Text input method, device, system, storage medium and program product

Abstract

Embodiments of the present application provide a text input method, device, system, storage medium and program product. In the embodiments of the present application, a prediction model trained by using text corpus of a vertical field as a sample is used to determine a candidate text corresponding to a currently input character key code and an accuracy of the candidate text according to the currently input character key code and a context thereof; further, the candidate text can be displayed according to the accuracy of the candidate text. Since the prediction model is trained by using the text corpus of the vertical field as the sample, the accuracy of a candidate item belonging to a vocabulary of the vertical field in the candidate text corresponding to the input character key code determined by using the prediction model is high, therefore, the candidate text is displayed according to the accuracy of the candidate text, the candidate text adapted to the vertical field can be preferentially recommended to a user, the time for the user to browse the candidate text to select a desired text can be reduced, and thus the text input efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a text input method, device, system, storage medium, and program product. Background Technology

[0002] Input methods, as the most basic tool for human-computer interaction, play a vital role in the computer field. General input methods typically use words and phrases as input units. Based on word frequency statistics from a large amount of general text, general input methods provide candidate words to the user based on the characters entered by the user on the keyboard, and the user then selects the correct candidate word to complete the input.

[0003] However, in vertical industries, this approach based on general input methods has significant limitations. Specialized terms in vertical industries often overlap with general vocabulary. Using a general text input method, specialized terms from vertical industries receive lower priority for recommendations, forcing users to constantly browse candidate words and determine the correct option, wasting considerable time. Therefore, improving text input efficiency in vertical industries is a pressing technical problem for computer professionals. Summary of the Invention

[0004] This application provides a text input method, device, system, storage medium, and computer program product to improve text input efficiency in vertical industries.

[0005] This application provides a data processing system, including: a first terminal and a second terminal; the first terminal and the second terminal are communicatively connected;

[0006] The first terminal is configured to acquire the currently input character key and the context of the currently input character key; input the currently input character key and the context into a prediction model; the prediction model is trained using text corpus from a vertical domain as samples; in the prediction model, based on the currently input character key and the context, determine at least one candidate text corresponding to the currently input character key and the accuracy of the at least one candidate text; display the at least one candidate text according to the accuracy of the at least one candidate text; and, in response to a selection event for the at least one candidate text, provide the target candidate text selected by the selection event to the second terminal.

[0007] The second terminal is used to input the target candidate text into a designated area.

[0008] This application embodiment also provides a text input method adapted to a first terminal, including:

[0009] Obtain the currently input character key code and the context of the currently input character key code;

[0010] The currently input character key and the context are input into the prediction model; the prediction model is trained using text corpora from the vertical domain as samples.

[0011] In the prediction model, based on the currently input character key and the context, at least one candidate text corresponding to the currently input character key and the accuracy of the at least one candidate text are determined.

[0012] The at least one candidate text is displayed according to its accuracy.

[0013] This application also provides a text input method, including:

[0014] Obtain the currently input character key code and the context of the currently input character key code;

[0015] Speech recognition is performed on the sound data currently picked up by the microphone to determine the text information reflected by the sound data;

[0016] The currently input character key, the context, and the text information are input into the prediction model; the prediction model is trained using text corpora from the vertical domain and the corresponding audio corpora as samples.

[0017] In the prediction model, based on the currently input character key code, the context, and the text information, at least one candidate text corresponding to the currently input character key code and the accuracy of the at least one candidate text are determined.

[0018] The at least one candidate text is displayed according to its accuracy.

[0019] This application also provides a computer device, including: a memory, a processor, and a display component;

[0020] The memory is used to store computer programs;

[0021] The processor is coupled to the memory and the display component to execute the computer program for performing the steps in the above-described text input methods.

[0022] This application also provides a computer-readable storage medium storing computer instructions, which, when executed by one or more processors, cause one or more processors to perform the steps in the above-described text input methods.

[0023] This application also provides a computer program product, including a computer program; when the computer program is executed by one or more processors, it can implement the steps in the above-described text input methods.

[0024] In this embodiment, the currently input character key and its context can be obtained; the currently input character key and its context are then input into a prediction model trained using text corpus from a vertical domain as samples. In this prediction model, candidate texts corresponding to the currently input character key and their accuracy can be determined based on the currently input character key and its context. Furthermore, candidate texts can be displayed according to their accuracy. On the one hand, since the prediction model is trained using text corpus from a vertical domain as samples, the candidate texts corresponding to the input character key determined by this prediction model have higher accuracy for words belonging to the vertical domain, and a greater probability that they are the text the user expects. Therefore, displaying candidate texts according to their accuracy prioritizes recommending candidate texts that are suitable for the vertical domain to the user, reducing the time the user spends browsing candidate texts to select the desired text, thereby improving text input efficiency.

[0025] On the other hand, when performing text prediction on the currently input character key code, this embodiment takes into account the contextual information of the currently input character key code, which helps to more accurately identify the user's input intent and improve the accuracy of the output candidate text. Attached Figure Description

[0026] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0027] Figure 1 This application provides a schematic diagram of the structure of a data processing system and a schematic diagram of the text input process.

[0028] Figure 2 This is another schematic diagram of a text input process provided in an embodiment of this application;

[0029] Figure 3 A schematic diagram illustrating the process of selecting a template using keyboard shortcuts, provided in an embodiment of this application.

[0030] Figure 4 and Figure 5 A flowchart illustrating the text input method provided in an embodiment of this application;

[0031] Figure 6 A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0032] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0033] There are duplicate codes between specialized terms in vertical industries and general vocabulary. When using a general text input method, specialized terms in vertical industries are recommended with lower priority, causing users to constantly browse candidate words to determine the correct option, which wastes a lot of time.

