Enterprise Public Opinion Analysis Method, Device, Equipment and Storage Medium

Abstract

The present invention relates to artificial intelligence, and provides an enterprise public opinion analysis method, device, equipment and storage medium. The method identifies the topic popularity according to the topic collection source and topic information of enterprise topics, identifies the comment popularity according to the comment collection source and comment information, sorts the enterprise topics and topic comments, generates a topic comment list, divides the enterprise topics based on a text classification model to obtain topic indicators, performs sentiment prediction on the topic comments, obtains a negative sentiment score and generates an indicator score, extracts negative words to construct a negative word cloud map, so as to improve the accuracy of the public opinion result. In addition, the present invention also relates to blockchain technology, and the public opinion result can be stored in the blockchain.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular, to a method, device, equipment and storage medium for enterprise public opinion analysis. Background Art

[0002] With the rapid development of social networks, the spread and fermentation speed of online public opinion is very fast. In order to maintain the social image of enterprises, enterprises need to constantly pay attention to relevant public opinion information. In traditional public opinion monitoring solutions, it is usually necessary to rely on the experience of experts to analyze and judge the severity of public opinion fermentation. Due to the lack of a quantitative standard, it is impossible to accurately analyze the public opinion results of enterprises. Summary of the Invention

[0003] In view of the above, it is necessary to provide a method, device, equipment and storage medium for enterprise public opinion analysis, which can solve the technical problem of being unable to accurately analyze the public opinion results of enterprises.

[0004] On the one hand, the present invention proposes a method for enterprise public opinion analysis, and the method for enterprise public opinion analysis includes:

[0005] Collect the enterprise topics of the enterprise to be monitored and the topic comments of the enterprise topics;

[0006] Identify the topic heat of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic, and identify the comment heat of the topic comment according to the comment collection source of the topic comment and the comment information of the topic comment;

[0007] Sort the enterprise topics and the topic comments according to the topic heat and the comment heat to generate a topic comment list;

[0008] Divide the enterprise topics based on a pre-trained text classification model to obtain the topic indicators where the enterprise topics are located;

[0009] Perform sentiment prediction on the topic comments to obtain a negative sentiment score;

[0010] Generate index scores of the enterprise to be monitored on multiple topic indicators based on the comment heat corresponding to the topic comments of the enterprise topics in the topic indicators and the negative sentiment score;

[0011] Extract negative words from the topic comments according to the comment heat and the negative sentiment score to construct a negative word cloud map;

[0012] Determine the topic comment list, the index scores and the negative word cloud map as the public opinion results of the enterprise to be monitored.

[0013] According to a preferred embodiment of the present invention, the topic information includes the number of forwards, the number of likes, the number of reads of the enterprise topic, and the number of comments of the topic comments. Identifying the topic popularity of the enterprise topic based on the topic collection source of the enterprise topic and the topic information includes:

[0014] Obtain the source weight of the topic collection source, and obtain the information weight of the topic information on the topic collection source;

[0015] Perform a weighted sum operation on the number of forwards, the number of likes, the number of reads, and the number of comments based on the source weight and the information weight to obtain the topic popularity. The calculation formula for the topic popularity is: hot = k×(x 1 y 1 +x 2 y 2 +x 3 y 3 +x 4 y 4 ), y 1 >y 4 >y 2 >y 3 , where hot represents the topic popularity, k represents the source weight, x 1 represents the number of forwards, y 1 represents the first information weight corresponding to the number of forwards, x 2 represents the number of likes, y 2 represents the second information weight corresponding to the number of likes, x 3 represents the number of reads, y 3 represents the third information weight corresponding to the number of reads, x 4 represents the number of comments, y 4 represents the fourth information weight corresponding to the number of comments.

[0016] According to a preferred embodiment of the present invention, sorting the enterprise topic and the topic comments according to the topic popularity and the comment popularity to generate a topic comment list includes:

[0017] Select popular topics from the enterprise topics based on the topic popularity;

[0018] Sort the popular topics according to the topic popularity to obtain a first list;

[0019] Select popular comments corresponding to the popular topics from the topic comments based on the comment popularity;

[0020] Sort the popular comments according to the comment popularity to obtain a second list;

[0021] Construct the topic comment list according to the second list and the first list.

[0022] According to a preferred embodiment of the present invention, the text classification model includes a keyword extraction network, an encoding network, a convolutional network, and a mapping network. The enterprise topics are divided based on the pre-trained text classification model, and the topic indicators where the enterprise topics are located include:

[0023] Extract topic keywords from the enterprise topics based on the keyword extraction network;

[0024] Encode the topic keywords based on the encoding network to obtain a topic encoding vector;

[0025] Perform convolutional processing on the topic encoding vector based on the convolutional network to obtain a feature vector;

[0026] Perform mapping processing on the feature vector based on the mapping network to obtain a probability vector;

[0027] Identify the topic indicators according to the dimension where the element with the largest value in the probability vector is located.

[0028] According to a preferred embodiment of the present invention, the multiple topic indicators include multiple first-type indicators and second-type indicators, the indicator scores include a first score and a second score, and generating the indicator scores of the enterprise to be monitored on the multiple topic indicators based on the comment popularity and the negative sentiment score corresponding to the topic comments of the enterprise topics in the topic indicators includes:

[0029] Calculate the first score of the enterprise to be monitored on each first-type indicator according to the comment popularity and the negative sentiment score. The calculation formula for the first score is: score j = max(m, n - ∑ i∈{第j个第一类型指标的话题评论} hot i × sc i ), n > m, where score j represents the first score of the j-th first-type indicator, hot i represents the comment popularity of the i-th topic comment in the j-th first-type indicator, and sc i represents the negative sentiment score of the i-th topic comment in the j-th first-type indicator;

[0030] Calculate the total score of the enterprise to be monitored on the multiple first-type indicators according to the multiple first scores;

[0031] Identify the enterprise ranking of the enterprise to be monitored in the preset enterprise library according to the total score as the second score.

