Keyword extraction method, device and equipment and computer readable storage medium

Abstract

This application discloses a text information processing method, apparatus, device, and computer-readable storage medium. Specifically, it involves extracting the initial text vocabulary of the target text, obtaining a set of historical keywords associated with similar historical texts, and determining candidate keywords matching the initial text vocabulary from the historical keyword set. The candidate keyword containing the most initial text vocabulary is then determined as the target keyword corresponding to the target text. Therefore, this solution can first extract the initial text vocabulary of the target text and query the target historical texts, then select candidate keywords matching the initial text vocabulary from the set of historical keywords associated with the target historical texts, and finally determine the target keyword of the target text based on the candidate keywords. This improves the accuracy of text keyword extraction, allowing users to more easily obtain the key semantics of the target text and enhancing the user experience.

Description

Technical Field

[0001] This application relates to the field of computer technology, specifically to a keyword extraction method, apparatus, device, and computer-readable storage medium. Background Technology

[0002] With the continuous development of internet technology, keyword detection functions are becoming increasingly common in software products. Traditional keyword detection functions mainly rely on manual configuration by personnel. When the amount of text is large, the workload for configuration personnel becomes excessive, and duplicate keywords are prone to being extracted. Related technologies can segment words in text through keyword extraction algorithms to achieve the purpose of keyword extraction.

[0003] However, while the relevant technology can extract corresponding keywords from text, when the text in the speech is divided too sparsely, the extracted keywords will also become too scattered, causing the keywords to not reflect the semantics of the text well, thereby reducing the accuracy of keyword extraction and the user experience. Summary of the Invention

[0004] This application provides a text information processing method, apparatus, device, and computer-readable storage medium that can accurately extract keywords from target text and improve the user experience.

[0005] This application provides a text information processing method, including:

[0006] Obtain the target text and perform word segmentation on the target text to obtain at least one initial text vocabulary;

[0007] Query the historical text collection for target historical texts that are similar to the target text;

[0008] Obtain the set of historical keywords associated with the target historical text, and determine candidate keywords from the set of historical keywords that match the initial text vocabulary;

[0009] The candidate keyword that contains the most words in the initial text is determined as the target keyword corresponding to the target text.

[0010] Accordingly, embodiments of this application provide a text information processing apparatus, including:

[0011] An extraction unit is used to acquire target text and perform word segmentation on the target text to obtain at least one initial text vocabulary;

[0012] The query unit is used to query target historical texts that are similar to the target text from the historical text set;

[0013] A matching unit is used to obtain a set of historical keywords associated with the target historical text, and to determine candidate keywords from the set of historical keywords that match the initial text vocabulary;

[0014] The determining unit is used to determine the candidate keyword that contains the most words in the initial text as the target keyword corresponding to the target text.

[0015] In some embodiments, the query unit is further configured to:

[0016] Determine the text similarity between the target text and each historical text in the historical text set;

[0017] Based on the text similarity relationship, target historical texts similar to the target text are selected from the historical text set.

[0018] In some embodiments, the query unit is further configured to:

[0019] The historical texts in the historical text set are sorted in descending order of text similarity to obtain a historical text sequence.

[0020] Select the target historical text that is first sorted from the historical text sequence.

[0021] In some embodiments, the determining unit is further configured to:

[0022] Identify whether there is a coverage relationship between any two candidate keywords;

[0023] If it is determined that there is no overlap relationship between any two candidate keywords, then all of the candidate keywords are determined as target keywords;

[0024] If two candidate keywords with a covering relationship are identified, the target candidate keyword with the longest word information is determined from the two candidate keywords with a covering relationship, and the target candidate keyword is determined as the target keyword.

[0025] In some embodiments, the text information processing device further includes a display unit for:

[0026] The target keywords are identified as the core keywords of the target text;

[0027] The core keywords of the target text are displayed on the text configuration page.

[0028] In some embodiments, the text information processing device further includes a display unit for:

[0029] From the plurality of initial text words, identify the initial text words that do not match the candidate keywords;

[0030] The initial text words that did not match the candidate keywords are identified as other keywords of the target text, wherein the other keywords are displayed in the vocabulary window corresponding to more components in the text configuration page.

[0031] In some embodiments, the text information processing apparatus further includes an extension unit for:

[0032] Input the target keyword into a preset vocabulary expansion model to obtain the expanded keywords corresponding to the target keyword;

[0033] The extended keywords are identified as the extended vocabulary of the target text, and the extended vocabulary is displayed in the vocabulary window corresponding to the extended component in the text configuration page.

[0034] In some embodiments, the extension unit is further configured to:

[0035] Obtain at least one preset combination word that has a combination relationship with the target keyword from the preset combination word model, and combine the target keyword with the preset combination word to obtain at least one extended word group;

[0036] Based on the word similarity between the target keyword and each reference word in the preset similar word model, the reference words with word similarity greater than the preset word similarity threshold are selected as the extended similar words of the target keyword;

[0037] The extended phrases and extended similar words are used as extended keywords for the target text.

[0038] In some embodiments, the text information processing apparatus further includes a storage unit for:

[0039] The target keywords are associated with the target text.

[0040] The target text, after being associated with the target keyword, is saved as historical text to the historical text collection.

[0041] Furthermore, this application also provides a computer device, including a processor and a memory, wherein the memory stores a computer program, and the processor is used to run the computer program in the memory to implement the steps in the text information processing method provided in this application.

[0042] Furthermore, embodiments of this application also provide a computer-readable storage medium storing a plurality of instructions adapted for loading by a processor to execute steps in any of the text information processing methods provided in embodiments of this application.

[0043] Furthermore, embodiments of this application also provide a computer program product, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the steps of any of the text information processing methods provided in embodiments of this application.

