A method and system for character annotation and automatic extraction of story plots

Abstract

This invention relates to a method and system for character annotation and automatic plot extraction. The method includes: transcribing and replacing personal pronouns in the story text with corresponding character names; identifying the protagonist from the character names; annotating story-related words associated with the protagonist, including story variable words, story slope words, and story intercept words; generating a story variable word matrix vector using the name, direction, and magnitude of the story variable words as columns; calculating plot scores based on the relative positions and magnitudes of the story variable words in the story variable word matrix vector; plotting a plot curve based on the relative positions and plot scores of the story variable words in the story text; calculating the influence of characters in the story text on the protagonist; and identifying and annotating the positive and negative roles of each character in the story based on the influence on the protagonist. This invention is widely used in the fields of plot extraction, story type identification, and character annotation.

Description

Technical Field

[0001] This invention relates to the field of natural language processing technology, and in particular to a method and system for automatically extracting character annotations and storylines. Background Technology

[0002] A story is a narrative information product primarily intended to persuade, record, and entertain others. It focuses on plot and character narration, using conflict as a driving force and employing narrative and detailed description to provide a meaningful experience for its target audience. Stories emphasize conciseness, intricate plots, singular conflict, vivid details, and ease of understanding and memorability, thereby enhancing the recall and recognition of the information. Compared to novels, stories are shorter, generally under 5000 words. The narrative typically revolves around the unfolding of the plot, which serves as the main thread of the story. While story plots are often complex, patterns emerge between different stories. For example, each story's plot typically includes a beginning, development, climax, and ending.

[0003] Automated extraction of story plots can provide a basis for rapid story comprehension. A plot is the organization of the "events" involved in a story. Typically, the storylines describing these events are called plots. Researchers have found that story plots are relatively fixed, with three common story models: Freytag's Pyramid proposed by German playwright Gustav Freytag, the hero's journey structure proposed by American author Joseph Campbell et al., and the famous "Boy Gets Girl" plot model proposed by Kurt Vonnegut. Automated extraction of story plots can provide story audiences with story lines, providing a basis for rapid understanding and in-depth study of the story content.

[0004] The automatic extraction of storylines also provides a basis and evaluation indicators for the automatic generation of data stories by computers. By extracting storylines, we can analyze whether the data stories automatically generated based on big data analysis conform to the general patterns of stories. Furthermore, we can find more big data stories that match the story, and then refine and optimize the data stories through related stories to achieve the goal of generating better stories.

[0005] In the area of ​​automated storytelling, existing research focuses on how to automatically generate a new story (including its plot), while research on how to automatically extract story plots remains unexplored. Regarding story plot extraction, related research has been limited to how to manually identify and optimize story plots, while methods and tools for automatic story plot extraction have not yet emerged. Therefore, how to achieve automatic story plot extraction has become a technical problem that needs to be solved. Summary of the Invention

[0006] To address the aforementioned problems, the purpose of this invention is to provide a method and system for character annotation and automatic extraction of storylines, thereby enabling automatic extraction and annotation of storylines.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] One aspect of the present invention is to provide a method for character annotation and automatic extraction of story plots, the method comprising the following steps:

[0009] Extract story-related terms associated with the protagonist from the story text; the story-related terms include at least story variable terms.

[0010] Generate a story variable word matrix vector using the name, direction, and magnitude of the story variable words as columns;

[0011] The plot score is calculated based on the relative position of the story variable words in the story variable word matrix vector and the size of the story variable words.

[0012] Plot a storyline curve based on the relative positions of story variables in the story text and their plot scores.

[0013] Preferably, the step of obtaining story-related keywords associated with the protagonist from the story text includes:

[0014] Translate and replace personal pronouns in the story text with the corresponding character names;

[0015] Identify the protagonist from the names of the characters in the story text;

[0016] Mark the story-related keywords in the story text that are associated with the protagonist.

[0017] Preferably, the step of transcribing and replacing personal pronouns in the story text with corresponding character names specifically includes: starting from the first sentence of the story text, replacing personal pronouns in the story text with corresponding character names according to the principle of proximity; wherein, the personal pronouns include first-person, second-person, and third-person pronouns in Chinese and English.

