Text processing method, apparatus, device, and medium

Abstract

The present disclosure relates to the technical field of computers, and discloses a text processing method, device, equipment and medium. The method comprises: obtaining an initial description text of a target object, the initial description text being used to describe initial characteristics of the target object; and outputting an extended description text obtained by revising the initial description text, the extended description text being used to describe the initial characteristics of the target object and at least partial label characteristics, the label characteristics being characteristics of one or more object groups to which the target object belongs, and the extended description text being used to search for the target object. The searching accuracy can be improved.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more specifically to text processing methods, apparatus, devices, and media. Background Technology

[0002] In some applications, users typically search for objects that match their descriptions. These objects can be people or things. However, a current problem is that the user-input descriptions are often inaccurate, leading to inaccurate object searches. Summary of the Invention

[0003] In view of this, the present disclosure provides a text processing method, a text processing apparatus, an electronic device, a computer-readable storage medium, and a computer program product that can improve search accuracy.

[0004] In a first aspect, this disclosure provides a text processing method, the method comprising:

[0005] Obtain the initial description text of the target object, wherein the initial description text is used to describe the initial characteristics of the target object;

[0006] Output the extended description text obtained by rewriting the initial description text. The extended description text is used to describe the initial features and at least some label features of the target object. The label features are the features of one or more object groups to which the target object belongs. The extended description text is used to find the target object.

[0007] Secondly, this disclosure provides a text processing apparatus, the apparatus comprising:

[0008] The text acquisition module is used to acquire the initial description text of the target object, wherein the initial description text is used to describe the initial characteristics of the target object;

[0009] The output module is used to output an extended description text obtained by rewriting the initial description text. The extended description text is used to describe the initial features and at least some label features of the target object. The label features are features of one or more object groups to which the target object belongs. The extended description text is used to find the target object.

[0010] Thirdly, this disclosure provides an electronic device, including: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the method described above.

[0011] Fourthly, this disclosure provides a computer-readable storage medium storing computer instructions for causing a computer to perform the above-described methods.

[0012] Fifthly, this disclosure provides a computer program product, including computer instructions for causing a computer to perform the methods described above.

[0013] In some embodiments of this disclosure, after obtaining the initial description text for the target object, the initial description text is rewritten, and at least some of the tag features of the target object are added to the initial description text to obtain extended description text. In this way, the extended description text can include more feature descriptions of the target object, allowing for more accurate constraints on the characteristics of the target object when searching for the target object based on the extended description text, thereby obtaining more accurate search results. This achieves the goal of improving search accuracy. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the specific embodiments or related technologies of this disclosure, the accompanying drawings used in the description of the specific embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart illustrating a text processing method provided in some embodiments of this disclosure;

[0016] Figure 2 This is a schematic diagram of the architecture of a text processing system provided in some embodiments of this disclosure;

[0017] Figure 3 This is a schematic diagram of a text processing apparatus provided in some embodiments of the present disclosure;

[0018] Figure 4 These are schematic diagrams of the structure of an electronic device provided in some embodiments of this disclosure. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0020] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0021] In the description of embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The terms "some embodiments" or "this embodiment" should be understood as "at least some embodiments". The term "some embodiments" should be understood as "at least some embodiments". Other explicit and implicit definitions may also be included below.

[0022] In this document, unless explicitly stated otherwise, performing a step in response to A does not mean that the step is performed immediately after A, but may include one or more intermediate steps.

[0023] It is understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition, use, storage or deletion of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0024] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, relevant users should be informed of the type, scope of use, and usage scenarios of the information involved in this disclosure through appropriate means in accordance with relevant laws and regulations, and authorization should be obtained from the relevant users. Among them, relevant users may include any type of rights holder, such as individuals, enterprises, and groups.

[0025] For example, in response to receiving an active request from a user, a prompt message is sent to the relevant user to clearly inform the user that the requested operation will require obtaining and using the user's information, thereby enabling the relevant user to choose whether to provide information to the software or hardware such as the electronic device, application, server, or storage medium that performs the operation of the technical solution disclosed herein based on the prompt message.