[0044] This application embodiment obtains target text and performs word segmentation on the target text to obtain at least one initial text vocabulary; it then queries a set of target historical texts similar to the target text from a historical text collection; it obtains a set of historical keywords associated with the target historical texts and determines candidate keywords matching the initial text vocabulary from the historical keyword set; finally, it determines the target keyword corresponding to the target text as the candidate keyword containing the most initial text vocabulary. Therefore, this solution first extracts the initial text vocabulary of the target text, then queries similar target historical texts, matches corresponding candidate keywords from the set of historical keywords associated with the target historical texts based on the initial text vocabulary, and thus determines the target keyword of the target text based on the candidate keywords. This improves the accuracy of text keyword extraction, allowing users to more easily obtain the key semantics of the target text and enhancing the user experience. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of a scenario for the text information processing system provided in an embodiment of this application;

[0047] Figure 2 This is a flowchart illustrating the steps of the text information processing method provided in the embodiments of this application;

[0048] Figure 3 This is a schematic flowchart of another step of the text information processing method provided in the embodiments of this application;

[0049] Figure 4 This is a block flowchart illustrating the text information processing method provided in the embodiments of this application;

[0050] Figure 5 This is a schematic diagram of the structure of the text information processing device provided in the embodiments of this application;

[0051] Figure 6 This is a schematic diagram of the structure of the computer device provided in the embodiments of this application. Detailed Implementation

[0052] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0053] This application provides a text information processing method, apparatus, device, and computer-readable storage medium. This application will describe the text information processing apparatus from the perspective of a text information processing device, which can be integrated into a computer device. This computer device can be a terminal device, specifically a terminal device mounted on a text information processing tool, i.e., a voice terminal. Furthermore, the terminal device can also be other types of devices, such as a television, smartphone, tablet computer, laptop computer, desktop computer, smart speaker, smartwatch, smart wearable device, etc.; however, it is not limited to these.

[0054] For example, see Figure 1 This is a schematic diagram of a scenario for a text information processing system provided in an embodiment of this application. The scenario includes a terminal or a server.

[0055] Specifically, the terminal can be a voice terminal, used to acquire target text, perform word segmentation on the target text to obtain at least one initial text vocabulary; query target historical texts similar to the target text from the historical text set; acquire a set of historical keywords associated with the target historical text, and determine candidate keywords that match the initial text vocabulary from the historical keyword set; and determine the candidate keyword that contains the most initial text vocabulary as the target keyword corresponding to the target text.

[0056] The following sections provide detailed descriptions of each example. It should be noted that the order in which the embodiments are described is not intended to limit the preferred order of the embodiments.

[0057] In this embodiment, the description will be from the perspective of a text information processing device, which can be specifically integrated into a computer device such as a terminal device or a server. See also Figure 2 , Figure 2 The following is a flowchart illustrating the steps of a text information processing method provided in this application embodiment. Taking a terminal device as an example, the terminal device is a terminal mounted on a text information processing tool. When the processor on the terminal device executes the program corresponding to the text information processing method, the specific flow of the text information processing method is as follows:

[0058] 101. Obtain the target text and perform word segmentation on the target text to obtain at least one initial text vocabulary.

[0059] In this embodiment of the application, in order to obtain the corresponding keyword information from the current target text, the target text can be segmented to obtain at least one initial text vocabulary, and then the keyword information of the target text can be inferred based on the initial text vocabulary to improve the accuracy of the keyword information.

[0060] The target text can be text information from which the user needs to extract keyword information, such as a piece of text or a string of characters.

[0061] The initial text vocabulary can be the vocabulary that makes up the target text. For example, if the content of target text A is "It is a century-old brand", then the initial text vocabulary of target text A can be "century-old" and "brand".

[0062] Specifically, users can generate target text through keyboard or voice input. After obtaining the target text, it is segmented into words. Common text segmentation methods include unary segmentation, bigram segmentation, and maximum matching. Taking maximum matching as an example, first, the maximum length of the target text words can be determined, for example, 5 characters. Then, 5 characters are extracted from the target text one by one from beginning to end to form a string. It is determined whether it is a word. If it is, it is extracted; otherwise, its last character is removed, and the previous step is repeated until the entire text is processed. The words that can form words are determined as the initial vocabulary of the target text.

[0063] For example, the target text A is "I love Beijing's food". The target text A is segmented as follows: the maximum length is 3. First, "I love Beijing", "love Beijing", "Beijing's", and "food" are extracted. Then, word judgment is performed. Finally, the initial text vocabulary result of the target text A is "I", "love", "Beijing", "of", and "food".

[0064] By using the above methods, at least one initial text vocabulary can be obtained by segmenting the target text. Then, the keyword information of the target text can be inferred based on the initial text vocabulary, thereby improving the accuracy of the keyword information.

[0065] 102. Query the target historical text that is similar to the target text from the historical text collection.

[0066] In this embodiment of the application, in order to obtain the corresponding keyword information from the current target text, the target historical text similar to the target text can be queried from the historical text set. Thus, the initial text vocabulary of the target text can be improved based on the keyword information associated with the target historical text, so that more accurate keywords can be extracted from the target text.

[0067] The historical text collection can be a collection of texts that save reference texts, or it can be a collection of texts that users set up in advance to save historical texts. After extracting the corresponding keywords from the target text, the target text can be saved into this historical text collection.

[0068] The target historical text can be a historical text that has a high degree of semantic and / or lexical similarity to the target text.

[0069] Specifically, after segmenting the target text to obtain its initial vocabulary, historical texts with high similarity to the target text can be queried from a pre-defined historical text set as associated target historical texts. These historical texts can also be segmented to extract their initial historical text vocabulary. Several initial historical text vocabularys (e.g., 20) are extracted from each historical text and merged into a set. To avoid differences in text length, the relative word frequencies of several initial historical text vocabularys can be used to calculate the word frequency of each historical text relative to the words in this set, generating a word frequency vector for each historical text. Similarly, the word frequency vector of the target text is calculated using this method. The cosine similarity between the two vectors is then calculated; a higher value indicates greater similarity. In this way, target historical texts similar to the target text are obtained from the historical text set, providing a reference for subsequent extraction of target keywords from the target text.

[0070] In some implementations, the target historical text can be selected based on the text similarity between the target text and each historical text in the historical text set. For example, step 102, "querying target historical texts similar to the target text from the historical text set," may include:

[0071] (102.1) Determine the text similarity between the target text and each historical text in the historical text set;

[0072] (102.2) Based on the relationship of text similarity, select target historical texts that are similar to the target text from the historical text set.

[0073] Text similarity can be the degree of similarity between the constituent words in the target text and the constituent words in the historical text. For example, if each character in the target text is the same as each character in the historical text, then the text similarity between the two is 1.

