, to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed;

[0008] According to the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on multiple Internet devices to be analyzed to output target device status information corresponding to each of the Internet devices to be analyzed, and the target device status information is used to reflect whether the status of the corresponding Internet device to be analyzed is abnormal.

[0009] In some preferred embodiments, in the above-mentioned Internet-based status analysis method, the step of predicting, for each Internet device to be analyzed, predicted device status description data corresponding to the device status data cluster to be analyzed corresponding to the Internet device to be analyzed includes:

[0010] Extracting a device status data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the device status data cluster to be analyzed includes device status data to be analyzed corresponding to at least one data dimension other than the first data dimension;

[0011] A first state data analysis neural network corresponding to a first data dimension is extracted; the first state data analysis neural network performs network optimization on a second state data analysis neural network based on the first exemplary state data to form a network; the first exemplary state data includes a first exemplary device state data cluster to be analyzed and corresponding exemplary device state description data in the first data dimension, and the first exemplary device state data cluster to be analyzed includes the device state data to be analyzed corresponding to the first data dimension; the second state data analysis neural network performs network optimization on a third state data analysis neural network based on the second exemplary state data; the second exemplary state data includes a combination of second exemplary device state data clusters to be analyzed and corresponding exemplary device state description data corresponding to a plurality of exemplary data dimensions;

[0012] The device status data cluster to be analyzed is loaded, so as to utilize the first status data analysis neural network to analyze and output the predicted device status description data of the Internet device to be analyzed in the first data dimension.

[0013] In some preferred embodiments, in the above-mentioned Internet-based state analysis method, the step of performing feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed includes:

[0014] Performing text segmentation processing on the predicted device state description data corresponding to the Internet device to be analyzed, so as to output a plurality of description data segmentation words corresponding to the predicted device state description data;

[0015] The predicted device state description data and the description data segmentation terms are subjected to feature mapping processing, and a target description data mapping feature representation corresponding to the Internet device to be analyzed is output.

[0016] In some preferred embodiments, in the above-mentioned Internet-based state analysis method, the step of performing feature mapping processing on the predicted device state description data and the description data segmentation words, and outputting the target description data mapping feature representation corresponding to the Internet device to be analyzed, includes:

[0017] Performing feature mapping processing on the predicted device state description data to output a first description data mapping feature representation corresponding to the Internet device to be analyzed;

[0018] Performing feature mapping processing on each of the description data segmentation words respectively to output a word mapping feature representation corresponding to each of the description data segmentation words, and sequentially splicing the word mapping feature representation corresponding to each of the description data segmentation words according to the distribution position information of the corresponding description data segmentation words in the predicted device state description data to form a second description data mapping feature representation corresponding to the Internet device to be analyzed;

[0019] The first description data mapping feature representation corresponding to the Internet device to be analyzed and the second description data mapping feature representation corresponding to the Internet device to be analyzed are aggregated to form a target description data mapping feature representation corresponding to the Internet device to be analyzed.

[0020] In some preferred embodiments, in the above-mentioned Internet-based state analysis method, the step of aggregating the first description data mapping feature representation corresponding to the Internet device to be analyzed and the second description data mapping feature representation corresponding to the Internet device to be analyzed to form a target description data mapping feature representation corresponding to the Internet device to be analyzed includes:

[0021] Determine at least one description data segmentation word from the multiple description data segmentation words as the corresponding key description data segmentation word, and select a local second description data mapping feature representation corresponding to the key description data segmentation word from the second description data mapping feature representation corresponding to the Internet device to be analyzed;

[0022] The local second description data mapping feature representation and the first description data mapping feature representation corresponding to the Internet device to be analyzed are concatenated to form a target description data mapping feature representation corresponding to the Internet device to be analyzed.

[0023] In some preferred embodiments, in the above-mentioned Internet-based status analysis method, the step of performing device status analysis processing on a plurality of Internet devices to be analyzed based on the target description data mapping feature representation corresponding to each Internet device to be analyzed to output target device status information corresponding to each Internet device to be analyzed includes:

[0024] Extracting target device type information corresponding to each of the Internet devices to be analyzed;

[0025] According to the corresponding target device type information, a plurality of the Internet devices to be analyzed are subjected to device classification processing to form at least one device classification cluster corresponding to the plurality of Internet devices to be analyzed, wherein the target device type information corresponding to the Internet devices to be analyzed included in each of the device classification clusters belongs to the same device type information, and the target device type information corresponding to the Internet devices to be analyzed included in different device classification clusters belongs to different device type information;

[0026] Based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the Internet devices to be analyzed included in the device classification cluster within each of the device classification clusters to output target device status information corresponding to each of the Internet devices to be analyzed included in the device classification cluster.

[0027] In some preferred embodiments, in the above-mentioned Internet-based status analysis method, the step of performing device status analysis processing on the Internet devices to be analyzed included in the device classification cluster based on the target description data mapping feature representation corresponding to each Internet device to be analyzed within each device classification cluster to output target device status information corresponding to each Internet device to be analyzed included in the device classification cluster includes:

[0028] For any device classification cluster, count the number of Internet devices to be analyzed included in the device classification cluster to output the device statistics corresponding to the device classification cluster;

[0029] Based on a pre-configured reference classification ratio and the device statistical quantity, a multiplication operation is performed to output a corresponding first device statistical quantity, and for each Internet device to be analyzed in the device classification cluster, a feature representation matching degree between a target description data mapping feature representation corresponding to the Internet device to be analyzed and a target description data mapping feature representation corresponding to each other Internet device to be analyzed in the device classification cluster is calculated, and then from the feature representation matching degrees between the target description data mapping feature representation corresponding to the Internet device to be analyzed and the target description data mapping feature representation corresponding to each other Internet device to be analyzed in the device classification cluster, the largest feature representation matching degree of the first device statistical quantity is selected as a first feature representation matching degree set corresponding to the Internet device to be analyzed, and the first feature representation matching degrees included in the first feature representation matching degree set are averaged to obtain an average first feature representation matching degree corresponding to the Internet device to be analyzed;

[0030] Based on the mean first feature representation matching degree corresponding to each Internet device to be analyzed in the device classification cluster, device status analysis processing is performed on each Internet device to be analyzed included in the device classification cluster to output target device status information corresponding to each Internet device to be analyzed included in the device classification cluster.

