A Visual Analysis Method for Multi-Pole Emotional Intelligence in Poetry Composition

Abstract

The present invention discloses a multi-polar emotional intelligent visual analysis method for wine-poetry, which belongs to the field of wine-poetry culture and visualization. The method comprises the following steps: obtaining wine-poetry text data, creating a Word2Vec associated word search model, and performing Chinese character association on the wine-poetry text data based on the Word2Vec associated word search model to obtain associated words; marking the emotion of the poem based on the associated words to obtain the emotion label of the poem to be analyzed; and determining the emotion of the poem and the theme word of the poem based on the emotion label of the poem to be analyzed. The present invention can effectively improve the accuracy of the emotion analysis of wine-poetry, and at the same time construct a more intuitive visual analysis method of spatiotemporal emotion association to support multi-polar wine-poetry text analysis and assist users in understanding the emotion of the poem text.

Description

Technical Field

[0001] The present invention belongs to the field of poetry and wine culture and visualization technology, and in particular relates to a multi-polar emotional intelligent visual analysis method for poetry about wine. Background Art

[0002] In exploring the combination of Chinese classical literature and visualization technology, current existing technologies have not only revealed the internal structure and emotional expression of literary works by applying visualization technology to the fields of poetry and wine culture, Song Dynasty lyrics, and ancient texts, but also provided new perspectives and methods for literary research. For example, through visual analysis methods, researchers can have a deeper understanding of the rhythm, emotional tone, and life trajectory of historical figures in poetry. At the same time, some studies have also tried to combine machine learning with manual proofreading to improve the efficiency and accuracy of classical document processing. However, despite certain achievements, existing technologies still have the following problems:

[0003] (1) There is a gap in the integration of digital science and literary research, especially the lack of research on "poetry and wine culture".

[0004] (2) Existing technologies have been involved in the analysis of poetry sentiment, and there have been some achievements in applying visualization techniques to poetry. However, existing technologies are highly dependent on dictionaries. In sentiment analysis, if the sentiment dictionary is incomplete or inaccurate, it will lead to deviations in the results of sentiment analysis. In addition, the simple dictionary addition method cannot accurately judge the sentiment tendency and theme.

[0005] Based on the above technical problems, the present invention provides a multi-polar emotional intelligent visual analysis method for drinking poems. Summary of the invention

[0006] In order to solve the above technical problems, the present invention proposes a multi-polar emotional intelligent visual analysis method for wine poetry to solve the problems existing in the above-mentioned prior art.

[0007] To achieve the above object, the present invention provides a multi-polar emotional intelligent visual analysis method for wine poetry, comprising:

[0008] Acquire poetry and wine text data, create a Word2Vec associated word search model, and associate Chinese characters with the poetry and wine text data based on the Word2Vec associated word search model to obtain associated words;

[0009] Marking the emotion of the poem based on the associated words to obtain the emotion label of the poem to be analyzed;

[0010] The emotion and theme words of the poem are determined based on the emotion tags of the poem to be analyzed.

[0011] Optionally, the process of associating Chinese characters with the poetry and wine text data based on the Word2Vec associated character search model includes:

[0012] Preprocessing the poetry and wine text data to obtain preprocessed text data;

[0013] Creating a Word2Vec associated word search model, and converting the preprocessed text data into a high-dimensional vector based on the Word2Vec associated word search model;

[0014] Define the emotion tags of poems and lyrics, generate corresponding associated words for each emotion tag based on the Word2Vec associated word search model and give the degree of association.

[0015] Optionally, the process of obtaining the emotion label of the poem to be analyzed based on the associated word tagging of the poem emotion includes:

[0016] Calculating a correlation score of the high-dimensional vector based on the correlation word and the correlation degree;

[0017] The sentiment label of the poem to be analyzed is determined based on the association score.

[0018] Optionally, the expression for calculating the association score is:

[0019] ;

[0020] In the formula, value (x i ) e It is the Chinese character x i For the association degree of class e label, x i is a Chinese character, n is the length of the poem, and e is set to be the emotion category label number.

[0021] Optionally, the process of determining the emotion of the poem and the theme words of the poem based on the emotion tag of the poem to be analyzed includes:

[0022] Analyze the emotion tags of the poem to be analyzed based on the LSTM deep learning algorithm model to determine the emotion of the poem;

[0023] The theme words of the poem to be analyzed are extracted based on the implicit Dirichlet distribution algorithm to obtain the theme words of the poem.

