Information health assessment method, device, equipment and storage medium

Abstract

The present invention relates to the field of artificial intelligence technology, and discloses an information health assessment method, device, equipment and storage medium for improving the accuracy of information health assessment. The information health assessment method includes: performing multi-element classification processing on the information information of the subject to be assessed to obtain at least one original element category information corresponding to the information information; performing sentiment tendency classification on the information information according to a single original element category information to obtain the sentiment tendency classification probability corresponding to each original element category information, and determining at least one target element category information, performing information health calculation on the sentiment tendency classification probability corresponding to each target element category information, obtaining the element information health corresponding to each target element category information, and determining the comprehensive information health of the subject to be assessed. In addition, the present invention also relates to blockchain technology, and the comprehensive information health can be stored in a blockchain node.

Description

Technical Field

[0001] The present invention relates to the field of natural language processing technology, and in particular to an information health assessment method, device, equipment and storage medium. Background Art

[0002] With the development of Internet technology, online information has exploded. The content of information often revolves around a certain subject, such as comments on a certain network event, assessment descriptions of a certain enterprise's capabilities, etc. Through these information, the information health of the description subject can be evaluated.

[0003] The existing technology usually implements the information health of a subject through text-based sentiment classification. This method can only determine whether a text is positive or negative. The lack of a multi-dimensional evaluation method leads to inaccurate assessment of information health. Summary of the invention

[0004] The present invention provides an information health evaluation method, device, equipment and storage medium for improving the accuracy of information health evaluation.

[0005] A first aspect of the present invention provides an information health assessment method, comprising:

[0006] Obtain information about the subject to be evaluated;

[0007] Perform multi-element classification processing on the information through a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information;

[0008] Based on at least one original element category information corresponding to the information information, classify the sentiment tendency of a single original element category information of the information information to obtain a sentiment tendency classification probability corresponding to each original element category information;

[0009] Determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and calculate the information health of the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information;

[0010] According to the element information health corresponding to each target element category information, the comprehensive information health of the subject to be evaluated is determined.

[0011] Optionally, in a first implementation of the first aspect of the present invention, the multi-element classification model includes label classification models corresponding to a plurality of elements respectively, and the multi-element classification processing of the information information by the pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information information includes:

[0012] Performing label classification processing on the information through the label classification model corresponding to each element to obtain classification label information corresponding to each element, where the classification label information corresponding to each element is used to indicate the probability that the information belongs to the corresponding element;

[0013] By using the classification label information corresponding to the multiple elements, it is determined whether the probability that the information belongs to the corresponding element is greater than a preset probability threshold;

[0014] If the probability that the information belongs to the corresponding element is greater than a preset probability threshold, the corresponding element is determined as the original element category information corresponding to the information, and at least one original element category information corresponding to the information is obtained.

[0015] Optionally, in a second implementation of the first aspect of the present invention, the label classification model corresponding to each element includes an input layer, a convolution layer, a pooling layer, and a first fully connected layer, and the label classification model corresponding to each element is used to perform label classification processing on the news information to obtain classification label information corresponding to each element, including:

[0016] The information is segmented through the input layer of the label classification model corresponding to each element to obtain the segmentation results of each label classification model, and the segmentation results of each label classification model are vectorized to obtain a sparse matrix of each label classification model;

[0017] The element feature extraction of the sparse matrix of each label classification model is performed through the convolution layer in each label classification model to obtain the element feature information of each label classification model;

[0018] The element feature information of each label classification model is processed with a fixed length through the pooling layer in each label classification model to obtain the fixed length feature information of each label classification model;

[0019] The fixed-length feature information of each label classification model is label classified through the first fully connected layer in each label classification model to obtain the classification label information corresponding to each element.

[0020] Optionally, in a third implementation of the first aspect of the present invention, based on at least one original element category information corresponding to the information information, the information information is classified according to the sentiment tendency of a single original element category information to obtain the sentiment tendency classification probability corresponding to each original element category information, including:

[0021] Obtaining a sentiment tendency classification model corresponding to each original element category information, wherein the sentiment tendency classification model corresponding to each original element category information includes a knowledge distillation bidirectional encoding representation converter and a second fully connected layer;

[0022] Extracting sentiment tendency features of the information information through the knowledge distillation bidirectional encoding representation converter in the sentiment tendency classification model corresponding to each original element category information, obtaining sentiment tendency feature information of each original element category information;

[0023] The sentiment tendency feature information is subjected to sentiment tendency classification through the second fully connected layer in the sentiment tendency classification model corresponding to each original element category information, so as to obtain the sentiment tendency classification probability corresponding to each original element category information.

[0024] Optionally, in a fourth implementation of the first aspect of the present invention, determining at least one target element category information by the sentiment tendency classification probability corresponding to each original element category information, and performing information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information, including:

[0025] The sentiment tendency classification probability corresponding to each original element category information is compared with a preset sentiment tendency classification probability threshold to obtain target element category information, wherein the target element category information is used to indicate the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold;

[0026] By using a preset score conversion formula, the sentiment tendency classification probability corresponding to the target element category information is converted into score data to obtain the element information healthiness corresponding to each target element category information.

[0027] Optionally, in a fifth implementation of the first aspect of the present invention, determining the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information includes:

[0028] Counting the number of all element category information to obtain the total number of elements, and calculating the total score of the elements according to the total number of elements;

[0029] Calculate the total score of the element information health corresponding to all target element category information to obtain the total score of the element category;

[0030] The total score of the element category and the total score of the element are averaged to obtain the comprehensive information health of the subject to be evaluated.