[0026] As an optional but non-restrictive implementation, in response to a user's active request, a prompt message can be sent to the user, such as a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide information to the electronic device.

[0027] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0028] Currently, some technologies allow users to input descriptive information about their target object into a search system. This description can describe the object's characteristics, such as its location, height, weight, and occupation. Based on this description, the search system generates a query. Executing the query retrieves objects from the database that match the description, which are then used as the user's target.

[0029] For example, suppose a user enters "people in city C who frequently send beef," the search system can automatically generate a Structured Query Language (SQL) statement, specifying city C and beef as the type of item sent. By executing the SQL query, the system can find the group of people in the database who send beef in city C and return the results to the user. In this way, even if a user lacks the skills to write SQL queries, they can still retrieve their desired information from the database.

[0030] However, some current problems include the possibility that user-inputted descriptions may be inaccurate or incomplete, leading to inaccurate SQL queries and ultimately resulting in searches that do not reflect the user's true intentions. For example, in the description above, the user inputs "people in city C who frequently send beef." Because the word "frequently" is vague, the SQL query may not specify the frequency of beef shipments. This could result in the search results including individuals who only send beef once. Clearly, such individuals do not match the user's description and should not be filtered out, thus reducing search accuracy.

[0031] Therefore, this disclosure provides a text processing method that can automatically recommend descriptive features with clear constraints based on user-input natural language descriptions. This enables the data platform to locate and provide feedback on the content, improving the accuracy and efficiency of data analysis. The text processing method can be applied to text processing systems, or to electronic devices running text processing systems. Electronic devices may include, but are not limited to, tablets, desktop computers, laptops, servers, etc. (See also...) Figure 1 This is a flowchart illustrating a text processing method provided in some embodiments of this disclosure. Figure 1 In Chinese, text processing methods include the following steps:

[0032] Step S101: Obtain the initial description text of the target object. The initial description text is used to describe the initial characteristics of the target object.

[0033] Specifically, the target object is the object the user needs to find, such as a certain group of people or a certain type of item. The initial description text can be text entered by the user to describe the characteristics of the target object, and the characteristics of the target object described in the initial description text can be brief or complex.

[0034] In this embodiment, the target object features described in the initial description text are used as the initial features of the target object.

[0035] Step S102: Output the extended description text obtained after rewriting the initial description text. The extended description text is used to describe the initial features and at least some label features of the target object. The label features are the features of one or more object groups to which the target object belongs. The extended description text is used to find the target object.

[0036] Specifically, in this embodiment, after obtaining the initial description text of the target object, the initial description text can be rewritten. Rewriting can involve adding text describing the characteristics of the target object to the initial description text. Of course, the sentence structure of the initial description text can also be readjusted, and inaccurate or erroneous content in the initial description text can be corrected. This disclosure does not impose any limitations on this.

[0037] In this embodiment, the tags of the target object can be determined based on the information provided in the initial description text, and the tag features of the tags can be added to the initial description text. Specifically, the tags of the target object are used to characterize the object group to which the target object belongs. Depending on different classification dimensions, the target object can have one or more tags. For example, if the target object has the tag "e-commerce customer," it means that the target object belongs to the object group of "e-commerce customer." Another example is if the target object has the tag "City C," which means that the target object is located in City C. Each tag can have its own associated tag features. Tag features are used to describe the characteristics of the object group represented by the tag, including but not limited to indicators and time constraints set for the object group. For example, the object group of "e-commerce customer" typically has the indicator "e-commerce platform transaction activity," so the indicator "e-commerce platform transaction activity" can be used as a tag feature associated with the tag "e-commerce customer."

[0038] By supplementing the initial description text with at least some of the target object's label features, the resulting extended description text can have a richer feature description of the target object; that is, the extended description text can more accurately describe the constraints for the target object. Thus, when generating a query statement for the target object based on the extended description text, the search conditions in the query statement can be more accurate, and consequently, the target object retrieved by executing the query statement can be more accurate.

