Classification result display method and apparatus

Abstract

This application provides a method and apparatus for displaying classification results. The method includes: obtaining text categories for multiple texts, at least one core word corresponding to each text category, and text features extracted during the text classification process; determining first distribution points of the text features of each text in each text category within a set display space based on the text features of each text in the text category; determining a distribution area formed by the first distribution points of each text in the text category; and displaying a text classification feature map based on the distribution area of ​​the text category. The text classification feature map shows the distribution area of ​​the text category, and the first distribution points of the texts under the text category and at least one core word under the text category are marked in the distribution area of ​​the text category. The solution of this application can more comprehensively reflect the classification basis information of text classification.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method and apparatus for displaying classification results. Background Technology

[0002] Text classification refers to determining the category to which a text belongs. In text classification scenarios, it is often necessary to view the text classification results in order to analyze the reasonableness or accuracy of the text classification.

[0003] Currently, most text classification results are presented in the form of statistical charts such as bar charts or pie charts. These charts only show the number of texts under each text category. Therefore, the information on the basis of text classification can be limited based on the statistical charts of text classification results, and it is impossible to have a comprehensive understanding of the basis of text classification. Summary of the Invention

[0004] This application provides a method and apparatus for displaying classification results.

[0005] One method for displaying classification results includes:

[0006] Obtain the classification results of classifying multiple texts and the classification feature information of the multiple texts. The classification results include: the text category to which the text belongs. The classification feature information includes: at least one core word corresponding to the text category and the text features of the text extracted during the classification process. The core word of the text category belongs to a set number of word segments in each text segment under the text category whose importance meets the conditions.

[0007] Based on the text features of each text in the text category, determine the first distribution location point of the text features of each text in the text category within the set display space;

[0008] Determine the distribution area formed by the first distribution position points of each text in the text category;

[0009] Based on the distribution area of ​​the text category, a text classification feature map is displayed. The text classification feature map shows the distribution area of ​​the text category, and the first distribution position point of the text under the text category and at least one of the core words under the text category are marked in the distribution area of ​​the text category.

[0010] In one possible implementation, before displaying the text classification feature map, the following is also included:

[0011] Based on the first distribution position of each text in the text category within the distribution area of ​​the text category, determine the second distribution position of each core word in the text category within the distribution area of ​​the text category;

[0012] The distribution region based on the text category displays a text classification feature map, including:

[0013] By combining the distribution area of ​​the text category, the first distribution position of each text, and the second distribution position of each core word, a text classification feature map is displayed. The distribution area of ​​the text category in the text classification feature map marks the first distribution position of each text under the text category and each core word under the text category.

[0014] In another possible implementation, the distribution area of ​​the text category in the text classification feature map is marked with the first distribution location point of each text under the text category and each core word under the text category;

[0015] Following the display of the text classification feature map, the following is also included:

[0016] If the input point is detected to be located within the distribution area of ​​the displayed text category and within the first set range of the first distribution position point of the target text in the text category, the core word belonging to the target text in the text category is determined, and the core word of the target text in the text classification feature map is switched from the first display effect to the second display effect, wherein the target text is any text under the text category.

[0017] In another possible implementation, the distribution area of ​​the text category in the text classification feature map is marked with the first distribution location point of each text under the text category and each core word under the text category;

[0018] Following the display of the text classification feature map, the following is also included:

[0019] If the input point is detected to be located within the distribution area of ​​the displayed text category and within the second set range of the second distribution position point where the target core word is located in the text category, at least one target text with the target core word under the text category is determined, and the first distribution position point of the target text in the text classification feature map is switched from the third display effect to the fourth display effect, wherein the target core word is any core word under the text category.

[0020] In another possible implementation, before displaying the text classification feature map, the following is also included:

[0021] The interface displays a category selection interface, which shows text display options and core word display options for each of at least one text category corresponding to the multiple texts. The text display options are used to indicate the text under the text category within the distribution area of ​​the text category, and the core word display options indicate the core words under the text category within the distribution area of ​​the text category.

[0022] Determine at least one target display option for the target text category selected by the user in the category selection interface, wherein the target text category belongs to the at least one text category, and the target display item belongs to the text display option or the core word display option;

[0023] The distribution region based on the text category displays a text classification feature map, including:

[0024] Based on the distribution area of ​​the target text category, a text classification feature map is displayed. The text classification feature map shows the distribution area of ​​the target text category, and the first distribution position point of the core word or text to be displayed by each of the at least one target display option is marked in the distribution area of ​​the target text category.

[0025] One of the classification result display devices includes:

[0026] An information acquisition unit is used to acquire classification results of multiple texts and classification feature information of the multiple texts. The classification results include: the text category to which the text belongs. The classification feature information includes: at least one core word corresponding to the text category and the text features of the text extracted during the classification process. The core word of the text category belongs to a set number of word segments in each text segment under the text category whose importance meets the conditions.

[0027] The first position determination unit is used to determine the first distribution position point of the text features of each text in the text category within a set display space based on the text features of each text in the text category.

[0028] A region determination unit is used to determine the distribution region formed by the first distribution location points of each text in the text category;

[0029] The image display unit is used to display a text classification feature map based on the distribution area of ​​the text category. The text classification feature map shows the distribution area of ​​the text category, and the first distribution position point of the text under the text category and at least one of the core words under the text category are marked in the distribution area of ​​the text category.

[0030] As can be seen from the above, in this embodiment, not only is information about multiple text categories obtained, but also multiple core words corresponding to each text category and text features of each text determined during the text classification process are obtained. Based on the text features of each text in each text category, the distribution points of the text features of each text in that text category within the set display space and the distribution area formed by the distribution points of each text can be determined. On this basis, this application can display the distribution area of ​​the text category in the text classification feature map, and the distribution points of each text in that text category and one or two core words under that text category can be marked within the displayed distribution area of ​​the text category. This allows the text classification feature map to reflect the text contained in the text category and the situation of the core words that affect the determination of the text category, thereby allowing the text classification feature map to more comprehensively reflect the classification basis information of the text classification. Attached Figure Description

[0031] 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 embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0032] Figure 1 This paper illustrates a flowchart of a classification result display method provided in an embodiment of this application.

[0033] Figure 2 This diagram illustrates the distribution area of ​​text categories in this application.

[0034] Figure 3 This is a schematic diagram of a text classification feature map of this application;

[0035] Figure 4 This illustration shows another flowchart of the classification result display method provided in the embodiments of this application;

[0036] Figure 5 This diagram illustrates the distribution area of ​​a text category in a text classification feature map when an input point exists within the distribution area of ​​that text category.

[0037] Figure 6 This diagram illustrates another way to display the distribution area of ​​a text category when an input point exists within the distribution area of ​​a text category in the text classification feature map.

[0038] Figure 7 An example diagram of a category selection interface is shown in an embodiment of this application;

[0039] Figure 8 This illustration shows a flowchart of one implementation process for determining core words under a text category in an embodiment of this application;

[0040] Figure 9 This illustration shows another flowchart of the classification result display method in the embodiments of this application;

[0041] Figure 10 This paper shows a schematic diagram of the composition structure of a classification result display device in an embodiment of this application;

[0042] Figure 11 A schematic diagram of the component architecture of an electronic device according to an embodiment of this application is shown. Detailed Implementation

[0043] 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 of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0044] like Figure 1 The illustration shows a flowchart of a classification result display method provided in an embodiment of this application. The method of this embodiment can be applied to electronic devices, such as personal computer devices such as laptops or desktop computers, servers, or nodes in server clusters or distributed systems, without any limitation.

[0045] The method in this embodiment may include:

[0046] S101, obtain the classification results of multiple texts and the classification feature information of multiple texts.