[0031] The embodiment of the present invention further provides an Internet-based status analysis system, comprising:

[0032] a state description data prediction module, configured to predict, for each Internet device to be analyzed, predicted device state description data corresponding to a device state data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the predicted device state description data belongs to the predicted device state description data of the device state data cluster to be analyzed in a first data dimension, and there are multiple Internet devices to be analyzed, and the device state data cluster to be analyzed includes multiple pieces of device state data to be analyzed;

[0033] A feature mapping processing module, used to perform feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed, so as to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed;

[0034] The device status analysis module is used to perform device status analysis on multiple Internet devices to be analyzed based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, so as to output target device status information corresponding to each of the Internet devices to be analyzed, and the target device status information is used to reflect whether the status of the corresponding Internet device to be analyzed is abnormal.

[0035] In some preferred embodiments, in the above-mentioned Internet-based status analysis system, the feature mapping processing module is specifically used to:

[0036] Performing text segmentation processing on the predicted device state description data corresponding to the Internet device to be analyzed, so as to output a plurality of description data segmentation words corresponding to the predicted device state description data;

[0037] The predicted device state description data and the description data segmentation terms are subjected to feature mapping processing, and a target description data mapping feature representation corresponding to the Internet device to be analyzed is output.

[0038] In some preferred embodiments, in the above-mentioned Internet-based status analysis system, the device status analysis module is specifically used to:

[0039] Extracting target device type information corresponding to each of the Internet devices to be analyzed;

[0040] According to the corresponding target device type information, a plurality of the Internet devices to be analyzed are subjected to device classification processing to form at least one device classification cluster corresponding to the plurality of Internet devices to be analyzed, wherein the target device type information corresponding to the Internet devices to be analyzed included in each of the device classification clusters belongs to the same device type information, and the target device type information corresponding to the Internet devices to be analyzed included in different device classification clusters belongs to different device type information;

[0041] Based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the Internet devices to be analyzed included in the device classification cluster within each of the device classification clusters to output target device status information corresponding to each of the Internet devices to be analyzed included in the device classification cluster.

[0042] The embodiment of the present invention provides a state analysis method and system based on the Internet. For each Internet device to be analyzed, the predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed is predicted; the predicted device state description data corresponding to each of the multiple Internet devices to be analyzed is subjected to feature mapping processing to output the target description data mapping feature representation corresponding to each Internet device to be analyzed; and the device state analysis processing is performed on the multiple Internet devices to be analyzed according to the target description data mapping feature representation corresponding to each Internet device to be analyzed, so as to output the target device state information corresponding to each Internet device to be analyzed. Based on this, since the predicted device state description data is determined first, compared with the conventional technical means of directly performing device state analysis processing based on the device state data cluster to be analyzed, the analysis basis can be more concise and the invalid information is less, so that the accuracy of the device state analysis can be improved to a certain extent, thereby improving the problem of low accuracy in the prior art.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a structural block diagram of an Internet-based status analysis platform provided in an embodiment of the present invention.

[0045] Figure 2 A flowchart of the steps of the Internet-based status analysis method provided in an embodiment of the present invention.

[0046] Figure 3 A schematic diagram of various modules included in the Internet-based status analysis system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0048] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] like Figure 1 As shown, an embodiment of the present invention provides an Internet-based status analysis platform, wherein the Internet-based status analysis platform may include a memory and a processor.

[0050] In detail, the memory and the processor are electrically connected directly or indirectly to achieve data transmission or interaction. For example, they can be electrically connected to each other through one or more communication buses or signal lines. The memory can store at least one software function module (computer program) that can exist in the form of software or firmware. The processor can be used to execute the executable computer program stored in the memory, thereby realizing the Internet-based status analysis method provided by the embodiment of the present invention.

[0051] It can be selected that in some embodiments, the memory can be, but not limited to, random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electric erasable programmable read-only memory (EEPROM), etc. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a system on chip (SoC), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0052] Optionally, in some implementations, the Internet-based status analysis platform may be a server with data processing capabilities.

[0053] Combination Figure 2 The embodiment of the present invention further provides an Internet-based state analysis method, which can be applied to the above-mentioned Internet-based state analysis platform. The method steps defined in the process related to the Internet-based state analysis method can be implemented by the Internet-based state analysis platform. Figure 2 The specific process shown is explained in detail.

[0054] Step S100: For each Internet device to be analyzed, predict the predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed.

[0055] In an embodiment of the present invention, the Internet-based state analysis platform can predict, for each Internet device to be analyzed, predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed. The predicted device state description data belongs to the predicted device state description data of the device state data cluster to be analyzed in the first data dimension, there are multiple Internet devices to be analyzed, and the device state data cluster to be analyzed includes multiple pieces of device state data to be analyzed.

[0056] Step S200 , performing feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed, so as to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed.