[0018] Preferably, identifying the protagonist from the character names in the story text specifically includes: statistically obtaining the frequency of each character name, and selecting one or more character names from the character names arranged in descending order of frequency as the protagonist.

[0019] Preferably, marking story-related keywords associated with the protagonist in the story text specifically includes:

[0020] Story variable words, story slope words, and story intercept words are extracted from the story text. Among them, story variable words are predicate verbs with the protagonist as the subject or object; story slope words are adverbs that modify the predicate; and story intercept words are adjectives that modify the protagonist.

[0021] The selected story variable words, story slope words, and story intercept words are labeled, and the labeling content includes direction and degree information;

[0022] Directional information refers to the direction relative to the protagonist. Directions that are beneficial to the protagonist are called "positive directions," those that are detrimental to the protagonist are called "negative directions," and those that have no relation to the protagonist are called "no directions."

[0023] The severity information includes two levels: "strong" and "weak".

[0024] Preferably, the step of calculating the plot score based on the relative position of the story variable word in the story variable word matrix vector and the size of the story variable word specifically involves: using the relative position of the story variable word in the variable word matrix vector as a dividing line, taking the difference between the proportion of reverse words before the dividing line and the proportion of reverse words after the dividing line as a coefficient, and multiplying this coefficient by the size of the story variable word to obtain the plot score of the story variable word.

[0025] Preferably, the step of drawing the plot curve based on the relative position and plot score of the story variable words in the story text specifically includes: drawing the plot curve in a Cartesian coordinate system with the relative position of the story variable words in the story text as the horizontal axis, the plot score of the story variable words as the vertical axis, and the points corresponding to the horizontal and vertical axes representing the story variable words.

[0026] Preferably, the relative position of the story variable words in the story text is: the position of the story variable words in the story text divided by the total number of valid words in the story.

[0027] Preferably, the method further includes: calculating the influence of characters in the story text on the protagonist, and identifying and labeling the positive and negative roles of each character in the story based on the influence on the protagonist, specifically:

[0028] The influence of the i-th character on the protagonist. iCalculated using the following formula:

[0029]

[0030] Among them, w i The word indicating the slope of the story, x i b is a variable word indicating the story. i Words indicating the intercept of a story;

[0031] Determine the influence of the i-th character on the protagonist. i If the value is greater than zero, the character is marked as a positive character from the protagonist's perspective; if it is less than zero, the character is marked as a negative character from the protagonist's perspective; if it is equal to zero, the character is someone who has no interest in the protagonist.

[0032] Another aspect of the present invention provides a system for character annotation and automatic extraction of storylines, the system comprising:

[0033] The related word recognition module is used to obtain story related words associated with the protagonist from the story text, and the story related words include at least story variable words;

[0034] The vector generation module is used to generate a story variable word matrix vector with the name, direction, and size of the story variable words as columns.

[0035] The score calculation module is used to calculate the plot score based on the relative position of the story variable words in the story variable word matrix vector and the size of the variable words.

[0036] The plotting module is used to draw plot curves based on the relative positions of story variable words in the story text and plot scores.

[0037] The present invention has the following advantages due to the adoption of the above technical solutions:

[0038] (1) High flexibility. By setting the direction of the verb predicate, this invention transforms multi-theme perspective problems into single-protagonist perspective problems. It can not only support multi-character scenes, but also automatically handle complex stories such as those with uncertain relationships between characters.

[0039] (2) Economic efficiency. This invention is a lightweight application that does not require a large amount of training data and computing resources, and can analyze a single document.

[0040] (3) Objectivity. The plot calculation in this invention is based on text analysis in the story context, which can objectively describe plot changes and truthfully reflect the story narrator's narrative ability.

[0041] (4) High scalability. This invention can be used not only for small data but also for big data, and it also supports deep learning, which can be incorporated into this invention.

[0042] (5) Fully automated. The extraction of storylines in this invention is a fully automated process that does not require human intervention.

[0043] (6) The story creators' word usage habits are highly adaptable. When calculating the part of speech of a story, only words of the same type appearing in the story are compared, and they are divided into strong and weak levels, which fully takes into account the habitual characteristics and shortcomings of the creators' word usage.