[0074] Specifically, the target text is segmented to obtain its initial vocabulary. Then, the text similarity between the target text and each historical text in the historical text set is calculated. Based on the text similarity between the target text and each historical text, the corresponding target historical text is selected. It is worth noting that, to avoid selecting invalid historical texts, a text similarity threshold can be preset. Historical texts with a text similarity lower than the threshold are excluded, and only historical texts with a text similarity higher than the threshold are selected as the target historical texts. This avoids selecting historical texts with very low similarity. When multiple target historical texts need to be selected, they can be selected in descending order of their text similarity.

[0075] In some implementations, a sequence of historical texts can be obtained based on the text similarity between the target text and each historical text in the historical text set. Then, the target historical text that appears first in the sequence can be selected. For example, step (102.2), "selecting target historical texts similar to the target text from the historical text set based on the magnitude of text similarity," can include:

[0076] (102.2.1) Sort the historical texts in the historical text set in descending order of text similarity to obtain a historical text sequence;

[0077] (102.2.2) Select the target historical text that is ranked first from the historical text sequence.

[0078] The historical text sequence can be a text sequence obtained by sorting the various historical texts in the historical text collection.

[0079] Specifically, after calculating the text similarity between the target text and each historical text in the historical text set, the historical texts can be sorted from largest to smallest according to the text similarity to obtain a historical text sequence. When it is necessary to select the target historical text of the target text, starting from the first historical text in the historical text sequence, the corresponding number of historical texts are selected as the target historical texts in sequence.

[0080] By using the above methods, we can query historical texts similar to the target text from the historical text collection. This allows us to refine the initial vocabulary of the target text based on the keyword information associated with the target historical text, enabling us to extract more accurate keywords from the target text.

[0081] 103. Obtain the set of historical keywords associated with the target historical text, and determine the candidate keywords that match the initial text vocabulary from the set of historical keywords.

[0082] In this embodiment of the application, in order to obtain more accurate keywords from the current target text, a set of historical keywords associated with the target historical text can be obtained, and candidate keywords that match the initial text vocabulary can be determined from the set of historical keywords. Thus, the initial text vocabulary of the target text can be adjusted by manually extracting the configured reference keywords to obtain more accurate keywords.

[0083] Among them, the historical keyword set can be a set of words that have keyword association with the target historical text, which can represent the semantics of the target historical text, and the historical keywords can be keywords that the user has configured for the target historical text in advance.

[0084] Among them, candidate keywords can be historical keywords that contain the initial text words in the historical keyword set. For example, if the initial text words are "century" and "brand", then the candidate keyword can be "century-old brand".

[0085] Specifically, after retrieving a target historical text similar to the target text from the historical text set, a set of historical keywords associated with the target historical text is obtained from the keyword set. Based on the initial text vocabulary of the target text, a query is performed from the historical keyword set to see if there are any historical keywords containing the initial text vocabulary. If the historical keywords contain the corresponding initial text vocabulary, then the historical keywords containing the corresponding initial text vocabulary are identified as candidate keywords.

[0086] For example, the set of historical keywords associated with target historical text A contains three historical keywords: "century-old brand", "world-class", and "trustworthy". The initial text words corresponding to target text a have three words: "century", "brand", and "first-class". Therefore, the candidate keyword for the target text is "century-old brand".

[0087] By using the above methods, we can obtain a set of historical keywords associated with the target historical text, and determine candidate keywords that match the initial text vocabulary from the set of historical keywords. In this way, we can manually extract the configured reference keywords to adjust the initial text vocabulary of the target text and obtain more accurate keywords.

[0088] 104. The candidate keyword that contains the most words from the initial text is identified as the target keyword for the target text.

[0089] In this embodiment of the application, in order to obtain more accurate keywords from the current target text, the candidate keyword containing the most words in the initial text can be determined as the target keyword corresponding to the target text. This allows the target keyword to better display the key information of the target text, thereby making it easier for users to configure the keyword information of the target text, improving the accuracy of text information processing, and enhancing the user experience.

[0090] Among them, target keywords can be words that best represent the key information of the target text. For example, the target keyword for the target text "Company A is a century-old company" can be "century-old company".

[0091] Specifically, based on the set of historical keywords associated with the target historical text, after obtaining candidate keywords that match the initial text words, if there are multiple candidate keywords and multiple candidate keywords contain the same initial keyword, then query the number of initial text words contained in each candidate keyword. The more initial text words contained in a candidate keyword, the more accurately it reflects the key information of the target text. The candidate keyword containing the most initial text words is determined as the target keyword corresponding to the target text.

[0092] For example, if the target text A is "Xiaoming won the league championship", the initial text words are "league" and "championship", and the candidate keywords are "league championship" and "championship" respectively, then the candidate keyword "league championship" contains two initial text words, "league" and "championship", while the candidate keyword "championship" contains only one initial text word, "championship". Therefore, the candidate keyword "league championship" better represents the key information of the target text A, and the candidate keyword "league championship" is the target keyword of the target text A.

[0093] In some implementations, there may be multiple candidate keywords. The target keyword can be selected based on its length by determining whether there is an overlap relationship between any two candidate keywords. For example, step 104, "determining the candidate keyword containing the most words from the initial text as the target keyword corresponding to the target text," may include:

[0094] (104.1) Identify whether there is a covering relationship between any two candidate keywords;

[0095] (104.2) If it is found that there is no covering relationship between any two candidate keywords, then all candidate keywords are determined as target keywords;

[0096] (104.3) If two candidate keywords with a covering relationship are identified, the target candidate keyword with the longest word information is determined from the two candidate keywords with a covering relationship, and the target candidate keyword is determined as the target keyword.

[0097] The word information length can be either the number of words in the initial text contained in the candidate keywords, or the length of the words in the candidate keywords.

[0098] Specifically, after identifying candidate keywords that match the initial text vocabulary from the historical keyword set, if multiple candidate keywords are matched, any two candidate keywords are selected, and it is determined whether there is a word coverage relationship between the two candidate keywords. That is, it is determined whether all the initial text vocabulary in one candidate keyword exists in another candidate keyword. For example, candidate keyword A

[0099] There is a coverage relationship between "century-old" and candidate keyword B "century-old brand," but there is no coverage relationship between candidate keyword B "century-old brand" and candidate keyword C "century-old talent cultivation." If there is no coverage relationship between any two selected candidate keywords (i.e., neither candidate keyword is covered by the other), then all candidate keywords without coverage relationships are determined as target keywords. If a coverage relationship is identified between any two candidate keywords, then the word information length of the corresponding candidate keywords is obtained, and the candidate keyword with the largest word information length between the two candidate keywords is taken as the target keyword.