[0057] In an embodiment of the present invention, the Internet-based state analysis platform can perform feature mapping processing on the predicted device state description data corresponding to each of the multiple Internet devices to be analyzed, so as to output a target description data mapping feature representation corresponding to each of the Internet devices to be analyzed.

[0058] Step S300 , performing device status analysis processing on a plurality of Internet devices to be analyzed according to the target description data mapping feature representation corresponding to each Internet device to be analyzed, so as to output target device status information corresponding to each Internet device to be analyzed.

[0059] In an embodiment of the present invention, the Internet-based status analysis platform can perform device status analysis on a plurality of Internet devices to be analyzed based on the target description data mapping feature representation corresponding to each Internet device to be analyzed, so as to output target device status information corresponding to each Internet device to be analyzed. The target device status information is used to reflect whether the status of the corresponding Internet device to be analyzed is abnormal.

[0060] Based on this, since the predicted equipment status description data is determined first, compared with the conventional technical means of directly performing equipment status analysis and processing based on the equipment status data cluster to be analyzed, the analysis basis can be more concise and contain less invalid information, thereby improving the accuracy of equipment status analysis to a certain extent, thereby improving the problem of low accuracy in the existing technology.

[0061] Optionally, in some implementations, the above step S100 may further include the following specific sub-steps, such as step S110, step S120 and step S130.

[0062] Step S110, extracting a device status data cluster to be analyzed corresponding to the Internet device to be analyzed.

[0063] In an embodiment of the present invention, the Internet-based state data processing platform can extract a device state data cluster to be analyzed corresponding to the Internet device to be analyzed. The device state data cluster to be analyzed includes device state data to be analyzed corresponding to other data dimensions other than at least one first data dimension (exemplarily, the first data dimension refers to the data dimension corresponding to the predicted device state description data required to be predicted for the Internet device to be analyzed, such as data transmission dimension, data reception dimension, data calculation dimension, data storage dimension, etc.).

[0064] Step S120, extracting the first state data analysis neural network corresponding to the first data dimension.

[0065] In an embodiment of the present invention, the Internet-based state data processing platform can extract a first state data analysis neural network corresponding to a first data dimension. The first state data analysis neural network is formed by optimizing the network of the second state data analysis neural network based on the first exemplary state data; the first exemplary state data includes a first exemplary device state data cluster to be analyzed and corresponding exemplary device state description data (i.e., real data) in the first data dimension, and the first exemplary device state data cluster to be analyzed includes the device state data to be analyzed corresponding to the first data dimension; the second state data analysis neural network is formed by optimizing the network of the third state data analysis neural network based on the second exemplary state data; the second exemplary state data includes a second exemplary device state data cluster to be analyzed corresponding to multiple exemplary data dimensions and a corresponding exemplary device state description data combination (that is, the exemplary device state description data combination includes a corresponding multiple exemplary device state description data; in addition, since the first state data analysis neural network is formed by optimizing the network of the second state data analysis neural network based on the first exemplary state data, the first exemplary state data includes the first exemplary device state data cluster to be analyzed A device status data cluster and exemplary device status description data of a first data dimension, wherein the first exemplary device status data cluster to be analyzed includes the device status data to be analyzed corresponding to the first data dimension, and the second status data analysis neural network is formed by network optimization of the third status data analysis neural network based on the second exemplary status data, wherein the second exemplary status data includes a combination of second exemplary device status data clusters to be analyzed and corresponding exemplary device status description data corresponding to multiple exemplary data dimensions, so that the first status data analysis neural network can learn data-related information between multiple data dimensions, and learn the correspondence between the exemplary device status description data of other data dimensions and the predicted device status description data of the first data dimension, and when the device status data cluster to be analyzed corresponding to the Internet device to be analyzed is loaded into the first status data analysis neural network, the output predicted device status description data is based on the data-related information between the other data dimensions and the first data dimension, so that the accuracy of the output predicted device status description data can be higher, thereby achieving a highly reliable prediction).

[0066] Step S130, loading the device status data cluster to be analyzed, and using the first status data analysis neural network to analyze and output the predicted device status description data of the Internet device to be analyzed in the first data dimension.

[0067] In an embodiment of the present invention, the Internet-based status data processing platform can load the device status data cluster to be analyzed, so as to utilize the first status data analysis neural network (based on learning of exemplary data, so that the first status data analysis neural network has the function of predicting device status description data), analyze and output the predicted device status description data of the Internet device to be analyzed in the first data dimension.

[0068] Based on the aforementioned steps S110, S120 and S130, since the second exemplary state data includes second exemplary device state data clusters to be analyzed corresponding to multiple exemplary data dimensions, the overfitting degree of the formed second state data analysis neural network is low, and the analysis reliability of the first state data analysis neural network formed based on the optimization of the second state data analysis neural network is high, thereby improving the reliability of the state description data analysis to a certain extent, thereby improving the problem of poor reliability of the state description data analysis existing in the prior art.

[0069] Optionally, in some embodiments, the first exemplary state data is obtained based on the following steps:

[0070] A first exemplary cluster of device operation data to be analyzed and a second exemplary cluster of device operation data to be analyzed of the first data dimension are extracted, wherein the device operation data to be analyzed included in the first exemplary cluster of device operation data to be analyzed is formed earlier than the device operation data to be analyzed included in the second exemplary cluster of device operation data to be analyzed, and the first exemplary cluster of device operation data to be analyzed includes the device operation data to be analyzed corresponding to the first data dimension (exemplarily, the device operation data to be analyzed can be obtained based on parsing operation log data);

[0071] Based on the first exemplary device operation data cluster to be analyzed, a corresponding first exemplary device status data cluster to be analyzed is analyzed and output (exemplarily, the first exemplary device operation data cluster to be analyzed can be directly used as the first exemplary device status data cluster to be analyzed);

[0072] Based on the second exemplary cluster of equipment operation data to be analyzed, exemplary equipment status description data of the first data dimension is analyzed and output (exemplarily, the equipment status description data corresponding to the second exemplary cluster of equipment operation data to be analyzed can be used as exemplary equipment status description data, and the equipment status description data can be manually configured).

