Text data processing method, device, and medium

Abstract

The embodiments of the present application provide a text data processing method, device, and medium, which relate to the fields of computer and artificial intelligence technology. The method includes: obtaining a search text input by a user to be pushed; when it is detected that the text type of the search text belongs to the text type defined in a preset vocabulary, obtaining a set of associated texts associated with the search text, wherein the associated text set includes associated texts; selecting at least one associated text from the associated text set as a recommended text, and obtaining text attribute information of the recommended text; and pushing the recommended text and the text attribute information to the user to be pushed. The technical solution of the embodiments of the present application can improve the efficiency of user text search.

Description

Technical Field

[0001] The present application relates to the field of computer and artificial intelligence technology, and more specifically, to a text data processing method, device, and medium. Background Art

[0002] In a text data processing scenario, such as an application scenario for processing search text data input by a user, a search phrase including the keyword is usually obtained based on the keyword in the user's search text, and the search phrase is displayed to the user for selection. However, in this solution, when the user needs to search for other texts, the user still needs to re-enter the other texts, which makes the user's text search inefficient. Based on this, how to improve the efficiency of user text search is a technical problem that needs to be solved urgently. Summary of the invention

[0003] The embodiments of the present application provide a text data processing method, apparatus, computer program product or computer program, and computer-readable medium, which can improve the efficiency of user text search at least to a certain extent.

[0004] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0005] According to one aspect of an embodiment of the present application, a text data processing method is provided, including: obtaining a search text input by a user to be pushed; when it is detected that the text type of the search text belongs to the text type defined in a preset vocabulary, obtaining an associated text set associated with the search text, the associated text set including associated texts; selecting at least one associated text from the associated text set as a recommended text, and obtaining text attribute information of the recommended text; and pushing the recommended text and the text attribute information to the user to be pushed.

[0006] According to one aspect of an embodiment of the present application, a text data processing device is provided, comprising: a first acquisition unit, used to acquire a search text input by a user to be pushed; a second acquisition unit, used to acquire an associated text set associated with the search text when it is detected that the text type of the search text belongs to a text type defined in a preset vocabulary, the associated text set including associated texts; a selection unit, used to select at least one associated text from the associated text set as a recommended text, and acquire text attribute information of the recommended text; and a push unit, used to push the recommended text and the text attribute information to the user to be pushed.

[0007] In some embodiments of the present application, based on the aforementioned scheme, the push unit is configured to: obtain a recommendation reference value for each recommended text, wherein the recommendation reference value is used to characterize the popularity of the recommended text being searched; and display the recommended text, as well as the recommendation reference value and text attribute information corresponding to the recommended text in an interface.

[0008] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit is configured to: identify the target text type of the search text; based on the target text type, acquire associated text associated with the search text to obtain the associated text set.

[0009] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit is further configured to: acquire the text entered by the user to be pushed within a predetermined historical time interval as the associated text to be selected; determine the associated text from the associated text to be selected to obtain the associated text set.

[0010] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit is further configured to: acquire text that is similar to the search text in a first text feature as a candidate associated text, the first text feature including at least one of a semantic feature, a writing form feature, and a pronunciation feature; determine the associated text from the candidate associated texts to obtain the associated text set.

[0011] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit is further configured to: obtain the text entered by the user to be pushed within a predetermined historical time interval, and the text that is similar to the search text in the second text feature, as the selected associated text, the second text feature including at least one of a semantic feature, a writing form feature, and a pronunciation feature; determine the associated text from the selected associated text to obtain the associated text set.

[0012] In some embodiments of the present application, based on the above-mentioned solution, the second acquisition unit includes: a duplicate filtering unit, which is used to determine repeated texts in the candidate associated texts and filter the repeated texts to obtain the associated texts.

[0013] In some embodiments of the present application, based on the aforementioned scheme, the selection unit is configured to: sort each associated text in the associated text set through a pre-trained sorting model to obtain reference sorting information corresponding to each associated text, and the reference sorting information is used to characterize the degree of association of the associated text; based on the reference sorting information, select at least one associated text from the associated text set as a recommended text.

[0014] In some embodiments of the present application, based on the aforementioned scheme, the selection unit is further configured to: obtain a third text feature corresponding to the search text and each associated text, the third text feature including at least one of a semantic feature, a writing form feature, and a pronunciation feature; based on the third text feature corresponding to the search text and each associated text, determine a recommendation index corresponding to each associated text through the sorting model; and sort each associated text in the associated text set according to the recommendation index.

[0015] In some embodiments of the present application, based on the aforementioned scheme, the sorting model includes a first sub-model and a second sub-model, the first sub-model has the ability to memorize the third text feature corresponding to the search text, and the second sub-model has the ability to generalize the third text feature corresponding to the search text, and the selection unit is also configured to: input the third text feature corresponding to the search text and each associated text into the first sub-model to output a first recommendation index corresponding to each associated text; input the third text feature corresponding to the search text and each associated text into the second sub-model to output a second recommendation index corresponding to each associated text; for each associated text, perform weighted calculation on the first sub-recommendation index and the second recommendation index to obtain the recommendation index corresponding to the associated text.

[0016] In some embodiments of the present application, based on the aforementioned scheme, the first sub-model includes a bilinear transformation model, and the second sub-model is a neural network model.