[0032] According to a preferred embodiment of the present invention, the obtaining of the negative sentiment score by performing sentiment prediction on the topic comment includes:

[0033] Performing encoding processing on the topic comment based on a preset sentiment dictionary to obtain a sentiment encoding vector;

[0034] Processing the sentiment encoding vector based on a pre-trained binary classification model to obtain an output probability;

[0035] Determine the probability value corresponding to the output probability in the negative dimension as the negative sentiment score.

[0036] According to a preferred embodiment of the present invention, the extracting of negative words from the topic comment according to the comment heat and the negative sentiment score to construct a negative word cloud map includes:

[0037] Selecting target comments from the topic comments according to the comment heat and the negative sentiment score;

[0038] Screening out the words that match successfully in the preset sentiment dictionary from the target comments as the negative words;

[0039] Generating a word label for the negative word according to the comment heat and the negative sentiment score of the target comment where the negative word is located;

[0040] Merging the word labels to obtain the negative word cloud map.

[0041] On the other hand, the present invention also proposes an enterprise public opinion analysis device, and the enterprise public opinion analysis device includes:

[0042] A collection unit, configured to collect enterprise topics of the enterprise to be monitored and topic comments of the enterprise topics;

[0043] An identification unit, configured to identify the topic heat of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic, and identify the comment heat of the topic comment according to the comment collection source of the topic comment and the comment information of the topic comment;

[0044] A generation unit, configured to sort the enterprise topics and the topic comments according to the topic heat and the comment heat, and generate a topic comment list;

[0045] A division unit, configured to divide the enterprise topic based on a pre-trained text classification model to obtain the topic index where the enterprise topic is located;

[0046] A prediction unit for performing sentiment prediction on the topic comments to obtain a negative sentiment score;

[0047] The generation unit is further configured to generate index scores of the enterprise to be monitored on multiple topic indexes based on the comment popularity corresponding to the topic comments of the enterprise topic in the topic indexes and the negative sentiment score;

[0048] A construction unit for extracting negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud diagram;

[0049] A determination unit for determining the topic comment list, the index score, and the negative word cloud diagram as the public opinion result of the enterprise to be monitored.

[0050] On the other hand, the present invention also proposes an electronic device, which includes:

[0051] A memory for storing computer-readable instructions; and

[0052] A processor for executing the computer-readable instructions stored in the memory to implement the enterprise public opinion analysis method.

[0053] On the other hand, the present invention also proposes a computer-readable storage medium, in which computer-readable instructions are stored, and the computer-readable instructions are executed by a processor in an electronic device to implement the enterprise public opinion analysis method.

[0054] It can be seen from the above technical solutions that the present invention can accurately identify the topic popularity by combining the topic collection source and the topic information, and can accurately identify the comment popularity by combining the comment collection source and the comment information, thereby improving the accuracy of the topic comment list. Further, through the comment popularity corresponding to the topic index and the corresponding negative sentiment score, the public opinion of all aspects of the enterprise to be monitored can be monitored from different multiple topic indexes, and the index score can be accurately generated. By combining the comment popularity and the negative sentiment score to extract negative words from the topic comments, the accuracy of constructing the negative word cloud diagram can be improved. The present invention can intuitively display the public opinion result of the enterprise to be monitored by constructing a topic comment list and a negative word cloud diagram. Description of the Drawings

[0055] Figure 1 is a flowchart of a preferred embodiment of the enterprise public opinion analysis method of the present invention.

[0056] Figure 2 is a functional module diagram of a preferred embodiment of the enterprise public opinion analysis device of the present invention.

[0057] Figure 3 It is a schematic structural diagram of an electronic device that is a preferred embodiment for implementing the enterprise public opinion analysis method of the present invention. Detailed implementation manners

[0058] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0059] As Figure 1 shown, it is a flowchart of a preferred embodiment of the enterprise public opinion analysis method of the present invention. According to different requirements, the order of the steps in this flowchart can be changed, and some steps can be omitted.

[0060] The enterprise public opinion analysis method can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology, and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0061] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0062] The enterprise public opinion analysis method is applied to one or more electronic devices. The electronic device is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored computer-readable instructions. Its hardware includes, but is not limited to, microprocessors, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0063] The electronic device can be any electronic product that can perform human-computer interaction with users. For example, personal computers, tablet computers, smart phones, personal digital assistants (PDAs), game consoles, Internet Protocol Televisions (IPTVs), smart wearable devices, etc.

[0064] The electronic device may include a network device and / or a user device. Among them, the network device includes, but is not limited to, a single network electronic device, a group of electronic devices composed of multiple network electronic devices, or a cloud composed of a large number of hosts or network electronic devices based on cloud computing (Cloud Computing).

[0065] The network where the electronic device is located includes, but is not limited to: the Internet, wide area network, metropolitan area network, local area network, virtual private network (Virtual Private Network, VPN), etc.

[0066] 101. Collect the enterprise topics of the enterprise to be monitored and the topic comments of the enterprise topics.

[0067] In at least one embodiment of the present invention, the enterprise to be monitored may be any enterprise that needs to conduct public opinion analysis. The enterprise topic may be a post published by any user on multiple preset platforms. The topic comment may be opinion information and the like published by other users on the post. For example, the multiple preset platforms may include platforms such as Maimai, Zhihu, Weibo, Tieba, etc.

[0068] In at least one embodiment of the present invention, the electronic device collects posts related to the enterprise to be monitored from the multiple preset platforms as the enterprise topics based on the buried point algorithm, and collects the comment texts corresponding to the enterprise topics as the topic comments.

[0069] 102. Identify the topic popularity of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic, and identify the comment popularity of the topic comment according to the comment collection source of the topic comment and the comment information of the topic comment.