[0073] Optionally, in some embodiments, the first state data analysis neural network is obtained based on the following steps:

[0074] A pre-configured first dimension data screening ratio is extracted, and a pre-configured second dimension data screening ratio is extracted (the specific ratio values ​​of the first dimension data screening ratio and the second dimension data screening ratio are not limited and can be selected according to actual application requirements);

[0075] The neural network for analyzing the second state data is subjected to network optimization processing based on the first exemplary state data, wherein a basis for the network optimization processing includes:

[0076] In the first exemplary device status data cluster to be analyzed, the device status data to be analyzed corresponding to the first data dimension is screened out based on the first dimension data screening ratio (that is, the data quantity ratio of the device status data to be analyzed corresponding to the first data dimension screened out is equal to the first dimension data screening ratio);

[0077] In the first exemplary device status data cluster to be analyzed, the device status data to be analyzed corresponding to data dimensions other than the first data dimension are filtered out based on the second dimensional data filtering ratio (that is, the proportion of the device status data to be analyzed corresponding to data dimensions other than the first data dimension that are filtered out is equal to the second dimensional data filtering ratio).

[0078] Optionally, in some implementations, the step of loading the device state data cluster to be analyzed, analyzing and outputting the predicted device state description data of the Internet device to be analyzed in the first data dimension by using the first state data analysis neural network, may further include the following specific sub-steps:

[0079] The device state data cluster to be analyzed is loaded to be loaded into the feature space mapping model included in the first state data analysis neural network, and the feature space mapping model is used to perform feature mapping processing on the device state data cluster to be analyzed to output a device state data mapping feature representation corresponding to each data dimension (that is, the discrete device state data cluster to be analyzed can be represented by a continuous device state data mapping feature representation);

[0080] Using the feature representation aggregation model included in the first state data analysis neural network, the device state data mapping feature representation corresponding to each data dimension is aggregated to output the aggregated device state data mapping feature representation corresponding to the Internet device to be analyzed;

[0081] Using the first state data analysis neural network, in the reference device state description data cluster corresponding to the first data dimension, the reference device state description data corresponding to the aggregated device state data mapping feature representation is matched as the predicted device state description data of the Internet device to be analyzed in the first data dimension (exemplarily, the first state data analysis neural network may include a classification model, such as a softmax function, so that after the aggregated device state data mapping feature representation corresponding to the Internet device to be analyzed is determined using the first state data analysis neural network, the aggregated device state data mapping feature representation and the reference device state description data mapping feature representation of the reference device state description data are loaded into the classification model of the first state data analysis neural network, and the classification model is used to output the feature representation matching degree between the aggregated device state data mapping feature representation and the reference device state description data mapping feature representation corresponding to each reference device state description data. Then, the corresponding device state description data may be matched based on the feature representation matching degree. The reference device state description data are arranged in order of size. Among them, the reference device state description data corresponding to the reference device state description data mapping feature representation with a greater degree of matching with the feature representation of the aggregated device state data mapping feature representation can be in front, and the reference device state description data corresponding to the reference device state description data mapping feature representation with a lower degree of matching with the feature representation of the aggregated device state data mapping feature representation can be in the back. In this way, starting from the first reference device state description data, a certain number of reference device state description data can be selected as the predicted device state description data corresponding to the Internet device to be analyzed. Among them, the reference device state description data mapping feature representation is formed by performing feature mapping processing on the reference device state description data. In addition, the degree of feature representation matching between the aggregated device state data mapping feature representation and the reference device state description data mapping feature representation corresponding to each reference device state description data can be considered as the possibility of the aggregated device state data mapping feature representation being matched to each reference device state description data).

[0082] Optionally, in some embodiments, the feature representation aggregation model includes an internal feature representation aggregation sub-model, an external feature representation aggregation sub-model, a feature representation association sub-model, and a feature representation integration sub-model. Based on this, the step of using the feature representation aggregation model included in the first state data analysis neural network to aggregate the device state data mapping feature representation corresponding to each data dimension to output the aggregated device state data mapping feature representation corresponding to the Internet device to be analyzed may further include the following specific sub-steps:

[0083] Using the internal feature representation aggregation sub-model, respectively aggregate the device status data mapping feature representation corresponding to each of the data dimensions to form an intra-dimensional aggregate mapping feature representation corresponding to each data dimension (exemplarily, an intra-dimensional aggregate mapping feature representation corresponding to a data dimension refers to a feature representation of characteristics in the data dimension formed by aggregating all device status data mapping feature representations in the data dimension);

[0084] Using the external feature representation aggregation sub-model, each of the in-dimensional aggregate mapping feature representations is aggregated to form an out-dimensional aggregate mapping feature representation corresponding to the Internet device to be analyzed (exemplarily, the out-dimensional aggregate mapping feature representation refers to a feature representation that is formed by aggregating the in-dimensional aggregate mapping feature representations corresponding to multiple data dimensions, and can reflect both the features within each data dimension and the features between the data dimensions; based on this, the in-dimensional aggregate mapping feature representations of multiple data dimensions are aggregated by the external feature representation aggregation sub-model to generate an out-dimensional aggregate mapping feature representation, which can avoid the discreteness of the data while abstracting the data features, and is conducive to predicting the device state description data);