[0017] In some embodiments of the present application, based on the aforementioned scheme, the device also includes: a training unit, which is used to obtain the behavior data generated by the user to be pushed based on the recommended text and the text attribute information pushed to the user to be pushed; retrain the sorting model based on the behavior data to obtain a retrained sorting model, and the retrained sorting model is used for the next associated text sorting.

[0018] In some embodiments of the present application, based on the aforementioned scheme, the training unit is configured to: determine the expected sorting information corresponding to each associated text based on the behavioral data; and correct the hidden layer parameters in the sorting model by comparing the expected sorting information with the reference sorting information through gradient back transfer.

[0019] According to one aspect of the embodiments of the present application, a computer program product or a computer program is provided, the computer program product or the computer program includes computer instructions, the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the text data processing method as described in the above embodiments.

[0020] According to one aspect of an embodiment of the present application, a text data processing device is provided, characterized in that it includes a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include instructions for performing the text data processing method as described in the above embodiments.

[0021] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, in which at least one program code is stored. The at least one program code is loaded and executed by a processor to implement the operations performed by the text data processing method described in the above embodiment.

[0022] In the technical solutions provided in some embodiments of the present application, when it is detected that the text type of the search text input by the user belongs to the text type defined in the preset vocabulary, a set of associated texts associated with the search text is obtained, at least one associated text can be selected from the associated text set as a recommended text, and the text attribute information of the recommended text can be obtained, so as to push the recommended text and the text attribute information to the user to be pushed. Since the recommended text and the corresponding text attribute information associated with the search text are pushed to the user, a reference for searching other associated texts can be provided to the user. For example, when the user searches for a Chinese language, other Chinese languages ​​associated with the Chinese language are also provided to the user. In this way, when the user needs to search for other associated texts, there is no need to enter the search text again, thereby saving the user's search time and improving the user's text search efficiency.

[0023] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0025] Figure 1 A schematic diagram showing an exemplary system architecture to which the technical solution of the embodiments of the present application can be applied;

[0026] Figure 2 A flowchart of a text data processing method according to an embodiment of the present application is shown;

[0027] Figure 3 A detailed flowchart of pushing the recommended text and the text attribute information to the user to be pushed according to an embodiment of the present application is shown;

[0028] Figure 4 A schematic diagram of an interface of an application text data processing method according to an embodiment of the present application is shown;

[0029] Figure 5 A detailed flow chart of selecting at least one associated text as a recommended text from the associated text set according to an embodiment of the present application is shown;

[0030] Figure 6 A detailed flow chart of sorting the associated texts in the associated text set by using a pre-trained sorting model according to an embodiment of the present application is shown;

[0031] Figure 7 A detailed flow chart showing how to determine the recommendation index corresponding to each associated text by using the ranking model according to an embodiment of the present application is shown;

[0032] Figure 8 A schematic diagram of a model of the sorting model according to an embodiment of the present application is shown;

[0033] Fig. 9 A flowchart of a method after pushing the recommended text and the text attribute information to the user to be pushed according to an embodiment of the present application is shown;

[0034] Fig.10 A block diagram of a text data processing device according to an embodiment of the present application is shown;

[0035] Fig.11 A block diagram of a text data processing device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0036] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0037] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the present application.

[0038] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0039] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0040] It should be noted that the "multiple" mentioned in this article refers to two or more. "And / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0041] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the objects used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those shown or described.

[0042] The embodiments in this application involve technologies related to artificial intelligence, that is, fully automated processing of data (such as text data) is achieved through artificial intelligence. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. In other words, artificial intelligence is a comprehensive technology in computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a similar way to human intelligence. Artificial intelligence is to study the design principles and implementation methods of various intelligent machines so that the machines have the functions of perception, reasoning and decision-making.

[0043] Figure 1 A schematic diagram of an exemplary system architecture to which the technical solution of the embodiments of the present application can be applied is shown.

[0044] like Figure 1 As shown, the system architecture may include terminal devices (such as Figure 1 The embodiment of the present invention is a network 104 and a server 105. The network 104 is a medium for providing a communication link between the terminal device and the server 105. The network 104 may include various connection types, such as a wired communication link, a wireless communication link, etc.

[0045] In one embodiment of the present application, when a user needs to query or search for a text, the user can enter the search text on the terminal device. The server 105 obtains the search text entered on the terminal device through the network 104. When it is detected that the text type of the search text belongs to the text type defined in the preset vocabulary, an associated text set including associated texts associated with the search text is obtained. Then, the server 105 selects at least one associated text from the associated text set as a recommended text, and obtains text attribute information of the recommended text. Finally, the server 105 pushes the recommended text and the text attribute information to the user to be pushed through the terminal device.

[0046] In this embodiment, while providing the user with search text related data, by pushing recommended text and text attribute information associated with the search text to the user, the user can be provided with more text data choices, thereby improving the efficiency of the user's text search.

[0047] It should be noted that the text data processing method provided in the embodiment of the present application can be executed by the server 105, and accordingly, the text data processing device is generally arranged in the server 105. However, in other embodiments of the present application, the terminal device can also have similar functions as the server, so as to execute the text data processing solution provided in the embodiment of the present application.