[0034] To address the aforementioned technical problems, in some embodiments of this application, the currently input character key code and its context can be obtained; and the currently input character key code and its context can be input into a prediction model trained using text corpora from a vertical domain as samples; in this prediction model, candidate texts corresponding to the currently input character key code and their accuracy rates can be determined based on the currently input character key code and its context; furthermore, candidate texts can be displayed according to their accuracy rates. On the one hand, since the prediction model is trained using text corpora from a vertical domain as samples, the candidate texts corresponding to the input character key codes determined by this prediction model have higher accuracy rates for candidates belonging to words from the vertical domain, and a greater probability that they are the text the user expects. Therefore, displaying candidate texts according to their accuracy rates can prioritize recommending candidate texts that are suitable for the vertical domain to the user, reducing the time the user spends browsing candidate texts to select the desired text, thereby helping to improve text input efficiency.

[0035] On the other hand, when performing text prediction on the currently input character key code, this embodiment takes into account the contextual information of the currently input character key code, which helps to more accurately identify the user's input intent and improve the accuracy of the output candidate text.

[0036] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0037] It should be noted that the same reference numerals denote the same object in the following figures and embodiments. Therefore, once an object is defined in one figure or embodiment, it does not need to be discussed further in subsequent figures and embodiments.

[0038] Figure 1 This is a schematic diagram of the structure of the data processing system provided in an embodiment of this application. Figure 1 As shown, the data processing system includes a first terminal 10 and a second terminal 20. The first terminal 10 and the second terminal 20 are communicatively connected.

[0039] In this embodiment, the first terminal 10 may be configured with a keyboard 10a. The keyboard 10a may be a physical keyboard or a soft keyboard displayed on the screen of the first terminal 10. Accordingly, the screen of the first terminal 10 is a touchscreen. If the keyboard 10a is implemented as a physical keyboard, it may be an externally connected keyboard to the first terminal 10 or a keyboard integrated into the first terminal 10. For an externally connected keyboard, it may be connected to the first terminal 10 via a USB interface, but is not limited to this. Figure 1 The illustration uses keyboard 10a as an example of an externally connected keyboard, but this is not intended to limit the scope. In this embodiment, the specific implementation of the first terminal 10 is not limited. Optionally, the first terminal 10 can be a personal computer, electronic screen, tablet computer, mobile phone, or wearable device, etc. Of course, in addition to supporting keyboard input, the first terminal 10 can also support voice input, or it can also support writing input, such as handwriting input or stylus writing input.

[0040] The second terminal 20 refers to a computer device used by the user, which has the computing, internet access, and communication functions required by the user. For example, it can be a mobile phone, tablet computer, personal computer, or wearable device. Preferably, the second terminal 20 is a personal computer, such as a user's work computer.

[0041] In this embodiment, the first terminal 10 and the second terminal 20 are communicatively connected. For example, the first terminal 10 and the second terminal 20 can be communicatively connected via a serial data bus. The serial data bus can be a USB data bus or a PCIe data bus, etc. Alternatively, the first terminal 10 and the second terminal 20 can also be communicatively connected via a network. For example, the first terminal 10 and the second terminal 20 can also be communicatively connected via a wired network or a wireless network. Optionally, the first terminal 10 and the second terminal 20 can be communicatively connected via a mobile network. Accordingly, the mobile network standard can be any one of 2G (GSM), 2.5G (GPRS), 3G (WCDMA, TD-SCDMA, CDMA2000, UTMS), 4G (LTE), 4G+ (LTE+), 5G, WiMax, etc. Optionally, the first terminal 10 and the second terminal 20 can also be communicatively connected via Bluetooth, WiFi, or infrared, etc.

[0042] In this embodiment, a vertical domain refers to an application area that provides specific services to a limited group. A vertical domain is a professional field, a product of the increasingly refined division of labor in society. Examples include fields such as healthcare, security, electricity, environmental protection, and education. These fields all involve a significant amount of manual data entry. Especially in the medical field, clinicians spend a considerable amount of time writing medical records.

[0043] However, for existing general input methods, words and phrases are generally used as input units. General input methods can provide candidate words for users based on the character key codes input by users on the keyboard and the frequency statistics of words in a large amount of general text, and then the users can select the correct candidate words to complete the input. Character key codes refer to the characters input by users for indexing the text they expect to input. For example, for the full-pinyin input method, the character key codes are the full pinyin of the text expected to be input by users. For example, the character key codes of "pneumonia" are "fei yan", etc.

[0044] General input methods cannot take into account the particularity of grammar expressions in vertical fields. Especially when there are homophones between professional terms in vertical fields and general vocabulary, using general input methods cannot ensure that the special terms in vertical fields are preferentially recommended to users. General input methods cause users to continuously browse candidate words to determine the correct options, wasting a lot of time and resulting in low text input efficiency. Among them, the existence of homophones between professional terms and general vocabulary means that the character key codes input for professional terms are the same as those for general vocabulary. For example, for the full-pinyin input method, the input character key codes of "blood sugar" and "school" are both "xue tang", etc.

[0045] In order to improve the text input efficiency in vertical fields, in this embodiment, a professional prediction model can be trained using the text corpus in the vertical field as a sample, and this prediction model can be applied to actual text input. Among them, the model training process can be executed by the first terminal 10, or can be trained by other computer devices and then provided to the first terminal 10. The following is an exemplary description of the training process of the prediction model, and the main steps are as follows:

[0046] S1. Obtain the text corpus in the vertical field. For example, for the medical field, historical medical record data can be obtained as the text corpus, etc.

[0047] S2. Use punctuation marks for pauses as delimiters to segment the sentences in the text corpus, and obtain multiple segmented clauses. Among them, punctuation marks for pauses include but are not limited to: commas (,), periods (.), semicolons (;), or exclamation marks (!), etc.

[0048] Among them, using punctuation marks for pauses to segment sentences can make the length of each clause fall within a set range. For example, the length of a clause is usually greater than or equal to 5 Chinese characters and less than or equal to 20 Chinese characters. An appropriate clause length can, on the one hand, reduce the probability of homophones, and on the other hand, enable the user to avoid inputting overly long character key codes when entering text. Homophones refer to different words having the same input character key code. For example, for the full-pinyin input method, the probability of homophones for 2-character words is higher than that for 5-character words. For example, words with the character key code "fei yan" can be "pneumonia", "flying swallow", "flying eaves", "eye strain", etc.; while the word with the character key code "zuo fei ju bu gai hua" is only "local calcification of the left lung". Therefore, an appropriate clause length can reduce the probability of homophones.