[0070] In at least one embodiment of the present invention, the topic information includes the number of forwards, the number of likes, the number of reads of the enterprise topic, and the number of comments of the topic comment.

[0071] It can be understood that for the same enterprise topic, the topic collection source is the same as the comment collection source.

[0072] The comment information includes the number of likes, the number of reads of the topic comment, and the number of re-comments of the topic comment, etc.

[0073] In at least one embodiment of the present invention, the electronic device identifying the topic popularity of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic includes:

[0074] Obtain the source weight of the topic collection source, and obtain the information weight of the topic information on the topic collection source;

[0075] Perform a weighted sum operation on the number of forwards, the number of likes, the number of reads, and the number of comments based on the source weight and the information weight to obtain the topic heat. The calculation formula for the topic heat is: hot = k × (x 1 y 1 + x 2 y 2 + x 3 y 3 + x 4 y 4 ), where y 1 > y 4 > y 2 > y 3 , where hot represents the topic heat, k represents the source weight, x 1 represents the number of forwards, y 1 represents the first information weight corresponding to the number of forwards, x 2 represents the number of likes, y 2 represents the second information weight corresponding to the number of likes, x 3 represents the number of reads, y 3 represents the third information weight corresponding to the number of reads, x 4 represents the number of comments, y 4 represents the fourth information weight corresponding to the number of comments.

[0076] By combining the source weight and the information weight to perform a weighted sum operation on the number of forwards, the number of likes, the number of reads, and the number of comments, since the influence of different topic collection sources on the topic heat is distinguished, the accuracy of the topic heat can be improved.

[0077] In at least one embodiment of the present invention, the recognition method of the comment heat is similar to the recognition method of the topic heat, and the present invention will not elaborate on this.

[0078] 103. Sort the enterprise topics and the topic comments according to the topic heat and the comment heat to generate a topic comment list.

[0079] In at least one embodiment of the present invention, the topic comment list includes multiple hot topics and the hot comments of each hot topic, etc.

[0080] In at least one embodiment of the present invention, the electronic device sorts the enterprise topics and the topic comments according to the topic popularity and the comment popularity, and generates a topic comment list including:

[0081] Select popular topics from the enterprise topics based on the topic popularity;

[0082] Sort the popular topics according to the topic popularity to obtain a first list;

[0083] Select popular comments corresponding to the popular topics from the topic comments based on the comment popularity;

[0084] Sort the popular comments according to the comment popularity to obtain a second list;

[0085] Construct the topic comment list according to the second list and the first list.

[0086] Wherein, the popular topic refers to an enterprise topic whose topic popularity is greater than a first preset popularity threshold.

[0087] The popular comment refers to a topic comment whose comment popularity is greater than a second preset popularity threshold and corresponds to the enterprise topic.

[0088] By screening the popular comments after screening the popular topics, it is possible to avoid selecting the topic comments of all enterprise topics, thereby improving the construction efficiency of the second list, and further improving the construction efficiency of the topic comment list.

[0089] Specifically, the electronic device folds and displays the second list in the first list.

[0090] 104. Based on a pre-trained text classification model, divide the enterprise topics to obtain the topic indicators where the enterprise topics are located.

[0091] In at least one embodiment of the present invention, the text classification model includes a keyword extraction network, an encoding network, a convolutional network, and a mapping network.

[0092] The topic indicators include multiple first-type indicators and second-type indicators. The multiple first-type indicators include satisfaction indicators, growth indicators, consistency indicators, and professionalism indicators. The second-type indicators may include competitiveness indicators. Among them, the satisfaction indicators include dimensions such as salary and compensation, welfare guarantee, working environment, and employment relationship. The growth indicators include dimensions such as career, development and promotion, training, and interpersonal environment. The consistency indicators include dimensions such as brand connotation, development vision, value proposition, and social responsibility. The professionalism indicators include dimensions such as industry status, corporate reputation, technological innovation, and product innovation. The competitiveness indicators include dimensions such as competitive ranking.

[0093] In at least one embodiment of the present invention, the electronic device divides the enterprise topic based on a pre-trained text classification model, and the obtained topic indicators where the enterprise topic is located include:

[0094] Extract topic keywords from the enterprise topic based on the keyword extraction network;

[0095] Perform encoding processing on the topic keywords based on the encoding network to obtain a topic encoding vector;

[0096] Perform convolution processing on the topic encoding vector based on the convolution network to obtain a feature vector;

[0097] Perform mapping processing on the feature vector based on the mapping network to obtain a probability vector;

[0098] Identify the topic indicator according to the dimension where the element with the largest value in the probability vector is located.

[0099] Among them, the topic keywords may refer to topic words with a preset part of speech in the enterprise topic.

[0100] The convolution network includes multiple convolutional neural networks.

[0101] The topic indicator refers to the indicator corresponding to the dimension where the largest element is located.

[0102] Through the keyword extraction network, the topic keywords can be accurately extracted, and then the topic keywords are encoded through the encoding network. Since the encoding of interfering words is reduced, the encoding efficiency of the topic encoding vector is improved. By performing convolution on the topic encoding vector through the convolution network, the representation ability of the feature vector can be improved, thereby improving the accuracy of the topic indicator.

[0103] 105, perform sentiment prediction on the topic comment to obtain a negative sentiment score.

[0104] In at least one embodiment of the present invention, the negative sentiment score refers to the score corresponding to the topic comment in the negative dimension.

[0105] In at least one embodiment of the present invention, the electronic device performs sentiment prediction on the topic comment to obtain the negative sentiment score, including:

[0106] Performing encoding processing on the topic comment based on a preset sentiment dictionary to obtain a sentiment encoding vector;

[0107] Processing the sentiment encoding vector based on a pre-trained binary classification model to obtain an output probability;

[0108] Determining the probability value corresponding to the output probability in the negative dimension as the negative sentiment score.

[0109] Among them, multiple sentiment words are stored in the preset sentiment dictionary.