[0085] By using the feature representation association submodel, a feature representation association aggregation operation is performed on the local mapping feature representations in the out-of-dimensional aggregate mapping feature representation to form a corresponding associated aggregate mapping feature representation (exemplarily, each feature representation parameter of the out-of-dimensional aggregate mapping feature representation belongs to a local mapping feature representation of the out-of-dimensional aggregate mapping feature representation, and the out-of-dimensional aggregate mapping feature representation may include multiple local mapping feature representations. The associated aggregate mapping feature representation refers to a feature representation obtained by associating the local mapping feature representations in the out-of-dimensional aggregate mapping feature representation, which reflects the correlation between the local mapping feature representations. For example, any two or more of the out-of-dimensional aggregate mapping feature representations may be associated with each other. The local mapping feature representations are subjected to association processing, such as multiplication processing, etc. After the association processing, multiple associated local mapping feature representations that can represent the correlation between the local mapping feature representations can be obtained. Then, these associated local mapping feature representations can be combined to form an associated aggregate mapping feature representation corresponding to the Internet device to be analyzed; in addition, the multiplication processing can refer to performing dot multiplication and cross multiplication on the local mapping feature representations. Specifically, for the first local mapping feature representation and the second local mapping feature representation, the first local mapping feature representation and the second local mapping feature representation can be firstly dot-multiplied, and then the dot-multiplication result, the first local mapping feature representation and the second local mapping feature representation can be cross-multiplied, so that the associated aggregate mapping feature representation can be obtained);

[0086] The feature representation integration sub-model is used to integrate the out-of-dimensional aggregate mapping feature representation and the associated aggregate mapping feature representation to form an aggregate device status data mapping feature representation corresponding to the Internet device to be analyzed (exemplarily, the dimension of the associated aggregate mapping feature representation can be scaled to the same dimension as the out-of-dimensional aggregate mapping feature representation through the first feature representation integration sub-model included in the feature representation integration sub-model, and then spliced ​​with the out-of-dimensional aggregate mapping feature representation to form a spliced ​​feature representation to be processed; then, the spliced ​​feature representation to be processed can be mapped to an aggregate device status data mapping feature representation through the second feature representation integration sub-model included in the feature representation integration sub-model, specifically, the parameters included in the spliced ​​feature representation to be processed can be mapped to the interval 0-1).

[0087] Optionally, in some implementations, the step of using the internal feature representation aggregation sub-model to aggregate the device status data mapping feature representation corresponding to each of the data dimensions to form an intra-dimensional aggregate mapping feature representation corresponding to each data dimension may further include the following specific sub-steps:

[0088] Extracting, from the internal feature representation aggregation sub-model, the focused feature analysis influence parameters corresponding to each device state data mapping feature representation corresponding to each of the data dimensions (exemplarily, the specific calculation method of the focused feature analysis influence parameters can refer to the relevant prior art on focused feature analysis, which is not specifically limited herein. For example, for a device state data mapping feature representation, the device state data mapping feature representation and the spatial transformation result corresponding to the device state data mapping feature representation can be multiplied first, and then the multiplication result and a shift parameter formed by network optimization are superimposed, and then the superposition result can be nonlinearly mapped, and then the mapping result and a configuration feature representation formed by network optimization can be multiplied to obtain a multiplication result, and then the positive correlation value of the multiplication result can be determined based on any positive correlation function, and finally, the positive correlation value of the multiplication result can be normalized based on the positive correlation values ​​corresponding to each device state data mapping feature representation to obtain the focused feature analysis influence parameter corresponding to the device state data mapping feature representation);

[0089] By utilizing the internal feature representation aggregation sub-model, each device status data mapping feature representation corresponding to each data dimension is weighted aggregated based on the corresponding focused feature analysis influence parameter to form an intra-dimensional aggregated mapping feature representation corresponding to each data dimension.

[0090] Optionally, in some implementations, the step of using the external feature representation aggregation sub-model to aggregate each of the in-dimensional aggregate mapping feature representations to form an out-of-dimensional aggregate mapping feature representation corresponding to the Internet device to be analyzed may further include the following specific sub-steps:

[0091] Extracting the focus feature analysis influence parameters corresponding to the aggregated mapping feature representations in each dimension from the external feature representation aggregation sub-model (refer to the above description);

[0092] Using the external feature representation aggregation sub-model, weighted aggregation processing is performed on each of the in-dimensional aggregate mapping feature representations based on the corresponding focus feature analysis influence parameter to form an out-dimensional aggregate mapping feature representation corresponding to the Internet device to be analyzed.

[0093] Optionally, in some implementations, the above step S200 may further include the following specific sub-steps:

[0094] Performing text segmentation processing on the predicted device state description data corresponding to the Internet device to be analyzed, so as to output a plurality of description data segmentation words corresponding to the predicted device state description data (the specific processing method of the text segmentation processing can refer to the relevant prior art);

[0095] The predicted device state description data and the description data segmentation terms are subjected to feature mapping processing, and a target description data mapping feature representation corresponding to the Internet device to be analyzed is output.