[0047] The classification results include the text category to which each text belongs. It is understood that by classifying multiple texts, they can be assigned to at least one text category. Therefore, the classification results can include multiple text categories and the texts contained within each category.

[0048] Understandably, depending on the classification scenario, text can have multiple possibilities. For example, text can be in the form of articles or video descriptions, without any restrictions. Correspondingly, depending on the classification scenario and the specific text, the text category will also vary.

[0049] For example, when classifying multiple articles, the text categories that can be divided into categories such as literature, science and technology, and comics, etc.

[0050] In this application, the classification feature information of multiple texts includes: at least one core word corresponding to each text category, and the text features of the text extracted during the text classification process.

[0051] Among them, the core words of a text category are a set number of word segments from each text segment within that text category that meet the required importance criteria. For example, the core words of a text category can be word segments derived from texts within that text category that have a high impact on the process of segmenting that text category.

[0052] It is understandable that there are many possible implementations for classifying text. Any classification method that combines text segmentation with word segmentation to classify text is applicable to this embodiment. This application does not limit the specific classification method.

[0053] S102, for each text category, based on the text features of each text in that text category, determine the first distribution location point of the text features of each text in the text category within the set display space.

[0054] It is understandable that the text features of a text can be in the form of multi-dimensional vectors. In order to present the situation of various texts under the same text category, this application uses the text features of a text to represent the text. At the same time, the text features of the text are mapped to a set display space to present the positional distribution of the text features of each text under different text categories in the set display space.

[0055] The distribution point of text features in the designated display space is the coordinate point of the text features mapped to that designated display space. To facilitate differentiation from the distribution points of subsequent core words, the distribution point of the text features in the designated display space is referred to as the first distribution point.

[0056] The display space can be set as needed, and can be a three-dimensional or multi-dimensional space. The display space can also be a two-dimensional space.

[0057] Understandably, when the dimensions of the display space and the dimensions of the text features are different, the text features need to be converted into features with the same number of dimensions as the dimensions of the display space before being mapped to the feature display space. Generally, the dimensions of the display space are smaller than the dimensions of the text features; therefore, the text features can be dimensionality-reduced so that the dimensions of the reduced text features match the dimensions of the display space.

[0058] For example, if the display space is set to a two-dimensional plane, then the text features of each text need to be reduced to two-dimensional features. Based on this, the coordinates of the reduced text features in the two-dimensional plane can be determined, and these coordinates are the first distribution location points corresponding to the text features of that text.

[0059] S103, for each text category, determine the distribution area formed by the first distribution location points of each text in that text category.

[0060] The distribution area of ​​a text category refers to the area within the defined display space determined by the first distribution location point corresponding to each text in the text category.

[0061] For example, the distribution area of ​​a text category can be the smallest area within the defined display space that can contain the first distribution position points of each text in that text category; or, it can be the smallest smooth area that can contain the first distribution position points of each text in that text category. Of course, the distribution area can also be the smallest area with a defined shape that contains the first distribution position points of each text in that text category, etc., without any limitation.

[0062] like Figure 2 This diagram illustrates the distribution area of ​​text categories.

[0063] exist Figure 2 In this example, the display space is set as a two-dimensional plane, and three text categories are used as examples.

[0064] Depend on Figure 2 It can be seen that the first distribution location points 201 of each text within the same text category exhibit clustering, making the first distribution location points of each text within each text category constitute a region. Figure 2 The distribution area 202 for each text category is a closed area enclosed by a line, and the distribution areas of different text categories do not overlap with each other. Figure 2 Different text categories are represented by the different areas enclosed by the lines.

[0065] S104 displays the text classification feature map based on the distribution area of ​​text categories.

[0066] The text classification feature map shows the distribution area of ​​text categories, and within the distribution area of ​​text categories, it marks the first distribution location of the text under the text category and at least one of the core words under the text category.

[0067] Among them, the text classification feature map can be used to present the distribution of text and core words under each text category.

[0068] Understandably, in practical applications, the specific content to be displayed in the distribution area under the text category in the text classification feature map can be preset as needed or selected by the user.

[0069] For example, in one scenario, within the distribution area of ​​text categories displayed on the text classification feature map, only the first distribution location points of each text within a category can be shown. This scenario can be seen in [reference needed]. Figure 2 As shown.

[0070] In another scenario, within the distribution area of ​​text categories shown in the text classification feature map, only the core words of each text category can be identified.

[0071] In another case, the first distribution location of each text in the text category and each core word can be marked in the distribution area of ​​the text category displayed in the text classification feature map.

[0072] There are several ways to identify the core words of a text category within the distribution area of ​​the text category displayed in the text classification feature map. For example, each core word under a text category can be displayed in the distribution area; if the core word is "art," then "art" can be directly displayed in the respective area. Alternatively, for each text category, a unique word identifier can be assigned to each core word within that category.

[0073] like Figure 3 To illustrate this, we take the example of a text classification feature map where the distribution region of each text category simultaneously marks the first distribution location point of each text under that text category as well as each core word.

[0074] exist Figure 3 The text displays the distribution areas 301 for each of the three text categories. In addition to the first distribution position point 302 of each text under that text category, the distribution area of ​​each text category also displays the core word 303 at the second distribution position point corresponding to the core word.

[0075] like Figure 3 The leftmost text category distribution area not only displays the points corresponding to the text features of each text, but also shows eight core words. These five core words are K1-1, K1-2, K1-3, K1-4, K1-5, K1-6, K1-7, and K1-8. For the distribution areas of other text categories, characters starting with K2 or K3 also represent core words under that text category.

[0076] Understandably, the position of the core words marked in the distribution area of ​​text categories in the text classification feature map can be set according to actual needs. For example, when only core words are displayed in the distribution area of ​​text categories, each core word under that text category can be displayed arbitrarily. Alternatively, core words under that text category can be randomly displayed in other areas outside the first distribution position point of each text in the distribution area of ​​text categories.

[0077] In one possible implementation, in order to improve the display efficiency of the text classification feature map and reduce the risk of mutual occlusion between the core words and the first distribution position points of the text under the same text category in the text classification feature map, for each text category, the second distribution position points of each core word in the text category can be determined in the distribution area of ​​the text category based on the first distribution position points of each text in the text category's distribution area.

[0078] Understandably, in order to avoid the text distribution points and core words occluding each other, for a text category, the second distribution point of each core word in that text category is different, and the second distribution point is different from the second distribution point.

[0079] Correspondingly, the text classification feature map can be displayed by combining the distribution area of ​​the text category, the first distribution position of each text, and the second distribution position of each core word, so that the distribution area of ​​the text category in the text classification feature map is marked with the first distribution position of each text under the text category and each core word under the text category.

[0080] Of course, if the distribution area of ​​the text category in the text classification feature map only needs the core words of that text category, then it is also possible to combine the distribution area of ​​the text category and the second distribution position points of each core word under that text category to display the text classification feature map, so that the corresponding core words are marked at the second distribution position corresponding to the core words in the distribution area of ​​the text category.

[0081] It is understandable that when multiple texts are classified into multiple text categories, the text classification feature map in this application can display the distribution area of ​​all text categories, or it can be preset or selected by the user to display only the distribution area of ​​some text categories and the first distribution position of related core words or texts in the text classification feature map.

[0082] In one alternative approach, to distinguish the distribution areas of different text categories, the display effects of the distribution areas for different text categories can be set differently in the text classification feature map. The different display effects of the distribution areas can include at least one of the following:

[0083] The display effect of the edge lines of the distribution areas for different text categories is different;

[0084] The display effect of core words varies in the distribution areas of different text categories;

[0085] The display effect of the first distribution position point of the text differs in the distribution areas of different text categories.