[0110] The model structure of the binary classification model is similar to that of the text classification model, and the present invention will not elaborate on this.

[0111] Encoding the topic comment through the preset sentiment dictionary, since the encoding of words without emotional color in the topic comment is avoided, the representation ability of the sentiment encoding vector is improved, thereby improving the accuracy of the negative sentiment score.

[0112] 106. Generating the index scores of the enterprise to be monitored on multiple topic indicators based on the comment popularity corresponding to the topic comments of the enterprise topic in the topic indicator and the negative sentiment score.

[0113] In at least one embodiment of the present invention, the index scores include a first score and a second score.

[0114] In at least one embodiment of the present invention, the electronic device generates the index scores of the enterprise to be monitored on multiple topic indicators based on the comment popularity corresponding to the topic comments of the enterprise topic in the topic indicator and the negative sentiment score, including:

[0115] Calculating the first score of the enterprise to be monitored on each first type of indicator according to the comment popularity and the negative sentiment score. The calculation formula of the first score is: score j = max(m, n - ∑ i∈{第j个第一类型指标的话题评论} hot i × sc i ), n > m, where score j represents the first score of the jth first type of indicator, hot iIndicates the comment popularity of the i-th topic comment in the j-th first-type indicator, sc i Indicates the negative sentiment score of the i-th topic comment in the j-th first-type indicator;

[0116] Calculate the total score of the enterprise to be monitored on the multiple first-type indicators according to the multiple first scores;

[0117] Identify the enterprise ranking of the enterprise to be monitored in the preset enterprise library according to the total score as the second score.

[0118] Wherein, n is usually set to 100 and m is usually set to 0.

[0119] The preset enterprise library stores multiple enterprises with the same business as the enterprise to be monitored.

[0120] By combining the comment popularity and the negative sentiment score, the accuracy of the first score can be improved. In addition, by setting m and n, the situation where the first score appears negative can be avoided, thereby improving the accuracy of the second score.

[0121] In other embodiments, when it is detected that the index score is less than the monitoring threshold, the electronic device generates a prompt message.

[0122] 107. Extract negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud map.

[0123] In at least one embodiment of the present invention, the negative words refer to the words with higher popularity and more negative sentiment in the topic comments.

[0124] In at least one embodiment of the present invention, the electronic device extracts negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud map, including:

[0125] Select target comments from the topic comments according to the comment popularity and the negative sentiment score;

[0126] Screen out the words that match successfully in the preset sentiment dictionary from the target comments as the negative words;

[0127] Generate a word label for the negative word according to the comment popularity and the negative sentiment score of the target comment where the negative word is located;

[0128] Merge the word labels to obtain the negative word cloud map.

[0129] Wherein, the target comment refers to the topic comment whose comment popularity and negative sentiment score are greater than a certain setting.

[0130] The vocabulary tags include the negative vocabulary and the rendering effect of the negative vocabulary.

[0131] By combining the comment popularity and the negative sentiment score, the target comments can be accurately selected. Then, with the help of the preset sentiment dictionary, the negative vocabulary can be directly extracted from the target comments. Furthermore, by combining the comment popularity and the negative sentiment score, the negative vocabulary is rendered, improving the rendering effect of the vocabulary tags and thus enhancing the intuitiveness of the negative word cloud diagram.

[0132] 108. Determine the topic comment list, the index score, and the negative word cloud diagram as the public opinion result of the enterprise to be monitored.

[0133] It should be emphasized that to further ensure the privacy and security of the above public opinion result, the above public opinion result can also be stored in a node of a blockchain.

[0134] As can be seen from the above technical solutions, the present invention can accurately identify the topic popularity by combining the topic collection source and the topic information, and can accurately identify the comment popularity by combining the comment collection source and the comment information, thereby improving the accuracy of the topic comment list. Further, through the comment popularity corresponding to the topic index and the corresponding negative sentiment score, the public opinion of all aspects of the enterprise to be monitored can be monitored from different multiple topic indexes, and the index score can be accurately generated. By combining the comment popularity and the negative sentiment score to extract negative vocabulary from the topic comments, the construction accuracy of the negative word cloud diagram can be improved. The present invention can intuitively display the public opinion result of the enterprise to be monitored by constructing a topic comment list and a negative word cloud diagram.

[0135] Such as Figure 2 shown, is a functional module diagram of a preferred embodiment of the enterprise public opinion analysis device of the present invention. The enterprise public opinion analysis device 11 includes a collection unit 110, an identification unit 111, a generation unit 112, a division unit 113, a prediction unit 114, a construction unit 115, and a determination unit 116. The module / unit referred to in the present invention means a series of computer-readable instruction segments that can be acquired by a processor 13 and can complete fixed functions, and are stored in a memory 12. In this embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0136] The collection unit 110 collects the enterprise topics of the enterprise to be monitored and the topic comments of the enterprise topics.

[0137] In at least one embodiment of the present invention, the enterprise to be monitored may be any enterprise that needs to conduct public opinion analysis. The enterprise topic may be a post published by any user on multiple preset platforms. The topic comment may be opinion information and the like published by other users on the post. For example, the multiple preset platforms may include platforms such as Maimai, Zhihu, Weibo, Tieba, etc.

[0138] In at least one embodiment of the present invention, the acquisition unit 110 acquires posts related to the enterprise to be monitored from the multiple preset platforms as the enterprise topics based on the buried point algorithm, and acquires the comment text corresponding to the enterprise topics as the topic comments.

[0139] The recognition unit 111 identifies the topic popularity of the enterprise topic according to the topic acquisition source of the enterprise topic and the topic information of the enterprise topic, and identifies the comment popularity of the topic comment according to the comment acquisition source of the topic comment and the comment information of the topic comment.

[0140] In at least one embodiment of the present invention, the topic information includes the number of forwards, the number of likes, the number of reads of the enterprise topic, and the number of comments of the topic comment.