[0100] For example, based on the three initial text terms "century", "brand", and "enterprise" corresponding to the target text, three candidate keywords were determined from the set of historical keywords associated with the target historical text corresponding to the target text. These are candidate keyword A "century", candidate keyword B "century brand", and candidate keyword C "century enterprise". By judging the coverage relationship between the three candidate keywords, it can be concluded that: there is a coverage relationship between candidate keyword A "century" and candidate keyword B "century brand", and there is also a coverage relationship between candidate keyword A "century" and candidate keyword C "century enterprise". There is no coverage relationship between candidate keyword B "century brand" and candidate keyword C "century enterprise". Therefore, the target keywords are "century brand" and "century enterprise".

[0101] In some implementations, the text information processing method may further identify target keywords as the core keywords of the target text and display them on the text configuration page. For example, after step 104, "identifying the candidate keywords containing the most initial text words as the target keywords corresponding to the target text," it may also include:

[0102] (104.a.1) Identify the target keywords as the core keywords of the target text;

[0103] (104.a.2) Display the core keywords of the target text in the text configuration page.

[0104] The text configuration page can be a page on the user terminal used to display the keyword information corresponding to the target text. The keyword information of the target text can be viewed through the text configuration page.

[0105] Specifically, by determining whether there is an overlap relationship between multiple candidate keywords, the target keyword of the target text is identified. This target keyword is then designated as the core keyword of the target text, which is the word that best represents the target text. The core keyword is then sent to the text configuration page on the user's terminal and displayed to the user. In this way, the user can accurately view the core keyword of the target text, thus improving the user experience.

[0106] In some implementations, the number of initial text words may be multiple. The text information processing method may further include first identifying initial text words that do not match candidate keywords, and then identifying these initial text words as other keywords of the target text and displaying them on the text configuration page. For example, after step 104, "identifying the candidate keyword containing the most initial text words as the target keyword corresponding to the target text," the method may further include:

[0107] (104.b.1) From a plurality of initial text words, identify the initial text words that do not match the candidate keywords;

[0108] (104.b.2) Initial text words that do not match candidate keywords are identified as other keywords for the target text, wherein other keywords are displayed in the vocabulary window corresponding to more components in the text configuration page.

[0109] Other keywords can be secondary keywords of the target text, mainly displayed as reference words for the target keywords in the window corresponding to more components in the text configuration page.

[0110] Specifically, the target text contains multiple initial text words. When matching the initial text words with the set of historical keywords associated with the target historical text, some initial text words do not match the corresponding candidate keywords. In this case, the initial text words that do not match the corresponding candidate keywords are used as other keywords of the target text. When more components are selected in the selected text configuration page, the other keywords are displayed in the corresponding word window, so that users can view the other keywords of the target text through the user terminal.

[0111] For example, the three initial text words corresponding to the target text are "century", "brand" and "enterprise". The set of historical keywords associated with the target historical text has three historical keywords, namely historical keyword A "century" and historical keyword B "century brand". Since the initial text word "enterprise" cannot match any of the historical keywords, the initial text word "enterprise" will be displayed as other keywords in the word window corresponding to more components in the text configuration page.

[0112] By using the above methods, the candidate keywords that contain the most words from the initial text can be identified as the target keywords for the target text. This allows the target keywords to better display the key information of the target text, making it easier for users to configure the keyword information of the target text, improving the accuracy of text information processing, and enhancing the user experience.

[0113] In this embodiment of the application, the text information processing method further includes step 105, which may include:

[0114] (105.1) Input the target keyword into the preset vocabulary expansion model to obtain the expanded keywords corresponding to the target keyword;

[0115] (105.2) The extended keywords are identified as the extended vocabulary of the target text, and the extended vocabulary is displayed in the vocabulary window corresponding to the extended component in the text configuration page.

[0116] The preset vocabulary expansion model can be a vertical domain linguistic model or an existing multimodal large model, used to expand similar keywords for the target keyword.

[0117] Among them, extended keywords can be combinations of target keywords and other words, or words similar to target keywords.

[0118] Among them, the extended component can be a component in the text configuration page used to provide users with extended keyword services. When the service corresponding to the extended component is triggered, the extended vocabulary of the target text will be displayed in the vocabulary window corresponding to the extended component in the text configuration page.

[0119] Specifically, after identifying the target keywords for the target text, these keywords can be expanded based on the user's needs. The target keywords can be input into a pre-trained vocabulary expansion model, which can be a model trained specifically for the target text, such as a vertical domain linguistics model. Upon detecting the target keywords, the pre-trained vocabulary expansion model generates expanded keywords for that target keyword. These expanded keywords are then designated as expanded vocabulary for the target text and displayed in the vocabulary window corresponding to the expansion component on the text configuration page. In this way, when the content of the target text is too long, and the user is not entirely certain whether there are other expressions for the target keywords, some expanded vocabulary related to the target keywords can be provided to the user.

[0120] In some implementations, the preset extended vocabulary model includes a preset combined vocabulary model and a preset similar vocabulary model. The extended keywords include extended words combined with extended similar words. The target keyword can be combined with the preset combined vocabulary to obtain at least one extended word group. Based on the vocabulary similarity between the target keyword and each reference word in the preset similar vocabulary model, the extended similar words corresponding to the target keyword are selected. For example, step (105.1) may include:

[0121] (105.1.1) Obtain at least one preset combination word that has a combination relationship with the target keyword from the preset combination word model, and combine the target keyword with the preset combination word to obtain at least one extended word group;

[0122] (105.1.2) Based on the word similarity between the target keyword and each reference word in the preset similar word model, select the reference words with a word similarity greater than the preset word similarity threshold as the extended similar words of the target keyword;

[0123] (105.1.3) Extended phrases and extended similar words are used as extended keywords of the target text.

[0124] Among them, the preset combination vocabulary model can be a linguistic model trained on the target text content, which can generate keyword association combinations within the same language range as the target text. For example, if the current target keyword is "century", and the preset combination vocabulary model is a model whose content is mostly operation and promotion, then the combination vocabulary generated based on the target keyword may be "century-old brand".

[0125] Among them, the preset similar word model can be a multimodal large model for a general domain. The range of extended similar words generated based on the preset similar word model is wider. For example, if the target keyword is "we", the generated extended similar words may be "us", "us", etc.