[0048] It should also be noted that Figure 1 The number of terminal devices, networks and servers in the description is only for illustration. According to the implementation requirements, the server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0049] It should be explained that cloud computing as described above is a computing model that distributes computing tasks on a resource pool composed of a large number of computers, so that various application systems can obtain computing power, storage space and information services as needed. The network that provides resources is called a "cloud". The resources in the "cloud" are infinitely expandable in the eyes of users, and can be obtained at any time, used on demand, and expanded at any time. By establishing a cloud computing resource pool (referred to as a cloud platform, generally referred to as an IaaS (Infrastructure as a Service) platform, various types of virtual resources are deployed in the resource pool for external customers to choose to use. The cloud computing resource pool mainly includes: computing devices (virtualized machines, including operating systems), storage devices, and network devices.

[0050] The implementation details of the technical solution of the embodiment of the present application are described in detail below:

[0051] Figure 2 A flowchart of a text data processing method according to an embodiment of the present application is shown. Figure 2 As shown, the text data processing method at least includes steps 210 to 270, which are described in detail as follows:

[0052] Step 210: Obtain the search text input by the user to be pushed.

[0053] In this application, the user to be pushed may refer to a user who needs to query information related to the search text. Here, the search text can be a single character, a word, or a sentence. For example, the user enters a single character "磐" and hopes to query the attribute information related to the character "磐" (such as the pronunciation and definition of the character "磐"). Another example is that the user enters a word "勤奋" and hopes to query the attribute information related to the word "勤奋" (such as the explanation and synonyms of "勤奋"). Another example is that the user enters a proverb "王婆卖瓜,自卖自夸" and hopes to query the attribute information related to the proverb "王婆卖瓜,自卖自夸" (such as the allusion of this proverb). Step 230, when it is detected that the text type of the search text belongs to the text types defined in the preset vocabulary, obtain an associated text set associated with the search text, where the associated text set includes associated texts.

[0054] In this application, the described text data processing method can be applied to the search scenario of Chinese texts. For example, the user needs to search for a certain character, a certain word, a certain idiom, or a certain proverb or poem.

[0055] Specifically, a preset vocabulary can be pre-constructed, which stores Chinese of various text types collected in advance, such as Chinese texts of single-character type, word type, idiom type, proverb or poem type, and so on. It can be understood that the text types defined by the preset vocabulary can include at least one of single-character type, word type, idiom type, proverb type, and poem type.

[0056] Therefore, when it is detected that the text type of the search text belongs to the text types defined in the preset vocabulary, an associated text set associated with the search text is obtained.

[0057] In this application, the associated texts in the associated text set refer to texts that have an associated relationship with the search text. For example, there is a semantically similar associated relationship, or there is an associated relationship in the input time order, or there is a phonetically similar associated relationship.

[0058] Step 250, select at least one associated text from the associated text set as a recommended text, and obtain the text attribute information of the recommended text.

[0059] It can be understood that the number of associated texts in the associated text set is multiple. Therefore, a part of the associated texts in the associated text set can be selected as recommended texts, or all the associated texts can be used as recommended texts.

[0060] In this application, the type of the text attribute information of the recommended text may be the same as that of the text attribute information of the search text input by the user. For example, if the text attribute information of the search text "磐" input by the user is the pronunciation and definition of "磐", the text attribute information of the recommended text is also the pronunciation and definition.

[0061] In this application, the type of the text attribute information of the search text can be determined by recognizing the user's search intention. For example, when the user inputs the search text "磐", the user's search intention of querying the pronunciation and definition of "磐" can be recognized according to the user's historical search habits.

[0062] Step 270: Push the recommended text and the text attribute information to the user to be pushed.

[0063] In step 270, pushing the recommended text and the text attribute information to the user to be pushed can be executed according to the Figure 3 shown steps.

[0064] See Figure 3 , which shows the detailed flowchart of pushing the recommended text and the text attribute information to the user to be pushed according to an embodiment of the present application. Specifically, it includes step 271 to step 272:

[0065] Step 271: Obtain the recommended reference value for each recommended text, and the recommended reference value is used to characterize the popularity of the recommended text being searched.

[0066] Step 272: Display the recommended text, the recommended reference value corresponding to the recommended text, and the text attribute information on the interface.

[0067] To enable those skilled in the art to better understand the present application, the following will be combined with Figure 4 , and a specific application scenario will be used for illustration.

[0068] See Figure 4 , which shows the interface schematic diagram of the application text data processing method according to an embodiment of the present application.

[0069] As Figure 4 shown, after the user inputs a search text 402 "长" in the input box of the interface 401, in addition to displaying the text attribute information related to the text "长", the associated texts 403 "常", "尝", "厂", and "场" of the character "长" are also displayed in the interface 401, and the text attribute information corresponding to the characters "常", "尝", "厂", and "场" is displayed. Further, the recommended reference values corresponding to each associated text are also displayed. For example, the recommended reference values corresponding to the characters "常" and "尝" are 92%, and the recommended reference values corresponding to the characters "厂" and "场" are 88%.

[0070] In the present application, while information related to the user's search text is displayed in the interface, recommended text and corresponding recommended reference values ​​and text attribute information are also displayed. This can provide users with a reference for searching other text information, so that when the user needs to search for other related texts, there is no need to enter the search text again, thereby saving the user's search time and improving the user's text search efficiency.