[0141] It can be understood that for the same enterprise topic, the topic acquisition source is the same as the comment acquisition source.

[0142] The comment information includes the number of likes, the number of reads of the topic comment, and the number of re-comments of the topic comment, etc.

[0143] In at least one embodiment of the present invention, the recognition unit 111 identifying the topic popularity of the enterprise topic according to the topic acquisition source of the enterprise topic and the topic information of the enterprise topic includes:

[0144] Obtain the source weight of the topic acquisition source, and obtain the information weight of the topic information on the topic acquisition source;

[0145] Perform a weighted sum operation on the number of forwards, the number of likes, the number of reads, and the number of comments based on the source weight and the information weight to obtain the topic popularity. The calculation formula of the topic popularity is: hot = k×(x 1 y 1 +x 2 y 2 +x 3 y 3 +x 4 y 4 ), y 1 >y 4 >y 2 >y3 , where hot represents the topic popularity, k represents the source weight, x 1 represents the number of forwards, y 1 represents the first information weight corresponding to the number of forwards, x 2 represents the number of likes, y 2 represents the second information weight corresponding to the number of likes, x 3 represents the number of reads, y 3 represents the third information weight corresponding to the number of reads, x 4 represents the number of comments, y 4 represents the fourth information weight corresponding to the number of comments.

[0146] By combining the source weight and the information weight, a weighted sum operation is performed on the number of forwards, the number of likes, the number of reads, and the number of comments. Since the influence of different topic collection sources on the topic popularity is distinguished, the accuracy of the topic popularity can be improved.

[0147] In at least one embodiment of the present invention, the method for identifying the comment popularity is similar to the method for identifying the topic popularity, and the present invention will not elaborate on this.

[0148] The generating unit 112 sorts the enterprise topics and the topic comments according to the topic popularity and the comment popularity, and generates a topic comment list.

[0149] In at least one embodiment of the present invention, the topic comment list includes multiple hot topics and the hot comments of each hot topic, etc.

[0150] In at least one embodiment of the present invention, the generating unit 112 sorts the enterprise topics and the topic comments according to the topic popularity and the comment popularity, and the generated topic comment list includes:

[0151] Selecting hot topics from the enterprise topics based on the topic popularity;

[0152] Sorting the hot topics according to the topic popularity to obtain a first list;

[0153] Selecting hot comments corresponding to the hot topics from the topic comments based on the comment popularity;

[0154] Sorting the hot comments according to the comment popularity to obtain a second list;

[0155] Constructing the topic comment list according to the second list and the first list.

[0156] Among them, the popular topic refers to an enterprise topic whose topic popularity is greater than the first preset popularity threshold.

[0157] The popular comment refers to a topic comment whose comment popularity is greater than the second preset popularity threshold and corresponds to the enterprise topic.

[0158] By screening the popular topics and then screening the popular comments, it is possible to avoid selecting topic comments for all enterprise topics, thereby improving the construction efficiency of the second list and further improving the construction efficiency of the topic comment list.

[0159] Specifically, the generating unit 112 folds and displays the second list in the first list.

[0160] The partitioning unit 113 partitions the enterprise topics based on a pre-trained text classification model to obtain the topic indicators where the enterprise topics are located.

[0161] In at least one embodiment of the present invention, the text classification model includes a keyword extraction network, an encoding network, a convolutional network, and a mapping network.

[0162] The topic indicators include multiple first-type indicators and second-type indicators. The multiple first-type indicators include a satisfaction indicator, a growth indicator, a consistency indicator, and a professionalism indicator. The second-type indicator may include a competitiveness indicator. Among them, the satisfaction indicator includes dimensions such as salary and compensation, welfare guarantee, working environment, and employment relationship. The growth indicator includes dimensions such as career, development and promotion, training, and interpersonal environment. The consistency indicator includes dimensions such as brand connotation, development vision, value proposition, and social responsibility. The professionalism indicator includes dimensions such as industry status, corporate reputation, technological innovation, and product innovation. The competitiveness indicator includes dimensions such as competitive ranking.

[0163] In at least one embodiment of the present invention, the partitioning unit 113 partitions the enterprise topics based on a pre-trained text classification model, and the obtained topic indicators where the enterprise topics are located include:

[0164] Extracting topic keywords from the enterprise topic based on the keyword extraction network;

[0165] Encoding the topic keywords based on the encoding network to obtain a topic encoding vector;

[0166] Performing convolution processing on the topic encoding vector based on the convolutional network to obtain a feature vector;

[0167] Performing mapping processing on the feature vector based on the mapping network to obtain a probability vector;

[0168] Identify the topic index according to the dimension where the element with the largest value in the probability vector is located.

[0169] Wherein, the topic keyword may refer to a topic vocabulary with a preset part of speech in the enterprise topic.

[0170] The convolutional network includes a plurality of convolutional neural networks.

[0171] The topic index refers to the index corresponding to the dimension where the largest element is located.

[0172] Through the keyword extraction network, the topic keyword can be accurately extracted, and then the topic keyword is encoded by the encoding network. Since the encoding of interfering vocabulary is reduced, the encoding efficiency of the topic encoding vector is improved. By performing convolution on the topic encoding vector through the convolutional network, the representation ability of the feature vector can be improved, thereby improving the accuracy of the topic index.

[0173] The prediction unit 114 performs sentiment prediction on the topic comment to obtain a negative sentiment score.

[0174] In at least one embodiment of the present invention, the negative sentiment score refers to the score corresponding to the negative dimension of the topic comment.

[0175] In at least one embodiment of the present invention, the prediction unit 114 performs sentiment prediction on the topic comment to obtain a negative sentiment score, including:

[0176] Performing encoding processing on the topic comment based on a preset sentiment dictionary to obtain a sentiment encoding vector;

[0177] Processing the sentiment encoding vector based on a pre-trained binary classification model to obtain an output probability;

[0178] Determine the probability value corresponding to the negative dimension of the output probability as the negative sentiment score.