[0049] S3. Based on the multiple clauses obtained after segmentation, obtain the character key code corresponding to each clause. Taking Chinese input as an example, for full-pinyin input, the full pinyin of each clause can be obtained as the character key code corresponding to the clause; for simple-pinyin input, the full pinyin of each clause can be obtained and the first letter of the full pinyin of each clause can be intercepted as the character key code corresponding to the clause; for Wubi input, the radicals of each clause can be obtained as the character key code corresponding to the clause.

[0050] S4. With the goal of minimizing the loss function, use the character key codes corresponding to multiple clauses as samples for model training to obtain a prediction model.

[0051] Among them, the loss function is determined according to the difference between the text corresponding to the clause output by each model training and the true text of the clause.

[0052] S5. The model when the loss function meets the set requirements is used as the prediction model.

[0053] Among them, the prediction model can be a recurrent neural network (RNN) model, a convolutional neural network (CNN) model, a deep neural network (DNN) model, etc.

[0054] Furthermore, after the above prediction model training is completed, the first terminal 10 can store the prediction model and use the prediction model trained with the text corpus in the vertical field as a sample during the actual text input process. The specific implementation is as follows:

[0055] The first terminal 10 can acquire the currently input character key code. In some embodiments, the first terminal 10 can support keyboard input. Accordingly, the first terminal 10 can, in response to a trigger operation on the keyboard, acquire the character key code input through the keyboard 10a as the currently input character key code. Taking Chinese input as an example, for full Pinyin input, the currently input character key code can be the full Pinyin of the Chinese text the user expects to input; for abbreviated Pinyin input, the full Pinyin of each clause can be acquired, and the currently input character key code can be the first letter of the Pinyin of the Chinese text the user expects to input, etc.

[0056] In other embodiments, the first terminal 10 may support voice input. Accordingly, the first terminal 10 may acquire the sound data currently picked up by the microphone as the character key code for the current input.

[0057] In some other embodiments, the first terminal 10 may also support handwriting input. Accordingly, the first terminal 10 may also collect the motion trajectory of the target medium on the screen as the currently input character key code. The target medium may be a user's finger or a stylus, etc.

[0058] In this embodiment, to improve the accuracy of the prediction model in determining the text corresponding to the currently input character key code, the context of the currently input character key code can also be obtained. Optionally, the first terminal 10 can obtain the character key code that was input before the currently input character key code as the context of the currently input character key code. Specifically, the character key code that was input before the currently input character key code refers to the character key code that was input the previous time. For example, for Chinese abbreviation input, if the user expects to input the sentence "left lung local calcification, right lung local effusion", and the currently input character key code is the first letters of "right lung local effusion" "yfjbjs", and the previous input character key code is the first letters of "left lung local calcification" "zfjbgh", then "zfjbgh" can be obtained as the context of the currently input character key code.

[0059] In other embodiments, the first terminal 10 may also obtain the target text corresponding to the character key code that was entered before the currently entered character key code, as the context of the currently entered character key code. The target text is the text that the user expects to output when entering the character key code that was entered before the currently entered character key code; it is text that matches the vertical domain. For example, if the user expects to input the statement "local calcification in the left lung, local effusion in the right lung," and the currently entered character key code is the first letters of "local effusion in the right lung" "yfjbjs," the previously selected target text "local calcification in the left lung" can be obtained as the context of the currently entered character key code.

[0060] Furthermore, the first terminal 10 can input the currently input character key and its context into the aforementioned prediction model; in the prediction model, based on the currently input character key and its context, at least one candidate text corresponding to the currently input character key and the accuracy of at least one candidate text can be determined. The accuracy of at least one candidate text can, to some extent, reflect the matching degree between the corresponding candidate text and the vertical domain.

[0061] The output of the prediction model can be at least one candidate text corresponding to the currently input character key code; or it can be at least one candidate text corresponding to the currently input character key code and its context. In this embodiment, no limitation is made.

[0062] Furthermore, the first terminal 10 can display at least one candidate text based on the accuracy of at least one candidate text, allowing the user to browse and select. Since the prediction model is trained using text corpora from a vertical domain, texts matching the character key codes in the prediction model have higher scores and accuracy. Therefore, the first terminal 10 can prioritize displaying candidate texts with higher accuracy, recommending candidate texts that fit the vertical domain to the user, reducing the time the user spends browsing candidate texts to select the desired text, thereby improving text input efficiency.

[0063] On the other hand, in this embodiment, when performing text prediction on the currently input character key code, the contextual information of the currently input character key code is taken into account, which helps the prediction model to more accurately identify the user's input intent, thereby helping to improve the accuracy of the output candidate text.

[0064] Considering that the candidate texts output by the prediction model may contain texts that are significantly mismatched with the vertical domain, this embodiment can also perform post-processing on the candidate texts output by the prediction model. Specifically, before displaying at least one candidate text, the first terminal 10 can compare the at least one candidate text output by the prediction model with a lexicon of the vertical domain to determine the first candidate text that does not match the vertical domain. Optionally, the first terminal 10 can calculate the similarity between each candidate text and the commonly used texts of the vertical domain recorded in the lexicon, and determine the candidate texts whose similarity does not meet the set requirements as the first candidate texts that do not match the vertical domain. That is, the candidate texts whose similarity with any commonly used text recorded in the lexicon does not meet the set requirements are determined as the first candidate texts that do not match the vertical domain. For example, the candidate texts whose similarity with any commonly used text recorded in the lexicon is less than the set first similarity threshold are determined as the first candidate texts that do not match the vertical domain.