[0039] In summary, in the technical solutions of some embodiments of this disclosure, after obtaining the initial description text for the target object, the initial description text is rewritten, and at least some of the tag features of the target object are added to the initial description text to obtain extended description text. In this way, the extended description text can include more feature descriptions of the target object, allowing for more accurate constraints on the characteristics of the target object when searching for the target object based on the extended description text, thereby obtaining more accurate search results. This achieves the goal of improving search accuracy.

[0040] For details, please refer to the following: Figure 2 This is a schematic diagram of the architecture of a text processing system provided in some embodiments of this disclosure. Figure 2 In this text processing system, a front-end input module, a text processing module, and an output module are included. The front-end input module can include an input interface where users can input initial descriptive text of the target object. In practice, the front-end input module can use an HTTP concurrent service component to receive the initial descriptive text input by the user via the GET method. Based on the initial descriptive text, the text processing module can rewrite the initial descriptive text to obtain extended descriptive text. The extended descriptive text can be output through the output module. The output module can include an output interface or an output interface. If the output module includes an output interface, the rewritten extended descriptive text can be displayed there. If the output module includes an output interface, the output interface can interface with other systems, such as a data query system. After obtaining the extended descriptive text, the data query system can generate a query statement based on it. By executing the query statement, the target object to be searched by the user can be obtained from the database.

[0041] In some embodiments, the output module may output the extended description text in JSON format. Of course, other formats may also be used. This disclosure does not limit this.

[0042] In some embodiments, the text processing module may rewrite the initial description text based on the following text processing logic.

[0043] Among multiple candidate tags, filter the target tags that match the target object. The candidate tags represent pre-created groups of objects, and the target tags represent the group of objects to which the target object belongs.

[0044] The tag features of the target tag are added to the initial description text to obtain the expanded description text.

[0045] Specifically, candidate tags can reside in a pre-established tag index. This index can be pre-built from multiple dimensions. For example, taking logistics and transportation, tags can be built from multiple dimensions such as business behavior characteristics (e.g., order volume, spending amount, transportation mode preferences), basic information of logistics shippers (e.g., region, age, gender), and logistics service-related attributes (e.g., number of insurance claims, claims history). Thus, when filtering target tags that match the target object based on candidate tags, multiple target tags from different dimensions can be obtained. These target tags from different dimensions can have their own associated tag features, and the tag features of different target tags can be different. After adding these target tag features to the initial description text, the resulting extended description text can contain the characteristics of the target object in different dimensions, thus facilitating more accurate searching and filtering of the target object.

[0046] In some embodiments, candidate labels may have label description text. The label description text typically provides a brief description of the basic characteristics of the target group represented by the candidate label. For example, the label description text for candidate label A is "e-commerce customer," and the label description text for candidate label B is "international logistics customer." The process of filtering target labels that match the target object from multiple candidate labels may include:

[0047] The similarity of the candidate label description text with the initial description text is compared to obtain the target label description text whose similarity with the initial description text exceeds the similarity threshold;

[0048] Use the candidate tags corresponding to the target tag description text as the target tag.

[0049] For example, suppose the initial description text entered by the user is "Tianjin users with less than 500 insured shipments in the e-commerce sector," and the tag index contains candidate tags A and B, with candidate tag A having the tag description text "e-commerce customer" and candidate tag B having the tag description text "international logistics customer." Then, we can compare the similarity between the initial description text "Tianjin users with less than 500 insured shipments in the e-commerce sector" and the tag description text "e-commerce customer," and we can also compare the similarity between the initial description text "Tianjin users with less than 500 insured shipments in the e-commerce sector" and the tag description text "international logistics customer." If the similarity between the initial description text "Tianjin users with less than 500 insured shipments in the e-commerce sector" and the tag description text "e-commerce customer" exceeds the similarity threshold, then candidate tag A corresponding to the tag description text "e-commerce customer" can be used as the target tag. If the similarity between the initial description text "Tianjin users with less than 500 insured shipments in the e-commerce sector" and the tag description text "international logistics customer" does not exceed the similarity threshold, then candidate tag B does not need to be used as the target tag.

[0050] Specifically, in the similarity comparison process, similarity comparison can be performed based on algorithms such as cosine similarity.