[0031] Optionally, in a sixth implementation of the first aspect of the present invention, the obtaining of information about the subject to be evaluated includes:

[0032] Acquire multiple original information pieces of information, each of which includes an information subject tag;

[0033] The information subjects of the multiple original information information are classified by the information subject labels to obtain the original information information corresponding to each information subject label, and the information information of the subject to be evaluated is extracted from the original information information corresponding to each information subject label.

[0034] A second aspect of the present invention provides an information health assessment device, comprising:

[0035] The acquisition module is used to obtain information about the subject to be evaluated;

[0036] A processing module, used to perform multi-element classification processing on the information through a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information;

[0037] A classification module, configured to classify the sentiment tendency of a single original element category information of the information information based on at least one original element category information corresponding to the information information, and obtain a sentiment tendency classification probability corresponding to each original element category information;

[0038] A calculation module is used to determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information;

[0039] The determination module is used to determine the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information.

[0040] Optionally, in a first implementation of the second aspect of the present invention, the multi-classification model includes label classification models corresponding to a plurality of elements respectively, and the processing module includes:

[0041] A classification unit, used to perform label classification processing on the information by using a label classification model corresponding to each element to obtain classification label information corresponding to each element, wherein the classification label information corresponding to each element is used to indicate the probability that the information belongs to the corresponding element;

[0042] A judgment unit, used to judge whether the probability that the information belongs to the corresponding element is greater than a preset probability threshold through the classification label information corresponding to the multiple elements;

[0043] A determination unit is used to determine the corresponding element as the original element category information corresponding to the information information if the probability that the information information belongs to the corresponding element is greater than a preset probability threshold, so as to obtain at least one original element category information corresponding to the information information.

[0044] Optionally, in a second implementation of the second aspect of the present invention, the label classification model corresponding to each element includes an input layer, a convolution layer, a pooling layer and a first fully connected layer, and the classification unit is specifically used to:

[0045] The information is segmented through the input layer of the label classification model corresponding to each element to obtain the segmentation results of each label classification model, and the segmentation results of each label classification model are vectorized to obtain a sparse matrix of each label classification model;

[0046] The element feature extraction of the sparse matrix of each label classification model is performed through the convolution layer in each label classification model to obtain the element feature information of each label classification model;

[0047] The element feature information of each label classification model is processed with a fixed length through the pooling layer in each label classification model to obtain the fixed length feature information of each label classification model;

[0048] The fixed-length feature information of each label classification model is label classified through the first fully connected layer in each label classification model to obtain the classification label information corresponding to each element.

[0049] Optionally, in a third implementation of the second aspect of the present invention, the classification module is specifically used to:

[0050] Obtaining a sentiment tendency classification model corresponding to each original element category information, wherein the sentiment tendency classification model corresponding to each original element category information includes a knowledge distillation bidirectional encoding representation converter and a second fully connected layer;

[0051] Extracting sentiment tendency features of the information information through the knowledge distillation bidirectional encoding representation converter in the sentiment tendency classification model corresponding to each original element category information, obtaining sentiment tendency feature information of each original element category information;

[0052] The sentiment tendency feature information is subjected to sentiment tendency classification through the second fully connected layer in the sentiment tendency classification model corresponding to each original element category information, so as to obtain the sentiment tendency classification probability corresponding to each original element category information.

[0053] Optionally, in a fourth implementation manner of the second aspect of the present invention, the calculation module is specifically used to:

[0054] The sentiment tendency classification probability corresponding to each original element category information is compared with a preset sentiment tendency classification probability threshold to obtain target element category information, wherein the target element category information is used to indicate the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold;

[0055] By using a preset score conversion formula, the sentiment tendency classification probability corresponding to the target element category information is converted into score data to obtain the element information healthiness corresponding to each target element category information.

[0056] Optionally, in a fifth implementation manner of the second aspect of the present invention, the determining module is specifically configured to:

[0057] Counting the number of all element category information to obtain the total number of elements, and calculating the total score of the elements according to the total number of elements;

[0058] Calculate the total score of the element information health corresponding to all target element category information to obtain the total score of the element category;

[0059] The total score of the element category and the total score of the element are averaged to obtain the comprehensive information health of the subject to be evaluated.

[0060] Optionally, in a sixth implementation of the second aspect of the present invention, the acquisition module is specifically used to:

[0061] Acquire multiple original information pieces of information, each of which includes an information subject tag;

[0062] The information subjects of the multiple original information information are classified by the information subject labels to obtain the original information information corresponding to each information subject label, and the information information of the subject to be evaluated is extracted from the original information information corresponding to each information subject label.

[0063] A third aspect of the present invention provides an information health assessment device, comprising: a memory and at least one processor, wherein the memory stores a computer program; the at least one processor calls the computer program in the memory so that the information health assessment device executes the above-mentioned information health assessment method.

[0064] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, which, when executed on a computer, enables the computer to execute the above-mentioned information health assessment method.