[0179] Wherein, a plurality of sentiment words are stored in the preset sentiment dictionary.

[0180] The model structure of the binary classification model is similar to the model structure of the text classification model, and the present invention will not elaborate on this.

[0181] Encoding the topic comment through the preset sentiment dictionary, since the encoding of words without emotional color in the topic comment is avoided, the representation ability of the sentiment encoding vector is improved, thereby improving the accuracy of the negative sentiment score.

[0182] The generating unit 112 generates the index scores of the enterprise to be monitored on multiple topic indexes based on the comment popularity corresponding to the topic comments of the enterprise topic in the topic indexes and the negative sentiment scores.

[0183] In at least one embodiment of the present invention, the index scores include a first score and a second score.

[0184] In at least one embodiment of the present invention, the generating unit 112 generating the index scores of the enterprise to be monitored on multiple topic indexes based on the comment popularity corresponding to the topic comments of the enterprise topic in the topic indexes and the negative sentiment scores includes:

[0185] Calculating the first score of the enterprise to be monitored on each first-type index according to the comment popularity and the negative sentiment score, and the calculation formula of the first score is: score j = max(m, n - ∑ i∈{第j个第一类型指标的话题评论} hot i × sc i ), n > m, where score j represents the first score of the jth first-type index, hot i represents the comment popularity of the ith topic comment in the jth first-type index, and sc i represents the negative sentiment score of the ith topic comment in the jth first-type index;

[0186] Calculating the total score of the enterprise to be monitored on the multiple first-type indexes according to the multiple first scores;

[0187] Identifying the enterprise ranking of the enterprise to be monitored in the preset enterprise library as the second score according to the total score.

[0188] Among them, n is usually set to 100, and m is usually set to 0.

[0189] The preset enterprise library stores multiple enterprises having the same business as the enterprise to be monitored.

[0190] By combining the comment popularity and the negative sentiment score, the accuracy of the first score can be improved. In addition, by setting m and n, the situation where the first score appears negative can be avoided, so that the accuracy of the second score can be improved.

[0191] In other embodiments, when it is detected that the index score is less than the monitoring threshold, the generating unit 112 generates a prompt message.

[0192] The constructing unit 115 extracts negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud map.

[0193] In at least one embodiment of the present invention, the negative words refer to the words with relatively high popularity and relatively negative sentiment in the topic comments.

[0194] In at least one embodiment of the present invention, the building unit 115 extracts negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud diagram, including:

[0195] Select target comments from the topic comments according to the comment popularity and the negative sentiment score;

[0196] Screen out the words that match successfully in the preset sentiment dictionary from the target comments as the negative words;

[0197] Generate a word label for the negative word according to the comment popularity and the negative sentiment score of the target comment where the negative word is located;

[0198] Merge the word labels to obtain the negative word cloud diagram.

[0199] Wherein, the target comment refers to the topic comment whose comment popularity and negative sentiment score are greater than a certain setting.

[0200] The word label includes the negative word and the rendering effect of the negative word.

[0201] By combining the comment popularity and the negative sentiment score, the target comment can be accurately selected. Furthermore, with the help of the preset sentiment dictionary, the negative words can be directly extracted from the target comment. Then, by combining the comment popularity and the negative sentiment score, the negative words are rendered, improving the rendering effect of the word label, and thus improving the intuitiveness of the negative word cloud diagram.

[0202] The determination unit 116 determines the topic comment list, the index score value, and the negative word cloud diagram as the public opinion result of the enterprise to be monitored.

[0203] It should be emphasized that to further ensure the privacy and security of the above public opinion result, the above public opinion result can also be stored in a node of a blockchain.

[0204] As can be seen from the above technical solutions, the present invention can accurately identify the topic popularity by combining the topic collection source and the topic information, and can accurately identify the comment popularity by combining the comment collection source and the comment information, thereby improving the accuracy of the topic comment list. Further, through the comment popularity corresponding to the topic index and the corresponding negative sentiment score, the public opinion of all aspects of the enterprise to be monitored can be monitored from different multiple topic indexes, and the index score can be accurately generated. By combining the comment popularity and the negative sentiment score to extract negative words from the topic comments, the accuracy of constructing the negative word cloud map can be improved. By constructing the topic comment list and the negative word cloud map, the present invention can intuitively display the public opinion results of the enterprise to be monitored.

[0205] As Figure 3 shown, it is a schematic structural diagram of an electronic device according to a preferred embodiment of the method for realizing enterprise public opinion analysis of the present invention.

[0206] In an embodiment of the present invention, the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and computer-readable instructions stored in the memory 12 and executable on the processor 13, such as an enterprise public opinion analysis program.

[0207] Those skilled in the art can understand that the schematic diagram is only an example of the electronic device 1, and does not constitute a limitation on the electronic device 1. It may include more or fewer components than shown, or combine certain components, or different components. For example, the electronic device 1 may further include input / output devices, network access devices, buses, etc.

[0208] The processor 13 may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor 13 is the operation core and control center of the electronic device 1, connecting various parts of the entire electronic device 1 through various interfaces and lines, and executing the operating system of the electronic device 1 and various installed application programs, program codes, etc.

[0209] Exemplarily, the computer-readable instructions may be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 13 to implement the present invention. The one or more modules / units may be a series of computer-readable instruction segments capable of performing specific functions, and these computer-readable instruction segments are used to describe the execution process of the computer-readable instructions in the electronic device 1. For example, the computer-readable instructions may be divided into an acquisition unit 110, an identification unit 111, a generation unit 112, a division unit 113, a prediction unit 114, a construction unit 115, and a determination unit 116.