[0051] In the above embodiments, since the label description text describes the basic characteristics of the object group represented by the candidate labels, and the initial description text also describes the basic characteristics of the target object, a more accurate target label can be found by comparing the similarity between the initial description text and the label description text. This also improves the feasibility of the solution.

[0052] In some embodiments, multiple candidate tags can be divided into multiple tag groups, and each tag group has a tag group description text. The candidate tags in each tag group can be tags of the same type. The tag group description text is used to describe the basic characteristics of the candidate tags in the tag group. The above-mentioned comparison of the similarity between the tag description text of the candidate tags and the initial description text may include:

[0053] Based on the initial description text and the tag group description text, filter the target tag group that matches the target object from multiple tag groups;

[0054] The similarity of the tag description text of each candidate tag in the target tag group with the initial description text is compared.

[0055] For example, assuming there are tag group A and tag group B, the initial description text can be compared with the tag group description texts of tag group A and tag group B respectively to obtain the target tag group, such as tag group B, whose similarity to the initial description text exceeds a similarity threshold. Then, the tag description texts of each candidate tag within the target tag group can be compared with the initial description text. Specifically, within the target tag group, some candidate tags may have a high similarity to the initial description text, while others may have a low similarity. In this case, candidate tags with high similarity to the initial description text can be selected as target tags within the target tag group, while candidate tags with low similarity to the initial description text can be discarded.

[0056] In the above embodiments, the initial description text is first compared with the description text of the tag group, and then the target tags are further filtered within the selected target tag groups. This reduces the amount of data comparison. For example, if the description text of tag group A has a low similarity to the initial description text, then the candidate tags in tag group A can be directly eliminated, i.e., there is no need to compare the candidate tags in tag group A with the initial description text, thereby reducing the amount of data comparison.

[0057] In some embodiments, a keyword dictionary can be pre-built. The keyword dictionary may include common feature words and business terms from various fields. Based on the keyword dictionary, keywords can be extracted from the initial description text, and the extracted keywords can be compared with the tag group description text and the tag description text in terms of similarity, thus improving the accuracy of similarity comparison.

[0058] Based on the above description, in some embodiments, based on the initial description text and the tag group description text, a target tag group matching the target object is selected from multiple tag groups, including:

[0059] Extract keywords from the initial description text;

[0060] Based on the initial description text, determine the intent to find the target object;

[0061] The description text of keywords and intents is matched and compared with the description text of each tag group to obtain the target tag group.

[0062] Specifically, intent refers to the purpose for which a user searches for a target object. For example, in some logistics service recommendation scenarios, it is necessary to find target objects that meet specified conditions and recommend logistics services to these target objects via SMS or other means. Here, recommending logistics services to the target objects is the purpose of searching for the target objects.

[0063] In this embodiment, matching the description text of keywords with the description text of each tag group can be understood as searching for target tag groups that match the user's initial description text. However, the initial description text input by the user may not be accurate enough, leading to inaccurate or missing target tag groups. Therefore, by identifying the user's intent to find a target object based on keywords and matching the description text of that intent with the description text of each tag group, the shortcomings of the initial description text can be compensated for, making the found target tag groups more accurate. Furthermore, the characteristics of the target object can be more precisely constrained and limited in the extended description text.

[0064] In this embodiment, the description text of keywords and the description text of intent can be combined, and the combined description text can be matched and compared with the description text of each tag group to find the target tag group. The matching and comparison methods include, but are not limited to, fuzzy matching, exact matching, and a combination of fuzzy matching and exact matching.

[0065] In some embodiments, filtering target tags that match the target object from a plurality of candidate tags includes:

[0066] Determine the sentence structure of the initial description text;

[0067] The sentence structure of the initial description text is matched and compared with multiple preset sentence structure templates, and each sentence structure template has its own associated alternative tags.

[0068] Based on the matching results, a target sentence structure template that matches the sentence structure of the initial description text is determined, and the candidate tags associated with the target sentence structure template are used as the target tags.