[0024] Optionally, the process of extracting the subject words of the poem to be analyzed includes:

[0025] Construct document-term matrix;

[0026] Randomly assign a topic to each word in the document and calculate the proportion of the current word in the document assigned to the target topic and the proportion of the word in all documents assigned to the target topic;

[0027] The topic assignment is adjusted based on the proportion of the current word in the document assigned to the target topic and the proportion of the word in all documents assigned to the target topic to determine the topic word of the poem to be analyzed.

[0028] The present invention also provides a multi-polar emotional intelligent visual analysis system for wine poetry, the visual analysis system comprising a data layer, a data analysis and processing layer, a design layer and an interaction layer;

[0029] Wherein, the data layer is used to obtain poetry and wine text data;

[0030] The data analysis and processing layer is used to pre-process the poetry and wine text data and execute the poetry sentiment analysis algorithm and poetry theme word extraction algorithm model to determine the poetry sentiment and poetry theme words;

[0031] The design layer is used to visualize poetry and wine data from the perspectives of text, time, space and relationship;

[0032] The interaction layer is used to realize the interaction between the user and the visual interface.

[0033] The present invention also provides a computer terminal device, comprising:

[0034] one or more processors;

[0035] A memory, coupled to the processor, for storing one or more programs;

[0036] When the one or more programs are executed by the one or more processors, the one or more processors implement a multi-polar emotional intelligent visual analysis method for wine poetry.

[0037] The present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, a multi-polar emotional intelligent visual analysis method for wine-themed poems is implemented.

[0038] Compared with the prior art, the present invention has the following advantages and technical effects:

[0039] (1) The poetry and wine data are processed and analyzed by using text sentiment analysis methods and poetry theme sentiment extraction algorithms. At the same time, the results of the algorithm analysis and the poetry and wine data themselves are visually encoded from the two dimensions of time and space. Finally, the entire visual analysis system is realized by laying out the visual charts and setting up rich user interactions.

[0040] (2) This system can be used by different types of people. For literary researchers, this system is a convenient research tool that can be used in combination with traditional research to make the research results more accurate; for the general public, this system can be used as a channel for popularizing knowledge and understanding cultural knowledge; and for foreign friends, this system can be used to intuitively understand the changes in the emotions of poetry. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0042] Figure 1 This is the technical route of the poetry and wine data visual analysis system of the embodiment of the present invention;

[0043] Figure 2 A flowchart of poetry emotion calculation according to an embodiment of the present invention;

[0044] Figure 3 This is an architecture diagram of the poetry and wine data visual analysis system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0045] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0046] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0047] Embodiment 1.

[0048] This embodiment provides a method for visual analysis of multi-polar emotion intelligence in wine poetry. Figure 1 As shown, the method includes the following parts: data crawling, algorithm, and actual visualization implementation.

[0049] The data crawling part mainly uses Python language to crawl data with request library and BeautifulSoup library, and uses json library to store data. Requests is a short and direct third-party library for processing HTTP requests, which is often used to crawl and evaluate the server's response data. BeautifulSoup is a third-party software package that can parse and process HTML and XML.

[0050] The algorithm implementation part is mainly divided into two parts: text sentiment analysis algorithm: LSTM (Long Short-Term Memory) + Word2Vec; topic extraction algorithm: LDA (Latent Dirichlet Allocation). The algorithm part is implemented in Python.

[0051] Visualization is mainly divided into two parts: front-end interface drawing and visual chart drawing. The front-end interface drawing mainly uses the front-end three-piece set plus the Vue framework, and the visual chart drawing uses a third-party JS visualization library.