[0126] Specifically, after determining the target keywords corresponding to the target text, in response to the user's keyword expansion needs, the target keywords are input into a preset combination vocabulary model. This model retrieves at least one word that has a combinational relationship with the target keywords. The target keywords are then combined with this word to obtain corresponding extended phrases. For example, "century-old" and "brand" can be combined to form "century-old brand," thus expanding the target text with relevant extended phrases. Simultaneously, the target keywords are input into a preset similarity vocabulary model. This model calculates the vocabulary similarity between the target keywords and various reference words. Specifically, it can be done by calculating the Euclidean distance between the target keywords and reference words. Reference words with a similarity greater than a preset similarity threshold are identified as extended similar words of the target keywords. These extended phrases and extended similar words are then used as extended keywords for the target text, which can be subsequently displayed to the user through the text configuration page.

[0127] By inputting the target keyword into the preset vocabulary expansion model, the expanded keywords corresponding to the target keyword can be obtained. These expanded keywords are then displayed as expanded vocabulary for the target text in the vocabulary window corresponding to the expanded component on the text configuration page. This provides users with more keyword reference directions for the target text and improves the user experience.

[0128] In this embodiment of the application, the text information processing method further includes step 106, which may include:

[0129] (106.1) Associate the target keywords with the target text;

[0130] (106.2) Save the target text after it has been processed and associated with the target keyword as historical text to the historical text collection.

[0131] Specifically, after identifying the corresponding target keywords based on the target text, the target keywords and target text are associated, and the associated target text is saved as a historical reference text in the historical text collection. This increases the number of reference texts in the historical text collection, providing more reference information for the text information processing of the next text.

[0132] By using the above methods, the target text, after being processed and associated with the target keywords, can be saved as historical text to the historical text collection, thereby expanding the reference text.

[0133] By implementing any one or a combination of implementation methods in the embodiments of this application, the application scenarios of text information processing can be realized.

[0134] As can be seen from the above, the embodiments of this application can extract the first voiceprint feature of the first speech and the second voiceprint feature of the second speech, then obtain the preset fusion weight between the first speech and the second speech, and fuse the first voiceprint feature and the second voiceprint feature according to the preset fusion weight to obtain the third voiceprint feature. Then, the first feature similarity between the third voiceprint feature and the first voiceprint feature is determined, and the second feature similarity between the third voiceprint feature and the second voiceprint feature is determined. Then, the similarity ratio between the first feature similarity and the second feature similarity is calculated, and the fusion weight difference between the similarity ratio and the preset fusion weight is determined. Finally, when the fusion weight difference is less than or equal to the preset weight threshold, the fused speech is generated based on the third voiceprint feature. Thus, this solution can first extract the initial text vocabulary of the target text, then query the target historical text similar to the target text, and match the corresponding candidate keywords from the historical keyword set associated with the target historical text based on the initial text vocabulary, thereby determining the target keywords of the target text based on the candidate keywords. In this way, the accuracy of text keyword extraction is improved, making it easier for users to obtain the key semantics of the target text and improving the user experience.

[0135] Based on the method described in the above embodiments, the following examples will provide further detailed explanations.

[0136] This application uses a text information processing device as an example to further describe the text information processing method provided in this application. Wherein, Figure 3 This is a schematic flowchart of another step in the text information processing method provided in this application embodiment. For ease of understanding, this application embodiment combines... Figure 3 Describe it.

[0137] In this embodiment, the description will focus on a text information processing device, which can be integrated into a computer device. When the processor on the computer device executes the program instructions corresponding to the data transmission method, the specific flow of the text information processing method is as follows:

[0138] 201. The server obtains the target text and performs word segmentation on the target text to obtain multiple initial text words.

[0139] The target text can be text information from which the user needs to extract keyword information, such as a piece of text or a string of characters.

[0140] The initial text vocabulary can be the vocabulary that makes up the target text. For example, if the content of target text A is "It is a century-old brand", then the initial text vocabulary of target text A can be "century-old" and "brand".

[0141] Specifically, users can generate target text through keyboard or voice input. After receiving the target text, the server performs word segmentation. Common word segmentation methods include unary segmentation, bigram segmentation, and maximum matching. Taking maximum matching as an example, first, the maximum length of the target text words can be determined, for example, 5 characters. Then, 5 characters are extracted from the target text one by one from beginning to end to form a string. It is determined whether it is a word. If it is, it is extracted; otherwise, the last character is removed, and the previous step is repeated until the entire text is processed. The words that can form words are determined as the initial vocabulary of the target text.

[0142] For example, the target text A is "I love Beijing's food". The target text A is segmented as follows: the maximum length is 3. First, "I love Beijing", "love Beijing", "Beijing's", and "food" are extracted. Then, word judgment is performed. Finally, the initial text vocabulary result of the target text A is "I", "love", "Beijing", "of", and "food".

[0143] 202. The server calculates the text similarity between the target text and each historical text in the historical text set, and sorts the historical texts in the historical text set in descending order of text similarity to obtain a historical text sequence.

[0144] The historical text collection can be a collection of texts that save reference texts, or it can be a collection of texts that users set up in advance to save historical texts. After extracting the corresponding keywords from the target text, the target text can be saved into this historical text collection.

[0145] Text similarity can be the degree of similarity between the constituent words in the target text and the constituent words in the historical text. For example, if each character in the target text is the same as each character in the historical text, then the text similarity between the two is 1.

[0146] The historical text sequence can be a text sequence obtained by sorting the various historical texts in the historical text collection.

[0147] Specifically, the server performs word segmentation on the target text to obtain the initial vocabulary of the target text. Then, it calculates the text similarity between the target text and each historical text in the historical text set. Specifically, the historical text can also be segmented to extract the initial historical text vocabulary. Several initial historical text vocabulary words (e.g., 20 words) are extracted from each historical text and merged into a set. To avoid differences in text length, the relative word frequencies of several initial historical text vocabulary words can be used to calculate the word frequency of each historical text for the words in this set, generating a word frequency vector for each historical text. Similarly, the word frequency vector of the target text is calculated in the same way. The cosine similarity between the two vectors is calculated. The larger the value, the more similar they are. Then, based on the text similarity between the target text and each historical text, the historical texts are sorted from largest to smallest to obtain the historical text sequence.

[0148] 203. The server selects multiple target historical texts that are sorted first from the historical text sequence.