[0065] For a first candidate text that does not match the vertical domain, in this embodiment, the similarity between the first candidate text and commonly used texts in the vertical domain recorded in the thesaurus can be calculated; candidate commonly used texts whose similarity to the first candidate text meets a set requirement can be obtained from the thesaurus. For example, M commonly used texts can be obtained from the thesaurus as candidate commonly used texts in descending order of similarity to the first candidate text. Where M≥2 and is an integer. Alternatively, candidate commonly used texts whose similarity to the first candidate text is greater than or equal to a set second similarity threshold can be obtained from the thesaurus. Where the second similarity threshold is less than the aforementioned first similarity threshold.

[0066] Furthermore, the first candidate text can be modified based on the frequency of selection of candidate idiomatic texts. Optionally, the candidate idiomatic text with the highest selection frequency can be selected, and the first candidate text can be modified to be the candidate idiomatic text with the highest selection frequency. This post-prediction processing method can supplement the situations that the prediction model cannot capture, further improve the matching degree between the displayed candidate text and the vertical domain, and improve the accuracy of the candidate text.

[0067] Furthermore, the first terminal 10 can display at least one candidate text according to the accuracy of at least one candidate text corresponding to the currently input character key code. Optionally, the first terminal 10 can display at least one candidate text in descending order of the accuracy of at least one candidate text corresponding to the currently input character key code.

[0068] For the user, at least one candidate text can be browsed to select the text to be entered. For the first terminal 10, in response to a selection event for at least one candidate text, the target candidate text selected by the selection event can be provided to the second terminal 20. The second terminal 20 receives the target candidate text and inputs the target candidate text into a designated area. The designated area refers to the area indicated by the input cursor. Figure 1 The target candidate text is illustrated as "Candidate Text 1", but this does not constitute a limitation.

[0069] In this embodiment, the specific implementation of the user selecting the desired input text is not limited. In some embodiments, for keyboard input, keyboard 10a can be configured with shortcut keys 10a1 corresponding to candidate texts. The user can select the target candidate text using the shortcut keys. For the first terminal 10, key press detection can be performed on the shortcut keys 10a1 corresponding to at least one candidate text; if the shortcut key 10a1 corresponding to at least one candidate text is detected to be triggered, a selection event for at least one candidate text is determined; and the candidate text corresponding to the triggered shortcut key is determined to be the target candidate text.

[0070] If keyboard 10a is implemented as a physical keyboard, users can select target candidate text by touching shortcut keys. If keyboard 10a is implemented as a soft keyboard, users can select target candidate text by touching the screen area corresponding to the shortcut keys.

[0071] In this embodiment, the first terminal 10 can also time the display duration of at least one candidate text; if no selection event for at least one candidate text is detected after the display duration has reached the set duration, the candidate text ranked first can be provided to the second terminal 20. The second terminal 20 can receive the candidate text and input it into the designated area.

[0072] The above description uses the first terminal 10 as an example of keyboard input to illustrate the text input method of the data processing system provided in this application. In some embodiments, the first terminal 10 may support composite text input, including multiple text input methods. For example, in some embodiments, the first terminal 10 may support voice input in addition to keyboard input. The following describes an exemplary implementation of keyboard and voice composite text input on the first terminal 10.

[0073] For complex text input methods, the prediction model can be trained using text corpora from a vertical domain and their corresponding speech corpora. For example, in the medical field, medical record texts can be obtained as the text corpus, and the speech data of communication between doctors and patients during the process of entering medical record texts can be obtained as the speech corpus corresponding to the text corpus. Furthermore, speech recognition can be performed on the speech corpus to obtain the corresponding text information; then, the model is trained using the text corpus and the text information corresponding to the speech corpus as samples to obtain the prediction model.

[0074] In practical use, the first terminal 10 can obtain the character key code currently input on the keyboard and the context of the character key code currently input on the keyboard. For specific implementation details, please refer to the relevant content of the above embodiments, which will not be repeated here.

[0075] Accordingly, such as Figure 2As shown, the first terminal 10 also performs speech recognition on the sound data currently picked up by the microphone to determine the text information reflected in the sound data. Further, the first terminal 10 can input the currently input character keycode, context, and text information reflected in the sound data into a prediction model. In the prediction model, based on the currently input character keycode, context, and text information reflected in the sound data, at least one candidate text corresponding to the currently input character keycode and the accuracy of at least one candidate text can be determined. In this embodiment, when using the prediction model to calculate the text corresponding to the character keycode, combining the sound data with the currently input character keycode provides multi-dimensional reference information for determining the currently input character keycode, which helps improve the accuracy of the determined candidate text. On the other hand, in this embodiment, when performing text prediction on the currently input character keycode, the context information of the currently input character keycode is taken into account, which helps the prediction model to more accurately identify the user's input intent, thereby helping to improve the accuracy of the output candidate text.

[0076] Furthermore, the first terminal 10 can display at least one candidate text based on its accuracy. Since the prediction model is trained using text corpora from a vertical domain and corresponding speech corpora, candidate texts matching the vertical domain in the model's output have higher scores and accuracy. Therefore, the first terminal 10 can prioritize displaying candidate texts with higher accuracy, recommending those suitable for the vertical domain to the user. This reduces the time users spend browsing candidate texts and selecting their desired text, thereby improving text input efficiency.

[0077] For details regarding the specific implementation of the first terminal 10 displaying at least one candidate text and the user selecting the desired target candidate text, please refer to the relevant content of the above embodiments, which will not be repeated here. It is worth noting that, in this embodiment, in addition to selecting the desired candidate text via keyboard shortcuts, a text input template used in a vertical field, such as a medical record template, can also be selected via shortcuts. In this embodiment, for ease of description and distinction, the shortcut key for selecting the desired candidate text can be defined as a first type of shortcut key; and the shortcut key for selecting a text input template can be defined as a second type of shortcut key. Figure 3As shown, the keyboard can be configured with a second type of shortcut key 10a2. Users can select the corresponding text input template by triggering the second type of shortcut key 10a2. For the first terminal 10, in response to the triggering operation of the second type of shortcut key 10a2 corresponding to the text template used in the vertical domain, the text input template corresponding to the triggered second type of shortcut key 10a2 can be retrieved; and this text input template can be provided to the second terminal 20. The second terminal 20 can display the text input template. For users, text input can be performed based on this text input template, further improving text input efficiency. For example, for a medical record template, users can input medical records based on the medical record template.