[0071] The above-mentioned text data processing scheme will be specifically described below through further embodiments.

[0072] In such Figure 2 In one embodiment of step 210, obtaining a set of associated texts associated with the search text may be performed according to the following steps:

[0073] Step 1: Identify the target text type of the search text.

[0074] Step 2: based on the target text type, obtain the associated text associated with the search text to obtain the associated text set.

[0075] In this embodiment, the text type may include a single word type, a phrase type, an idiom type, a sentence type, and the like.

[0076] In this embodiment, based on the target text type, the associated text associated with the search text is obtained. For example, if the search text type is a word type, the text type of the associated text is also a word type. By making the type of the associated text consistent with the type of the search text, it can be ensured that the associated text has a high degree of relevance with the search text, thereby improving the user's text search efficiency in the subsequent process.

[0077] In such Figure 2 In another embodiment of step 210, obtaining a set of associated texts associated with the search text may be performed according to the following steps:

[0078] Step 1: obtaining text input by the user to be pushed within a predetermined historical time interval as the associated text to be selected.

[0079] Step 2: determine the associated text from the candidate associated texts to obtain the associated text set.

[0080] In this embodiment, the predetermined historical time interval can be within the past minute or within the past thirty seconds. For example, the user enters the character "长" in the input box of the interface search engine to query the pronunciation and definition of "长". And within the past minute, the user has entered "重", "常", "行有几个读音", and "涨是什么意思". Then, these entered texts can be obtained from the operation log record, and the key texts can be determined from the entered texts. For example, for the input text "行有几个读音", the keyword "行" can be obtained as the associated text. Also, for the input text "涨是什么意思", the keyword "涨" can be obtained as the associated text. Finally, "重", "常", "行", and "涨" are used as the associated texts of "长", and an associated text set composed of "重", "常", "中", and "涨" is obtained.

[0081] It should be noted that the predetermined historical time interval can be determined according to actual needs and is not limited to those listed above.

[0082] In this embodiment, since the texts entered by the user before and after within a certain time period generally have a certain correlation, using the texts entered by the user within the predetermined historical time interval as the candidate associated texts can ensure a high degree of relevance between the associated texts and the search text, and thus can improve the user's text search efficiency in the subsequent process.

[0083] In Figure 2 In another embodiment of step 210 as shown, to obtain an associated text set associated with the search text, the following steps can be executed:

[0084] Step 1, obtain texts that are similar to the search text in the first text feature as the candidate associated texts, where the first text feature includes at least one of semantic feature, writing form feature, and pronunciation feature.

[0085] Step 2, determine the associated texts from the candidate associated texts to obtain the associated text set.

[0086] In this embodiment, the candidate associated texts can be determined by determining the similarity between the text and the search text in the first text feature, that is, texts with a similarity exceeding the similarity threshold can be used as the candidate associated texts. For example, if the similarity threshold is set to 70%, and the similarity between the text "常" and the search text "长" in the pronunciation feature is 100%, then the text "常" can be determined as the candidate associated text of the search text "长". If the similarity between the text "高" and the search text "长" in the semantic feature is 71%, then the text "高" can also be determined as the candidate associated text of the search text "长".

[0087] In this embodiment, a text library can be established for each text, and the text library includes one or more texts similar to the corresponding text, for example, synonymous texts and antonymous texts of the search text, etc. When it is detected that a user has entered a text, the text in the similar text library corresponding to the text can be used as the candidate associated text for the text.

[0088] In this embodiment, by selecting texts similar to the search text in the first text feature as candidate associated texts, it is also possible to ensure that the associated texts have a high degree of association with the search text, thereby improving the user's text search efficiency in the subsequent process.

[0089] In such Figure 2 In another embodiment of step 210, obtaining a set of associated texts associated with the search text may be performed according to the following steps:

[0090] Step 1, obtaining the text entered by the user to be pushed within a predetermined historical time period, and the text similar to the search text in a second text feature, as the selected associated text, wherein the second text feature includes at least one of a semantic feature, a writing form feature, and a pronunciation feature.

[0091] Step 2: determine the associated text from the candidate associated texts to obtain the associated text set.

[0092] In this embodiment, the texts entered by the user to be pushed within a predetermined historical time period, as well as the texts similar to the search text in the second text feature, are simultaneously selected as associated texts to be selected, thereby expanding the number of associated texts and further improving the user's text search efficiency in the subsequent text data processing process.

[0093] In step 2 of the above embodiments, the associated text is determined from the candidate associated texts, and repeated texts in the candidate associated texts are first determined, and the repeated texts are filtered to obtain the associated texts.

[0094] In this embodiment, since there may be multiple repeated texts in the selected associated texts, directly processing the selected associated texts will increase the text processing load. Therefore, filtering the repeated texts in the selected associated texts can reduce the text processing load and save computer resources.

[0095] In such Figure 2 In one embodiment of step 250, at least one associated text is selected from the associated text set as a recommended text, and the recommended text may be selected as follows: Figure 5 Follow the steps shown.

[0096] See also Figure 5, shows a detailed flow chart of selecting at least one associated text from the associated text set as a recommended text according to an embodiment of the present application. Specifically, it includes steps 251 and 252:

[0097] Step 251 , sorting each associated text in the associated text set by using a pre-trained sorting model to obtain reference sorting information corresponding to each associated text, wherein the reference sorting information is used to characterize the degree of association of the associated text.