[0065] In the technical solution provided by the present invention, information information of a subject to be evaluated is obtained; multi-element classification processing is performed on the information information through a pre-trained multivariate classification model to obtain at least one original element category information corresponding to the information information; based on at least one original element category information corresponding to the information information, sentiment tendency classification is performed on the information information for a single original element category information to obtain a sentiment tendency classification probability corresponding to each original element category information; at least one target element category information is determined through the sentiment tendency classification probability corresponding to each original element category information, and information health calculation is performed on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information; based on the element information health corresponding to each target element category information, the comprehensive information health of the subject to be evaluated is determined. In an embodiment of the present invention, in order to conduct a multi-faceted evaluation of the information health of the subject to be evaluated, the information information is first classified into multiple elements through a multivariate classification model to determine at least one original element category information mainly described by the information information, and then the information information is classified into sentiment tendency classification of a single original element category information to determine the sentiment tendency of the information information in each main descriptive element, and finally, the proportion of the sentiment tendency of the main descriptive element in the overall sentiment tendency is calculated to obtain the comprehensive information health of the subject to be evaluated. The present invention can improve the accuracy of information health evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram of an embodiment of a method for evaluating information health in an embodiment of the present invention;

[0067] Figure 2 It is a schematic diagram of another embodiment of the method for evaluating information health in an embodiment of the present invention;

[0068] Figure 3 A schematic diagram of an embodiment of an information health evaluation device according to an embodiment of the present invention;

[0069] Figure 4 It is a schematic diagram of another embodiment of the information health evaluation device in the embodiment of the present invention;

[0070] Figure 5 FIG. 4 is a schematic diagram of an embodiment of an information health evaluation device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0071] The embodiments of the present invention provide an information health evaluation method, device, equipment and storage medium for improving the accuracy of information health evaluation.

[0072] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention 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 interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0073] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0074] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0075] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0076] It is understandable that the execution subject of the present invention can be an information health assessment device, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0077] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 An embodiment of the information health evaluation method in the embodiment of the present invention includes:

[0078] 101. Obtain information about the subject to be evaluated;

[0079] It should be noted that the information information of the subject to be evaluated is used to indicate information text information with the subject to be evaluated as the description center, such as information text information with a social event as the description center, information text information with a certain enterprise as the description center, etc. In one embodiment, the server reads the information information of the subject to be evaluated from the preset information library, so that the efficiency of obtaining information information is improved, thereby improving the efficiency of information health assessment.

[0080] In one embodiment, after the server obtains all the information within a preset time period, it uses a pre-trained evaluation subject identification model to identify the evaluation subject of all the information information, obtains the evaluation subject information corresponding to each piece of information, and then extracts the information information of the subject to be evaluated from all the information information based on the evaluation subject information corresponding to each piece of information. For example, if the information is "the products produced by Enterprise A are hot-selling in the market", then the evaluation subject information of the information information is "Enterprise A". Assuming that the subject to be evaluated is also "Enterprise A", then the information information is the information information of the subject to be evaluated. This embodiment can improve the efficiency and accuracy of obtaining the information information of the subject to be evaluated, thereby improving the efficiency and accuracy of information health assessment.

[0081] 102. Perform multi-element classification processing on the information through a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information;

[0082] In this embodiment, a pre-trained multivariate classification model is used to perform multi-element classification processing on information, and each element is used to indicate an indicator for measuring the information health of the subject to be evaluated. For example, assuming that the subject to be evaluated is "Enterprise A", then the multiple elements may include information health measurement indicators such as profitability, financing ability, bargaining power, investment expansion ability, and capacity utilization ability. Assuming that the subject to be evaluated is "Product A", then the multiple elements may include pricing ability, competitiveness of similar products, user repurchase ability, shipment ability, and functional stability ability. Information health measurement indicators, which are not specifically limited here. It should be noted that the multiple elements corresponding to the multivariate classification model are multiple preset elements, which are used to classify specific elements of information information, thereby improving the accuracy of element classification, and then improving the accuracy of information health assessment.

[0083] In one embodiment, the server performs multi-element classification processing on the information information through a pre-trained multi-element classification model to obtain the classification result of each element corresponding to the information information, and then determines at least one original element category information corresponding to the information information through the classification result of each element corresponding to the information information. For example, the multi-element of the multi-element classification model may include element a, element b, element c, element d, and element e. The multi-element classification processing is performed on the information information of "Enterprise A" through the multi-element classification model to obtain the classification result of each element. For example, the classification result of element a is "belongs to the classification", the classification result of element b is "does not belong to the classification", the classification result of element c is "does not belong to the classification", the classification result of element d is "belongs to the classification", and the classification result of element e is "does not belong to the classification". Then, it is finally determined that at least one original element category information corresponding to the information information includes "element a" and "element d", indicating that the description object of the information information is these two elements of "Enterprise A". Subsequently, the information health of these two elements of "Enterprise A" is evaluated through the information information, which can more accurately evaluate the information health of the subject to be evaluated in a certain aspect, thereby improving the accuracy of the information health evaluation.

[0084] 103. Based on at least one original element category information corresponding to the information information, classify the sentiment tendency of a single original element category information on the information information to obtain a sentiment tendency classification probability corresponding to each original element category information;

[0085] In this implementation, in order to improve the accuracy of information health assessment, the server obtains the sentiment tendency classification model corresponding to each original element category information, and classifies the sentiment tendency of each original element category information of the information through the sentiment tendency classification model corresponding to each original element category information, and obtains the sentiment tendency classification probability corresponding to each original element category information. Among them, the sentiment tendency classification probability is used to indicate the probability of the corresponding original element category information being biased towards positive sentiment and the probability of being biased towards negative sentiment. For example, at least one original element category information corresponding to the information information includes original element category information a, original element category information b and original element category information c, the sentiment tendency classification model A corresponding to the original element category information a, the sentiment tendency classification model B corresponding to the original element category information b, and the sentiment tendency classification model C corresponding to the original element category information c, then, the sentiment tendency classification model A, the sentiment tendency classification model B and the sentiment tendency classification model C are used to classify the sentiment tendency of the information information corresponding to the original element category information, respectively, and the classification result output by the sentiment tendency classification model A is determined as the sentiment tendency classification probability corresponding to the original element category information a, the classification result output by the sentiment tendency classification model B is determined as the sentiment tendency classification probability corresponding to the original element category information b, and the classification result output by the sentiment tendency classification model C is determined as the sentiment tendency classification probability corresponding to the original element category information c. This embodiment can accurately identify the sentiment tendency of a certain aspect of the information information based on the sentiment tendency characteristics of different elements, thereby improving the accuracy of the information health assessment.