[0078] In addition to the system embodiments described above, this application also provides a text input method, which will be described below as an example.

[0079] Figure 4 This is a flowchart illustrating the text input method provided in an embodiment of this application. For details regarding the implementation of the first terminal, please refer to the relevant content of the above system embodiments; it will not be repeated here. Figure 4 As shown, the text input method includes:

[0080] 401. Get the key code of the currently entered character and the context of the currently entered character key code.

[0081] 402. Input the currently input character key and context into the prediction model; the prediction model is trained using text corpora from the vertical domain as samples.

[0082] 403. In the prediction model, based on the current input character key and context, determine at least one candidate text corresponding to the current input character key and the accuracy of at least one candidate text.

[0083] 404. Display at least one candidate text based on the accuracy of at least one candidate text.

[0084] To improve the efficiency of text input in vertical domains, this embodiment utilizes text corpora from those domains to train a specialized prediction model, which is then applied to actual text input. The model training process can be performed by the terminal or trained by other computer devices and then provided to the terminal. For details on the training process of the prediction model, please refer to the relevant content in the above embodiments; it will not be repeated here.

[0085] Furthermore, after the aforementioned prediction model is trained, the terminal can store the prediction model and use it for text input during actual text input. The specific implementation is as follows:

[0086] In step 401, the currently input character key code can be obtained. In this embodiment, the specific implementation of obtaining the currently input character key code is not limited. In some embodiments, the terminal may support keyboard input. Accordingly, the terminal may, in response to a trigger operation on the keyboard, obtain the character key code input via the keyboard as the currently input character key.

[0087] In other embodiments, the terminal may support voice input. Accordingly, the sound data currently picked up by the microphone can be acquired as the character key code for the current input.

[0088] In some other embodiments, the terminal may also support handwritten input. Accordingly, the motion trajectory of the target medium on the screen can be captured as the currently input character key code. The target medium may be the user's finger or a stylus, etc.

[0089] Of course, for terminals that use multiple input methods for payment, the character key codes input through various methods can also be acquired as the currently input character key codes. For example, the character key codes input via the keyboard and the sound data currently picked up by the microphone can be acquired as the keyboard input for that moment, and so on.

[0090] To improve the accuracy of the prediction model in determining the text corresponding to the currently input character keycode, the context of the currently input character keycode can also be obtained in step 401. Optionally, the first terminal can obtain character keycodes that have been input before the currently input character keycode as the context of the currently input character keycode. Specifically, the character keycodes input before the currently input character keycode refer to the character keycode input in the previous instance.

[0091] In other embodiments, the first terminal may also obtain the target text corresponding to the character key codes that were entered before the currently entered character key codes, as the context of the currently entered character key codes. The target text is the text that the user expects to output when entering the character key codes that were entered before the currently entered character key codes, and it is text that matches the vertical domain.

[0092] Further, in step 402, the currently input character key and its context can be input into the prediction model; then, in step 403, the prediction model can determine at least one candidate text corresponding to the currently input character key and the accuracy of at least one candidate text based on the currently input character key and its context. The accuracy of at least one candidate text can, to some extent, reflect the matching degree between the corresponding candidate text and the vertical domain.

[0093] The output of the prediction model can be at least one candidate text corresponding to the currently input character key code; or it can be at least one candidate text corresponding to the currently input character key code and its context. In this embodiment, no limitation is made.

[0094] Furthermore, in step 404, at least one candidate text can be displayed according to its accuracy, allowing the user to browse and select. Since the prediction model is trained using text corpora from a vertical domain, texts matching the vertical domain have higher scores and accuracy when calculating text corresponding to character keys. Therefore, prioritizing the display of candidate texts with higher accuracy and recommending those suitable for the vertical domain to the user reduces the time spent browsing candidate texts and selecting the desired text, thereby improving text input efficiency.

[0095] On the other hand, in this embodiment, when performing text prediction on the currently input character key code, the contextual information of the currently input character key code is taken into account, which helps the prediction model to more accurately identify the user's input intent, thereby helping to improve the accuracy of the output candidate text.

[0096] Considering that the candidate text output by the prediction model may contain text that is significantly mismatched with the vertical domain, this embodiment can also perform post-processing on the candidate text output by the prediction model. Specifically, before step 404, at least one candidate text output by the prediction model can be compared with a thesaurus of the vertical domain to determine the first candidate text that does not match the vertical domain. For specific implementation details on determining the first candidate text that does not match the vertical domain, please refer to the relevant content of the above embodiments, which will not be repeated here.

[0097] In this embodiment, for a first candidate text that does not match the vertical domain, the similarity between the first candidate text and the commonly used texts in the vertical domain recorded in the thesaurus can be calculated; and candidate commonly used texts whose similarity with the first candidate text meets the set requirements can be obtained from the thesaurus.

[0098] Furthermore, the first candidate text can be modified based on the frequency of selection of candidate idiomatic texts. Optionally, the candidate idiomatic text with the highest selection frequency can be selected, and the first candidate text can be modified to be the candidate idiomatic text with the highest selection frequency. This post-prediction processing method can supplement the situations that the prediction model cannot capture, further improve the matching degree between the displayed candidate text and the vertical domain, and improve the accuracy of the candidate text.

[0099] Furthermore, at least one candidate text can be displayed according to the accuracy of at least one candidate text corresponding to the currently input character key code. Optionally, at least one candidate text can be displayed in descending order of accuracy of at least one candidate text corresponding to the currently input character key code.

[0100] For a user, at least one candidate text can be browsed to select the text to be entered. For a terminal, in response to a selection event for at least one candidate text, the target candidate text selected by the selection event can be entered into a first designated area. Alternatively, the target candidate text selected by the selection event can be provided to other terminals with which it is communicating. The other terminals receive the target candidate text and enter it into a second designated area.