[0052] As a specific implementation of this embodiment, the method includes the following steps: obtaining poetry and wine text data, creating a Word2Vec associated word search model, and performing Chinese character association on the poetry and wine text data based on the Word2Vec associated word search model to obtain associated words; marking the emotion of the poem based on the associated words to obtain the emotion label of the poem to be analyzed; and determining the emotion of the poem and the theme word of the poem based on the emotion label of the poem to be analyzed. Further, the following contents are included:

[0053] 1. Implementation of algorithm for visual analysis system of poetry and wine data;

[0054] This system mainly uses three algorithms: Word2Vec, LSTM, and LDA. Word2Vec and LSTM algorithms are mainly used to analyze the emotions expressed in poetry, while the LDA algorithm is mainly used to extract poetry keywords. This example will explain the specific execution process of the poetry and wine data sentiment analysis method, such as Figure 2 As shown. First, add the experimental corpus prepared in advance to the word vector model, and use the model to obtain the associated word data set of five emotions. Next, you need to merge the associated word data set to obtain the training set for executing the algorithm, and then introduce the LSTM model, training set and test set at the same time in the steps of executing the algorithm. After the algorithm is executed, the result obtained includes the emotion of the entire poem and the emotion of each sentence in the poem. If the correlation between the poem and its five emotions is not high, it is expressed as no emotion.

[0055] 1.1 Classification of poetry sentiment;

[0056] 1.1.1 Related word search based on Word2Vec;

[0057] The process of performing Chinese character association on poem and wine text data based on the Word2Vec associated word search model includes: preprocessing the poem and wine text data to obtain preprocessed text data; creating a Word2Vec associated word search model, and converting the preprocessed text data into high-dimensional vectors based on the Word2Vec associated word search model; defining poem sentiment labels, and generating corresponding associated words and giving the association degrees for each sentiment label based on the Word2Vec associated word search model.

[0058] First, perform preprocessing operations on all the poems in the corpus. The purpose of preprocessing is to preprocess the original data into the input data of the model. The original data is poem and wine text data. Since this algorithm only needs to use the data in the dimension of poem content, it is necessary to extract the original data. When extracting the data, it is necessary to filter and delete irrelevant content such as punctuation marks, special characters, and classical Chinese function words in the poems. Then, take each sentence of each poem as a word vector, and extract vectors with a length greater than 1 from all the sentences in the corpus as the training data of the Word2Vec model.

[0059] Secondly, create a Word2Vec associated word search model. Here, it is necessary to introduce the training data, and then it is necessary to set the model parameters.

[0060] Then, set the poem sentiment labels. Divide the sentiment labels of wine poems into five categories: joy, happiness, anger, sorrow, and thought. Since the character "怒" in classical Chinese is often expressed as getting angry or having a strong momentum, which has a certain discrepancy in meaning with the emotion of "anger", in order to improve the accuracy of the model, in this algorithm, "恚" is used to replace "怒". Next, execute the Word2Vec associated word search model. Putting the sentiment labels into the Word2Vec model one by one will obtain sets of Chinese characters with different degrees of association, and the association values are distributed between 0 and 1.

[0061] 1.1.2 Mark the sentiment of poems based on associated words;

[0062] The process of obtaining the sentiment label of the poem to be analyzed by marking the sentiment of the poem based on associated words includes: calculating the association score of the high-dimensional vector based on the associated words and the association degree; determining the sentiment label of the poem to be analyzed based on the association score.

[0063] Mark the sentiment of the poem sentences and poems according to the associated words. First, convert the set of associated words into a data format for subsequent operations. Traverse all the associated words of the five sentiment labels. If there are duplicate associated words between different sentiment labels, then only keep the item with the largest association score, and delete the rest of the duplicate items. The formula calculation is as follows:

[0064] ;

[0065] ;

[0066] ;

[0067] x i is a Chinese character, n is the length of the poem, set e to be the emotional category label number, value (x i ) e It is the Chinese character x i For the association degree of class e label.

[0068] The threshold T of each sentiment label is determined according to the dictionary of associated words corresponding to the five categories of labels. e , which is used to judge the emotion corresponding to the poem.

[0069] 1.1.3 Determine the emotion of the poem and extract the poem keywords;

[0070] The process of determining the emotion and theme words of the poem based on the emotion label of the poem to be analyzed includes: analyzing the emotion label of the poem to be analyzed based on the LSTM deep learning algorithm model to determine the emotion of the poem; extracting the theme words of the poem to be analyzed based on the latent Dirichlet distribution algorithm to obtain the theme words of the poem.