[0210] The memory 12 may be used to store the computer-readable instructions and / or modules. By running or executing the computer-readable instructions and / or modules stored in the memory 12, and by invoking the data stored in the memory 12, the processor 13 realizes various functions of the electronic device 1. The memory 12 may mainly include a program storage area and a data storage area. Among them, the program storage area may store an operating system, application programs required for at least one function (such as a sound playback function, an image playback function, etc.); the data storage area may store data created according to the use of the electronic device. The memory 12 may include non-volatile and volatile memories, such as: a hard disk, a memory, a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a Flash Card, at least one magnetic disk storage device, a flash memory device, or other storage devices.

[0211] The memory 12 may be an external memory and / or an internal memory of the electronic device 1. Further, the memory 12 may be a memory in physical form, such as a memory stick, a TF card (Trans-flash Card), etc.

[0212] If the modules / units integrated in the electronic device 1 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above method embodiments of the present invention, it may also be completed by computer-readable instructions instructing relevant hardware. The computer-readable instructions may be stored in a computer-readable storage medium, and when executed by the processor, the steps of the above method embodiments can be realized.

[0213] Among them, the computer-readable instructions include computer-readable instruction codes, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer-readable medium can include: any entity or device capable of carrying the computer-readable instruction codes, recording media, USB flash drives, mobile hard disks, magnetic disks, optical disks, computer memories, read-only memories (ROM, Read-Only Memory), and random access memories (RAM, Random Access Memory).

[0214] The blockchain referred to in the present invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, and encryption algorithms. Blockchain, in essence, is a decentralized database, a string of data blocks generated by using cryptographic methods, and each data block contains information about a batch of network transactions, used to verify the validity (anti-counterfeiting) of the information and generate the next block. The blockchain can include the blockchain underlying platform, the platform product service layer, and the application service layer, etc.

[0215] Combined Figure 1 , the memory 12 in the electronic device 1 stores computer-readable instructions to implement an enterprise public opinion analysis method, and the processor 13 can execute the computer-readable instructions to thereby implement:

[0216] Collect the enterprise topics of the enterprise to be monitored and the topic comments of the enterprise topics;

[0217] Identify the topic heat of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic, and identify the comment heat of the topic comment according to the comment collection source of the topic comment and the comment information of the topic comment;

[0218] Sort the enterprise topics and the topic comments according to the topic heat and the comment heat to generate a topic comment list;

[0219] Divide the enterprise topics based on a pre-trained text classification model to obtain the topic indicators where the enterprise topics are located;

[0220] Perform sentiment prediction on the topic comments to obtain a negative sentiment score;

[0221] Generate the index scores of the enterprise to be monitored on multiple topic indicators based on the comment heat corresponding to the topic comments of the enterprise topics in the topic indicators and the negative sentiment score;

[0222] Extract negative words from the topic comments according to the comment heat and the negative sentiment score to construct a negative word cloud map;

[0223] Determine the topic comment list, the index score, and the negative word cloud map as the public opinion result of the enterprise to be monitored.

[0224] Specifically, for the specific implementation method of the above computer-readable instructions by the processor 13, reference can be made to Figure 1 the description of the relevant steps in the corresponding embodiments, which will not be elaborated here.

[0225] In several embodiments provided by the present invention, it should be understood that the disclosed system, device, and method can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0226] The computer-readable storage medium stores computer-readable instructions, where the computer-readable instructions, when executed by the processor 13, are used to implement the following steps:

[0227] Collect the enterprise topics of the enterprise to be monitored and the topic comments of the enterprise topics;

[0228] Identify the topic popularity of the enterprise topics according to the topic collection source of the enterprise topics and the topic information of the enterprise topics, and identify the comment popularity of the topic comments according to the comment collection source of the topic comments and the comment information of the topic comments;

[0229] Sort the enterprise topics and the topic comments according to the topic popularity and the comment popularity to generate a topic comment list;

[0230] Divide the enterprise topics based on a pre-trained text classification model to obtain the topic indicators where the enterprise topics are located;

[0231] Perform sentiment prediction on the topic comments to obtain a negative sentiment score;

[0232] Generate the index scores of the enterprise to be monitored on multiple topic indicators based on the comment popularity corresponding to the topic comments in the topic indicators and the negative sentiment score;

[0233] Extract negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud map;

[0234] Determine the topic comment list, the index score, and the negative word cloud map as the public opinion result of the enterprise to be monitored.

[0235] The module described as a separation component may or may not be physically separated. The component shown as a module may or may not be a physical unit, that is, it may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0236] In addition, in each embodiment of the present invention, each functional module can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated unit can be implemented in the form of hardware, or in the form of a combination of hardware and software functional modules.

[0237] Therefore, from any perspective, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be embraced by the present invention. Any reference signs in the claims should not be regarded as limiting the claims involved.

[0238] In addition, obviously, the term "including" does not exclude other units or steps, and the singular does not exclude the plural. The above-mentioned multiple units or devices can also be implemented by one unit or device through software or hardware. The terms such as "first" and "second" are used to indicate names and do not represent any specific order.

[0239] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An enterprise public opinion analysis method, characterized in that, the enterprise public opinion analysis method includes: collecting enterprise topics of the enterprise to be monitored and topic comments of the enterprise topics; identifying the topic popularity of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic, and identifying the comment popularity of the topic comment according to the comment collection source of the topic comment and the comment information of the topic comment; sorting the enterprise topics and the topic comments according to the topic popularity and the comment popularity to generate a topic comment list; dividing the enterprise topics based on a pre-trained text classification model to obtain the topic indicators where the enterprise topics are located; performing sentiment prediction on the topic comments to obtain a negative sentiment score; generating index scores of the enterprise to be monitored on multiple topic indicators based on the comment popularity corresponding to the topic comments of the enterprise topics in the topic indicators and the negative sentiment score. The multiple topic indicators include multiple first-type indicators and second-type indicators. The index scores include a first score and a second score. Generating the index scores of the enterprise to be monitored on multiple topic indicators based on the comment popularity corresponding to the topic comments of the enterprise topics in the topic indicators and the negative sentiment score includes: calculating the first score of the enterprise to be monitored on each first-type indicator according to the comment popularity and the negative sentiment score. The calculation formula of the first score is: ， ， Among them, represents the first score of the th first type of indicator, represents the comment popularity of the th topic comment among the th first type of indicators, represents the negative sentiment score of the th topic comment among the th first type of indicators; calculate the total score of the enterprise to be monitored on the multiple first type of indicators according to the multiple first scores; identify the enterprise ranking of the enterprise to be monitored in the preset enterprise library as the second score according to the total score; extracting negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud diagram; determining the topic comment list, the index scores and the negative word cloud diagram as the public opinion result of the enterprise to be monitored.