[0086] In one embodiment, the support vector machine (SVM) model corresponding to each original element category information is trained by using sample information with element category information labels and sentiment tendency labels, and the SVM model parameters are corrected by the loss function to obtain the sentiment tendency classification model corresponding to each original element category information. This embodiment can improve the efficiency of sentiment tendency classification through a binary classification model, thereby improving the efficiency of information health assessment.

[0087] 104. Determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information;

[0088] In one embodiment, the emotion tendency classification probability corresponding to each original element category information is used to indicate the probability of the corresponding original element category information being biased toward positive emotion and the probability of being biased toward negative emotion. The server compares the probability of being biased toward positive emotion and the probability of being biased toward negative emotion in the emotion tendency classification probability to obtain the target emotion bias and the target emotion bias probability corresponding to each original element category information. The target emotion bias probability is used to indicate the emotion bias with the highest probability in the emotion tendency classification probability. For example, the emotion tendency classification probability corresponding to a certain original element category information has a probability of being biased toward positive emotion of 88% and a probability of being biased toward negative emotion of 9%. Since 88%>9%, the target emotion bias corresponding to the original element category information is biased toward positive emotion, and the target emotion bias probability is 88%. Next, the server determines the original element category information with a positive emotion bias as the target element category information through the target emotion bias corresponding to each original element category information. The target element category information includes at least one original element category information. Then, according to the target emotion bias probability corresponding to each target element category information, the element information health corresponding to each target element category information is calculated. For example, if the target emotion bias probability corresponding to a certain target element category information is 88%, then the element information health corresponding to the target element category information can be 100*88%=88 points on a 100-point system, or 20*88%=17.6 points on a 20-point system. The specifics are not limited here. This embodiment can accurately identify the factors of positive emotions in information information, so as to facilitate the subsequent determination of the proportion of positive emotion factors in the overall information information, thereby accurately determining the information health of the subject to be evaluated.

[0089] 105. Determine the comprehensive information health of the subject to be evaluated based on the element information health corresponding to each target element category information.

[0090] In one embodiment, the target element category information is used to indicate the original element category information that the probability of classifying the emotion tendency tending toward positive emotion is greater than the probability of classifying the emotion tendency tending toward negative emotion, that is, the target element category information is used to indicate the original element category information that tends toward positive emotion. Therefore, in order to determine the proportion of positive emotion elements in the overall information information, the server superimposes the element information health corresponding to each target element category information to obtain the target emotion score, and then calculates the total score based on the total number of elements, and calculates the ratio between the target emotion score and the total score to obtain the comprehensive information health of the subject to be evaluated. For example, assuming that the multivariate classification model includes 5 total elements when performing multi-element classification processing on information, and the scoring system of element information health is a percentage system, then the total score is 5*100=500 points. Assuming that the target element category information includes 2, the element information health corresponding to each target element category information is 88 and 90 respectively, 88+90=178 is the target sentiment score, and finally, 178 / 500=35.6% is the comprehensive information health of the subject to be evaluated, indicating that the proportion of positive sentiment factors in the information information of the subject to be evaluated is less than 50%, and negative sentiment factors dominate. This implementation method can accurately determine the proportion of positive sentiment factors in the information information of the subject to be evaluated, thereby accurately determining the information health of the subject to be evaluated.

[0091] Furthermore, the server stores the comprehensive information health in the blockchain database, which is not specifically limited here.

[0092] In an embodiment of the present invention, in order to conduct a multi-faceted evaluation of the information health of the subject to be evaluated, the information information is first classified into multiple elements through a multivariate classification model to determine at least one original element category information mainly described by the information information, and then the information information is classified into sentiment tendency classification of a single original element category information to determine the sentiment tendency of the information information in each main descriptive element, and finally, the proportion of the sentiment tendency of the main descriptive element in the overall sentiment tendency is calculated to obtain the comprehensive information health of the subject to be evaluated. The present invention can improve the accuracy of information health evaluation.

[0093] See also Figure 2 Another embodiment of the information health evaluation method in the embodiment of the present invention includes:

[0094] 201. Obtain information about the subject to be evaluated;

[0095] Specifically, step 201 includes: obtaining multiple original information information, each of which includes an information subject tag; classifying the multiple original information information by information subject through the information subject tag, obtaining the original information information corresponding to each information subject tag, and extracting the information information of the subject to be evaluated from the original information information corresponding to each information subject tag. In this embodiment, in order to improve the efficiency of obtaining information information of the subject to be evaluated, the server reads multiple original information information from the preset information library, and extracts the information information of the subject to be evaluated from the multiple original information information through the information subject tag in each original information information, thereby improving the efficiency of information health evaluation.