[0071] Through the existing sentiment tags, combined with the deep learning framework, the text features and semantic information are refined to make judgments on the emotions of the poems: prepare the data set for the LSTM model; perform LSTM model preprocessing operations; create and execute the LSTM deep learning algorithm model; evaluate the model accuracy; use the model to obtain the emotions of wine poems; use the latent Dirichlet allocation (LDA) algorithm to extract poetry keywords. Di represents the document-word matrix, the words are represented by Ws, and the topics are represented by the letter k. During the first iteration, the topic is randomly assigned to each word in the document. Therefore, in the corpus, the words in the document will be associated with some random topics.

[0072] ;

[0073] ;

[0074] .

[0075] Based on those correct topic word assignments, LDA will iterate over each document Di and each word Ws by calculating two probabilities: P1 (the current assignment of document Di to topic K) and P2 (the current assignment of document Di to topic K). k word) and P2 (the proportion of words Ws among all documents assigned to topic K k The ratio of ) is used to correct and adjust the topic assignment of the current word. Finally, the model is run and the data is processed.

[0076] .

[0077] The present embodiment also discloses a visual analysis system for multi-polar emotional intelligent visual analysis of poetry about wine, which includes a data layer, a data analysis and processing layer, a design layer and an interaction layer; wherein the data layer is used to obtain poetry and wine text data; the data analysis and processing layer is used to pre-process the poetry and wine text data and execute the poetry sentiment analysis algorithm and the poetry keyword extraction algorithm model to determine the poetry sentiment and poetry keyword; the design layer is used to visualize the poetry and wine data from the perspectives of text, time, space and relationship; the interaction layer is used to realize the interaction between the user and the visual interface.

[0078] 2. Design of visual analysis system for poetry and wine data;

[0079] This system is divided into four layers: data layer, data analysis and processing, design layer, and interaction layer. Figure 3 As shown in the figure, it is divided into data acquisition, data processing, visualization design and interaction design. Data analysis and processing mainly include processing, analysis and extraction of raw data, and then selecting appropriate visualization encoding methods for the processed data. The visualization design layer mainly designs different visualization charts for different categories of data. Since the system of this subject is mainly based on text data, it is necessary to rely on selection, brightness and multi-view collaboration as the basic interaction to achieve the research objectives. In addition, it is necessary to consider the linkage of views, so a reasonable view layout should be arranged.

[0080] 2.1 Visualization data layer;

[0081] The data layer of the entire system contains three parts of raw data: text data in the existing database, poet relationship database and other data. The text data in the existing database is mainly the poetry data of Tang Dynasty poets such as Li Bai and Du Fu. The data dimension is relatively complete, but the data volume is small. The poet relationship database is the "China Biographical Database Project" (CBDB), which is a relational database that collects data on the life and social relations of most literati from the pre-Qin period to modern times. For this database, it can be used to extract relationship data between poets.

[0082] Conduct data analysis and processing. Data preprocessing, first evaluate each part of the data to understand the entire data set and determine the relevance and consistency of the data with this topic. Secondly, perform data cleaning, data conversion, and data reduction (data subpackaging) to lay the foundation for the smooth progress of subsequent experiments; execute poetry sentiment analysis algorithms and poetry theme word extraction algorithm models; perform data integration to update the data layer of the entire system, which contains three parts of data: basic information data of poems, poet relationship data, and poetry theme and sentiment data.

[0083] 2.2 Establish visual design layer;

[0084] The five major data types in the present invention, namely text data, time series data, spatial data, relational data, and other types of data, are visually encoded. Among them, text type data mainly includes the text of the poem content, the theme of the poem, the emotion of each sentence in the poem, and the emotion of the whole poem. This part of the data can use color coding to represent the emotion category or use the form of word cloud to present different emotion categories and proportions; since poetry can be divided according to different dynasties from the time dimension, it has a strong time series characteristic. For this type of data, a river map or area map with a time axis can be used to display time and emotion data in multiple directions; similarly, the creation locations of poems are mostly different from the spatial dimension, which also forms the spatial distribution of the poetry creation locations. At this time, a map can be used to visualize this type of data; relational data in this system specifically refers to the relationship data between poets. There may be more or less communication between poets in the same period. Under normal circumstances, a force-directed graph can be used to present the data; other types of data can use bar charts, bar graphs, etc. for statistical analysis, and pie charts are used to show the emotional proportion, etc.