[0044] The method and system for character annotation and automatic plot extraction of the present invention can be widely used in the fields of plot extraction, story type identification, and character annotation. Attached Figure Description

[0045] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Throughout the drawings, the same reference numerals denote the same parts. In the drawings:

[0046] Figure 1 This is a flowchart of the method for character annotation and automatic story extraction according to an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram illustrating the specific implementation process of character annotation and automatic story extraction in an embodiment of the present invention;

[0048] Figure 3 This is a schematic diagram of the generated story x word matrix vector X according to an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram of the storyline curve of an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0051] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0052] To address the lack of automated methods for extracting storylines, this invention proposes a method and system for character annotation and automatic storyline extraction, which enables automatic extraction and annotation of storylines and supports the extraction of storylines from any story.

[0053] Example 1

[0054] like Figure 1 As shown, the present invention provides a method for character annotation and automatic extraction of storylines, comprising the following steps:

[0055] Step S1: Translate and replace the personal pronouns in the story text with the corresponding character names;

[0056] Step S2: Identify the protagonist from the character names in the story text;

[0057] Step S3: Mark the story-related words associated with the protagonist in the story text. The story-related words include story variable words, story slope words, and story intercept words.

[0058] Step S4: Generate a story variable word matrix vector using the name, direction, and size of the story variable words as columns;

[0059] Step S5: Calculate the plot score based on the relative position of the story variable words in the story variable word matrix vector and the size of the variable words;

[0060] Step S6: Draw a storyline curve based on the relative positions of story variable words in the story text and the plot score;

[0061] Step S7: Calculate the influence of the characters in the story text on the protagonist, and identify and label the positive and negative roles of each character in the story based on the influence of the characters on the protagonist.

[0062] In the method, story-related keywords associated with the protagonist are marked through steps S1 to S3.

[0063] like Figure 2 As shown, the specific implementation process of the method for character annotation and automatic extraction of story plot is further given.

[0064] In step S1, the personal pronouns in the story text D are transcribed and replaced with their corresponding character names. Specifically, starting from the first sentence of the story text, the personal pronouns in the story text are replaced with their corresponding character names according to the principle of proximity. The personal pronouns include first-person, second-person, and third-person pronouns in both Chinese and English. More specifically, starting from the first sentence of the story text D, the personal pronouns in the story text are replaced with their corresponding character names according to the principle of proximity. For Chinese personal pronouns, singular / plural forms are not distinguished. For English, the nominative / accusative case, singular / plural forms, and possessive adjectives / possessive nouns are all replaced with their corresponding character names. The specific Chinese and English personal pronouns to be processed are shown in Table 1.

[0065] Table 1: Personal pronouns that need to be replaced

[0066]

[0067] In step S2, after transing and replacing personal pronouns in the story text with their corresponding character names, the frequency of each character name is counted and sorted in descending order. The top-k characters with the highest frequency are designated as the protagonist. When counting character names, a switch option can be set to include character names appearing in spoken content (including self-talk) in the count. The pseudocode for the algorithm to identify the top-k characters is shown below:

[0068] Algorithm for identifying top-k characters

[0069] / / Input: string D. D is the story text.

[0070] / / Output: A list of top-kCharacters. The elements of top-kCharacters are the k main characters in story D.

[0071] Set the value of k

[0072] Define an empty list cList, i.e., cList = []

[0073] For c in D:

[0074] cList=cList.append(D.count(c))

[0075] cList.sort(ascending = False)

[0076] top-kCharacters = cList[0:k]

[0077] return top-k Characters

[0078] In step S3, story-related words associated with the protagonist are marked in the story text. Story-related words include story variable words, story slope words, and story intercept words; predicate verbs with the protagonist as the subject or object are selected from the story text as story variable words; adverbs modifying the predicate are selected from the story text as story slope words; and adjectives modifying the protagonist are selected from the story text as story intercept words.

[0079] In the specific annotation process, the annotation of story-related words is limited to three parts of speech: story x words (variable words), story w words (slope words), and story b words (intercept words). These represent predicate verbs with the protagonist as the subject or object, adverbs modifying the predicate, and adjectives modifying the protagonist in the story text, respectively. Other words are not processed or annotated. The content of story-related word annotation has two types of information: direction and degree. Among them, "direction" refers to the direction relative to the protagonist. Those that are beneficial to the protagonist are called "positive direction," those that are detrimental to the protagonist are called "negative direction," and those that have no relation to the protagonist are called "no direction." Degree is divided into two levels: "strong" and "weak." The method of degree classification is to perform a binary analysis of all adjectives (or adverbs) appearing in the story, dividing them into strong and weak levels.