[0086] The display effect of the edge lines of the distribution area can include the thickness and color of the edge lines, etc. For example... Figure 2 and Figure 3 The thickness of the edge lines corresponding to the distribution areas of different text categories varies.

[0087] The display effect of core words includes differences in their color, size, font, and positional distribution.

[0088] The display effect of the first distribution point of the text includes one or more of the following: the shape, structure, color, and brightness of the first distribution point. For example... Figure 2 and Figure 3 In the text, the representation of the first distribution position of the text in the distribution area of ​​different text categories is also different. Some text categories use black dots to represent the first distribution position, some use white dots, and some use circular dots to represent the first distribution position.

[0089] As can be seen from the above, in this embodiment, not only is information about multiple text categories obtained, but also multiple core words corresponding to each text category and text features of each text determined during the text classification process are obtained. Based on the text features of each text in each text category, the distribution points of the text features of each text in that text category within the set display space and the distribution area formed by the distribution points of each text can be determined. On this basis, this application can display the distribution area of ​​the text category in the text classification feature map, and the distribution points of each text in that text category and one or two core words under that text category can be marked within the displayed distribution area of ​​the text category. This allows the text classification feature map to reflect the text contained in the text category and the situation of the core words that affect the text category, thereby providing a more comprehensive reflection of the classification basis information of the classification category.

[0090] Understandably, in order to enable users to obtain more information on the classification criteria of text categories based on the text classification feature map, or to flexibly view the text categories, the core words under the text categories, and the mutual influence information between the texts under the text categories, the first distribution position point of each text under the text category and each core word are marked in the distribution area of ​​the text category in the text classification feature map. Furthermore, the content information or display effect of the displayed text category distribution area can be adjusted based on the user's operation behavior in the distribution area of ​​the text category in the text classification feature map.

[0091] The following section, using a flowchart, introduces one implementation of adjusting text classification feature maps based on user input.

[0092] like Figure 4 The diagram illustrates a flowchart of a classification result display method provided in this embodiment. The method of this embodiment may include:

[0093] S401, obtain the classification results of multiple texts and the classification feature information of multiple texts.

[0094] The classification results include the text category to which the text belongs.

[0095] The classification feature information includes: at least one core word corresponding to the text category, and the text features of each text under the text category. The text features are the features of the text extracted during the classification process.

[0096] The core word of a text category belongs to a set number of word segments that meet the importance criteria among the word segments of each text segment under that text category.

[0097] S402, Based on the text features of each text in the text category, determine the first distribution location point of the text features of each text in the text category within the set display space.

[0098] S403, determine the distribution area formed by the first distribution location points of each text in the text category.

[0099] For details of steps S401 to S403 above, please refer to the relevant descriptions in the previous embodiments, which will not be repeated here.

[0100] S404, based on the distribution area of ​​text categories, displays the text classification feature map.

[0101] This text classification feature map shows the distribution areas of text categories. Specifically, the text category features can show the distribution area of ​​each text category from multiple texts, or it can show only the distribution areas of a subset of text categories from multiple texts.

[0102] However, the distribution area of ​​each text category in the text classification feature map simultaneously marks the first distribution location point of the text under that text category and the core words of that text category.

[0103] Similar to the previous embodiments, before executing step S404, for each text category, this application may first determine the second distribution position points of each core word in the text category by combining the first distribution position points of each text in the text category. Based on this, when displaying the text classification feature map, this application can mark the core words in the distribution area of ​​the text category according to the second distribution position points of each core word in the text category.

[0104] In this embodiment, in addition to determining the second distribution location of each core word in the text category within the distribution area of ​​the text category, the orientation and relative angle of each core word in the text category within the distribution area of ​​the text category can also be determined as needed, without limitation.

[0105] S405, if the input point is detected to be located within the distribution area of ​​the displayed text category and within the first set range of the first distribution position point of the target text in the text category, determine the core word belonging to the target text in the text category.

[0106] The input point can be the location of the operation point in the display unit. For example, the input point can be the coordinate position of the cursor under mouse control, or the touch position point of the user's touch operation on the display unit.

[0107] Understandably, after the text classification feature map shows the distribution area of ​​text categories, the user can control the input point to enter the distribution area of ​​any text category in the text classification feature map.

[0108] Correspondingly, if the input point enters the distribution area of ​​a certain text category, and the input point is located at or close to the first distribution position of a certain text in that text category, it indicates that the user is currently interested in that text and hopes to learn more about the details of the text classification or to see more intuitively the information that affects the text classification.

[0109] Based on this, in order to determine which text the user is focusing on, this application sets a first set range. This first set range can be set as needed and is not limited thereto. For example, the first set range can be the distance from the first distribution point of the text as the target distance value.

[0110] For ease of distinction, in this application, the text that the user expects to focus on, determined based on the input point, is referred to as the target text. The target text can be any text from the text category to which the distribution area where the input point is located belongs.

[0111] For example, if the input point is located within the distribution area of ​​text category 1 in the text classification feature map, and the input point is within the first set range of the first distribution position point of text 1-1 in text category 1, then text 1-1 is the target text.

[0112] It is understandable that the core words in a text category are derived from the word segments of each text in the text category. Therefore, this application can determine which core words in a text category belong to the target text and obtain the core words corresponding to the target text.

[0113] S406, switch the core words of the target text in the text classification feature map from the first display effect to the second display effect.

[0114] The first display effect of the core words refers to the initial display effect of the core words when displaying the text classification feature map.

[0115] The second display effect of core words differs from their first display effect. By switching the core words of the target text in a text category from the first display effect to the second display effect, the core words of the text category can be distinguished from other core words within the distribution area of ​​the text category. This allows users to intuitively understand the core words related to the target text and those that have a significant impact on identifying the text category.

[0116] Furthermore, compared to the first display effect, the second display effect in this application is designed to better highlight the core word. Accordingly, switching the core word of the target text from the first display effect to the second display effect can more effectively highlight the core word of the target text.

[0117] Both the first and second display effects can be customized as needed. The differences between the first and second display effects can be reflected in the distribution, color, size, and brightness of the core words, among other display features.

[0118] For example, relative to the first display effect of the core words of the target text, the second display effect of the core words of the target text may have some or all of the following characteristics:

[0119] The core words of the target text are closer to their relative first distribution positions in the target text;

[0120] The core words of the target text are distributed around the first distribution position points of the target text;

[0121] The core words in the target text are highlighted more highly;

[0122] The font size of the core words in the target text is larger;

[0123] The core words in the target text are in a darker color.

[0124] In particular, this application can also change the display effect of the first distribution position point of the target text, such as making the first distribution position point of the target text larger, brighter, blinking, or darker in color, etc.

[0125] For example:

[0126] by Figure 3 For example, in Figure 3 The display effect of the first distribution position point of each text in the presented text classification feature map is consistent.

[0127] Furthermore, the display effect of the core keywords is also consistent, such as... Figure 3 Each core word is evenly distributed within the blank area of ​​the text category's distribution area, and the value of each core word is the first numerical value. The display effect of the core words at this point can be called the first display effect.

[0128] When a user moves their mouse (or touch point, etc.) within the distribution area of ​​the leftmost text category, and the mouse is positioned on or around the first distribution point of a particular text, then the first distribution point of that text will increase in size. For example... Figure 5 As shown, it illustrates a schematic diagram of the distribution area corresponding to a text category after an input point exists within the distribution area of ​​a text category in the text classification feature map.

[0129] Figure 5 The input point is located at Figure 3 Let's take the distribution area of ​​the leftmost text category as an example.