[0096] 202. Perform label classification processing on the information by using the label classification model corresponding to each element to obtain classification label information corresponding to each element, where the classification label information corresponding to each element is used to indicate the probability that the information belongs to the corresponding element;

[0097] Specifically, the multi-classification model includes label classification models corresponding to multiple elements, and the label classification model corresponding to each element includes an input layer, a convolution layer, a pooling layer and a first fully connected layer. Step 202 includes: performing word segmentation processing on the information through the input layer in the label classification model corresponding to each element to obtain the word segmentation results of each label classification model, and vectorizing the word segmentation results of each label classification model to obtain the sparse matrix of each label classification model; performing element feature extraction on the sparse matrix of each label classification model through the convolution layer in each label classification model to obtain the element feature information of each label classification model; performing fixed-length processing on the element feature information of each label classification model through the pooling layer in each label classification model to obtain the fixed-length feature information of each label classification model; performing label classification on the fixed-length feature information of each label classification model through the first fully connected layer in each label classification model to obtain the classification label information corresponding to each element.

[0098] In this embodiment, the multivariate classification model is composed of multiple label classification models, each of which is used to classify a single element of information. For example, the multivariate classification model includes label classification models corresponding to three elements: label classification model A corresponding to element a, label classification model B corresponding to element b, and label classification model C corresponding to element c. Then, label classification model A is used to classify element a of information, label classification model B is used to classify element b of information, and label classification model C is used to classify element c of information. Each label classification model includes an input layer, a convolution layer, a pooling layer, and a first fully connected layer. Since each label classification model corresponds to different elements, the sample data used in model training is also different, so the internal structures of the input layer, convolution layer, pooling layer, and first fully connected layer in each label classification model can be different. For example, the convolution layer in the label classification model corresponding to different elements is used to extract element feature information corresponding to different elements, which is not limited here.

[0099] In this implementation, since the lengths (number of words) of different information may be different, the server performs element feature extraction on the sparse matrix of each label classification model through the convolutional layer in each label classification model, and the lengths of the element feature information of each label classification model may also be different. Therefore, in order to avoid the inability of the first fully connected layer to label due to feature information of different lengths, the server performs fixed-length processing on the element feature information of each label classification model through the pooling layer in each label classification model to obtain fixed-length feature information of each label classification model. The fixed-length feature information of each label classification model has the same length and size, which reduces the error rate of subsequent label classification, thereby improving the accuracy of information health assessment.

[0100] 203. Judging, based on the classification label information corresponding to the multiple elements, whether the probability that the information belongs to the corresponding element is greater than a preset probability threshold;

[0101] It should be noted that, since the classification label information corresponding to each element is used to indicate the probability that the information information belongs to the corresponding element, in order to determine the element category to which the information information belongs, the server determines whether the probability that the information information belongs to the corresponding element is greater than the preset probability threshold through the classification label information corresponding to each element, thereby determining whether the information information belongs to the corresponding element. For example, the classification label information corresponding to multiple elements includes 70% of the classification label information corresponding to element a, 99% of the classification label information corresponding to element b, and 2% of the classification label information corresponding to element c. Assuming that the preset probability threshold is 82%, then only the classification label information 99% corresponding to element b is greater than 82%, then element b can be determined as the original element category information corresponding to the information information, and at least one original element category information corresponding to the information information is obtained. This embodiment can accurately determine which aspects of the subject to be evaluated are mainly described by the information information, thereby calculating the information health of the subject to be evaluated from different aspects, and improving the accuracy of the information health evaluation.

[0102] 204. If the probability that the information belongs to the corresponding element is greater than a preset probability threshold, the corresponding element is determined as the original element category information corresponding to the information, and at least one original element category information corresponding to the information is obtained;

[0103] It should be noted that if the probability that the information belongs to the corresponding element is greater than the preset probability threshold, it means that the information describes the corresponding element of the subject to be evaluated. Then, other elements described by non-information information can be left unprocessed to avoid introducing dirty data while reducing computing efficiency. Therefore, the server determines the elements whose probability of belonging to the corresponding element is greater than the preset probability threshold as original element category information, and obtains at least one original element category information corresponding to the information information, thereby improving the accuracy of information health assessment.

[0104] 205. Based on at least one original element category information corresponding to the information information, classify the sentiment tendency of a single original element category information on the information information to obtain a sentiment tendency classification probability corresponding to each original element category information;

[0105] Specifically, step 205 includes: obtaining a sentiment tendency classification model corresponding to each original element category information, wherein the sentiment tendency classification model corresponding to each original element category information includes a knowledge distillation bidirectional encoding representation converter and a second fully connected layer; extracting sentiment tendency features from information information through the knowledge distillation bidirectional encoding representation converter in the sentiment tendency classification model corresponding to each original element category information, and obtaining sentiment tendency feature information of each original element category information; classifying sentiment tendency feature information through the second fully connected layer in the sentiment tendency classification model corresponding to each original element category information, and obtaining the sentiment tendency classification probability corresponding to each original element category information. In this embodiment, the sentiment tendency classification model corresponding to each original element category information is a miniaturized and lightweight knowledge distillation bidirectional encoding representation conversion model (Distill Bert) based on a bidirectional encoding representation converter (Bidirectional Encoder Representation from Transformers, BERT), which reduces the volume by 40%, increases the speed by 60%, and maintains 97% of the language comprehension ability compared to BERT, so that the efficiency of sentiment tendency classification is higher while not being greatly distorted. Among them, the sentiment tendency classification model corresponding to each original element category information includes a knowledge distillation bidirectional encoding representation converter and a second fully connected layer. Knowledge distillation is a model compression technology. It teaches another model B through a trained model A. These two models are called teacher-student models. Usually, model A has more parameters and stronger performance. With the help of model A, model B can break through itself and learn better. The bottom layer of the sentiment tendency classification model corresponding to each original element category information uses distilled Bert to represent semantics, and the second layer of full connection performs Softmax classification, which finally realizes sentiment tendency classification and obtains the sentiment tendency classification probability corresponding to each original element category information.