[0069] Sentence structure, in this context, refers to dividing descriptive text into modules based on its parts of speech (such as nouns, verbs, prepositions, etc.). For example, "Beijing mooncakes" and "Shenzhen lychees" can correspond to the same sentence structure template: city+class. Similarly, "mooncake users in Beijing" can correspond to the template: city+n+class+n, where "city" represents the city, "n" represents the region, "class" represents the category of the item, and "n" represents the part of speech of the user.

[0070] Specifically, a knowledge base can be pre-built, and multiple sentence structure templates and associated candidate tags can be set within the knowledge base. This allows for the selection of target tags based on the sentence structure of the initial description text.

[0071] This disclosure has thus far described two different methods for filtering target tags. One method compares the sentence structure of the initial description text with a sentence structure template based on similarity to obtain the target tags. The other method compares the initial description text with the tag description text of candidate tags based on similarity to obtain the target tags. Using two different filtering methods can prevent issues such as omissions. For example, the first filtering method might not be able to filter out target tag A, but the other method can, thus solving the problem of omissions.

[0072] In some embodiments, the target tags obtained from these two filtering methods can be merged to form the final filtered target tags.

[0073] In some embodiments, when multiple target tags exist, adding the tag features of the target tags to the initial description text includes:

[0074] Based on the initial description text, determine the intent to find the target object;

[0075] Determine the correlation between each target label and the intent;

[0076] The tag features of each target tag are added to the initial description text in descending order of relevance.

[0077] Specifically, after adding the tag features of each target tag to the initial description text in descending order of relevance, the subsequent query system can prioritize target tags with high relevance to the intent, thereby ensuring search accuracy.

[0078] Specifically, a multi-factor weighted sorting algorithm can be used to sort the target labels.

[0079] In some embodiments, based on historical text processing results, some target tags may be associated with relatively accurate descriptive text. Therefore, adding the tag features of the target tags to the initial descriptive text includes:

[0080] If the target label does not have an associated description text, the label features of the target label are added to the initial description text to obtain the expanded description text;

[0081] If the target tag has associated descriptive text, then the associated descriptive text of the target tag will be used as the extended descriptive text.

[0082] This avoids regenerating the extended description text and reduces the amount of data processing.

[0083] In some embodiments, the extended description text can be organized in a JSON data format. For ease of understanding, an example of extended description text is given below:

[0084] {"input":"E-commerce section price-protected items less than 500 tickets Tianjin","min":0.7999,"topk":10,"part":"add_tag","data":[{"sim":0.8,"name":"rule","hot":"{"template":"add_tagadd_city","relationType":"AND","seg":"zhiye / e-commercenz / e-commercen / sectionn / price-protectedbaojia / price-protecteduj / ofq / item package / itemn / quantityv / less than number / 500n / ticket city / Tianjin","isRelation":false,"tagList":[{"tagValue":"500","tagName":"cc_ecom_service_2_cnt","tagCompare":"lt"},{"tagValue":"TianJin","tagName":"consignor_city","tagCompare":"in"}]}"}]}

[0085] In the extended description text mentioned above, the value corresponding to the "input" field is the initial description text input by the user, while the values ​​corresponding to the "data" and "tagList" fields can be tag features for adding target tags. The tag features can be sorted according to the relevance of each target tag to the intent, placing tag features associated with target tags with a high degree of relevance to the intent first, and tag features associated with target tags with a low degree of relevance to the intent later. This allows the query system receiving the extended description text to determine the importance of the tag features based on their order.

[0086] In some embodiments, the extended description text may also include, but is not limited to, information such as the tag description text of each target tag, the importance of the target tag, the similarity between the tag description text of each target tag and the initial description text, the similarity threshold, and the number of target tags.

[0087] As can be seen from the examples of extended descriptive text above, extended descriptive text can include target object description text with clear conditions, complete constraints, and sufficient boundaries. However, the content of extended descriptive text is usually extensive and difficult to understand. Therefore, in some embodiments, the extended descriptive text can also be input into a trained language model. The language model can summarize the extended descriptive text to obtain a summarized descriptive text. For example, after summarizing the extended descriptive text in the above example, the language model might obtain the summarized descriptive text as: "A highly active group in the e-commerce sector with fewer than 500 insured items located in Tianjin." This makes it easier to understand.