[0101] In this embodiment, the specific implementation of the user selecting the desired input text is not limited. In some embodiments, the keyboard can be configured with shortcut keys corresponding to candidate texts. The user can select the target candidate text using shortcut keys. For the first terminal, key press detection can be performed on at least one type of shortcut key corresponding to at least one candidate text; if at least one type of shortcut key corresponding to at least one candidate text is detected to be triggered, a selection event for at least one candidate text is determined; and the candidate text corresponding to the triggered type of shortcut key is determined to be the target candidate text.

[0102] If the keyboard is implemented as a physical hardware keyboard, users can select target candidate text by touching shortcut keys. If the keyboard is implemented as a software keyboard, users can select target candidate text by touching the screen area corresponding to the shortcut keys.

[0103] In this embodiment, the display duration of at least one candidate text can also be timed; if no selection event for at least one candidate text is detected after the display duration has reached the set duration, the candidate text ranked first can be input into a first designated area; or, the candidate text ranked first can be provided to other terminals communicating with it. Other terminals can receive the candidate text and input it into the second designated area.

[0104] The above description uses keyboard input as an example to illustrate the text input method of the data processing system provided in this application embodiment. In some embodiments, in addition to keyboard input, the first terminal may also support voice input. The following describes an exemplary implementation of compound text input using the first terminal.

[0105] For complex text input methods, the prediction model can be trained using text corpora from a vertical domain and their corresponding speech corpora. For example, in the medical field, medical record texts can be obtained as the text corpus, and the speech data of communication between doctors and patients during the process of entering medical record texts can be obtained as the speech corpus corresponding to the text corpus. Furthermore, speech recognition can be performed on the speech corpus to obtain the corresponding text information; then, the model is trained using the text corpus and the text information corresponding to the speech corpus as samples to obtain the prediction model.

[0106] For this prediction model, embodiments of this application also provide another text input method. For example... Figure 5 As shown, the method mainly includes:

[0107] 501. Get the key code of the currently entered character and the context of the currently entered character key code.

[0108] 502. Perform speech recognition on the sound data currently picked up by the microphone to determine the text information reflected in the sound data.

[0109] 503. Input the currently input character key, context, and text information into the prediction model; the prediction model is trained using text corpora from the vertical domain and the corresponding audio corpora as samples.

[0110] 504. In the prediction model, based on the currently input character key code, context, and text information, determine at least one candidate text corresponding to the currently input character key code and the accuracy of at least one candidate text.

[0111] 505. Display at least one candidate text based on the accuracy of at least one candidate text.

[0112] The specific implementation of step 501 can be found in the relevant content of step 401 above, and will not be repeated here. In step 502, the first terminal can perform speech recognition on the sound data currently picked up by the microphone to determine the text information reflected by the sound data. Further, in step 503, the currently input character key code, context, and text information reflected by the sound data can be input into the prediction model; then, in step 504, the prediction model can determine at least one candidate text corresponding to the currently input character key code and the accuracy of at least one candidate text based on the currently input character key code, context, and text information reflected by the sound data. In this embodiment, when using the prediction model to calculate the text corresponding to the character key code, the sound data is combined with the currently input character key code, providing multi-dimensional reference information for determining the currently input character key code, which helps to improve the accuracy of the determined candidate text. On the other hand, in this embodiment, when performing text prediction on the currently input character key code, the context information of the currently input character key code is taken into account, which helps the prediction model to more accurately identify the user's input intention, thereby helping to improve the accuracy of the output candidate text.

[0113] Furthermore, in step 505, the at least one candidate text can be displayed according to its accuracy. Since the prediction model is trained using text corpora from the vertical domain and their corresponding speech corpora, candidate texts matching the vertical domain in the model's output have higher scores and accuracy. Therefore, prioritizing the display of candidate texts with higher accuracy and recommending those suitable for the vertical domain to the user reduces the time spent browsing candidate texts and selecting the desired text, thereby improving text input efficiency.

[0114] For details on how to display at least one candidate text and how the user selects the desired target candidate text, please refer to the relevant content of the above embodiments, which will not be repeated here.

[0115] It is worth noting that, in this embodiment, in addition to selecting the desired candidate text via keyboard shortcuts, a text input template used in a vertical field, such as a medical record template, can also be selected via shortcuts. In this embodiment, for ease of description and distinction, the shortcut for selecting the desired candidate text can be defined as a first type of shortcut; and the shortcut for selecting a text input template can be defined as a second type of shortcut. The keyboard can be configured with the second type of shortcut. The user can select the corresponding text input template by triggering the second type of shortcut. For the first terminal, in response to the triggering operation of the second type of shortcut corresponding to the text template used in the vertical field, the text input template corresponding to the triggered second type of shortcut can be retrieved and displayed for the user to input text based on the text input template; or, the text input template can be provided to other terminals communicating with it. Other terminals can display the text input template. For the user, text input can be performed based on the text input template, further improving text input efficiency. For example, for a medical record template, the user can input medical records based on the medical record template.

[0116] It should be noted that the execution subject of each step of the method provided in the above embodiments can be the same device, or the method can be executed by different devices. For example, the execution subject of steps 401 and 402 can be device A; or the execution subject of step 401 can be device A, and the execution subject of step 402 can be device B; and so on.

[0117] Furthermore, some processes described in the above embodiments and accompanying drawings include multiple operations that appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or they may be executed in parallel. The operation numbers, such as 401, 402, etc., are merely used to distinguish different operations and do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel.

[0118] Accordingly, embodiments of this application also provide a computer-readable storage medium storing computer instructions, which, when executed by one or more processors, cause one or more processors to perform the steps in the above-described text input methods.

[0119] This application also provides a computer program product, including a computer program; when the computer program is executed by one or more processors, it can implement the steps in the above-described text input methods. This computer program product can be implemented as a text input method application, a plugin, or other software integrating text input method functionality.

[0120] Figure 6 A schematic diagram of the structure of a computer device provided in an embodiment of this application. For example...Figure 6 As shown, the computer device includes: a memory 60a, a processor 60b, and a display component 60c.