[0106] 206. Determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information;

[0107] Specifically, step 206 includes: comparing the sentiment tendency classification probability corresponding to each original element category information with the preset sentiment tendency classification probability threshold, obtaining the target element category information, the target element category information is used to indicate the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold; converting the sentiment tendency classification probability corresponding to the target element category information into score data through the preset score conversion formula, and obtaining the element information health degree corresponding to each target element category information. In this embodiment, in order to obtain the element category information of the target sentiment tendency, the server compares the sentiment tendency classification probability corresponding to each original element category information with the preset sentiment tendency classification probability threshold, thereby extracting the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold, obtaining the target element category information, and then converting the sentiment tendency classification probability corresponding to the target element category information into score data through the preset score conversion formula, and obtaining the element information health degree corresponding to each target element category information, wherein the preset score conversion formula can be a percentage conversion formula, or a conversion formula of different scoring standards such as a twenty-point conversion formula, which is not specifically limited here. This embodiment can determine the proportion of information information containing the target emotional tendency through the emotional tendency classification probability, thereby accurately determining the health of the information.

[0108] 207. Determine the comprehensive information health of the subject to be evaluated based on the element information health corresponding to each target element category information.

[0109] Specifically, step 207 includes: counting the number of all element category information to obtain the total number of elements, and calculating the total score of the elements through the total number of elements; calculating the total score of the element information health corresponding to all target element category information to obtain the total score of the element category; averaging the total score of the element category and the total score of the element to obtain the comprehensive information health of the subject to be evaluated. In this embodiment, in order to calculate the proportion of the target emotional tendency in the information element category in the emotional tendency of the overall element category, the server first counts the total number of elements and calculates the total score of the elements, then calculates the total score of the element category of the element information health corresponding to all target element category information, and finally divides the total score of the element category by the total score of the element to obtain the comprehensive information health of the subject to be evaluated. For example, the total number of elements is 5, and each is scored on a percentage basis (determined by the preset score conversion formula). Then, the total score of the elements is 5*100=500. Assuming that the element information health corresponding to the two target element category information is 88 and 70 respectively, then the total score of the element category is 88+70=158. Finally, 158 / 500=31.6 is the comprehensive information health of the subject to be evaluated.

[0110] In an embodiment of the present invention, in order to conduct a multi-faceted evaluation of the information health of the subject to be evaluated, the information information is first classified into multiple elements through a label classification model corresponding to multiple elements in a multi-classification model, thereby determining at least one original element category information mainly described by the information information, and then the information information is classified into sentiment tendency classification for a single original element category information, and the sentiment tendency of the information information in each main descriptive element is determined. Finally, the proportion of the sentiment tendency of the main descriptive element in the overall sentiment tendency is calculated to obtain the comprehensive information health of the subject to be evaluated. The present invention can improve the accuracy of information health evaluation.

[0111] The above describes the information health evaluation method in the embodiment of the present invention. The following describes the information health evaluation device in the embodiment of the present invention. Figure 3 In one embodiment of the present invention, an information health evaluation device includes:

[0112] The acquisition module 301 is used to acquire information of the subject to be evaluated;

[0113] The processing module 302 is used to perform multi-element classification processing on the information by using a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information;

[0114] A classification module 303 is used to classify the sentiment tendency of a single original element category information of the information information based on at least one original element category information corresponding to the information information, and obtain the sentiment tendency classification probability corresponding to each original element category information;

[0115] The calculation module 304 is used to determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information;

[0116] The determination module 305 is used to determine the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information.

[0117] Furthermore, the comprehensive information health is stored in the blockchain database, which is not specifically limited here.

[0118] In an embodiment of the present invention, in order to conduct a multi-faceted evaluation of the information health of the subject to be evaluated, the information information is first classified into multiple elements through a multivariate classification model to determine at least one original element category information mainly described by the information information, and then the information information is classified into sentiment tendency classification of a single original element category information to determine the sentiment tendency of the information information in each main descriptive element, and finally, the proportion of the sentiment tendency of the main descriptive element in the overall sentiment tendency is calculated to obtain the comprehensive information health of the subject to be evaluated. The present invention can improve the accuracy of information health evaluation.

[0119] See also Figure 4 Another embodiment of the information health evaluation device in the embodiment of the present invention includes:

[0120] The acquisition module 301 is used to acquire information of the subject to be evaluated;

[0121] The processing module 302 is used to perform multi-element classification processing on the information by using a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information;

[0122] A classification module 303 is used to classify the sentiment tendency of a single original element category information of the information information based on at least one original element category information corresponding to the information information, and obtain the sentiment tendency classification probability corresponding to each original element category information;

[0123] The calculation module 304 is used to determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information;

[0124] The determination module 305 is used to determine the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information.

[0125] Optionally, the multi-classification model includes label classification models corresponding to a plurality of elements respectively, and the processing module 302 includes:

[0126] The classification unit 3021 is used to perform label classification processing on the information by using the label classification model corresponding to each element to obtain classification label information corresponding to each element, and the classification label information corresponding to each element is used to indicate the probability that the information belongs to the corresponding element;

[0127] A judgment unit 3022, configured to judge whether the probability that the information belongs to a corresponding element is greater than a preset probability threshold value based on the classification label information corresponding to the multiple elements;

[0128] The determination unit 3023 is used to determine the corresponding element as the original element category information corresponding to the information information if the probability that the information information belongs to the corresponding element is greater than a preset probability threshold, and obtain at least one original element category information corresponding to the information information.