[0085] 2.3 Carry out visual layout design;

[0086] 2.4 Visualization interaction layer;

[0087] This system mainly focuses on text data. After presenting the data in different types of visual charts, it is necessary to use interactive means to make the entire system more vivid and have stronger scalability. The interactive design of the system is mainly divided into interactions within a single view and linkage interactions between multiple views. The interactions within a single view are mainly divided into mouse click events, scroll bar zoom events, highlighting of selected charts, hovering highlights and other interactive operations. Table 1 is a comparison table of interactions and charts within a single view.

[0088] Table 1

[0089] Interaction Visualization charts involved Mouse click event Map, poet relationship-oriented map, keyword cloud, area map Mouse scroll zoom Map, Poet Relationship Directed Map Highlight Poet relationship-oriented graph, keyword cloud, and area graph Mouseover Statistical bar chart, keyword cloud, area chart

[0090] Embodiment 2.

[0091] This example provides a method for intelligent visual analysis of multi-polar emotions in wine poems, which is divided into two parts. The first part is the algorithm implementation of the visual analysis system of poetry and wine data, including the following steps:

[0092] Step 1: Classification of poem sentiment. The map used is the map of the People's Republic of China. The regional map is used to present the spatial distribution of wine poem sentiment. Different color codes are used to distinguish different categories of wine poem sentiment, and then presented using a heat map.

[0093] Step 1.1: Word2Vec-based related word search. First, preprocess all the poems in the corpus; second, create a Word2Vec related word search model; then, set the emotion tags of the poems. The emotion tags of wine poems are divided into five categories: joy, happiness, anger, sadness, and thought.

[0094] Step 1.2: Tag the emotion of the poem based on the associated words. Convert the associated word set into data format, and then calculate the association score of each poem corresponding to the five types of emotion tags. The calculation formula is as follows:

[0095] ;

[0096] ;

[0097] ;

[0098] x i is a Chinese character, n is the length of the poem, set e to be the emotional category label number, value (x i ) e It is the Chinese character x i For the association degree of class e label.

[0099] Step 1.3: Determine the threshold T of each sentiment label based on the dictionary of associated words corresponding to the five categories of labels e , which is used to judge the emotion corresponding to the poem.

[0100] Step 2: Determine the emotion of the poem. Through the existing emotion tags, combined with the deep learning framework, the text features and semantic information are extracted to make a judgment on the emotion of the poem. This example uses the LSTM deep learning algorithm model to implement text emotion analysis. In the early stage, prepare the data set for the LSTM model, perform the LSTM model preprocessing operation, and then create and execute the LSTM deep learning algorithm model, and then evaluate the model accuracy and use the model to obtain the emotion of the wine poem.

[0101] Step 3: Use Latent Dirichlet Allocation (LDA) algorithm for poetry topic word extraction. LDA generates probabilities for words used to form topics and ultimately classifies topics into documents. The entire process of poetry topic word extraction performs two tasks: finding topics from a corpus (a collection of text documents) and simultaneously assigning these topics to documents present in the same corpus.

[0102] Let Di denote the document-word matrix, words are denoted by Ws, and topics are denoted by the letter k. During the first iteration, a topic is randomly assigned to each word in the document. Therefore, in the corpus, the words in the document will be associated with some random topics.

[0103] ;

[0104] ;

[0105] .

[0106] Based on the correct topic word assignments, LDA will iterate over each document Di and each word Ws by calculating two probabilities: P1 (the current probability assigned to topic K in document Di) and P2 (the probability assigned to topic K in document Di). k word) and P2 (the proportion of words Ws among all documents assigned to topic K k ratio) to correct and adjust the topic assignment of the current word.

[0107] .

[0108] Step 4: Run the model and process the data.

[0109] The second part is the design of the poetry and wine data visual analysis system. The steps are as follows:

[0110] Step 1: Establish a visualization data layer to extract the relationship data between poets;

[0111] Step 2: Perform data analysis and processing;

[0112] Step 2.1: Data preprocessing: First, evaluate each part of the data to understand the entire data set and determine the relevance and consistency of the data with this topic. Then, perform data cleaning, data conversion, and data reduction (data subpackaging) to lay the foundation for the smooth progress of subsequent experiments.

[0113] Step 2.2: Execute the poetry sentiment analysis algorithm and poetry theme word extraction algorithm model;

[0114] Step 2.3: Perform data integration to update the data layer of the entire system, which contains three parts of data: basic information data of poems, poet relationship data, and poem theme and emotion data.