[0130] Will Figure 3 The distribution area of ​​the leftmost text category is... Figure 5 A comparison shows that after the input point is located at the first position distribution point 501 of a text within the text classification distribution area, the first position distribution point 501 of that text becomes significantly larger. Of course, in practical applications, other effects such as highlighting and blinking can be added to the first distribution point of the text; there are no restrictions on this. Meanwhile, in Figure 5 The core words belonging to this text category will be enlarged and highlighted with boxes, so that users can intuitively understand the core words related to this text.

[0131] Moreover, through comparison Figure 3 and Figure 5It can be seen that after the input point is located at the first distribution location point, the distribution positions of the core words 502 associated with the text represented by the first distribution location point 501 also change, such as... Figure 5 The core words K1-1, K1-3, and K1-7 are distributed around the first distribution point 501. Figure 5 The core words of the text will be displayed around the first distribution location point 501 to highlight the correlation between these core words and the first distribution location point 501.

[0132] Furthermore, this application can also obtain the first classification contribution weight of each core word corresponding to the target text, which reflects the degree of influence or importance of the core words in determining the text category of the target text.

[0133] In this case, after the input point is within the first set range of the first distribution position point of the target text, this application can also control the second display effect of each core word of the target text according to the first classification contribution weight of each core word in the target text, so that when the first classification contribution weight of the core words of the target text is different, the second display effect of the core words of the target text is also different.

[0134] For example, for each core word corresponding to the target text, the higher the first category contribution weight of the core word, the closer the core word is to the first distribution position point of the target text, the higher the brightness of the core word or the more obvious the color brightness, etc.

[0135] In this application, when the input point enters the first set range of the first distribution position of the target text, if the user wishes to view further details of the target text, they can click on or otherwise select the first distribution position of the target text. Correspondingly, upon detecting a selection operation on the first distribution position of the target text, this application can also display the content of the target text. The content of the target text can be its detailed content, such as the information contained within the target text, or it can be the author or source of the target text.

[0136] Furthermore, in response to the selection operation, this application can also determine the updated display effect of each core word of the target text based on the first category contribution weight of each core word of the target text, and display each core word of the target text according to the updated display effect corresponding to each core word of the target text.

[0137] The first category contribution weight of the core words in the target text varies, resulting in different display effects after the core words are updated. The updated display effect of the core words refers to a new display effect that differs from the first and second display effects. In this application, to distinguish it from the first, second, and subsequent fourth display effects of the first distribution position points of the target text, the updated display effect of the core words of the target text can also be referred to as the fifth display effect.

[0138] For example, relative to the second display effect of the core words of the target text, the fifth display effect of the core words of the target text could be that the color of the core words is different, the brightness is increased, or the flashing frequency is increased, etc., without any restrictions.

[0139] Understandably, before classifying text, at least one word segment of the text can be identified. For a given text, even if some words do not belong to the core words of the text category to which the text belongs, the influence weight of the text category to which the text belongs may be relatively high, i.e., the first classification contribution weight of the word is relatively high. Based on this, this application can also identify multiple important words from at least one word of the text whose first classification contribution weights satisfy the weight conditions before displaying the classified text feature map. For example, a specified number of words with high first classification contribution weights corresponding to the text can be selected as important words of the text, and this specified number can be set as needed.

[0140] Accordingly, when the input point is detected to be within a first set range of the first distribution location point of the target text, this application can also mark the important words of the target text in the distribution area of ​​the text category to which the target text belongs in the text classification feature map. The important words of the target text can all be displayed using the second display effect.

[0141] Furthermore, in order to distinguish the core words belonging to the text category of the target text from the important word segments that do not belong to the core words, this application can use a specified display effect different from that of the core words to display the important word segments of the target text, so that users can intuitively distinguish the core words of the target text from the important word segments other than the core words. The specified display effect is different from the first display effect and the second display effect corresponding to the core words, etc.

[0142] like Figure 5 As shown, the first distribution point of the text is surrounded by the core word 502, which belongs to the text and is within the text category. Figure 5 In addition to K1-1, L1-3, and K1-7, 503 important word segments belonging to this text but not belonging to its text category are also displayed. Figure 5It can be seen that the border of the important word segment 503 is a dotted line, and the font size and other display effects are also different from the display effects of other core words 502. Figure 5 Taking the important word segmentation K1-16 as an example.

[0143] It is understood that, in this embodiment, after the input point moves out of the first set range of the first distribution position point of the target text, the core word corresponding to the target text will return to the first display effect. The first distribution position point of the target text will also return to its initial display effect. Of course, if the distribution area of ​​the text category of the target text displays important words in the target text that do not belong to the core word, then the display of such important words can be canceled.

[0144] S407, if the input point is detected to be located within the distribution area of ​​the displayed text category and within the second set range of the second distribution position point where the target core word is located in the text category, then at least one target text with the target core word under the text category is determined.

[0145] The first and second setting ranges can be set as needed. Here, they are referred to as the first and second setting ranges respectively to facilitate the distinction between the setting ranges corresponding to the distribution locations of the core words and the text.

[0146] It is understandable that the input point can be located in any text category in the text classification feature map. If the input point is located within the second set range of the second distribution position of a core word in a certain text category, then that core word in this text category is determined as the target core word. Therefore, the target core word can be any core word under the text category to which the distribution area where the input point is located belongs.

[0147] Understandably, the core words of a text category originate from the word segments of each text within that category. A core word in a text category may be a word segment from one or more texts. Therefore, to identify a target core word, it is necessary to determine which texts contain that target core word in their word segments. For ease of distinction, texts containing the target core word within the text category to which the distribution region of the input point belongs are referred to as target texts.

[0148] S408, switch the first distribution location point of the target text in the text classification feature map from the third display effect to the fourth display effect.

[0149] The third display effect of the first distribution position point of the target text refers to the display effect of the first distribution position point of the target text when the input point is located before the second set range where the second distribution position point of the target core word is located. For example... Figure 3 As shown, Figure 3The display effect of the first distribution position point of each target text in each text category is the third display effect of the first distribution position point of the target text.

[0150] The fourth display effect differs from the third display effect, similar to the difference between the first and second display effects. The third and fourth display effects can differ in at least one dimension, such as the color, size, flickering, and brightness of the text's distribution points.

[0151] Compared to the third display effect of the first distribution position point of the target text, the fourth display effect of the first distribution position point of the target text can highlight the first distribution position point of the target text more and can distinguish the first distribution position point of the target text from other texts.

[0152] For example, relative to the third display effect of the first distribution position point of the target text, the fourth display effect of the first distribution position point of the target text may have some or all of the following characteristics:

[0153] The brightness of the first distribution point in the target text is higher;

[0154] The size of the first distribution location point in the target text is larger;

[0155] The first distribution point of the target text is darker in color;

[0156] The first distribution point of the target text exhibits a flickering characteristic.

[0157] Furthermore, to allow users to intuitively perceive the target keyword at the input point, this application can switch the target keyword from the first display effect to the sixth display effect. This sixth display effect differs from the first display effect. The sixth display effect will highlight the target keyword more effectively; for example, compared to the first display effect, the font in the sixth display effect will be larger, brighter, darker, or will include dynamic flashing.

[0158] like Figure 6 As shown, it illustrates the display effect when the input point is located on a core word in the distribution area of ​​a text category.

[0159] Figure 6 Also based on the input point being located Figure 3 Let's take the distribution area of ​​the leftmost text category as an example.

[0160] contrast Figure 3 The distribution area of ​​the leftmost text category is related to this. Figure 6 As can be seen, if the input point is located within the second set range of the target core word 601, then the target core word will be marked with a box to highlight the target core word corresponding to the input point.