[0149] The target historical text can be a historical text that has a high degree of semantic and / or lexical similarity to the target text.

[0150] Specifically, after the server calculates the historical text sequence, it selects the corresponding number of historical texts as the target historical texts, starting from the first historical text in the sequence, according to the required number of historical texts.

[0151] 204. The server obtains a set of historical keywords associated with multiple target historical texts, and determines multiple candidate keywords from the historical keyword set that match the initial text vocabulary.

[0152] Among them, the historical keyword set can be a set of words that have keyword association with the target historical text, which can represent the semantics of the target historical text, and the historical keywords can be keywords that the user has configured for the target historical text in advance.

[0153] Among them, candidate keywords can be historical keywords that contain the initial text words in the historical keyword set. For example, if the initial text words are "century" and "brand", then the candidate keyword can be "century-old brand".

[0154] Specifically, after the server finds a target historical text similar to the target text from the historical text set, it obtains a set of historical keywords associated with the target historical text from the keyword set. Based on the initial text vocabulary of the target text, it checks whether there are historical keywords containing the initial text vocabulary in the historical keyword set. If the historical keywords contain the corresponding initial text vocabulary, then the historical keywords containing the corresponding initial text vocabulary are determined as candidate keywords.

[0155] For example, the set of historical keywords associated with target historical text A contains three historical keywords: "century-old brand", "world-class", and "trustworthy". The initial text words corresponding to target text a have three words: "century", "brand", and "first-class". Therefore, the candidate keyword for the target text is "century-old brand".

[0156] 205. When the server identifies that there is no overlap between any two candidate keywords, it will determine all of the candidate keywords as the target keyword.

[0157] Among them, target keywords can be words that best represent the key information of the target text. For example, the target keyword for the target text "Company A is a century-old company" can be "century-old company".

[0158] Specifically, after the server identifies candidate keywords matching the initial text vocabulary from the historical keyword set, if multiple candidate keywords are matched, it selects any two candidate keywords and identifies whether there is a word coverage relationship between them. That is, it determines whether all the initial text vocabulary in one candidate keyword exists in the other candidate keyword. For example, there is a coverage relationship between candidate keyword A "century" and candidate keyword B "century brand", but there is no coverage relationship between candidate keyword B "century brand" and candidate keyword C "century-old talent cultivation". If there is no coverage relationship between any two randomly selected candidate keywords, that is, neither candidate keyword is covered by the other, then all candidate keywords without coverage relationship are determined as target keywords.

[0159] 206. When the server identifies a covering relationship between any two candidate keywords, it determines the target candidate keyword with the longest word information from the two candidate keywords with the covering relationship, and identifies the target candidate keyword as the target keyword.

[0160] The word information length can be either the number of words in the initial text contained in the candidate keywords, or the length of the words in the candidate keywords.

[0161] Specifically, if the server identifies an overlap relationship between any two candidate keywords, it obtains the word information length corresponding to the candidate keywords. Specifically, it can query the number of initial text words contained in these two candidate keywords. The more initial text words contained in the candidate keywords, the more accurately the key information of the target text can be reflected, and the longer the corresponding word information length, the candidate keyword with the largest word information length between the two candidate keywords is taken as the target keyword.

[0162] For example, based on the three initial text terms "century", "brand", and "enterprise" corresponding to the target text, three candidate keywords were determined from the set of historical keywords associated with the target historical text corresponding to the target text. These are candidate keyword A "century", candidate keyword B "century brand", and candidate keyword C "century enterprise". By judging the coverage relationship between the three candidate keywords, it can be concluded that: there is a coverage relationship between candidate keyword A "century" and candidate keyword B "century brand", and there is also a coverage relationship between candidate keyword A "century" and candidate keyword C "century enterprise". There is no coverage relationship between candidate keyword B "century brand" and candidate keyword C "century enterprise". Therefore, the target keywords are "century brand" and "century enterprise".

[0163] 207. The server inputs the target keyword into the preset vocabulary expansion model to obtain the expanded keywords corresponding to the target keyword.

[0164] The preset vocabulary expansion model can be a vertical domain linguistic model or an existing multimodal large model, used to expand similar keywords for the target keyword.

[0165] Specifically, after determining the target keywords corresponding to the target text, the server can expand the target keywords according to the user's expansion needs. The target keywords can be input into a preset vocabulary expansion model, which can be a model trained specifically for the current target text, such as a vertical domain linguistics model. After detecting the target keywords, the preset vocabulary expansion model generates expanded keywords for that target keyword. These expanded keywords are then identified as expanded vocabulary for the target text and displayed as such in the vocabulary window corresponding to the expansion component on the text configuration page. It's worth noting that the preset expanded vocabulary model includes a preset combined vocabulary model and a preset similar vocabulary model. Expanded keywords include combinations of expanded words and expanded similar words. When the target keyword is input into the preset combined vocabulary model, it obtains at least one word that has a combinational relationship with the target keyword. The target keyword is then combined with this word to obtain corresponding expanded phrases. For example, "century-old" and "brand" can be combined to form "century-old brand," thus expanding the target text with relevant expanded phrases. Simultaneously, the target keyword is input into a preset similar vocabulary model, and the vocabulary similarity between the target keyword and each reference word is calculated using the preset similar vocabulary model. Specifically, the vocabulary similarity can be calculated by calculating the Euclidean distance between the target keyword and the reference words, and the reference words with a vocabulary similarity greater than a preset similarity threshold are identified as extended similar words of the target keyword; the extended phrases and extended similar words are used as extended keywords of the target text.

[0166] To facilitate understanding of the embodiments of this application, specific application scenario examples will be used to describe the embodiments of this application. Specifically, by performing the above steps 201-207, and in conjunction with... Figure 5 This document describes an example of the application scenario. Taking the text information processing process as an example, the specific flow of this text information processing scenario example is as follows:

[0167] On the text configuration page, the user enters the target text "The new EQE does not use the Burmester self-sound system, and its total output power is 710 watts" in the dialogue box. Clicking the "Automatically Generate Keywords" component in the lower right corner of the dialogue box extracts the initial text words "Berlin," "EQE," "power," "new," and "710 watts" from the target text. Simultaneously, after matching historical keywords, the initial text words "Berlin," "EQE," and "power" are displayed as target keywords below the dialogue box. Clicking the "More" component in the lower right corner of the dialogue box displays the initial text words "new" and "710 watts" as other keywords in the lower right window of the dialogue box. After determining the target text keywords, clicking the "Expand Keywords" component in the lower right corner of the dialogue box displays the similar word "Dynaudio" for the target keyword "Berlin" and the word combination "power output" for the target keyword "power."