[0098] Step 252: Select at least one associated text from the associated text set as a recommended text based on the reference ranking information.

[0099] In the present application, the pre-trained ranking model has the ability to determine the degree of association between the associated text and the search text, wherein the degree of association of the associated text can be characterized by referring to the ranking information.

[0100] In the present application, reference sorting information may include a sorting number, for example, a text with a higher sorting number indicates a higher degree of association. In addition, the sorting information may also include a sorting weight, wherein a higher sorting weight of a text indicates a higher degree of association of the text.

[0101] In the present application, at least one associated text is selected from the associated text set as a recommended text based on the reference sorting information, and at least one associated text with a high degree of association is selected from the associated text set as a recommended text based on the reference sorting information.

[0102] In the present application, the texts may be sorted according to the click-through rates and search volumes of each search text or recommended text in history, the statistical features, text features, and keyword features of each text may be mined, and a sorting model may be trained based on the statistical features, text features, and keyword features of each text, and the trained sorting model may be finally applied to sort each related text;

[0103] In such Figure 5 In one embodiment of step 251, the associated texts in the associated text set are sorted by a pre-trained sorting model, as follows: Figure 6 Follow the steps shown.

[0104] See also Figure 6 , shows a detailed flow chart of sorting the associated texts in the associated text set by using a pre-trained sorting model according to an embodiment of the present application. Specifically, it includes steps 2511 and 2513:

[0105] Step 2511, obtaining third text features corresponding to the search text and each associated text, wherein the third text features include at least one of semantic features, writing form features, and pronunciation features.

[0106] Step 2512: Based on the third text features corresponding to the search text and each associated text, determine the recommendation index corresponding to each associated text through the ranking model.

[0107] Step 2513: sort the associated texts in the associated text set according to the recommendation index.

[0108] In the present application, the third text feature may include at least one of a semantic feature, a writing form feature, and a pronunciation feature, and the text semantic feature, writing form feature, and pronunciation feature can be specifically represented by a feature vector. Different text features of different texts correspond to different feature vectors. For example, the semantic feature of the text "long" can be represented by the feature vector "[24, 12, 34, 31, 10, 31, 45, 30, 10, 12, 24, 5, 9, 24, 12, 24]" and the pronunciation feature of the text "long" can be represented by the feature vector "[41, 2, 5, 71, 1, 47, 14, 7, 8, 24, 77, 31, 4, 7, 55, 23]".

[0109] In the present application, the recommendation index determined by the ranking model may indicate the degree to which the associated text is associated.

[0110] In the present application, the sorting model may be obtained by training based on a single model, for example, may be obtained by training based on a machine learning model.

[0111] In the present application, the ranking model may also be obtained by training based on a combination model of multiple single models. For example, the ranking model disclosed in the present application may include a first sub-model and a second sub-model, wherein the first sub-model has the ability to memorize the third text feature corresponding to the search text, and the second sub-model has the ability to generalize the third text feature corresponding to the search text.

[0112] Based on this, Figure 6 In one embodiment of step 2512, based on the third text features corresponding to the search text and each associated text, the recommendation index corresponding to each associated text is determined by the ranking model, which can be as follows: Figure 7 Follow the steps shown.

[0113] See also Figure 7, shows a detailed flow chart of determining the recommendation index corresponding to each associated text by using the ranking model according to an embodiment of the present application. Specifically, it includes steps 25121 and 25123:

[0114] Step 25121: input the third text features corresponding to the search text and each associated text into the first sub-model to output a first recommendation index corresponding to each associated text.

[0115] Step 25122: Input the third text features corresponding to the search text and each associated text into the second sub-model to output a second recommendation index corresponding to each associated text.

[0116] Step 25123: For each associated text, weighted calculation is performed on the first sub-recommendation index and the second recommendation index to obtain the recommendation index corresponding to the associated text.

[0117] In order to enable those skilled in the art to better understand the sorting model, Figure 8 The proposed sorting model in this embodiment is briefly described as follows:

[0118] See also Figure 8 , shows a model schematic diagram of the sorting model according to an embodiment of the present application.

[0119] like Figure 8 As shown, the ranking model may include an input layer, a hidden layer, and an output layer. In the input layer, the third text feature 801 of the search text "X" and the third text feature 804 of the associated text "a, b, c, d, e" are input to the hidden layers of the first sub-model 802 and the second sub-model 805. The hidden layers of the first sub-model 802 and the second sub-model 805 respectively learn the third text features corresponding to the search text and each associated text, and output the first recommendation index 803 and the second recommendation index 806 corresponding to the associated text "a, b, c, d, e" in the output layer. Finally, the first recommendation index 803 and the second recommendation index 806 are weighted to obtain the recommendation index 807 corresponding to the associated text "a, b, c, d, e".

[0120] In this embodiment, the first sub-model may include a bilinear transformation model.

[0121] In this embodiment, the second sub-model may include a neural network model. Specifically, the neural network model may include multiple hidden layers, each layer is fully connected, and the activation function is a Relu activation function.

[0122] In this application, a ranking model is composed of a first sub-model with memory ability and a second sub-model with generalization ability, which enables the ranking model to have both memory ability and generalization ability, so that the output result of the ranking model has both accuracy and scalability, thereby further improving the user's text search efficiency in the subsequent text data processing process.