[0088] In some embodiments, both the extended description text and the summary description text of the extended description text can be output to the backend query system simultaneously.

[0089] In summary, in some embodiments, the extended description text of the target object can be obtained by following these steps:

[0090] 1) Load keyword dictionaries, intent recognition models, language models, etc., and establish connections with tag index libraries, knowledge bases, etc.;

[0091] 2) Receive initial description text input by the user, such as "E-commerce platform, less than 500 items for Tianjin users with price protection";

[0092] 3) Based on the keyword dictionary, segment the initial description text and extract keywords from the initial description text, such as: e-commerce / sector / price protection / of / quantity / less than / 500 / ticket / Tianjin / user;

[0093] 4) Input the initial description text into the intent recognition model to obtain the intent to find the target object;

[0094] 5) Based on intent and keywords, find target tag groups, such as "e-commerce customers" and "price protection behavior";

[0095] 6) Based on similarity comparison, filter target tags from the target tag group;

[0096] 7) Add at least some of the label features of the target label to the initial description text to obtain the expanded description text;

[0097] 8) Input the extended descriptive text into the language model to obtain a generalized descriptive text of the extended descriptive text.

[0098] See also Figure 3 This is a schematic diagram of a text processing apparatus provided in some embodiments of the present disclosure. Figure 3 In the text processing device, the following are included:

[0099] The text acquisition module 301 is used to acquire the initial description text of the target object, which is used to describe the initial characteristics of the target object.

[0100] Output module 302 is used to output extended description text obtained by rewriting the initial description text. The extended description text is used to describe the initial features and at least some label features of the target object. The label features are the features of one or more object groups to which the target object belongs. The extended description text is used to find the target object.

[0101] In some embodiments, the text processing apparatus further includes a text processing module 303, which is specifically used for:

[0102] Among multiple candidate tags, filter the target tags that match the target object. The candidate tags represent pre-created groups of objects, and the target tags represent the group of objects to which the target object belongs.

[0103] The tag features of the target tag are added to the initial description text to obtain the expanded description text.

[0104] In some embodiments, the alternative labels have label description text; the text processing module 303 is specifically used for:

[0105] The similarity of the candidate label description text with the initial description text is compared to obtain the target label description text whose similarity with the initial description text exceeds the similarity threshold;

[0106] Use the candidate tags corresponding to the target tag description text as the target tag.

[0107] In some embodiments, multiple candidate tags can be divided into multiple tag groups, and each tag group has tag group description text; the text processing module 303 is specifically used for:

[0108] Based on the initial description text and the tag group description text, filter the target tag group that matches the target object from multiple tag groups;

[0109] The similarity of the tag description text of each candidate tag in the target tag group with the initial description text is compared.

[0110] In some embodiments, the text processing module 303 is specifically used for:

[0111] Extract keywords from the initial description text;

[0112] Based on the initial description text, determine the intent to find the target object;

[0113] The description text of keywords and intents is matched and compared with the description text of each tag group to obtain the target tag group.

[0114] In some embodiments, the text processing module 303 is specifically used for:

[0115] Determine the sentence structure of the initial description text;

[0116] The sentence structure of the initial description text is matched and compared with multiple preset sentence structure templates, and each sentence structure template has its own associated alternative tags.

[0117] Based on the matching results, a target sentence structure template that matches the sentence structure of the initial description text is determined, and the candidate tags associated with the target sentence structure template are used as the target tags.

[0118] In some embodiments, the text processing module 303 is specifically used for:

[0119] Based on the initial description text, determine the intent to find the target object;

[0120] Determine the correlation between each target label and the intent;

[0121] The tag features of each target tag are added to the initial description text in descending order of relevance.

[0122] In some embodiments, the text processing module 303 is specifically used for:

[0123] If the target label does not have an associated description text, the label features of the target label are added to the initial description text to obtain the expanded description text;

[0124] If the target tag has associated descriptive text, then the associated descriptive text of the target tag will be used as the extended descriptive text.

[0125] This disclosure also provides an electronic device having the above-described features. Figure 3 The text processing device shown.