[0161] Furthermore, the first distribution position point 602 of each target text to which the target core word can be assigned will also be... Figure 3 Switch the display effect (i.e., the third display effect) to Figure 6 The display effect (sixth display effect). For example... Figure 6 It can be seen that the first distribution location point 602 of the target text is relative to Figure 3 The size of the first distribution point is larger. Of course, in practical applications, it can be set... Figure 6 The first distribution point of the target text may have a flashing effect or be lit up, without any restrictions.

[0162] Furthermore, upon detecting a selection operation targeting the core keyword, while displaying the core keyword using the sixth display effect, a seventh display effect can also be used to display the first distribution position points of each text containing the core keyword within the text category to which the core keyword belongs. This seventh display effect can differ from the third display effect.

[0163] For example, in Figure 3 The first distribution position points of each text containing the target core word (i.e., the target text) are displayed statically. However, after a selection operation targeting the target core word is detected, the first distribution position points of the text containing the target core word can be highlighted and set to a blinking effect, etc.

[0164] Similar to the previous one, the selection operation for the target keyword can be clicking on the target keyword or other selection operations, without any restrictions.

[0165] Understandably, if the input point moves outside the second set range of the target core word, or if a click operation on the target core word is detected after the user selects the target core word, the display effect of the target core word can be restored to the first display effect, and the display effect of the first distribution position points of each text associated with the target core word can be restored to the third display effect.

[0166] As can be seen from the above solution in this embodiment, after displaying the text classification feature map, this application can not only intuitively see the distribution of each text and each core word in the text category, but also view the core words that affect the text classification, the detailed content of the text, and the information that each core word can affect the text in various dimensions as needed. Thus, it not only provides users with more comprehensive information on the basis of text classification, but also flexibly displays the relevant information required for text classification to users.

[0167] It is understood that in this application, the distribution area of ​​one or more classified text categories can be displayed through text classification feature maps, and the first distribution position point of the text of each text category can be selected as needed to display one or two of the core words in the distribution area of ​​each text category.

[0168] To allow users to more flexibly select or set the text categories to be displayed, the text related to the text categories, and the core words of the text categories, this application can also display a category selection interface before displaying the text category feature map.

[0169] For example, if a user's request to display a selection interface is detected, the text category interface will be displayed.

[0170] Of course, it could also be that after detecting the launch of a category display application or after completing the text categorization, the category selection interface is actively displayed, and there are no restrictions on this.

[0171] The category selection interface displays text display options and core keyword display options for each of the at least one text category corresponding to multiple texts.

[0172] This text display option indicates that the text under that text category is displayed within the distribution area of ​​that text category. Accordingly, if a user selects a text category's text display option in the category selection interface, it means that the text under that text category needs to be displayed within the distribution area of ​​that text category.

[0173] The core keyword display option indicates that the core keywords for that text category will be displayed within the distribution area of ​​that text category. Conversely, if a user selects the core keyword display option for a specific text category, it means that the core keywords for that text category should be displayed within the distribution area for that text category.

[0174] Based on this, this application can determine at least one target display option for the target text category selected by the user in the category selection interface. This target text category belongs to at least one text category displayed in the category selection interface. One of the target display items is either a text display option or a core keyword display option.

[0175] Accordingly, a text classification feature map can be displayed based on the determined distribution area of ​​the target text category. This text classification feature map shows the distribution area of ​​the target text category, and within the distribution area of ​​the target text category, the first distribution location point of the core word or text to be displayed for each of the at least one target display option is marked.

[0176] The target text category can be one or more. Depending on the number of target text categories selected by the user, the number of distribution areas of the text categories shown in the text classification feature map will also vary.

[0177] To facilitate understanding, the following explanation uses an example image of a category selection interface.

[0178] like Figure 7 As shown in the classification selection interface, the classification selection interface includes two option areas: one option area 701 is for text distribution points, and the other option area 702 is for core words.

[0179] The text distribution point options area includes text display options 703 for various text categories. The core word options area includes core word display options 704 for various text categories.

[0180] exist Figure 7 Taking the example of multiple texts ultimately being categorized into three text categories, the option area for each text distribution point displays text display options for text category 1, text category 2, and text category 3, respectively. Similarly, the option area for each core word displays core word options for text category 1, text category 2, and text category 3, respectively.

[0181] Based on this, let's assume that the user selected the text display options for text category 1 and text category 2 in the text distribution point option area, and also selected the core word display option for text category 1 in the core word option area.

[0182] Both text category 1 and text category 2 belong to the target text category. However, text category 1 has two target display options: one is a text display option, and the other is a keyword display option. Text category 2, on the other hand, only has one target display option: a text display option.

[0183] Correspondingly, the displayed text classification feature map will show the distribution areas of text category 1 and text category 2. The distribution area of ​​text category 1 not only indicates the first distribution position of each text under text category 1, but also indicates each core word under text category 1; while the distribution area of ​​text category 2 only indicates the first distribution position of each text under text category 2.

[0184] It is understandable that, in this application, the core word display options for each text category in the core word option area can also be associated with a core word quantity configuration area. For example... Figure 7 This example uses the top 8 keywords for each text category. In practical applications, users can also select or configure the number of keywords to be displayed for each text category as needed.

[0185] Understandably, in order to more flexibly display the text classification feature map, this application can also construct a text distribution layer and a core word distribution layer before displaying the classification selection interface.

[0186] Correspondingly, the category selection interface can also be used to set the layer overlay order between the text distribution layer and the core word distribution layer.

[0187] The category selection interface allows you to set the layer overlay order of the text distribution layer and the core word distribution layer. This can be done as needed and is not restricted.

[0188] like Figure 7 As shown, the category selection interface also includes a stacking order switching option 705, currently available in... Figure 7 If the text distribution point option area 701 is located above the core word option area 702, it means that the text distribution layer should be located above the core word distribution layer. However, if the user clicks the overlay order toggle option 705, the vertical order of the text distribution point option area 701 and the core word option area 702 will be swapped, indicating that the core word distribution layer will be set above the text distribution layer.

[0189] certainly, Figure 7 This is just one example; in real-world applications, there are other ways to set the layer stacking order.

[0190] Understandably, for a given target text category, if a user only selects to display the text and core keywords of that target text category, for example... Figure 7 If a user selects only one text display option or core word display option under a text category, then the overlay of the text distribution layer and the core word distribution layer under that text category will not be involved.

[0191] However, if a user wants to display both the text and core words under a certain target text category at the same time, then when displaying the text classification feature map, it is necessary to combine the layer overlay order to display the text distribution layer and the core word distribution image.

[0192] Specifically, for each target text category selected by the user, if the text and core keywords of that target text category need to be displayed based on at least one target display option selected by the user, then the following operations need to be performed in sequence:

[0193] First, the distribution area of ​​the target text category is constructed on the text distribution layer and the core word distribution layer, respectively.

[0194] Then, within the distribution area of ​​the target text category in the text distribution layer, the first distribution position point of each text in the target text category is marked, and within the distribution area of ​​the target text category in the core word distribution layer, each core word in the target text category is marked.

[0195] Finally, based on the layer overlay order set by the user in the category selection interface, the text distribution layer and the core word distribution layer are overlaid and displayed.

[0196] Understandably, when the text category feature map displays both the core words and the first distribution point of the text in the distribution area of ​​the text category, the distance between the core words and the first distribution point of the text in the text category may be relatively small due to the limited display area of ​​the text category distribution area. In this case, based on the position of the input point in the distribution area of ​​the text category, it may not be able to accurately determine whether the user wants to view the core words or the relevant information of the text in that text category.

[0197] Based on this, this application can also combine the overlay order of the text distribution layer and the core word distribution layer to comprehensively determine the preceding... Figure 4 The relative positional relationship between the input point, the core word, and the first distribution point of the text in the embodiment.