[0123] In this application, after pushing the recommended text and the text attribute information to the user to be pushed, the following steps can also be executed as Fig. 9 shown.

[0124] See Fig. 9 , which shows a flowchart of a method after pushing the recommended text and the text attribute information to the user to be pushed according to an embodiment of the present application. Specifically, it includes steps 281 and 282:

[0125] Step 281, obtain the behavior data generated by the user to be pushed based on the pushed recommended text and the text attribute information.

[0126] Step 282, re-train the ranking model based on the behavior data to obtain a re-trained ranking model, and the re-trained ranking model is used for the next associated text ranking.

[0127] In an embodiment of step 282 as Fig. 9 shown, to re-train the ranking model based on the behavior data, the following steps can be executed:

[0128] Step 1, based on the behavior data, determine the expected ranking information corresponding to each associated text.

[0129] Step 2, through the comparison between the expected ranking information and the reference ranking information, correct the hidden layer parameters in the ranking model by gradient backpropagation.

[0130] In this application, the behavior data may refer to the data generated by the user's click behavior on the recommended text and the text attribute information, and the expected ranking information is determined based on the user's behavior data. For example, please refer to Figure 4 , the user clicks on the character "field" and the text attribute information corresponding to the character "field", so it can be determined that the text attribute information corresponding to the character "field" has a higher expected ranking (for example, ranked first among all recommended texts).

[0131] In this application, please continue to refer to Figure 4, for example, the expected sorting information corresponding to each associated text determined based on the behavior data should be the sorting order of "场", "常", "尝", and "厂", that is, the character "场" and the text attribute information corresponding to the character "场" should be recommended first, while the reference sorting information determined by the sorting model is the sorting order of "常", "尝", "厂", and "场". It can be seen that there is a difference between the expected sorting information and the reference sorting information.

[0132] Based on this, based on this difference, the hidden layer parameters in the sorting model can be corrected through gradient backpropagation to obtain a retrained sorting model.

[0133] In this application, after the sorting model is applied to sort each associated text, it is trained with the generated behavior data, so that the parameters in the sorting model are continuously optimized and updated. As a result, when sorting each associated text next time, more appropriate sorting information can be determined through the sorting model. Therefore, the embodiments of this application can, to a certain extent, improve the accuracy of sorting each associated text, thereby improving the text search efficiency of users.

[0134] The technical solution provided by this application, when detecting that the text type of the input search text of the user belongs to the text type defined in the preset word list, obtains a set of associated texts related to the search text, can select at least one associated text from the set of associated texts as a recommended text, and obtains the text attribute information of the recommended text, so as to push the recommended text and the text attribute information to the user to be pushed. Since recommended texts associated with the search text and corresponding text attribute information are pushed to the user, it can provide a reference for the user to search for other associated texts. When the user has a need to search for other associated texts, there is no need to input the search text again, thereby saving the user's search time and improving the user's text search efficiency.

[0135] The following introduces the device embodiments of this application, which can be used to execute the text data processing method in the above embodiments of this application. For the details not disclosed in the device embodiments of this application, please refer to the embodiments of the text data processing method above of this application.

[0136] Fig.10 The block diagram of a text data processing device according to an embodiment of this application is shown.

[0137] Referring to Fig.10 As shown, a text data processing device 1000 according to an embodiment of this application includes: a first acquisition unit 1001, a second acquisition unit 1002, a selection unit 1003, and a push unit 1004.

[0138] Among them, the first acquisition unit 1001 is used to acquire the search text input by the user to be pushed; the second acquisition unit 1002 is used to acquire a set of associated texts associated with the search text when it is detected that the text type of the search text belongs to the text type defined in the preset vocabulary, and the associated text set includes associated texts; the selection unit 1003 is used to select at least one associated text from the associated text set as a recommended text, and acquire text attribute information of the recommended text; the push unit 1004 is used to push the recommended text and the text attribute information to the user to be pushed.

[0139] In some embodiments of the present application, based on the aforementioned scheme, the push unit 1004 is configured to: obtain a recommendation reference value for each recommended text, wherein the recommendation reference value is used to characterize the popularity of the recommended text being searched; and display the recommended text, as well as the recommendation reference value and text attribute information corresponding to the recommended text in the interface.

[0140] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit 1002 is configured to: identify the target text type of the search text; based on the target text type, acquire associated text associated with the search text to obtain the associated text set.

[0141] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit 1002 is further configured to: acquire the text entered by the user to be pushed within a predetermined historical time interval as the associated text to be selected; determine the associated text from the associated text to be selected to obtain the associated text set.

[0142] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit 1002 is also configured to: acquire text that is similar to the search text in a first text feature as a candidate associated text, wherein the first text feature includes at least one of a semantic feature, a writing form feature, and a pronunciation feature; determine the associated text from the candidate associated text to obtain the associated text set.

[0143] In some embodiments of the present application, based on the aforementioned scheme, the second acquisition unit 1002 is also configured to: obtain the text entered by the user to be pushed within a predetermined historical time interval, and the text that is similar to the search text in the second text feature, as the selected associated text, the second text feature including at least one of a semantic feature, a writing form feature, and a pronunciation feature; determine the associated text from the selected associated text to obtain the associated text set.