[0126] See also Figure 4 This is a schematic diagram of the structure of an electronic device provided in some embodiments of this disclosure. For example... Figure 4As shown, the electronic device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise as required. The processors can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 4 Take a processor 10 as an example.

[0127] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0128] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0129] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0130] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0131] The electronic device also includes a communication interface 30 for communicating with other devices or communication networks.

[0132] This disclosure also provides a computer-readable storage medium in which the methods described in this disclosure can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium after being downloaded over a network. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium may be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium may also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code that, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0133] A portion of this disclosure can be applied to computer program products, such as computer program instructions, which, when executed by a computer, can invoke or provide methods and / or technical solutions according to this disclosure through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, and installation package files. Accordingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions; the computer compiling the instructions and then executing the corresponding compiled program; the computer reading and executing the instructions; or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0134] Although embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A text processing method, characterized in that, The method includes: Obtain the initial description text of the target object, wherein the initial description text is used to describe the initial characteristics of the target object; Output the extended description text obtained by rewriting the initial description text. The extended description text is used to describe the initial features and at least some label features of the target object. The label features are the features of one or more object groups to which the target object belongs. The extended description text is used to find the target object.

2. The method according to claim 1, characterized in that, Rewriting the initial description text includes: Among multiple candidate tags, target tags that match the target object are selected, where the candidate tags represent pre-created object groups and the target tag represents the object group to which the target object belongs; The tag features of the target tag are added to the initial description text to obtain the extended description text.

3. The method according to claim 2, characterized in that, The alternative labels have label description text; The step of filtering target tags that match the target object from multiple candidate tags includes: The similarity of the label description text of the candidate label is compared with that of the initial description text to obtain the target label description text whose similarity to the initial description text exceeds the similarity threshold; The candidate tags corresponding to the target tag description text are used as the target tag.

4. The method according to claim 3, characterized in that, The multiple candidate tags can be divided into multiple tag groups, and each tag group has tag group description text; The step of comparing the similarity between the label description text of the candidate labels and the initial description text includes: Based on the initial description text and the tag group description text, a target tag group that matches the target object is selected from the plurality of tag groups; The similarity of the tag description text of each candidate tag in the target tag group with the initial description text is compared.

5. The method according to claim 4, characterized in that, The step of filtering target tag groups that match the target object from the plurality of tag groups based on the initial description text and the tag group description text includes: Extract keywords from the initial description text; Based on the initial description text, the intent to locate the target object is determined; The keyword and the description text of the intent are matched and compared with the description text of each of the tag groups to obtain the target tag group.

6. The method according to claim 2, characterized in that, The step of filtering target tags that match the target object from multiple candidate tags includes: Determine the sentence structure of the initial description text; The sentence structure of the initial description text is matched and compared with multiple preset sentence structure templates, and each sentence structure template has its own associated candidate tags; Based on the matching results, a target sentence structure template that matches the sentence structure of the initial description text is determined, and the candidate tags associated with the target sentence structure template are used as the target tags.

7. The method according to claim 2, characterized in that, In the presence of multiple target tags, adding the tag features of the target tags to the initial description text includes: Based on the initial description text, the intent to locate the target object is determined; Determine the degree of association between each of the target tags and the intent; The tag features of each target tag are added to the initial description text in descending order of relevance.

8. The method according to claim 2, characterized in that, Adding the tag features of the target tag to the initial description text includes: If the target label does not have an associated description text, then the label features of the target label are added to the initial description text to obtain the extended description text; If the target tag has associated descriptive text, then the descriptive text associated with the target tag will be used as the extended descriptive text.

9. A text processing device, characterized in that, The device includes: The text acquisition module is used to acquire the initial description text of the target object, wherein the initial description text is used to describe the initial characteristics of the target object; The output module is used to output an extended description text obtained by rewriting the initial description text. The extended description text is used to describe the initial features and at least some label features of the target object. The label features are the features of one or more object groups to which the target object belongs. The extended description text is used to find the target object.

10. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the text processing method of any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the text processing method according to any one of claims 1 to 8.

Images