[0121] Memory 60a is used to store computer programs.

[0122] The processor 60b is coupled to a memory and a display component for executing a computer program to: acquire the currently input character key and the context of the currently input character key; input the currently input character key and the context into a prediction model; the prediction model is trained using text corpus from a vertical domain as samples; in the prediction model, based on the currently input character key and the context, determine at least one candidate text corresponding to the currently input character key and the accuracy of at least one candidate text; and display at least one candidate text through the display component 60c according to the accuracy of at least one candidate text.

[0123] In this embodiment, since the prediction model is trained using text corpora from a vertical domain, texts matching the character key codes in the prediction model have higher scores and accuracy when calculating the text. Therefore, candidate texts with higher accuracy can be prioritized and recommended to the user, reducing the time spent browsing candidate texts and selecting the desired text, thus improving text input efficiency.

[0124] On the other hand, in this embodiment, when performing text prediction on the currently input character key code, the contextual information of the currently input character key code is taken into account, which helps the prediction model to more accurately identify the user's input intent, thereby helping to improve the accuracy of the output candidate text.

[0125] In some embodiments, when the processor 60b acquires the currently input character key code, it specifically performs at least one of the following operations:

[0126] In response to a keyboard trigger, the key code of the character entered via the keyboard is obtained and used as the key code of the currently entered character.

[0127] Obtain the sound data currently picked up by the microphone as the key code for the currently input character;

[0128] The motion trajectory of the target medium on the screen is captured and used as the character key code for the current input.

[0129] In some embodiments, when the processor 60b obtains the context of the currently input character key code, it is specifically used to: obtain the character key code of the previous input character key code as the context of the currently input character key code; or, obtain the target text corresponding to the previously input character key code as the context of the currently input character key code; the target text is matched with the vertical domain.

[0130] In some embodiments, the processor 60b is further configured to: input the target candidate text selected by the selection event into a first designated area in response to a selection event for at least one candidate text; or, in response to a selection event for at least one candidate text, provide the target candidate text selected by the selection event to a second terminal communicatively connected to the first terminal via the communication component 60d, so that the second terminal can input the target candidate text into a second designated area.

[0131] In other embodiments, the processor 60b is further configured to: perform key detection on a first type of shortcut key corresponding to at least one candidate text; if the first type of shortcut key is detected to be triggered, determine that a selection event for at least one candidate text is generated; and determine that the candidate text corresponding to the triggered first type of shortcut key is the target candidate text.

[0132] Optionally, the processor 60b is further configured to: time the display duration of at least one candidate text; if no selection event is detected when the display duration reaches a set duration, input the candidate text ranked first into a first designated area, or provide the candidate text ranked first to a second terminal via the communication component 60d so that the second terminal can input the candidate text ranked first into a second designated area.

[0133] Optionally, the processor 60b is further configured to: compare at least one candidate text in a lexicon of the vertical domain before displaying at least one candidate text according to the matching degree between at least one candidate text and the vertical domain, to determine a first candidate text that does not match the vertical domain; calculate the similarity between the first candidate text and the conventional text of the vertical domain recorded in the lexicon; obtain from the lexicon candidate conventional text whose similarity to the first candidate text meets the set requirements; and correct the first candidate text according to the frequency of selection of the candidate conventional text.

[0134] In other embodiments, the processor 60b is further configured to: in response to a triggering operation of a second type of shortcut key corresponding to a text template used in a vertical domain, retrieve a text input template corresponding to the second type of shortcut key; display the text input template through the display component 60c for the user to input text based on the text input template, or provide the text input template to a second terminal that is communicatively connected to the first terminal through the communication component 60d for the second terminal to display the text input template.

[0135] In some embodiments of this application, the processor 60b is further configured to: acquire the currently input character key code and the context of the currently input character key code; perform speech recognition on the sound data currently picked up by the microphone 60e to determine the text information reflected by the sound data; input the currently input character key code, context and text information into a prediction model; the prediction model is trained using text corpus in a vertical domain and the corresponding sound corpus as samples; in the prediction model, based on the currently input character key code, context and text information, determine at least one candidate text corresponding to the currently input character key code and the accuracy of at least one candidate text; and display at least one candidate text through the communication component 60d according to the accuracy of at least one candidate text.

[0136] In this embodiment, when using the prediction model to calculate the text corresponding to the character keycode, the sound data is combined with the currently input character keycode, providing multi-dimensional reference information for determining the currently input character keycode, which helps improve the accuracy of the determined candidate text. On the other hand, in this embodiment, when predicting the text of the currently input character keycode, the contextual information of the currently input character keycode is taken into account, which helps the prediction model to more accurately identify the user's input intent, thereby helping to improve the accuracy of the output candidate text.

[0137] Furthermore, the computer device can display at least one candidate text based on the accuracy of at least one candidate text. Since the prediction model is trained using text corpora from a vertical domain and their corresponding speech corpora, candidate texts that match the vertical domain in the model's output have higher scores and accuracy. Therefore, the computer device can prioritize displaying candidate texts with higher accuracy, recommending candidate texts that fit the vertical domain to the user. This reduces the time users spend browsing candidate texts and selecting the desired text, thus helping to improve text input efficiency.

[0138] In some alternative implementations, such as Figure 6 As shown, the computer device may also include: a power supply component 60f, a speaker 60g, or a sensor 60h, etc. The sensor 60h may include components such as a vision sensor, a distance sensor, a light sensor, and a Hall effect sensor. Figure 6 The diagram only shows some components and does not mean that the computer device must contain them. Figure 6 The inclusion of all components does not imply that a computer device can only include... Figure 6 The components shown.

[0139] In this embodiment, the memory is used to store computer programs and can be configured to store various other data to support operation on its host device. The processor can execute the computer programs stored in the memory to implement corresponding control logic. The memory 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 storage, flash memory, magnetic disk, or optical disk.