[0080] As shown in Table 2, the specific scope, direction, and magnitude of story-related keywords associated with the protagonist are defined.

[0081] Table 2. Tags of stories related to the protagonist.

[0082]

[0083]

[0084] The pseudocode for the story-related keyword tagging algorithm is shown below:

[0085] Algorithm Story Related Word Tagging Method

[0086] / / Input: string D. D is the story text.

[0087] / / Output: The string D' with embedded story-related keywords

[0088]

[0089] In step S4, a story variable word matrix vector X is generated using the name, direction, and magnitude of the story variable words (x words) as columns. The shape of the matrix vector X is (N, 3), and the 0th to 2nd columns represent the name x of the i-th x word, respectively. i The direction of word x The size of x words ||x i ||.

[0090] The direction of the i-th word x Represented as:

[0091] Where m represents the modifier x in the story text. i The total number of adverbs, w j The direction value of the verb modified by the j-th adverb can be {-1, +1};

[0092] The size of the i-th word x ||x i || is represented as:

[0093] ||x i ||=w j ×x i +b j

[0094] Among them, w j Indicates the strength or weakness of a modifier of the predicate verb, x i The predicate verb indicating that the subject or object is the protagonist, b j It indicates the degree of strength or weakness of the protagonist.

[0095] The direction of word x The size of x words ||x i The calculation results are as follows: Figure 3 As shown, the direction of word x takes a value from the set {+1, 0, -1}, where the values ​​represent favorable, unfavorable, and no benefit to the protagonist, respectively. The magnitude of word x is ||x i The value of || indicates the degree to which it modifies the predicate and the protagonist.

[0096] In step S5, the plot score is calculated based on the relative position of the story variable word (x word) in the matrix vector X and the size of the related words. The plot score of the story variable word (x word) is determined by two factors: its relative position in the matrix vector X and its size. Specifically, using the relative position of the story variable word (x word) in the matrix vector X as a dividing line, the difference between the proportion of related words before and after the dividing line is used as a coefficient. This coefficient is multiplied by the size of the story variable word (x word) to obtain its plot score. The plot score of story word x is pScore(x). i The following is represented:

[0097] pScore(x i ) = (Percentage of previous antonyms - Percentage of subsequent antonyms) × ||xi ||

[0098] In step S6, a storyline curve is plotted based on the relative position of the story variable word (x word) in the story text and its plot score. In a Cartesian coordinate system, the storyline curve is plotted with the relative position of the story variable word (x word) in the story text as the abscissa and the plot score of the story variable word (x word) as the ordinate, with the points corresponding to the abscissa and ordinate representing the story variable word (x word).

[0099] A schematic diagram of the storyline curve is shown below. Figure 4 As shown. Points on the Cartesian coordinate system represent story variable words (x words), the horizontal axis (xLOC) represents the relative position of the story variable word (x word) in the story text, and the vertical axis (yLOC) is the plot score pScore of the story variable word (x word). The relative position xLOC of the story variable word (x word) is calculated as: the position of the story variable word (x word) in the story / the total number of valid words in the story.

[0100] The x-coordinate (xLOC) of the story variable word (x word) is represented as follows:

[0101]

[0102] Where, order(x) i ) represents the relative position of the story variable word (x word) in the story text, and count(D) represents the total number of valid words in the story text D.

[0103] The y-coordinate (yLOC) of the story variable word (x word) is represented as follows:

[0104] yLOC(x i ) = pScore(x i ), where pScore(x) i ) represents the plot score of the story variable word (x word).

[0105] according to Figure 4 As can be seen, all story variable words (x words) related to the protagonist are distributed along the curve. x1 is the first word appearing in the story text; at this point, the plot has not yet unfolded, and its relative position on the vertical axis is low. As the plot progresses, story variable words (x words) appear one by one. At the highest point of the curve, the plot score of the story variable word (x word) is the largest, and at this point, x... i The related word matrix vector X is in a relatively central position, and x i It has a strong modifying effect. As the story gradually comes to an end, the intensity of the words also gradually decreases, but compared to the first appearance of x1, it still has a stronger modifying effect than x1.