[0168] Through the above application scenarios, the following effects can be achieved: After obtaining the target text, the corresponding initial text words can be initially extracted. Then, target historical texts similar to the target text are selected, and candidate keywords matching the initial text words are selected from the keyword set associated with the target historical texts. Based on the candidate keywords, the target keywords of the target text are determined, thereby improving the accuracy of target text keyword extraction. This allows users to more easily understand the semantics of the target text through the target keywords, thus enhancing the user experience.

[0169] Therefore, this solution can first extract the initial text vocabulary of the target text, then query the target historical texts similar to the target text, and match the corresponding candidate keywords from the set of historical keywords associated with the target historical texts based on the initial text vocabulary, thereby determining the target keywords of the target text based on the candidate keywords. In this way, the accuracy of text keyword extraction is improved, making it easier for users to obtain the key semantics of the target text and enhancing the user experience.

[0170] To better implement the above methods, this application also provides a text information processing device that can be integrated into a computer device, such as a voice terminal or other computer device.

[0171] For example, such as Figure 5 As shown, the text information processing device may include an extraction unit 301, a query unit 302, a matching unit 303, and a determination unit 304.

[0172] Extraction unit 301 is used to acquire target text and perform word segmentation on the target text to obtain at least one initial text vocabulary;

[0173] Query unit 302 is used to query target historical texts that are similar to the target text from the historical text collection;

[0174] The matching unit 303 is used to obtain a set of historical keywords associated with the target historical text, and to determine candidate keywords that match the initial text vocabulary from the set of historical keywords;

[0175] The determining unit 304 is used to determine the candidate keyword that contains the most words in the initial text as the target keyword corresponding to the target text.

[0176] In some embodiments, the query unit 302 is further configured to:

[0177] Determine the text similarity between the target text and each historical text in the historical text set;

[0178] Based on the similarity of texts, target historical texts that are similar to the target text are selected from the historical text set.

[0179] In some embodiments, the query unit 302 is further configured to:

[0180] The historical texts in the historical text set are sorted in descending order of text similarity to obtain a historical text sequence.

[0181] Select the target historical text that is first sorted from the historical text sequence.

[0182] In some embodiments, the determining unit 304 is further configured to:

[0183] Identify whether there is a coverage relationship between any two candidate keywords;

[0184] If it is determined that there is no overlap between any two candidate keywords, then all candidate keywords will be identified as target keywords;

[0185] If two candidate keywords with a covering relationship are identified, the target candidate keyword with the longest word information is determined from the two candidate keywords with a covering relationship, and the target candidate keyword is determined as the target keyword.

[0186] In some embodiments, the text information processing apparatus further includes a display unit for:

[0187] The target keywords are identified as the core keywords of the target text;

[0188] Display the core keywords of the target text on the text configuration page.

[0189] In some embodiments, the text information processing apparatus further includes a display unit for:

[0190] From multiple initial text words, identify the initial text words that do not match the candidate keywords;

[0191] The initial text words that did not match the candidate keywords are identified as other keywords for the target text. These other keywords are displayed in the vocabulary window corresponding to more components on the text configuration page.

[0192] In some embodiments, the text information processing apparatus further includes an extension unit for:

[0193] Input the target keyword into the preset vocabulary expansion model to obtain the expanded keywords corresponding to the target keyword;

[0194] The extended keywords are identified as the extended vocabulary of the target text, and the extended vocabulary is displayed in the vocabulary window corresponding to the extended component in the text configuration page.

[0195] In some embodiments, the extension unit is further configured to:

[0196] Obtain at least one pre-defined combination word that has a combination relationship with the target keyword from the pre-defined combination word model, and combine the target keyword with the pre-defined combination word to obtain at least one extended word phrase;

[0197] Based on the word similarity between the target keyword and each reference word in the preset similar word model, reference words with a word similarity greater than the preset word similarity threshold are selected as extended similar words of the target keyword.

[0198] Expanded phrases and expanded similar words are used as expanded keywords for the target text.

[0199] In some embodiments, the text information processing apparatus further includes a storage unit for:

[0200] Associate the target keywords with the target text;

[0201] The target text, after being associated with the target keywords, is saved as historical text to the historical text collection.

[0202] As shown above, this solution first extracts the initial text vocabulary of the target text, then queries the target historical texts similar to the target text, and matches the corresponding candidate keywords from the set of historical keywords associated with the target historical texts based on the initial text vocabulary. Thus, the target keywords of the target text are determined based on the candidate keywords. In this way, the accuracy of text keyword extraction is improved, making it easier for users to obtain the key semantics of the target text and enhancing the user experience.

[0203] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0204] This application also provides a computer device, such as... Figure 6 As shown, it illustrates a structural schematic diagram of the computer device involved in the embodiments of this application, specifically:

[0205] The computer device may include components such as a processor 401 with one or more processing cores, a memory 402 with one or more computer-readable storage media, a power supply 403, and an input unit 404. Those skilled in the art will understand that... Figure 6 The computer device structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:

[0206] The processor 401 is the control center of the computer device. It connects various parts of the computer device via various interfaces and lines, and performs various functions and processes data by running or executing software programs and / or modules stored in the memory 402, and by calling data stored in the memory 402, thereby providing overall monitoring of the computer device. Optionally, the processor 401 may include one or more processing cores; preferably, the processor 401 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 401.

[0207] The memory 402 can be used to store software programs and modules. The processor 401 executes various functional applications and text information processing by running the software programs and modules stored in the memory 402. The memory 402 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as voice playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 402 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 402 may also include a memory controller to provide the processor 401 with access to the memory 402.

[0208] The computer device also includes a power supply 403 that supplies power to the various components. Preferably, the power supply 403 can be logically connected to the processor 401 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 403 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0209] The computer device may also include an input unit 404, which can be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.