[0198] Specifically, when the text distribution layer is above the core word distribution layer, if the input point is within the distribution area of ​​a certain text category in the text classification feature map, then it is confirmed that the object the user wants to operate on is the text in that text category. Naturally, it is necessary to pay attention to whether the input point is within the first set range of the first distribution position point of a certain text in that text category, rather than monitoring whether the input point is within the second set range of the second distribution position point of the core word in that text category.

[0199] Correspondingly, if the input point is within the first set range of the first distribution position point of the target text, the core words belonging to the target text in that text category can be switched from the first display effect to the second display effect. Of course, the same applies to other operations mentioned in step S406 above.

[0200] Similarly, when the core word distribution layer is located above the text distribution layer, if the input point is within the distribution area of ​​a certain text category in the text classification feature map, then it can be confirmed that the object the user wants to operate on is the core word in that text category. Naturally, it is only necessary to monitor whether the input point is within the second set range of the second distribution position point of the core word in that text category. Accordingly, if the input point is within the second set range of the second distribution position point of the target core word in the text category, then the relevant operations in steps S407 and S408 can be executed. Of course, for the case where the input point moves out of the second set range of the second distribution position point of the target core word... Figure 4 The related operations in the embodiments also apply.

[0201] It is evident that by setting the overlay order of the text distribution layer and the core word distribution layer, it is more effective to accurately determine the objects within the distribution area of ​​the text category that the user wants to manipulate, thereby presenting the user with category-related information in a more reasonable manner.

[0202] It is understood that in the embodiments of this application, the core words under each text category are determined based on the importance of word segmentation of each text under that text category, and the conditions that the importance of the core words needs to meet can be set as needed.

[0203] To facilitate understanding, the following example illustrates one method for determining the core words of a text category. For instance... Figure 8 This illustration shows a schematic diagram of an implementation process for determining core words under a text category in an embodiment of this application. The implementation process of this embodiment may include:

[0204] S801, obtain the first classification contribution weight of at least one word of the text.

[0205] Among them, the first segmentation contribution weight in the text is the influence weight of the extracted segmentation on determining the text category to which the text belongs during the text classification process.

[0206] Understandably, before classifying multiple texts in this application, each text needs to be segmented into words to obtain at least one word for each text. Based on this, this application will combine at least one word from each of the multiple texts to classify the multiple texts. In the process of classifying multiple texts, the influence weight of each word in each text on determining the text category to which the text belongs can be determined, i.e., the first classification contribution weight.

[0207] For example, in one possible implementation, this application can utilize a classification model to determine the respective text categories of multiple texts. This classification model includes an attention layer and a text feature extraction layer. Accordingly, the first classification contribution weight of at least one word segment of the text output by the attention layer is obtained.

[0208] S802, for any word segment contained in each text under each text category, determine the second classification contribution weight of the word segment to the text category based on the first classification contribution weight of the word segment to different texts in the text category.

[0209] The second-category contribution weight reflects the influence weight of a word in determining the text category to which it belongs. The second-category contribution weight of a word under a text category is determined by the combined first-category contribution weights of that word in all texts under that text category.

[0210] For example, in one possible implementation, for any word segment of a text under a text category, the sum of the first category contribution weights of the word segment in each text under that text category can be determined as the second category contribution weight of the word segment.

[0211] S803, for any text category derived from multiple text classifications, determine a set number of core words from at least one word segment of each text under that text category, such that the contribution weight of the second classification satisfies the condition.

[0212] Among them, the core words belong to the word segments of each text under the text category.

[0213] The conditions for the second category contribution weight can be set as needed. For example, the second category contribution weight exceeds a set threshold. In one optional approach, a set number of words with higher second category contribution weights can be identified as the core words of that text category.

[0214] This embodiment allows for the reasonable selection of core words from the word segmentation of each text in a text category, based on the contribution to determining that text category. This enables users to more accurately reflect the classification criteria of the text category based on the core words.

[0215] In this application, there are many possible implementations for classifying multiple texts and determining the classification feature information of the texts, and this application does not impose any restrictions on them.

[0216] To facilitate understanding, the following uses a specific implementation of text classification as an example to explain the process of obtaining classification results and classification feature information of multiple texts in this application, and introduces the method for displaying the classification results in this application.

[0217] like Figure 9The diagram illustrates another flowchart of the classification result display method in this embodiment. The method of this embodiment may include:

[0218] S901: Obtain multiple texts to be classified.

[0219] The text to be classified can be obtained through data collection or platform reporting, without any restrictions.

[0220] S902, use a classification model to determine the text category to which each of the multiple texts belongs.

[0221] Understandably, after obtaining multiple texts, this application may preprocess the texts to remove some texts that do not meet the requirements.

[0222] Furthermore, before determining the text category based on the classification model, this application also needs to segment each text separately to obtain at least one segmented word for each text. For each text, each segmented word can be converted into a word vector, and then the word vector of at least one segmented word of each text can be input into the classification model to obtain the text category to which each text belongs.

[0223] Understandably, this classification model is trained in advance using multiple text samples labeled with text categories, and there are no restrictions on the specific training process.

[0224] In this application, the classification model can have various possibilities, but in this application, the classification model introduces an attention layer and a text feature extraction layer. For example, the classification model can be any neural network that includes an attention mechanism, such as a Hierarchical Attention Network (HAN) model.

[0225] S903, obtain the first classification contribution weight of at least one word of the text output by the attention layer in the classification model, and the text features of each text proposed by the text feature extraction layer of the classification model.

[0226] For example, after inputting the word vectors of each word segment corresponding to each text into the classification model, the word vectors of each text are first input into the attention layer. This attention layer extracts the classification contribution weight of each word segment in the text, i.e., the first classification contribution weight. This application obtains the classification contribution weight of each text output by this attention layer.

[0227] Furthermore, the first classification contribution weight of each word segment in each text output by the attention layer is input to the text feature extraction layer. The text feature extraction layer determines the text features of each text based on the first classification contribution weight of each group of words corresponding to each text, and inputs the text features of each text into the normalization function layer of the classification features to obtain the text category to which each text belongs. However, in order to analyze the classification criteria of multiple text categories, this application requires obtaining the text features of each text output by the text feature extraction layer.

[0228] S904, for any word segment contained in each text under each text category, determine the second classification contribution weight of the word segment to the text category based on the first classification contribution weight of the word segment to different texts in the text category.

[0229] S905, for any text category derived from multiple text classifications, determine a set number of core words from at least one word segment of each text under that text category, based on the condition that the contribution weight of the second classification satisfies the given conditions.

[0230] Understandably, in order to display the important word segments associated with the first distribution position points of the text based on the input points after displaying the text classification feature map, this application can also determine a specified number of word segments in the first text whose first classification contribution weight meets the condition as important word segments of that text. For example, the first specified number of word segments with higher first classification contribution weights in the text can be determined as important word segments of that text.

[0231] The steps S904 to S905 above can be found in the relevant descriptions of the previous embodiments, and will not be repeated here.

[0232] It is understandable that the above is an example of one way to obtain classification results and classification feature information of multiple texts for ease of understanding, and there are no restrictions on other methods.

[0233] S906, Based on the text features of each text in the text category, determine the first distribution location point of the text features of each text in the text category in the two-dimensional coordinate space.

[0234] Since textual features are multidimensional, this application can reduce the dimensionality of textual features to convert them into two-dimensional features. For example, textual features can be converted into features represented by horizontal coordinates x and vertical coordinates y. Accordingly, based on the two-dimensional features of the text, the first distribution position point of the text in the two-dimensional coordinate plane can be determined.

[0235] Among them, the dimensionality reduction of text features can be performed using t-distributed stochastic neighbor embedding (t-SNE) or the unified manifold approximation and projection (UMAP) method, without any restrictions.