[0106] In step S7, the influence of the characters in the story text on the protagonist is calculated. Based on this influence, the positive and negative roles of each character in the story are identified and labeled. The story variable word (x-word) matrix X of all characters in the story text, excluding the protagonist, can be determined according to step 4. The influence of the i-th character on the protagonist is defined by Characters. i Calculated using the following formula:

[0107]

[0108] Among them, w i The word indicating the slope of the story, x i b is a variable word indicating the story. i Words indicating the truncated segment of a story.

[0109] Example 2

[0110] The above-described embodiment 1 provides a method for character annotation and automatic storyline extraction. Correspondingly, this embodiment provides a system for character annotation and automatic storyline extraction. The system provided in this embodiment can implement the method for character annotation and automatic storyline extraction of embodiment 1. This system can be implemented through software, hardware, or a combination of both. For example, the system may include integrated or separate functional modules or units to execute the corresponding steps in the methods of embodiment 1. Since the system in this embodiment is basically similar to the method embodiment, the description process in this embodiment is relatively simple. For relevant details, please refer to the description of embodiment 1. The system embodiment provided in this embodiment is merely illustrative.

[0111] The system for character annotation and automatic storyline extraction provided in this embodiment includes:

[0112] The related word recognition module is used to obtain story related words associated with the protagonist from the story text, and the story related words include at least story variable words;

[0113] The vector generation module is used to generate a story variable word matrix vector with the name, direction, and size of the story variable words as columns.

[0114] The score calculation module is used to calculate the plot score based on the relative position of the story variable words in the story variable word matrix vector and the size of the variable words.

[0115] The plotting module is used to draw plot curves based on the relative positions and plot scores of story variable words in the story text. The system further includes a character recognition module, used to calculate the influence of characters in the story text on the protagonist, and to identify and label each character's positive and negative roles based on this influence.

[0116] Specifically:

[0117] The influence of the i-th character on the protagonist. i Calculated using the following formula:

[0118]

[0119] Among them, w i The word indicating the slope of the story, x i b is a variable word indicating the story. i Words indicating the intercept of a story;

[0120] Determine the influence of the i-th character on the protagonist. i If the value is greater than zero, the character is marked as a positive character from the protagonist's perspective; if it is less than zero, the character is marked as a negative character from the protagonist's perspective; if it is equal to zero, the character is someone who has no interest in the protagonist.

[0121] The present invention has the following advantages due to the adoption of the above technical solutions:

[0122] (1) High flexibility. By setting the direction of the verb predicate, this invention transforms multi-theme perspective problems into single-protagonist perspective problems. It can not only support multi-character scenes, but also automatically handle complex stories such as those with uncertain relationships between characters.

[0123] (2) Economic efficiency. This invention is a lightweight application that does not require a large amount of training data and computing resources, and can analyze a single document.

[0124] (3) Objectivity. The plot calculation in this invention is based on text analysis in the story context, which can objectively describe plot changes and truthfully reflect the story narrator's narrative ability.

[0125] (4) High scalability. This invention can be used not only for small data but also for big data, and it also supports deep learning, which can be incorporated into this invention.

[0126] (5) Fully automated. The extraction of storylines in this invention is a fully automated process that does not require human intervention.

[0127] (6) The story creators' word usage habits are highly adaptable. When calculating the part of speech of a story, only words of the same type appearing in the story are compared, and they are divided into strong and weak levels, which fully takes into account the habitual characteristics and shortcomings of the creators' word usage.