[0210] Although not shown, the computer device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 401 in the computer device loads the executable files corresponding to the processes of one or more applications into the memory 402 according to the following instructions, and the processor 401 runs the applications stored in the memory 402 to realize various functions, as follows:

[0211] Obtain the target text and perform word segmentation on it to obtain at least one initial text vocabulary; query the target historical text that is similar to the target text from the historical text set; obtain the historical keyword set associated with the target historical text and determine the candidate keywords that match the initial text vocabulary from the historical keyword set; determine the candidate keyword that contains the most initial text vocabulary as the target keyword corresponding to the target text.

[0212] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0213] As shown above, this solution first extracts the initial text vocabulary of the target text, then queries the target historical texts similar to the target text, and matches the corresponding candidate keywords from the set of historical keywords associated with the target historical texts based on the initial text vocabulary. Thus, the target keywords of the target text are determined based on the candidate keywords. In this way, the accuracy of text keyword extraction is improved, making it easier for users to obtain the key semantics of the target text and enhancing the user experience.

[0214] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0215] Therefore, embodiments of this application provide a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute steps in any of the text information processing methods provided in embodiments of this application. For example, the instructions can execute the following steps:

[0216] Obtain the target text and perform word segmentation on it to obtain at least one initial text vocabulary; query the target historical text that is similar to the target text from the historical text set; obtain the historical keyword set associated with the target historical text and determine the candidate keywords that match the initial text vocabulary from the historical keyword set; determine the candidate keyword that contains the most initial text vocabulary as the target keyword corresponding to the target text.

[0217] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0218] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0219] This application also provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the text information processing methods provided in the various optional implementations of the above embodiments.

[0220] Since the instructions stored in the computer-readable storage medium can execute the steps of any of the text information processing methods provided in the embodiments of this application, the beneficial effects that any of the text information processing methods provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.

[0221] The foregoing has provided a detailed description of a text information processing method, apparatus, device, and computer-readable storage medium provided in the embodiments of this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method of processing text information, characterized by, include: Obtain the target text and perform word segmentation on the target text to obtain at least one initial text vocabulary; Query the historical text collection for target historical texts that are similar to the target text; Obtain a set of historical keywords associated with the target historical text, and determine candidate keywords that match the initial text vocabulary from the set of historical keywords. The set of historical keywords includes keywords configured for the target historical text, and the candidate keywords are keywords in the set of historical keywords that contain the initial text vocabulary. The candidate keyword containing the most initial text words is determined as the target keyword corresponding to the target text: identify whether there is a covering relationship between any two candidate keywords; if there is no covering relationship between any two candidate keywords, then all of the candidate keywords are determined as target keywords; if two candidate keywords with a covering relationship are identified, then the target candidate keyword with the largest word information length is determined from the two candidate keywords with a covering relationship, and the target candidate keyword is determined as the target keyword, wherein the covering relationship is: all initial text words in one candidate keyword exist in another candidate keyword.

2. The method according to claim 1, characterized in that, The step of querying target historical texts similar to the target text from the historical text set includes: Determine the text similarity between the target text and each historical text in the historical text set; Based on the text similarity relationship, target historical texts similar to the target text are selected from the historical text set.

3. The method according to claim 2, characterized in that, The step of selecting target historical texts similar to the target text from the historical text set based on the text similarity ranking includes: The historical texts in the historical text set are sorted in descending order of text similarity to obtain a historical text sequence. Select the target historical text that is first sorted from the historical text sequence.

4. The method according to claim 1, characterized in that, After determining the candidate keyword containing the most words from the initial text as the target keyword corresponding to the target text, the method further includes: The target keywords are identified as the core keywords of the target text; The core keywords of the target text are displayed on the text configuration page.

5. The method according to claim 1, characterized in that, The initial text vocabulary is multiple, and after determining the candidate keyword containing the most initial text vocabulary as the target keyword corresponding to the target text, the method further includes: From the plurality of initial text words, identify the initial text words that do not match the candidate keywords; The initial text words that did not match the candidate keywords are identified as other keywords of the target text, wherein the other keywords are displayed in the vocabulary window corresponding to more components in the text configuration page.

6. The method according to claim 1, characterized in that, The method further includes: Input the target keyword into a preset vocabulary expansion model to obtain the expanded keywords corresponding to the target keyword; The extended keywords are identified as the extended vocabulary of the target text, and the extended vocabulary is displayed in the vocabulary window corresponding to the extended component in the text configuration page.

7. The method according to claim 6, characterized in that, The preset vocabulary expansion model includes a preset combined vocabulary model and a preset similar vocabulary model. The expanded keywords include expanded word combinations and expanded similar words. Expanded keywords that match the target keywords are obtained from the preset vocabulary expansion model, and these expanded keywords are displayed as expanded vocabulary of the target text on the text configuration page. Obtain at least one preset combination word that has a combination relationship with the target keyword from the preset combination word model, and combine the target keyword with the preset combination word to obtain at least one extended word group; Based on the word similarity between the target keyword and each reference word in the preset similar word model, the reference words with word similarity greater than the preset word similarity threshold are selected as the extended similar words of the target keyword; The extended phrases and extended similar words are used as extended keywords for the target text.

8. The method according to claim 1, characterized in that, The method further includes: The target keywords are associated with the target text. The target text, after being associated with the target keywords, is saved as historical text to the historical text collection.

9. A text information processing device, characterized in that, include: An extraction unit is used to acquire target text and perform word segmentation on the target text to obtain at least one initial text vocabulary; The query unit is used to query target historical texts that are similar to the target text from the historical text set; A matching unit is configured to obtain a set of historical keywords associated with the target historical text, and to determine candidate keywords that match the initial text words from the set of historical keywords. The set of historical keywords includes keywords configured for the target historical text, and the candidate keywords are keywords in the set of historical keywords that contain the initial text words. The determining unit is configured to identify the candidate keyword containing the most initial text words as the target keyword corresponding to the target text: identify whether there is a covering relationship between any two candidate keywords; if there is no covering relationship between any two candidate keywords, then all of the candidate keywords are identified as target keywords; if two candidate keywords with a covering relationship are identified, then the target candidate keyword with the largest word information length is determined from the two candidate keywords with a covering relationship, and the target candidate keyword is determined as the target keyword, wherein the covering relationship is: all initial text words in one candidate keyword exist in another candidate keyword.

10. A computer device, characterized in that, It includes a processor and a memory, the memory storing a computer program, and the processor being used to run the computer program in the memory to implement the steps of the method of any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium is computer-readable and stores a plurality of instructions adapted for loading by a processor to perform the steps of the method of any one of claims 1 to 8.

Images