[0144] In some embodiments of the present application, based on the above-mentioned solution, the second acquisition unit 1002 includes: a duplicate filtering unit, which is used to determine repeated texts in the candidate associated texts and filter the repeated texts to obtain the associated texts.

[0145] In some embodiments of the present application, based on the aforementioned scheme, the selection unit 1003 is configured to: sort each associated text in the associated text set through a pre-trained sorting model to obtain reference sorting information corresponding to each associated text, and the reference sorting information is used to characterize the degree of association of the associated text; based on the reference sorting information, select at least one associated text from the associated text set as a recommended text.

[0146] In some embodiments of the present application, based on the aforementioned scheme, the selection unit 1003 is further configured to: obtain a third text feature corresponding to the search text and each associated text, the third text feature including at least one of a semantic feature, a writing form feature, and a pronunciation feature; based on the third text feature corresponding to the search text and each associated text, determine a recommendation index corresponding to each associated text through the sorting model; and sort each associated text in the associated text set according to the recommendation index.

[0147] In some embodiments of the present application, based on the aforementioned scheme, the sorting model includes a first sub-model and a second sub-model, the first sub-model has the ability to memorize the third text feature corresponding to the search text, and the second sub-model has the ability to generalize the third text feature corresponding to the search text, and the selection unit 1003 is also configured to: input the third text feature corresponding to the search text and each associated text into the first sub-model to output a first recommendation index corresponding to each associated text; input the third text feature corresponding to the search text and each associated text into the second sub-model to output a second recommendation index corresponding to each associated text; for each associated text, perform weighted calculation on the first sub-recommendation index and the second recommendation index to obtain the recommendation index corresponding to the associated text.

[0148] In some embodiments of the present application, based on the aforementioned scheme, the first sub-model includes a bilinear transformation model, and the second sub-model is a neural network model.

[0149] In some embodiments of the present application, based on the aforementioned scheme, the device also includes: a training unit, which is used to obtain the behavior data generated by the user to be pushed based on the recommended text and the text attribute information pushed to the user to be pushed; retrain the sorting model based on the behavior data to obtain a retrained sorting model, and the retrained sorting model is used for the next associated text sorting.

[0150] In some embodiments of the present application, based on the aforementioned scheme, the training unit is configured to: determine the expected sorting information corresponding to each associated text based on the behavioral data; and correct the hidden layer parameters in the sorting model by comparing the expected sorting information with the reference sorting information through gradient back transfer.

[0151] As another aspect, an embodiment of the present application also provides another text data processing device, including a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include instructions for performing the text data processing method as described in the above embodiments.

[0152] Fig.11 A block diagram of a text data processing device according to an embodiment of the present application is shown. For example, the device 1100 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0153] Reference Fig.11 , the device 1100 may include one or more of the following components: a processing component 1102 , a memory 1104 , a power component 1106 , a multimedia component 1108 , an audio component 1110 , an input / output (I / O) interface 1112 , a sensor component 1114 , and a communication component 1116 .

[0154] The processing component 1102 generally controls the overall operation of the device 1100, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 1102 may include one or more processors 1120 to execute instructions to perform all or part of the steps of the above-described method. In addition, the processing component 1102 may include one or more modules to facilitate interaction between the processing component 1102 and other components. For example, the processing component 1102 may include a multimedia module to facilitate interaction between the multimedia component 1108 and the processing component 1102.

[0155] The memory 1104 is configured to store various types of data to support operations on the device 1100. Examples of such data include instructions for any application or method operating on the device 1100, contact data, phone book data, messages, pictures, videos, etc. The memory 1104 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0156] The power supply component 1106 provides power to the various components of the device 1100. The power supply component 1106 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device 1100.

[0157] The multimedia component 1108 includes a screen that provides an output interface between the device 1100 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 1108 includes a front camera and / or a rear camera. When the device 1100 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.

[0158] The audio component 1110 is configured to output and / or input audio signals. For example, the audio component 1110 includes a microphone (MIC), and when the device 1100 is in an operation mode, such as a call mode, a recording mode, and a voice information processing mode, the microphone is configured to receive an external audio signal. The received audio signal can be further stored in the memory 1104 or sent via the communication component 1116. In some embodiments, the audio component 1110 also includes a speaker for outputting audio signals.

[0159] I / O interface 1112 provides an interface between processing component 1102 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0160] The sensor assembly 1114 includes one or more sensors for providing various aspects of the status assessment of the device 1100. For example, the sensor assembly 1114 can detect the open / closed state of the device 1100, the relative positioning of components, such as the display and keypad of the device 1100, the sensor assembly 1114 can also search for changes in the position of the search results display device 1100 or a component of the device 1100, the presence or absence of user contact with the device 1100, the orientation or acceleration / deceleration of the device 1100, and the temperature change of the device 1100. The sensor assembly 1114 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly 1114 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 1114 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0161] The communication component 1116 is configured to facilitate wired or wireless communication between the device 1100 and other devices. The device 1100 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 1116 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 1116 also includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency information processing (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

[0162] In an exemplary embodiment, the device 1100 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above methods.

[0163] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 1104 including instructions, and the instructions can be executed by the processor 1120 of the device 1100 to complete the above-mentioned text data processing method. For example, the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0164] As another aspect, the present application also provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the text data processing method described in the above embodiment.