[0096] Optionally, in some implementations, the step of performing feature mapping processing on the predicted device state description data and the description data segmentation terms, and outputting the target description data mapping feature representation corresponding to the Internet device to be analyzed, may further include the following specific sub-steps:

[0097] Performing feature mapping processing (as described above) on the predicted device state description data to output a first description data mapping feature representation corresponding to the Internet device to be analyzed;

[0098] Performing feature mapping processing on each of the description data segmentation words respectively to output a word mapping feature representation corresponding to each of the description data segmentation words, and sequentially splicing the word mapping feature representation corresponding to each of the description data segmentation words according to the distribution position information of the corresponding description data segmentation words in the predicted device state description data to form a second description data mapping feature representation corresponding to the Internet device to be analyzed;

[0099] The first description data mapping feature representation corresponding to the Internet device to be analyzed and the second description data mapping feature representation corresponding to the Internet device to be analyzed are aggregated to form a target description data mapping feature representation corresponding to the Internet device to be analyzed.

[0100] Optionally, in some implementations, the step of aggregating the first description data mapping feature representation corresponding to the Internet device to be analyzed and the second description data mapping feature representation corresponding to the Internet device to be analyzed to form a target description data mapping feature representation corresponding to the Internet device to be analyzed may further include the following specific sub-steps:

[0101] Determine at least one description data segmentation word from the multiple description data segmentation words as the corresponding key description data segmentation word (refer to any keyword determination method in the prior art, such as determining the key description data segmentation word based on the number of occurrences in the corresponding corpus), and select the local second description data mapping feature representation corresponding to the key description data segmentation word from the second description data mapping feature representation corresponding to the Internet device to be analyzed (the local second description data mapping feature representation corresponding to each key description data segmentation word);

[0102] The local second description data mapping feature representation and the first description data mapping feature representation corresponding to the Internet device to be analyzed are concatenated to form a target description data mapping feature representation corresponding to the Internet device to be analyzed.

[0103] Optionally, in some implementations, the above step S300 may further include the following specific sub-steps:

[0104] Extracting target device type information corresponding to each of the Internet devices to be analyzed;

[0105] According to the corresponding target device type information, a plurality of the Internet devices to be analyzed are subjected to device classification processing to form at least one device classification cluster corresponding to the plurality of Internet devices to be analyzed, wherein the target device type information corresponding to the Internet devices to be analyzed included in each of the device classification clusters belongs to the same device type information, and the target device type information corresponding to the Internet devices to be analyzed included in different device classification clusters belongs to different device type information;

[0106] Based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the Internet devices to be analyzed included in the device classification cluster within each of the device classification clusters to output target device status information corresponding to each of the Internet devices to be analyzed included in the device classification cluster.

[0107] Optionally, in some implementations, the step of performing device status analysis on the Internet devices to be analyzed included in the device classification cluster based on the target description data mapping feature representation corresponding to each Internet device to be analyzed within each device classification cluster to output target device status information corresponding to each Internet device to be analyzed included in the device classification cluster may further include the following specific sub-steps:

[0108] For any device classification cluster, count the number of Internet devices to be analyzed included in the device classification cluster to output the device statistics corresponding to the device classification cluster;

[0109] Based on a pre-configured reference classification ratio (the specific value of the reference classification ratio is not limited and can be configured according to actual application requirements) and the device statistical quantity, a multiplication operation is performed to output a corresponding first device statistical quantity, and for each Internet device to be analyzed in the device classification cluster, a feature representation matching degree (such as cosine similarity) between a target description data mapping feature representation corresponding to the Internet device to be analyzed and a target description data mapping feature representation corresponding to each other Internet device to be analyzed in the device classification cluster is calculated, and then from the feature representation matching degrees between the target description data mapping feature representation corresponding to the Internet device to be analyzed and the target description data mapping feature representation corresponding to each other Internet device to be analyzed in the device classification cluster, the largest feature representation matching degrees of the first device statistical quantity are selected as a first feature representation matching degree set corresponding to the Internet device to be analyzed, and the first feature representation matching degrees included in the first feature representation matching degree set are averaged to obtain an average first feature representation matching degree corresponding to the Internet device to be analyzed;

[0110] Based on the matching degree of the mean first feature representation corresponding to each Internet device to be analyzed in the device classification cluster, device status analysis processing is performed on each Internet device to be analyzed included in the device classification cluster to output target device status information corresponding to each Internet device to be analyzed included in the device classification cluster (exemplarily, for each of the device classification clusters, the target device status information of the Internet device to be analyzed whose corresponding mean first feature representation matching degree in the device classification cluster is too large or too small can be configured as abnormal, wherein the determination method of being too large or too small can be compared with the average value).

[0111] Combination Figure 3 The embodiment of the present invention further provides an Internet-based status analysis system, which can be applied to the above-mentioned Internet-based status analysis platform. The Internet-based status analysis system may include the following software function modules:

[0112] a state description data prediction module, configured to predict, for each Internet device to be analyzed, predicted device state description data corresponding to a device state data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the predicted device state description data belongs to the predicted device state description data of the device state data cluster to be analyzed in a first data dimension, and there are multiple Internet devices to be analyzed, and the device state data cluster to be analyzed includes multiple pieces of device state data to be analyzed;

[0113] A feature mapping processing module, used to perform feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed, so as to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed;

[0114] The device status analysis module is used to perform device status analysis on multiple Internet devices to be analyzed based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, so as to output target device status information corresponding to each of the Internet devices to be analyzed, and the target device status information is used to reflect whether the status of the corresponding Internet device to be analyzed is abnormal.

[0115] Optionally, in some embodiments, the feature mapping processing module is specifically used to:

[0116] Performing text segmentation processing on the predicted device state description data corresponding to the Internet device to be analyzed, so as to output a plurality of description data segmentation words corresponding to the predicted device state description data;

[0117] The predicted device state description data and the description data segmentation terms are subjected to feature mapping processing, and a target description data mapping feature representation corresponding to the Internet device to be analyzed is output.