[0140] In the embodiments of this application, the processor can be any hardware processing device capable of executing the above-described method logic. Optionally, the processor can be a central processing unit (CPU), a graphics processing unit (GPU), or a microcontroller unit (MCU); it can also be a field-programmable gate array (FPGA), a programmable array logic (PAL), a general array logic (GAL), a complex programmable logic device (CPLD), or other programmable devices; or it can be an advanced reduced instruction set (RISC) processor (ARM) or a system on chip (SOC), etc., but is not limited thereto.

[0141] In this embodiment, the communication component is configured to facilitate wired or wireless communication between its host device and other devices. The device housing the communication component can access wireless networks based on communication standards, such as WiFi, 2G or 3G, 4G, 5G, or combinations thereof. In one exemplary embodiment, the communication component receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In another exemplary embodiment, the communication component may also be implemented based on Near Field Communication (NFC), Radio Frequency Identification (RFID), Infrared Data Association (IrDA), Ultra Wideband (UWB), Bluetooth (BT), or other technologies.

[0142] In embodiments of this application, the display component may include a liquid crystal display (LCD) and a touch panel (TP). If the display component includes a touch panel, the display component may be implemented as a touchscreen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may not only sense the boundaries of touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operation.

[0143] In this embodiment, a power supply component is configured to provide power to various components of the device in which it resides. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device in which the power supply component resides.

[0144] In embodiments of this application, the audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC), which is configured to receive external audio signals when the device containing the audio component is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals can be further stored in memory or transmitted via a communication component. In some embodiments, the audio component also includes a speaker for outputting audio signals. For example, in devices with voice interaction capabilities, voice interaction with the user can be achieved through the audio component.

[0145] It should be noted that the terms "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0146] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0147] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0148] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0149] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0150] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0151] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0152] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0153] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0154] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A text input method, characterized in that, include: Obtain the currently input character key code and the context of the currently input character key code; The currently input character key and the context are input into the prediction model; the prediction model is trained using text corpora from the vertical domain as samples. In the prediction model, based on the currently input character key and the context, at least one candidate text corresponding to the currently input character key and the accuracy of the at least one candidate text are determined. Display the at least one candidate text according to its accuracy. Before displaying the at least one candidate text according to its matching degree with the vertical domain, the method further includes: The at least one candidate text is compared with the thesaurus of the vertical domain to determine the first candidate text that does not match the vertical domain. Calculate the similarity between the first candidate text and the commonly used texts in the vertical domain recorded in the thesaurus; From the lexicon, obtain candidate commonly used texts that have a similarity to the first candidate text that meets the set requirements; The first candidate text is modified based on the frequency with which the candidate commonly used text is selected.

2. The method according to claim 1, characterized in that, The context for obtaining the currently input character key code includes: The previously entered character key code is used as the context of the currently entered character key code; or, Obtain the target text corresponding to the previously input character key code, and use it as the context of the currently input character key code; the target text is matched with the vertical domain.

3. The method according to claim 1, characterized in that, Also includes: In response to a selection event for the at least one candidate text, the target candidate text selected by the selection event is input into a first designated area; or, In response to a selection event for the at least one candidate text, the target candidate text selected by the selection event is provided to other terminals so that the other terminals can input the target candidate text into a second designated area.

4. The method according to claim 3, characterized in that, Also includes: Key press detection is performed on the first type of shortcut key corresponding to the at least one candidate text; If the first type of shortcut key is detected to be triggered, it is determined that a selection event for the at least one candidate text is generated; The candidate text corresponding to the first type of shortcut key that was triggered is determined to be the target candidate text.

5. The method according to claim 3, characterized in that, Also includes: The display duration of the at least one candidate text is timed; If the selection event is not detected when the display duration reaches the set duration, the candidate text ranked first will be entered into the first designated area, or the candidate text ranked first will be provided to the other terminals so that the other terminals can enter the candidate text ranked first into the second designated area.

6. The method according to any one of claims 1-5, characterized in that, Also includes: In response to the triggering operation of the second type of shortcut key corresponding to the text template used in the vertical domain, the text input template corresponding to the second type of shortcut key is retrieved; The text input template can be displayed so that users can input text based on it, or the text input template can be provided to other terminals so that they can display it.

7. The method according to any one of claims 1-5, characterized in that, The method of obtaining the currently input character key code includes at least one of the following: In response to a trigger operation on the keyboard, the character key code input through the keyboard is obtained as the currently input character key code; Obtain the sound data currently picked up by the microphone as the key code for the currently input character; The motion trajectory of the target medium on the screen is captured and used as the character key code of the current input.

8. A data processing system, characterized in that, include: First terminal and second terminal; the first terminal and second terminal are connected in communication; The first terminal is used to obtain the currently input character key code and the context of the currently input character key code; The currently input character key and the context are input into the prediction model; the prediction model is trained using text corpus from a vertical domain as samples; in the prediction model, based on the currently input character key and the context, at least one candidate text corresponding to the currently input character key and the accuracy of the at least one candidate text are determined. Display the at least one candidate text according to its accuracy. And, in response to a selection event for the at least one candidate text, the target candidate text selected by the selection event is provided to the second terminal; The second terminal is used to input the target candidate text into a designated area; Before displaying the at least one candidate text according to its matching degree with the vertical domain, the first terminal is further configured to: compare the at least one candidate text with the lexicon of the vertical domain to determine a first candidate text that does not match the vertical domain; calculate the similarity between the first candidate text and the commonly used texts of the vertical domain recorded in the lexicon; obtain from the lexicon candidate commonly used texts whose similarity with the first candidate text meets a set requirement; and modify the first candidate text according to the frequency of selection of the candidate commonly used texts.

9. A computer device, characterized in that, include: Memory, processor, and display components; The memory is used to store computer programs; The processor is coupled to the memory and the display component for executing the computer program to perform the steps of the method according to any one of claims 1-7.

10. A computer program product, characterized in that, include: A computer program; which, when executed by one or more processors, performs the steps of the method according to any one of claims 1-7.

11. A computer-readable storage medium storing computer instructions, characterized in that, When the computer instructions are executed by one or more processors, they cause one or more processors to perform the steps of the method according to any one of claims 1-7.

Images