[0129] Optionally, the label classification model corresponding to each element includes an input layer, a convolution layer, a pooling layer and a first fully connected layer, and the classification unit 3021 is specifically used for:

[0130] The information is segmented through the input layer of the label classification model corresponding to each element to obtain the segmentation results of each label classification model, and the segmentation results of each label classification model are vectorized to obtain a sparse matrix of each label classification model;

[0131] The element feature extraction of the sparse matrix of each label classification model is performed through the convolution layer in each label classification model to obtain the element feature information of each label classification model;

[0132] The element feature information of each label classification model is processed with a fixed length through the pooling layer in each label classification model to obtain the fixed length feature information of each label classification model;

[0133] The fixed-length feature information of each label classification model is label classified through the first fully connected layer in each label classification model to obtain the classification label information corresponding to each element.

[0134] Optionally, the classification module 303 is specifically used for:

[0135] Obtaining a sentiment tendency classification model corresponding to each original element category information, wherein the sentiment tendency classification model corresponding to each original element category information includes a knowledge distillation bidirectional encoding representation converter and a second fully connected layer;

[0136] Extracting sentiment tendency features of the information information through the knowledge distillation bidirectional encoding representation converter in the sentiment tendency classification model corresponding to each original element category information, obtaining sentiment tendency feature information of each original element category information;

[0137] The sentiment tendency feature information is subjected to sentiment tendency classification through the second fully connected layer in the sentiment tendency classification model corresponding to each original element category information, so as to obtain the sentiment tendency classification probability corresponding to each original element category information.

[0138] Optionally, the calculation module 304 is specifically used for:

[0139] The sentiment tendency classification probability corresponding to each original element category information is compared with a preset sentiment tendency classification probability threshold to obtain target element category information, wherein the target element category information is used to indicate the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold;

[0140] By using a preset score conversion formula, the sentiment tendency classification probability corresponding to the target element category information is converted into score data to obtain the element information healthiness corresponding to each target element category information.

[0141] Optionally, the determining module 305 is specifically configured to:

[0142] Counting the number of all element category information to obtain the total number of elements, and calculating the total score of the elements according to the total number of elements;

[0143] Calculate the total score of the element information health corresponding to all target element category information to obtain the total score of the element category;

[0144] The total score of the element category and the total score of the element are averaged to obtain the comprehensive information health of the subject to be evaluated.

[0145] Optionally, the acquisition module 301 is specifically used for:

[0146] Acquire multiple original information pieces of information, each of which includes an information subject tag;

[0147] The information subjects of the multiple original information information are classified by the information subject labels to obtain the original information information corresponding to each information subject label, and the information information of the subject to be evaluated is extracted from the original information information corresponding to each information subject label.

[0148] In an embodiment of the present invention, in order to conduct a multi-faceted evaluation of the information health of the subject to be evaluated, the information information is first classified into multiple elements through a label classification model corresponding to multiple elements in a multi-classification model, thereby determining at least one original element category information mainly described by the information information, and then the information information is classified into sentiment tendency classification for a single original element category information, and the sentiment tendency of the information information in each main descriptive element is determined. Finally, the proportion of the sentiment tendency of the main descriptive element in the overall sentiment tendency is calculated to obtain the comprehensive information health of the subject to be evaluated. The present invention can improve the accuracy of information health evaluation.

[0149] above Figure 3 and Figure 4 The information health evaluation device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The information health evaluation device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0150] Figure 5 1 is a schematic diagram of the structure of an information health assessment device provided by an embodiment of the present invention. The information health assessment device 500 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 510 (for example, one or more processors) and a memory 520, and one or more storage media 530 (for example, one or more mass storage devices) storing application programs 533 or data 532. Among them, the memory 520 and the storage medium 530 can be temporary storage or permanent storage. The program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a series of computer program operations in the information health assessment device 500. Furthermore, the processor 510 may be configured to communicate with the storage medium 530 to execute a series of computer program operations in the storage medium 530 on the information health assessment device 500.

[0151] The information health evaluation device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and / or one or more operating systems 531, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 5 The illustrated structure of the information health assessment device does not constitute a limitation on the information health assessment device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0152] The present invention also provides a computer device, which includes a memory and a processor, wherein a computer-readable computer program is stored in the memory, and when the computer-readable computer program is executed by the processor, the processor executes the steps of the information health assessment method in the above-mentioned embodiments.

[0153] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program runs on a computer, the computer executes the steps of the information health assessment method.