[0118] Optionally, in some implementations, the device status analysis module is specifically used to:

[0119] Extracting target device type information corresponding to each of the Internet devices to be analyzed;

[0120] According to the corresponding target device type information, a plurality of the Internet devices to be analyzed are subjected to device classification processing to form at least one device classification cluster corresponding to the plurality of Internet devices to be analyzed, wherein the target device type information corresponding to the Internet devices to be analyzed included in each of the device classification clusters belongs to the same device type information, and the target device type information corresponding to the Internet devices to be analyzed included in different device classification clusters belongs to different device type information;

[0121] Based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the Internet devices to be analyzed included in the device classification cluster within each of the device classification clusters to output target device status information corresponding to each of the Internet devices to be analyzed included in the device classification cluster.

[0122] In summary, the present invention provides an Internet-based state analysis method and system, which predicts, for each Internet device to be analyzed, predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed; performs feature mapping processing on the predicted device state description data corresponding to each of the multiple Internet devices to be analyzed, respectively, to output a target description data mapping feature representation corresponding to each Internet device to be analyzed; performs device state analysis processing on the multiple Internet devices to be analyzed based on the target description data mapping feature representation corresponding to each Internet device to be analyzed, to output the target device state information corresponding to each Internet device to be analyzed. Based on this, since the predicted device state description data is determined first, compared with the conventional technical means of directly performing device state analysis processing based on the device state data cluster to be analyzed, the analysis basis can be more concise and have less invalid information, thereby improving the accuracy of device state analysis to a certain extent, thereby improving the problem of low accuracy in the prior art.

[0123] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A state analysis method based on the Internet, characterized in that: include: For each Internet device to be analyzed, predict the predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the predicted device state description data belongs to the predicted device state description data of the device state data cluster to be analyzed in the first data dimension, and there are multiple Internet devices to be analyzed, and the device state data cluster to be analyzed includes multiple pieces of device state data to be analyzed; Performing feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed, respectively, to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed; According to the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the multiple Internet devices to be analyzed to output target device status information corresponding to each of the Internet devices to be analyzed, wherein the target device status information is used to reflect whether the status of the corresponding Internet device to be analyzed is abnormal; The step of predicting, for each Internet device to be analyzed, predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed comprises: Extracting a device status data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the device status data cluster to be analyzed includes device status data to be analyzed corresponding to at least one data dimension other than the first data dimension; A first state data analysis neural network corresponding to a first data dimension is extracted; the first state data analysis neural network performs network optimization on a second state data analysis neural network based on the first exemplary state data to form a network; the first exemplary state data includes a first exemplary device state data cluster to be analyzed and corresponding exemplary device state description data in the first data dimension, and the first exemplary device state data cluster to be analyzed includes the device state data to be analyzed corresponding to the first data dimension; the second state data analysis neural network performs network optimization on a third state data analysis neural network based on the second exemplary state data; the second exemplary state data includes a combination of second exemplary device state data clusters to be analyzed and corresponding exemplary device state description data corresponding to a plurality of exemplary data dimensions; The device state data cluster to be analyzed is loaded, so as to analyze and output the predicted device state description data of the Internet device to be analyzed in the first data dimension by using the first state data analysis neural network; The step of performing device status analysis on a plurality of Internet devices to be analyzed based on the target description data mapping feature representation corresponding to each Internet device to be analyzed to output target device status information corresponding to each Internet device to be analyzed includes: Extracting target device type information corresponding to each of the Internet devices to be analyzed; According to the corresponding target device type information, a plurality of the Internet devices to be analyzed are subjected to device classification processing to form at least one device classification cluster corresponding to the plurality of Internet devices to be analyzed, wherein the target device type information corresponding to the Internet devices to be analyzed included in each of the device classification clusters belongs to the same device type information, and the target device type information corresponding to the Internet devices to be analyzed included in different device classification clusters belongs to different device type information; Based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the Internet devices to be analyzed included in the device classification cluster within each of the device classification clusters to output target device status information corresponding to each of the Internet devices to be analyzed included in the device classification cluster.

2. The Internet-based status analysis method according to claim 1, characterized in that: The step of performing feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed comprises: Performing text segmentation processing on the predicted device state description data corresponding to the Internet device to be analyzed, so as to output a plurality of description data segmentation words corresponding to the predicted device state description data; The predicted device state description data and the description data segmentation terms are subjected to feature mapping processing, and a target description data mapping feature representation corresponding to the Internet device to be analyzed is output.

3. The Internet-based status analysis method according to claim 2, characterized in that: The step of performing feature mapping processing on the predicted device state description data and the description data segmentation terms, and outputting the target description data mapping feature representation corresponding to the Internet device to be analyzed, comprises: Performing feature mapping processing on the predicted device state description data to output a first description data mapping feature representation corresponding to the Internet device to be analyzed; Performing feature mapping processing on each of the description data segmentation words respectively to output a word mapping feature representation corresponding to each of the description data segmentation words, and sequentially splicing the word mapping feature representation corresponding to each of the description data segmentation words according to the distribution position information of the corresponding description data segmentation words in the predicted device state description data to form a second description data mapping feature representation corresponding to the Internet device to be analyzed; The first description data mapping feature representation corresponding to the Internet device to be analyzed and the second description data mapping feature representation corresponding to the Internet device to be analyzed are aggregated to form a target description data mapping feature representation corresponding to the Internet device to be analyzed.