2. The enterprise public opinion analysis method according to claim 1, characterized in that, the topic information includes the number of forwards, the number of likes, the number of reads of the enterprise topic and the number of comments of the topic comment. Identifying the topic popularity of the enterprise topic according to the topic collection source of the enterprise topic and the topic information of the enterprise topic includes: obtaining the source weight of the topic collection source and obtaining the information weight of the topic information on the topic collection source; Perform a weighted sum operation on the number of forwards, the number of likes, the number of reads, and the number of comments based on the source weight and the information weight to obtain the topic popularity. The calculation formula for the topic popularity is: , wherein, represents the topic popularity, represents the source weight, represents the number of forwards, represents the first information weight corresponding to the number of forwards, represents the number of likes, represents the second information weight corresponding to the number of likes, represents the number of reads, represents the third information weight corresponding to the number of reads, represents the number of comments, represents the fourth information weight corresponding to the number of comments.

3. The enterprise public opinion analysis method according to claim 1, characterized in that, sorting the enterprise topics and the topic comments according to the topic popularity and the comment popularity to generate a topic comment list includes: selecting popular topics from the enterprise topics based on the topic popularity; sorting the popular topics according to the topic popularity to obtain a first list; selecting popular comments corresponding to the popular topics from the topic comments based on the comment popularity; sorting the popular comments according to the comment popularity to obtain a second list; constructing the topic comment list according to the second list and the first list.

4. The enterprise public opinion analysis method according to claim 1, characterized in that, The text classification model includes a keyword extraction network, an encoding network, a convolutional network, and a mapping network. The enterprise topics are classified based on the pre-trained text classification model, and the topic indicators where the enterprise topics are located include: Extract topic keywords from the enterprise topics based on the keyword extraction network; Perform encoding processing on the topic keywords based on the encoding network to obtain topic encoding vectors; Perform convolutional processing on the topic encoding vectors based on the convolutional network to obtain feature vectors; Perform mapping processing on the feature vectors based on the mapping network to obtain probability vectors; Identify the topic indicators according to the dimension where the element with the largest value in the probability vector is located.

5. The enterprise public opinion analysis method according to claim 1, characterized in that, The negative sentiment score obtained by performing sentiment prediction on the topic comments includes: Perform encoding processing on the topic comments based on a preset sentiment dictionary to obtain sentiment encoding vectors; Process the sentiment encoding vectors based on a pre-trained binary classification model to obtain output probabilities; Determine the probability value corresponding to the negative sentiment dimension of the output probability as the negative sentiment score.

6. The enterprise public opinion analysis method according to claim 5, characterized in that, The method of extracting negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud map includes: Select target comments from the topic comments according to the comment popularity and the negative sentiment score; Screen out the words that match successfully in the preset sentiment dictionary from the target comments as the negative words; Generate word labels for the negative words according to the comment popularity and the negative sentiment score of the target comments where the negative words are located; Merge the word labels to obtain the negative word cloud map.

7. An enterprise public opinion analysis device, characterized in that, The enterprise public opinion analysis device includes: A collection unit for collecting enterprise topics of the enterprise to be monitored and topic comments of the enterprise topics; An identification unit for identifying the topic popularity of the enterprise topics according to the topic collection source of the enterprise topics and the topic information of the enterprise topics, and identifying the comment popularity of the topic comments according to the comment collection source of the topic comments and the comment information of the topic comments; A generation unit for sorting the enterprise topics and the topic comments according to the topic popularity and the comment popularity, and generating a topic comment list; A division unit for dividing the enterprise topics based on a pre-trained text classification model to obtain the topic indicators where the enterprise topics are located; A prediction unit for performing sentiment prediction on the topic comments to obtain a negative sentiment score; The generating unit is further configured to generate index scores of the enterprise to be monitored on multiple topic metrics based on the comment heat corresponding to the topic comments of the enterprise topic in the topic metrics and the negative sentiment score. The multiple topic metrics include multiple first-type metrics and second-type metrics. The index scores include a first score and a second score. The generating of the index scores of the enterprise to be monitored on multiple topic metrics based on the comment heat corresponding to the topic comments of the enterprise topic in the topic metrics and the negative sentiment score includes: calculating the first score of the enterprise to be monitored on each first-type metric according to the comment heat and the negative sentiment score. The calculation formula of the first score is: , , where represents the first score of the th first-type metric, represents the comment heat of the th topic comment in the th first-type metric, represents the negative sentiment score of the th topic comment in the th first-type metric; calculating the total score of the enterprise to be monitored on the multiple first-type metrics according to the multiple first scores; identifying the enterprise ranking of the enterprise to be monitored in the preset enterprise database as the second score according to the total score; A construction unit for extracting negative words from the topic comments according to the comment popularity and the negative sentiment score to construct a negative word cloud map; A determination unit for determining the topic comment list, the index score, and the negative word cloud map as the public opinion result of the enterprise to be monitored.

8. An electronic device, characterized in that, The electronic device includes: A memory storing computer-readable instructions; and A processor that executes computer-readable instructions stored in the memory to implement the enterprise public opinion analysis method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions are executed by a processor in an electronic device to implement the enterprise public opinion analysis method according to any one of claims 1 to 6.

Images