[0236] For ease of understanding, this embodiment uses a two-dimensional coordinate space as an example for illustration, as described above. Figure 3 , Figure 5 as well as Figure 6 The examples used are text classification feature maps with a two-dimensional coordinate space as the display space. The same applies to setting the display space to a three-dimensional space, so they will not be elaborated on here.

[0237] S907, for each text category obtained from the classification, determine the distribution area formed by the first distribution location points of each text in the text category.

[0238] S908, based on the first distribution position point of each text in the text category within the distribution area of ​​the text category, determine the second distribution position point of each core word in the text category within the distribution area of ​​the text category.

[0239] S909 displays the category selection interface based on the category information display request.

[0240] The category selection interface displays text display options and core keyword display options for each of the multiple texts corresponding to at least one text category.

[0241] This embodiment illustrates the process of displaying a category selection interface after detecting a request to display category information. This embodiment also applies to displaying a category selection interface in other ways, and there are no limitations on this.

[0242] S910, determine at least one target display option for the target text category selected by the user in the category selection interface.

[0243] Among them, the target text category belongs to at least one text category, and a target display item belongs to either the text display option or the core word display option.

[0244] S911 displays the text classification feature map based on the distribution area of ​​each target text category, the first distribution location of the text under the target text category, and the second distribution location of the core words.

[0245] The text classification feature map shows the distribution area of ​​the target text category, and in the distribution area of ​​the target text category, at least one target display option indicates the first distribution location point or core word of the text to be displayed.

[0246] Corresponding to the classification result display method of this application, this application also provides a classification result display device.

[0247] like Figure 10 This diagram illustrates a structural composition of a classification result display device according to this application. The device in this embodiment may include:

[0248] Information acquisition unit 1001 is used to acquire classification results of classifying multiple texts and classification feature information of the multiple texts. The classification results include: the text category to which the text belongs. The classification feature information includes: at least one core word corresponding to the text category and the text features of the text extracted during the classification process. The core word of the text category belongs to a set number of word segments in each text segment under the text category whose importance meets the conditions.

[0249] The first position determination unit 1002 is used to determine the first distribution position point of the text features of each text in the text category within a set display space based on the text features of each text in the text category.

[0250] The region determination unit 1003 is used to determine the distribution region formed by the first distribution position points of each text in the text category;

[0251] The image display unit 1004 is used to display a text classification feature map based on the distribution area of ​​the text category. The text classification feature map shows the distribution area of ​​the text category, and the first distribution position point of the text under the text category and at least one of the core words under the text category are marked in the distribution area of ​​the text category.

[0252] In one possible implementation, the device further includes:

[0253] The second position determination unit is used to determine the second distribution position of each core word in the text category in the distribution area of ​​the text category based on the first distribution position of each text in the text category in the distribution area of ​​the text category before the graph display unit displays the text classification feature map;

[0254] The graph display unit is specifically used to display a text classification feature map by combining the distribution area of ​​the text category, the first distribution position point of each text, and the second distribution position point of each core word. The text classification feature map shows the first distribution position point of each text under the text category and each core word under the text category within the distribution area of ​​the text category.

[0255] In another possible implementation, the distribution area of ​​the text category in the text classification feature graph displayed by the graph unit is marked with the first distribution location point of each text under the text category and each core word under the text category;

[0256] The device also includes:

[0257] The first effect control unit is used to detect, after the text classification feature map is displayed in the image display unit, that the input point is located within the distribution area of ​​the displayed text category and within a first set range of the first distribution position point of the target text in the text category, determine the core word of the target text in the text category, and switch the core word of the target text in the text classification feature map from the first display effect to the second display effect, wherein the target text is any text under the text category.

[0258] In yet another possible implementation, the device further includes:

[0259] A content display unit is used to detect a selection operation for a first distribution location point of the target text and display the content of the target text.

[0260] The first selection control unit is used to obtain the first classification contribution weight of each core word of the target text to the target text, and determine the fifth display effect of each core word of the target text based on the first classification contribution weight of each core word of the target text. The fifth display effect of the core words of the target text is also different if the first classification contribution weight of the core words of the target text is different.

[0261] The word display unit is used to display each core word of the target text according to the fifth display effect corresponding to each core word of the target text.

[0262] In another possible implementation, the distribution area of ​​the text category in the text classification feature graph displayed by the graph unit is marked with the first distribution location point of each text under the text category and each core word under the text category;

[0263] The device also includes:

[0264] The second effect display unit is used to detect, after displaying the text classification feature map, that the input point is located within the distribution area of ​​the displayed text category and within the second set range of the second distribution position point where the target core word is located in the text category, determine at least one target text under the text category that has the target core word, and switch the first distribution position point of the target text in the text classification feature map from the third display effect to the fourth display effect, wherein the target core word is any core word under the text category.

[0265] In yet another possible implementation, the device further includes:

[0266] The second selection control unit is used to detect the selection operation for the target core word, display the target core word using the sixth display effect, and display the first distribution position points of each text containing the target core word in the text category to which the target core word belongs using the seventh display effect.

[0267] In yet another possible implementation, the device further includes:

[0268] The selection interface display unit is used to display a classification selection interface before the graph display unit displays the text classification feature map. The classification selection interface displays text display options and core word display options for each of at least one text category corresponding to the plurality of texts. The text display options are used to indicate the text under the text category within the distribution area of ​​the text category, and the core word display options indicate the core words under the text category within the distribution area of ​​the text category.

[0269] The content determination unit is used to determine at least one target display option of the target text category selected by the user in the category selection interface, wherein the target text category belongs to the at least one text category, and the target display item belongs to the text display option or the core word display option.

[0270] The diagram display unit includes:

[0271] The diagram shows a sub-unit for displaying a text classification feature map based on the distribution area of ​​the target text category. The text classification feature map displays the distribution area of ​​the target text category, and the distribution area of ​​the target text category is marked with the first distribution position point of the core word or text to be displayed by each of the at least one target display option.

[0272] In yet another possible implementation, the device further includes:

[0273] The layer building unit is used to build the text distribution layer and the core word distribution layer before the category selection interface is displayed in the image display unit;

[0274] The category selection interface is also used to set the layer overlay order between the text distribution layer and the core word distribution layer;

[0275] The diagram display unit includes:

[0276] A layer processing unit is used to determine, based on at least one target display option of the target text category selected by the user, the text and core words of the target text category to be displayed, and to construct the distribution area of ​​the target text category on the text distribution layer and the core word distribution layer, respectively.

[0277] The first marking unit is used to mark the first distribution position point of each text in the target text category within the distribution area of ​​the target text category in the text distribution layer;

[0278] The second labeling unit is used to label each core word in the target text category within the distribution area of ​​the target text category in the core word distribution layer;

[0279] The layer overlay unit is used to overlay the text distribution layer and the core word distribution layer based on the layer overlay order set by the user in the category selection interface.

[0280] In yet another possible implementation, this embodiment further includes:

[0281] The first weighting unit is used to obtain the first classification contribution weight of at least one word segment of the text. The first word segmentation contribution weight of the word segment in the text is the influence weight of the extracted word segment on determining the text category to which the text belongs during the process of classifying the text.

[0282] The second weight determination unit is used to determine the second classification contribution weight of the word segment to the text category based on the first classification contribution weight of the word segment to different texts in the text category for any word segment contained in each text under the text category;

[0283] The core word determination unit is used to determine a set number of core words that satisfy the condition of the second classification contribution weight from at least one word segment of each text under the text category, wherein the core words belong to the word segments of each text under the text category.

[0284] Furthermore, this application also provides an electronic device, such as Figure 11 As shown, it illustrates a schematic diagram of the composition structure of the electronic device. The electronic device can be any type of electronic device, and the electronic device includes at least a processor 1101 and a memory 1102.