[0154] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0155] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0156] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0157] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several computer programs to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0158] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating information health, characterized in that: The information health assessment method includes: Obtain information about the subject to be evaluated; Perform multi-element classification processing on the information through a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information; Based on at least one original element category information corresponding to the information information, classify the sentiment tendency of a single original element category information of the information information to obtain a sentiment tendency classification probability corresponding to each original element category information; Determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information; Determine the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information; The multi-element classification model includes label classification models corresponding to multiple elements respectively. The multi-element classification processing of the information information by the pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information information includes: Performing label classification processing on the information through the label classification model corresponding to each element to obtain classification label information corresponding to each element, where the classification label information corresponding to each element is used to indicate the probability that the information belongs to the corresponding element; By using the classification label information corresponding to the multiple elements, it is determined whether the probability that the information belongs to the corresponding element is greater than a preset probability threshold; If the probability that the information belongs to the corresponding element is greater than a preset probability threshold, the corresponding element is determined as the original element category information corresponding to the information, and at least one original element category information corresponding to the information is obtained; The method of determining at least one target element category information by the sentiment tendency classification probability corresponding to each original element category information, and calculating the information health of each target element category information for the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health of each target element category information includes: The sentiment tendency classification probability corresponding to each original element category information is compared with a preset sentiment tendency classification probability threshold to obtain target element category information, wherein the target element category information is used to indicate the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold; The sentiment tendency classification probability corresponding to the target element category information is converted into score data through a preset score conversion formula to obtain the element information healthiness corresponding to each target element category information; Determining the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information includes: Counting the number of all element category information to obtain the total number of elements, and calculating the total score of the elements according to the total number of elements; Calculate the total score of the element information health corresponding to all target element category information to obtain the total score of the element category; The total score of the element category and the total score of the element are averaged to obtain the comprehensive information health of the subject to be evaluated.

2. The information health evaluation method according to claim 1, characterized in that: The label classification model corresponding to each element includes an input layer, a convolution layer, a pooling layer and a first fully connected layer. The label classification model corresponding to each element performs label classification processing on the information to obtain classification label information corresponding to each element, including: The information is segmented through the input layer of the label classification model corresponding to each element to obtain the segmentation results of each label classification model, and the segmentation results of each label classification model are vectorized to obtain a sparse matrix of each label classification model; The element feature extraction of the sparse matrix of each label classification model is performed through the convolution layer in each label classification model to obtain the element feature information of each label classification model; The element feature information of each label classification model is processed with a fixed length through the pooling layer in each label classification model to obtain the fixed length feature information of each label classification model; The fixed-length feature information of each label classification model is label classified through the first fully connected layer in each label classification model to obtain the classification label information corresponding to each element.

3. The information health evaluation method according to claim 1, characterized in that: The step of classifying the sentiment tendency of a single original element category information of the information based on at least one original element category information corresponding to the information information to obtain the sentiment tendency classification probability corresponding to each original element category information includes: Obtaining a sentiment tendency classification model corresponding to each original element category information, wherein the sentiment tendency classification model corresponding to each original element category information includes a knowledge distillation bidirectional encoding representation converter and a second fully connected layer; Extracting sentiment tendency features of the information information through the knowledge distillation bidirectional encoding representation converter in the sentiment tendency classification model corresponding to each original element category information, obtaining sentiment tendency feature information of each original element category information; The sentiment tendency feature information is subjected to sentiment tendency classification through the second fully connected layer in the sentiment tendency classification model corresponding to each original element category information, so as to obtain the sentiment tendency classification probability corresponding to each original element category information.

4. The information health evaluation method according to any one of claims 1 to 3, characterized in that: The obtaining of information about the subject to be evaluated includes: Acquire multiple original information pieces of information, each of which includes an information subject tag; The information subjects of the multiple original information information are classified by the information subject labels to obtain the original information information corresponding to each information subject label, and the information information of the subject to be evaluated is extracted from the original information information corresponding to each information subject label.

5. An information health evaluation device, characterized in that: The information health evaluation device includes: The acquisition module is used to obtain information about the subject to be evaluated; A processing module, used to perform multi-element classification processing on the information through a pre-trained multi-element classification model to obtain at least one original element category information corresponding to the information; A classification module, configured to classify the sentiment tendency of a single original element category information of the information information based on at least one original element category information corresponding to the information information, and obtain a sentiment tendency classification probability corresponding to each original element category information; A calculation module is used to determine at least one target element category information through the sentiment tendency classification probability corresponding to each original element category information, and perform information health calculation on the sentiment tendency classification probability corresponding to each target element category information to obtain the element information health corresponding to each target element category information; A determination module, used to determine the comprehensive information health of the subject to be evaluated according to the element information health corresponding to each target element category information; The multivariate classification model includes label classification models corresponding to a plurality of elements, and the processing module is specifically used to: perform label classification processing on the information information through the label classification model corresponding to each element to obtain classification label information corresponding to each element, and the classification label information corresponding to each element is used to indicate the probability that the information information belongs to the corresponding element; through the classification label information corresponding to the plurality of elements, determine whether the probability that the information information belongs to the corresponding element is greater than a preset probability threshold; if the probability that the information information belongs to the corresponding element is greater than the preset probability threshold, determine the corresponding element as the original element category information corresponding to the information information, and obtain at least one original element category information corresponding to the information information; The calculation module is specifically used to: compare the sentiment tendency classification probability corresponding to each original element category information with a preset sentiment tendency classification probability threshold to obtain target element category information, wherein the target element category information is used to indicate the original element category information whose sentiment tendency classification probability is greater than the preset sentiment tendency classification probability threshold; convert the sentiment tendency classification probability corresponding to the target element category information into score data through a preset score conversion formula to obtain the element information health degree corresponding to each target element category information; The determination module is specifically used to: count the number of all element category information to obtain the total number of elements, and calculate the total element score based on the total number of elements; calculate the total score of the element information health corresponding to all target element category information to obtain the element category total score; calculate the average score of the element category total score and the element total score to obtain the comprehensive information health of the subject to be evaluated.

6. An information health assessment device, characterized in that: The information health evaluation device includes: a memory and at least one processor, wherein the memory stores a computer program; The at least one processor calls the computer program in the memory to enable the information health assessment device to execute the information health assessment method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the information health evaluation method as described in any one of claims 1 to 4 is implemented.

Images