[0165] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the device described in the above embodiment; or may exist independently without being assembled into the device. The computer-readable storage medium stores at least one program code, which is loaded and executed by the processor of the device to implement the operations performed by the text data processing method described in the above embodiment.

[0166] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0167] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the implementation methods of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the implementation methods of the present application.

[0168] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.

[0169] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A text data processing method, characterized in that: The method comprises: Get the search text entered by the user to be pushed; When it is detected that the text type of the search text belongs to the text type defined in the preset vocabulary, obtaining an associated text set associated with the search text, wherein the associated text set includes associated texts; Acquire a third text feature corresponding to the search text and each associated text, wherein the third text feature includes at least one of a semantic feature, a writing form feature, and a pronunciation feature; determine a recommendation index corresponding to each associated text by a pre-trained ranking model based on the third text feature corresponding to the search text and each associated text, wherein the ranking model includes a first sub-model and a second sub-model, wherein the first sub-model has the ability to memorize the third text feature corresponding to the search text, and the second sub-model has the ability to generalize the third text feature corresponding to the search text; Sort the associated texts in the associated text set by the recommendation index to obtain reference sorting information corresponding to each associated text, wherein the reference sorting information is used to characterize the degree of association of the associated texts; based on the reference sorting information, select at least one associated text from the associated text set as a recommended text, and obtain text attribute information of the recommended text; The recommended text and the text attribute information are pushed to the user to be pushed.

2. The method according to claim 1, characterized in that The step of pushing the recommended text and the text attribute information to the user to be pushed includes: Obtaining a recommendation reference value for each recommended text, wherein the recommendation reference value is used to represent the popularity of the recommended text in searches; The recommended text, as well as the recommended reference value and text attribute information corresponding to the recommended text are displayed in the interface.

3. The method according to claim 1, characterized in that The step of acquiring a set of associated texts associated with the search text comprises: Identifying a target text type of the search text; Based on the target text type, associated texts associated with the search text are acquired to obtain the associated text set.

4. The method according to claim 1, characterized in that: The step of acquiring a set of associated texts associated with the search text comprises: Acquire the to-be-selected associated texts, and determine the associated texts from the to-be-selected associated texts to obtain the associated text set; The obtaining of the associated text to be selected includes any of the following: Acquire the text entered by the user to be pushed within a predetermined historical time interval as the associated text to be selected; Acquire text similar to the search text in a first text feature as a candidate associated text, wherein the first text feature includes at least one of a semantic feature, a writing form feature, and a pronunciation feature; The text input by the user to be pushed within a predetermined historical time interval and the text similar to the search text in a second text feature are obtained as the associated text to be selected, wherein the second text feature includes at least one of a semantic feature, a writing form feature, and a pronunciation feature.

5. The method according to claim 1, characterized in that The method of determining the recommendation index corresponding to each associated text by using a pre-trained ranking model based on the third text features corresponding to the search text and each associated text includes: Inputting the third text features corresponding to the search text and each associated text into the first sub-model to output a first recommendation index corresponding to each associated text; Inputting the third text features corresponding to the search text and each associated text into the second sub-model to output a second recommendation index corresponding to each associated text; For each associated text, the first recommendation index and the second recommendation index are weightedly calculated to obtain a recommendation index corresponding to the associated text.

6. The method according to any one of claims 1 to 5, characterized in that: After pushing the recommended text and the text attribute information to the user to be pushed, the method further includes: Acquire behavior data generated by the user to be pushed according to the recommended text and the text attribute information; The sorting model is retrained based on the behavior data to obtain a retrained sorting model, and the retrained sorting model is used for the next associated text sorting.

7. The method according to claim 6, characterized in that The retraining of the sorting model based on the behavior data comprises: Based on the behavior data, determining the expected sorting information corresponding to each associated text; By comparing the expected sorting information with the reference sorting information, the hidden layer parameters in the sorting model are corrected through gradient back transfer.

8. A text data processing device, characterized in that: The device comprises: A first acquisition unit is used to acquire a search text input by a user to be pushed; A second acquisition unit is used to acquire an associated text set associated with the search text when it is detected that the text type of the search text belongs to the text type defined in the preset vocabulary, wherein the associated text set includes associated texts; A selection unit is used to obtain a third text feature corresponding to the search text and each associated text, wherein the third text feature includes at least one of a semantic feature, a writing form feature, and a pronunciation feature; based on the third text feature corresponding to the search text and each associated text, a recommendation index corresponding to each associated text is determined by a pre-trained sorting model, wherein the sorting model includes a first sub-model and a second sub-model, wherein the first sub-model has the ability to memorize the third text feature corresponding to the search text, and the second sub-model has the ability to generalize the third text feature corresponding to the search text; sorting each associated text in the associated text set by the recommendation index to obtain reference sorting information corresponding to each associated text, wherein the reference sorting information is used to characterize the degree of association of the associated text; based on the reference sorting information, selecting at least one associated text from the associated text set as a recommended text, and obtaining text attribute information of the recommended text; The push unit is used to push the recommended text and the text attribute information to the user to be pushed.

9. A text data processing device, characterized in that: The invention comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include instructions for performing the text data processing method as claimed in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the operations performed by the text data processing method according to any one of claims 1 to 7.

Images