[0285] The processor 1101 is used to execute the classification result display method in any of the above embodiments.

[0286] The memory 1102 is used to store the programs required for the processor to perform operations.

[0287] It is understood that the electronic device may also include a display unit 1103 and an input unit 1104.

[0288] Of course, the electronic device can also have more than Figure 11 There are no restrictions on the number of more or fewer components.

[0289] On the other hand, this application also provides a computer-readable storage medium storing at least one instruction, at least one program, code set, or instruction set, wherein the at least one instruction, the at least one program, the code set, or the instruction set is loaded and executed by a processor to implement the classification result display method as described in any of the above embodiments.

[0290] This application also proposes a computer program comprising computer instructions stored in a computer-readable storage medium. When run on an electronic device, the computer program performs the classification result display method as described in any of the above embodiments.

[0291] It is understood that in this application, the terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings are used to distinguish similar parts and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than that illustrated herein.

[0292] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Furthermore, the features described in the various embodiments of this specification can be substituted or combined with each other, enabling those skilled in the art to implement or use this application. For apparatus embodiments, since they are basically similar to method embodiments, the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0293] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0294] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0295] The above are merely preferred embodiments of this application. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for displaying classification results, comprising: Obtain the classification results of classifying multiple texts and the classification feature information of the multiple texts. The classification results include: the text category to which the text belongs. The classification feature information includes: at least one core word corresponding to the text category and the text features of the text extracted during the classification process. The core word of the text category belongs to a set number of word segments in each text segment under the text category whose importance meets the conditions. Based on the text features of each text in the text category, determine the first distribution location point of the text features of each text in the text category within the set display space; Determine the distribution area formed by the first distribution position points of each text in the text category; Based on the distribution area of ​​the text category, a text classification feature map is displayed. The text classification feature map shows the distribution area of ​​the text category, and the first distribution position point of the text under the text category and at least one of the core words under the text category are marked in the distribution area of ​​the text category. At least one core word corresponding to the text category is obtained through the following method: Obtain the first classification contribution weight of at least one word segment of the text, wherein the first word segmentation contribution weight of the word segment in the text is the influence weight of the extracted word segment on determining the text category to which the text belongs during the process of classifying the text; For any word segment contained in each text under the text category, the second classification contribution weight of the word segment to the text category is determined based on the first classification contribution weight of the word segment to different texts in the text category; From at least one word segment of each text under the text category, determine a set number of core words that satisfy the condition of the second classification contribution weight, wherein the core words belong to the word segments of each text under the text category.

2. The method according to claim 1, further comprising, before displaying the text classification feature map: Based on the first distribution position of each text in the text category within the distribution area of ​​the text category, determine the second distribution position of each core word in the text category within the distribution area of ​​the text category; The distribution region based on the text category displays a text classification feature map, including: By combining the distribution area of ​​the text category, the first distribution position of each text, and the second distribution position of each core word, a text classification feature map is displayed. The distribution area of ​​the text category in the text classification feature map marks the first distribution position of each text under the text category and each core word under the text category.

3. The method according to claim 1, wherein the distribution area of ​​the text category in the text classification feature map is marked with the first distribution location point of each text under the text category and each core word under the text category; Following the display of the text classification feature map, the following is also included: If the input point is detected to be located within the distribution area of ​​the displayed text category and within the first set range of the first distribution position point of the target text in the text category, the core word belonging to the target text in the text category is determined, and the core word of the target text in the text classification feature map is switched from the first display effect to the second display effect, wherein the target text is any text under the text category.

4. The method according to claim 1, wherein the distribution area of ​​the text category in the text classification feature map is marked with the first distribution location point of each text under the text category and each core word under the text category; Following the display of the text classification feature map, the following is also included: If the input point is detected to be located within the distribution area of ​​the displayed text category and within the second set range of the second distribution position point where the target core word is located in the text category, at least one target text with the target core word under the text category is determined, and the first distribution position point of the target text in the text classification feature map is switched from the third display effect to the fourth display effect, wherein the target core word is any core word under the text category.

5. The method according to any one of claims 1 to 4, further comprising, before displaying the text classification feature map: The interface displays a category selection interface, which shows text display options and core word display options for each of at least one text category corresponding to the multiple texts. The text display options are used to indicate the text under the text category within the distribution area of ​​the text category, and the core word display options indicate the core words under the text category within the distribution area of ​​the text category. Determine at least one target display option for the target text category selected by the user in the category selection interface, wherein the target text category belongs to the at least one text category, and the target display item belongs to the text display option or the core word display option; The distribution region based on the text category displays a text classification feature map, including: Based on the distribution area of ​​the target text category, a text classification feature map is displayed. The text classification feature map shows the distribution area of ​​the target text category, and the first distribution position point of the core word or text to be displayed by each of the at least one target display option is marked in the distribution area of ​​the target text category.

6. The method according to claim 5, further comprising, before displaying the category selection interface: Construct a text distribution layer and a core word distribution layer; The category selection interface is also used to set the layer overlay order between the text distribution layer and the core word distribution layer; The distribution region based on the target text category displays a text classification feature map, including: If, based on at least one target display option of the target text category selected by the user, it is determined that the text and core words of the target text category need to be displayed, a distribution area of ​​the target text category is constructed on the text distribution layer and the core word distribution layer, respectively. Within the distribution area of ​​the target text category in the text distribution layer, the first distribution location point of each text in the target text category is marked; Within the distribution area of ​​the target text category in the core word distribution layer, each core word in the target text category is marked; Based on the layer overlay order set by the user in the category selection interface, the text distribution layer and the core word distribution layer are overlaid and displayed.

7. The method according to claim 3, further comprising: Upon detecting a selection operation for a first distribution location point of the target text, the content of the target text is displayed. Obtain the first category contribution weight of each core word of the target text to the target text, and determine the fifth display effect of each core word of the target text based on the first category contribution weight of each core word of the target text. The fifth display effect of the core words of the target text is different if the first category contribution weight of the core words of the target text is different. According to the fifth display effect corresponding to each core word of the target text, each core word of the target text is displayed respectively.

8. The method according to claim 4, further comprising: Upon detecting a selection operation targeting the target core word, the target core word is displayed using a sixth display effect, and the first distribution position points of each text containing the target core word in the text category to which the target core word belongs are displayed using a seventh display effect.

9. A classification result display device, comprising: An information acquisition unit is used to acquire classification results of multiple texts and classification feature information of the multiple texts. The classification results include: the text category to which the text belongs. The classification feature information includes: at least one core word corresponding to the text category and the text features of the text extracted during the classification process. The core word of the text category belongs to a set number of word segments in each text segment under the text category whose importance meets the conditions. The first position determination unit is used to determine the first distribution position point of the text features of each text in the text category within a set display space based on the text features of each text in the text category. A region determination unit is used to determine the distribution region formed by the first distribution location points of each text in the text category; The image display unit is used to display a text classification feature map based on the distribution area of ​​the text category. The text classification feature map shows the distribution area of ​​the text category, and the first distribution position point of the text under the text category and at least one of the core words under the text category are marked in the distribution area of ​​the text category. At least one core word corresponding to the text category is obtained through the following method: Obtain the first classification contribution weight of at least one word segment of the text, wherein the first word segmentation contribution weight of the word segment in the text is the influence weight of the extracted word segment on determining the text category to which the text belongs during the process of classifying the text; For any word segment contained in each text under the text category, the second classification contribution weight of the word segment to the text category is determined based on the first classification contribution weight of the word segment to different texts in the text category; From at least one word segment of each text under the text category, determine a set number of core words that satisfy the condition of the second classification contribution weight, wherein the core words belong to the word segments of each text under the text category.

Images