[0115] Step 3: Establish a visualization design layer. Visual coding and visualization layout design are performed for the above five types of data, mainly including the following parts: spatial distribution of poetry emotions, temporal distribution of poetry emotions, visual analysis of poet relationships, display of wine poetry themes, and display of poetry content.

[0116] Step 4: Focus on text data, then present the data in different types of visual charts and use interactive means to make the entire system more vivid and more scalable.

[0117] Embodiment three.

[0118] This embodiment also provides a computer terminal device, including:

[0119] one or more processors;

[0120] A memory, coupled to the processor, for storing one or more programs;

[0121] When one or more programs are executed by one or more processors, the one or more processors implement a multi-polar emotional intelligent visual analysis method for drinking poems.

[0122] Embodiment 4.

[0123] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, a method for intelligent visual analysis of multi-polar emotions in wine-themed poems is implemented.

[0124] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A multi-polar emotional intelligent visual analysis method for wine poetry, characterized in that: The following steps are involved: Acquire poetry and wine text data, create a Word2Vec associated word search model, and associate Chinese characters with the poetry and wine text data based on the Word2Vec associated word search model to obtain associated words; Marking the emotion of the poem based on the associated words to obtain the emotion label of the poem to be analyzed; Determine the emotion of the poem and the theme words of the poem based on the emotion tag of the poem to be analyzed; The process of obtaining the emotion label of the poem to be analyzed based on the associated word tagging of the poem emotion includes: Calculate the association score of the high-dimensional vector based on the associated words and the association degree; Determine the sentiment label of the poem to be analyzed based on the association score; The expression for calculating the association score is: In the formula, value ( x i ) e It is a Chinese character x i For the relevance of class e labels, x i is a Chinese character, n is the length of the poem, and e is set to be the emotion category label number; The process of determining the emotion of the poem and the theme words of the poem based on the emotion tag of the poem to be analyzed includes: Based on the emotion tags of the poem to be analyzed, the data with the emotion tags are obtained; Input the merged data with emotion labels into the LSTM deep learning algorithm model to determine the emotion of the poem; The theme words of the poem to be analyzed are extracted based on the implicit Dirichlet distribution algorithm to obtain the theme words of the poem.

2. The multi-polar emotional intelligent visual analysis method for wine poetry according to claim 1 is characterized in that: The process of associating Chinese characters with the poetry and wine text data based on the Word2Vec associated word search model includes: Preprocessing the poetry and wine text data to obtain preprocessed text data; Creating a Word2Vec associated word search model, and converting the preprocessed text data into a high-dimensional vector based on the Word2Vec associated word search model; Define the emotion tags of poems and lyrics, generate corresponding associated words for each emotion tag based on the Word2Vec associated word search model and give the degree of association.

3. The multi-polar emotional intelligent visual analysis method for wine poetry according to claim 1 is characterized in that: The process of extracting the subject words of the poem to be analyzed includes: Construct document-term matrix; Randomly assign a topic to each word in the document and calculate the proportion of the current word in the document assigned to the target topic and the proportion of the word in all documents assigned to the target topic; The topic assignment is adjusted based on the proportion of the current word in the document assigned to the target topic and the proportion of the word in all documents assigned to the target topic to determine the topic word of the poem to be analyzed.

4. A multi-polar emotional intelligent visual analysis system for wine poetry, characterized in that: The system is used to implement the multi-polar emotional intelligent visual analysis method of wine poetry as described in any one of claims 1 to 3, and comprises: the visual analysis system comprises a data layer, a data analysis and processing layer, a design layer and an interaction layer; Wherein, the data layer is used to obtain poetry and wine text data; The data analysis and processing layer is used to pre-process the poetry and wine text data and execute the poetry sentiment analysis algorithm and poetry theme word extraction algorithm model to determine the poetry sentiment and poetry theme words; The design layer is used to visualize poetry and wine data from the perspectives of text, time, space and relationship; The interaction layer is used to realize the interaction between the user and the visual interface.

5. A computer terminal device, characterized in that: include: one or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the multi-polar emotional intelligent visual analysis method of drinking poetry as described in any one of claims 1-3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the multi-polar emotional intelligent visual analysis method of drinking poetry as described in any one of claims 1-3 is implemented.

Images