4. The Internet-based status analysis method according to claim 3, characterized in that: The step of aggregating the first description data mapping feature representation corresponding to the Internet device to be analyzed and the second description data mapping feature representation corresponding to the Internet device to be analyzed to form a target description data mapping feature representation corresponding to the Internet device to be analyzed includes: Determine at least one description data segmentation word from the multiple description data segmentation words as the corresponding key description data segmentation word, and select a local second description data mapping feature representation corresponding to the key description data segmentation word from the second description data mapping feature representation corresponding to the Internet device to be analyzed; The local second description data mapping feature representation and the first description data mapping feature representation corresponding to the Internet device to be analyzed are concatenated to form a target description data mapping feature representation corresponding to the Internet device to be analyzed.

5. The Internet-based status analysis method according to claim 1, characterized in that: The step of performing device status analysis processing on the Internet devices to be analyzed included in the device classification cluster based on the target description data mapping feature representation corresponding to each Internet device to be analyzed within each device classification cluster to output target device status information corresponding to each Internet device to be analyzed included in the device classification cluster includes: For any device classification cluster, count the number of Internet devices to be analyzed included in the device classification cluster to output the device statistics corresponding to the device classification cluster; Based on a pre-configured reference classification ratio and the device statistical quantity, a multiplication operation is performed to output a corresponding first device statistical quantity, and for each Internet device to be analyzed in the device classification cluster, a feature representation matching degree between a target description data mapping feature representation corresponding to the Internet device to be analyzed and a target description data mapping feature representation corresponding to each other Internet device to be analyzed in the device classification cluster is calculated, and then from the feature representation matching degrees between the target description data mapping feature representation corresponding to the Internet device to be analyzed and the target description data mapping feature representation corresponding to each other Internet device to be analyzed in the device classification cluster, the largest feature representation matching degree of the first device statistical quantity is selected as a first feature representation matching degree set corresponding to the Internet device to be analyzed, and the first feature representation matching degrees included in the first feature representation matching degree set are averaged to obtain an average first feature representation matching degree corresponding to the Internet device to be analyzed; Based on the mean first feature representation matching degree corresponding to each Internet device to be analyzed in the device classification cluster, device status analysis processing is performed on each Internet device to be analyzed included in the device classification cluster to output target device status information corresponding to each Internet device to be analyzed included in the device classification cluster.

6. A status analysis system based on the Internet, characterized in that: include: a state description data prediction module, configured to predict, for each Internet device to be analyzed, predicted device state description data corresponding to a device state data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the predicted device state description data belongs to the predicted device state description data of the device state data cluster to be analyzed in a first data dimension, and there are multiple Internet devices to be analyzed, and the device state data cluster to be analyzed includes multiple pieces of device state data to be analyzed; A feature mapping processing module, used to perform feature mapping processing on the predicted device state description data corresponding to each of the plurality of Internet devices to be analyzed, so as to output a mapping feature representation of the target description data corresponding to each of the Internet devices to be analyzed; A device status analysis module, configured to perform device status analysis on a plurality of Internet devices to be analyzed according to a target description data mapping feature representation corresponding to each Internet device to be analyzed, so as to output target device status information corresponding to each Internet device to be analyzed, wherein the target device status information is used to reflect whether the status of the corresponding Internet device to be analyzed is abnormal; The step of predicting, for each Internet device to be analyzed, predicted device state description data corresponding to the device state data cluster to be analyzed corresponding to the Internet device to be analyzed includes: Extracting a device status data cluster to be analyzed corresponding to the Internet device to be analyzed, wherein the device status data cluster to be analyzed includes device status data to be analyzed corresponding to at least one data dimension other than the first data dimension; A first state data analysis neural network corresponding to a first data dimension is extracted; the first state data analysis neural network performs network optimization on a second state data analysis neural network based on the first exemplary state data to form a network; the first exemplary state data includes a first exemplary device state data cluster to be analyzed and corresponding exemplary device state description data in the first data dimension, and the first exemplary device state data cluster to be analyzed includes the device state data to be analyzed corresponding to the first data dimension; the second state data analysis neural network performs network optimization on a third state data analysis neural network based on the second exemplary state data; the second exemplary state data includes a combination of second exemplary device state data clusters to be analyzed and corresponding exemplary device state description data corresponding to a plurality of exemplary data dimensions; The device state data cluster to be analyzed is loaded, so as to analyze and output the predicted device state description data of the Internet device to be analyzed in the first data dimension by using the first state data analysis neural network; The device status analysis process is performed on the plurality of Internet devices to be analyzed according to the target description data mapping feature representation corresponding to each Internet device to be analyzed, so as to output the target device status information corresponding to each Internet device to be analyzed, including: Extracting target device type information corresponding to each of the Internet devices to be analyzed; According to the corresponding target device type information, a plurality of the Internet devices to be analyzed are subjected to device classification processing to form at least one device classification cluster corresponding to the plurality of Internet devices to be analyzed, wherein the target device type information corresponding to the Internet devices to be analyzed included in each of the device classification clusters belongs to the same device type information, and the target device type information corresponding to the Internet devices to be analyzed included in different device classification clusters belongs to different device type information; Based on the target description data mapping feature representation corresponding to each of the Internet devices to be analyzed, device status analysis processing is performed on the Internet devices to be analyzed included in the device classification cluster within each of the device classification clusters to output target device status information corresponding to each of the Internet devices to be analyzed included in the device classification cluster.

7. The Internet-based status analysis system according to claim 6, characterized in that: The feature mapping processing module is specifically used for: Performing text segmentation processing on the predicted device state description data corresponding to the Internet device to be analyzed, so as to output a plurality of description data segmentation words corresponding to the predicted device state description data; The predicted device state description data and the description data segmentation terms are subjected to feature mapping processing, and a target description data mapping feature representation corresponding to the Internet device to be analyzed is output.

Images