[0128] The method and system for character annotation and automatic plot extraction of the present invention can be widely used in the fields of plot extraction, story type identification, and character annotation.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for character annotation and automatic extraction of story plot, characterized in that, Includes the following steps: Extract story-related terms associated with the protagonist from the story text; the story-related terms include at least story variable terms. Generate a story variable word matrix vector using the name, direction, and magnitude of the story variable words as columns; The plot score is calculated based on the relative position of the story variable words in the story variable word matrix vector and the size of the story variable words. Plot a storyline curve based on the relative positions of story variable words in the story text and their plot scores; The story's related keywords include: Story variable words, story slope words, and story intercept words are extracted from the story text. Among them, story variable words are predicate verbs with the protagonist as the subject or object; story slope words are adverbs that modify the predicate; and story intercept words are adjectives that modify the protagonist. The selected story variable words, story slope words, and story intercept words are labeled, and the labeling content includes direction and degree information; Directional information refers to the direction relative to the protagonist. Directions that are beneficial to the protagonist are called "positive directions," those that are detrimental to the protagonist are called "negative directions," and those that have no relation to the protagonist are called "no directions." The severity information includes two levels: "strong" and "weak". The plot score is calculated based on the relative position and magnitude of the story variable words in the story variable word matrix vector. Specifically, the plot score is obtained by using the relative position of the story variable word in the variable word matrix vector as a dividing line, taking the difference between the proportion of reverse words before and after the dividing line as a coefficient, and multiplying this coefficient by the magnitude of the story variable word to obtain the plot score of that story variable word. The plot score of the story variable word is expressed as follows: pScore(x i ) = (Percentage of previous antonyms - Percentage of subsequent antonyms) × ||x i || Where pScore(x) i ) is the story variable word x i The method further includes: calculating the influence of characters in the story text on the protagonist, and identifying and labeling the positive and negative roles of each character in the story based on the influence of these characters on the protagonist. Specifically, the influence of the i-th character on the protagonist is calculated using the Characters... i Calculated using the following formula: Among them, w i The word indicating the slope of the story, x i b is a variable word indicating the story. i Words indicating the truncated segment of a story.

2. The method for character annotation and automatic extraction of story plots according to claim 1, characterized in that, The process of obtaining story-related keywords associated with the protagonist from the story text includes: Translate and replace personal pronouns in the story text with the corresponding character names; Identify the protagonist from the names of the characters in the story text; Mark the story-related keywords in the story text that are associated with the protagonist.

3. The method for character annotation and automatic storyline extraction according to claim 2, characterized in that, The process of transcribing and replacing personal pronouns in the story text with corresponding character names specifically includes: starting from the first sentence of the story text, replacing personal pronouns in the story text with corresponding character names according to the principle of proximity; wherein, the personal pronouns include first-person, second-person, and third-person pronouns in both Chinese and English.

4. The method for character annotation and automatic extraction of storylines according to claim 2, characterized in that, The process of identifying the protagonist from the character names in the story text specifically includes: statistically obtaining the frequency of each character name, and selecting one or more character names from the character names arranged in descending order of frequency to be designated as the protagonist.

5. The method for character annotation and automatic extraction of story plots according to claim 1, characterized in that, The method of drawing a story plot curve based on the relative position and plot score of story variable words in the story text specifically includes: drawing a story plot curve in a Cartesian coordinate system with the relative position of story variable words in the story text as the horizontal axis, the plot score of story variable words as the vertical axis, and the points corresponding to the horizontal and vertical axes representing story variable words.

6. The method for character annotation and automatic extraction of storylines according to claim 1 or 5, characterized in that, The relative position of the story variable words in the story text is: the position of the story variable words in the story text divided by the total number of valid words in the story.

7. The method for character annotation and automatic extraction of story plots according to claim 1, characterized in that, The calculation of the influence of characters in the story text on the protagonist, and the identification and labeling of the positive and negative roles of each character in the story based on the influence on the protagonist, also includes: Determine the influence of the i-th character on the protagonist. i If the value is greater than zero, the character is marked as a positive character from the protagonist's perspective; if it is less than zero, the character is marked as a negative character from the protagonist's perspective; if it is equal to zero, the character is someone who has no interest in the protagonist.

8. A system for automatically extracting character annotations and storylines, used to implement the method for automatically extracting character annotations and storylines as described in any one of claims 1 to 7, characterized in that, include: The related word recognition module is used to obtain story related words associated with the protagonist from the story text, and the story related words include at least story variable words; The vector generation module is used to generate a story variable word matrix vector with the name, direction, and size of the story variable words as columns. The score calculation module is used to calculate the plot score based on the relative position of the story variable words in the story variable word matrix vector and the size of the variable words. The plotting module is used to draw plot curves based on the relative positions of story variable words in the story text and plot scores.

Images