Method and apparatus for processing personalized term

Abstract

Embodiments of the present application provide a personalized term processing method and device and a device for processing personalized terms, wherein the method specifically comprises: determining a user feature word from input content of a user; determining a related term corresponding to the user feature word according to a vector relationship; and determining a personalized term corresponding to the user according to the related term. Embodiments of the present application can meet the fine input requirement of a user and reduce storage space.

Description

TECHNICAL FIELD

[0001] The present application relates to the input technical field, and in particular to a personalized word processing method and device. BACKGROUND

[0002] With the popularization and development of computer technology and Internet technology, input method has become an important tool for users to carry out human-computer interaction. Generally, when evaluating the performance of the input method, the accuracy of the first word and the accuracy of the candidate sorting are two important evaluation indexes. The input method dictionary recording word information and word frequency information is an important factor affecting the two evaluation indexes. In the existing input method system, the system dictionary covers most of the commonly used words used by users. However, due to the differences in professional fields, interests and language use habits of different users, there is a personalized demand for the word of the input method.

[0003] In order to meet the personalized input needs of different users, the input method system provides a cell dictionary. The cell dictionary generally refers to a subdivided dictionary. A cell dictionary is a set of words in a fine classification. The classification of the cell dictionary can be a professional field (such as a medical field dictionary), a region (such as a Beijing place name dictionary), a game (such as a World of Warcraft vocabulary), etc.

[0004] The inventors found in the implementation of the embodiments of the present application that a cell dictionary usually contains a large number of words, which makes the cell dictionary occupy a large storage space. Moreover, the cell dictionary is usually constructed based on the cognition of domain experts on words, which makes the granularity of the cell dictionary coarse, and further causes the inability to meet the fine input needs of users. SUMMARY

[0005] The embodiments of the present application provide a personalized word processing method, device and device for input, which can meet the fine input needs of users and reduce the storage space.

[0006] In order to solve the above problem, the embodiments of the present application disclose a personalized word processing method, comprising:

[0007] determining a user feature word from the input content of the user;

[0008] determining a related word corresponding to the user feature word according to the vector relationship;

[0009] determining a personalized word corresponding to the user according to the related word.

[0010] On the other hand, the embodiments of the present application disclose a personalized word processing device, comprising:

[0011] The user feature word determination module is configured to determine a user feature word from input content of the user;

[0012] The related term determination module is configured to determine a related term corresponding to the user feature word according to a vector relationship; and

[0013] The personalized term determination module is configured to determine a personalized term corresponding to the user according to the related term.

[0014] In still another aspect, an embodiment of the present application discloses a device for processing a personalized term, comprising a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, and the one or more programs contain instructions for performing the following operations:

[0015] determining a user feature word from input content of the user;

[0016] determining a related term corresponding to the user feature word according to a vector relationship;

[0017] determining a personalized term corresponding to the user according to the related term.

[0018] In still another aspect, an embodiment of the present application discloses a machine readable medium having instructions stored thereon, which, when executed by one or more processors, cause an apparatus to perform the processing method of the personalized term as described in one or more of the preceding aspects.

[0019] Embodiments of the present application include the following advantages:

[0020] Embodiments of the present application utilize the concept of word pushing word, determine a related term corresponding to a user feature word according to a vector relationship, and determine a personalized term corresponding to the user according to the related term. Since the user feature word represents user features and user input requirements, embodiments of the present application can recommend personalized terms based on the word granularity corresponding to the user feature word, and thus the obtained personalized terms can meet the fine input requirements of the user.

[0021] Furthermore, embodiments of the present application provide corresponding personalized terms based on the word granularity corresponding to the user feature word. The number of the above-mentioned personalized terms is generally less than the number of terms included in the cell term library, and thus the storage space can be reduced. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the description of the embodiments of the present application. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0023] Figure 1 is a schematic diagram of an application environment of a personalized word processing method according to an embodiment of the present application;

[0024] Figure 2 is a step flow chart of the personalized word processing method according to an embodiment of the present application;

[0025] Figure 3 is a step flow chart of the personalized word processing method according to another embodiment of the present application;

[0026] Figure 4 is a structural block diagram of a personalized word processing device according to an embodiment of the present application;

[0027] Figure 5 is a block diagram of a device 800 for processing personalized words according to an embodiment of the present application; and

[0028] Figure 6 is a structural diagram of a server in some embodiments of the present application. DETAILED DESCRIPTION

[0029] The technical solutions of the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0030] In order to solve the technical problems of the cell word library occupying more storage space and failing to meet the fine input needs of users in the prior art, the embodiments of the present application provide a personalized word processing scheme. The scheme can determine user characteristic words from the input content of a user; determine relevant words corresponding to the user characteristic words according to a vector relationship; and determine personalized words corresponding to the user according to the relevant words.

[0031] In the embodiments of the present application, the onscreen operation refers to an operation of outputting and displaying the content in the preview window to the application window. Specifically, in the input method, the onscreen operation refers to an operation of outputting and displaying the candidate in the candidate window to the application window. The candidate corresponding to the onscreen operation can be referred to as the input content. The input content in the embodiments of the present application can include one or more onscreen contents. The input content can correspond to a language unit such as a word, a phrase, a sentence, etc.

[0032] The characteristic word can represent a word with a special characteristic. The embodiments of the present application represent the user characteristics and the user input requirements according to the user characteristic word in the input content.

[0033] The embodiments of the present application determine the related word corresponding to the user characteristic word according to the vector relationship, and determine the personalized word corresponding to the user according to the related word. Since the user characteristic word represents the user characteristics and the user input requirements, the embodiments of the present application can recommend the personalized word based on the word granularity corresponding to the user characteristic word, and thus the personalized word obtained can meet the fine input requirements of the user.

[0034] In addition, the embodiments of the present application provide the corresponding personalized word based on the word granularity corresponding to the user characteristic word. The number of the above-mentioned personalized word is usually less than the number of the word in the cell word library, and thus the storage space can be reduced.

[0035] The processing method of the personalized word provided by the embodiments of the present application can be applied to Figure 1 As shown in FIG. 1, the application environment includes a client 100 and a server 200. Figure 1 As shown in FIG. 1, the client 100 and the server 200 are located in a wired or wireless network, and the client 100 and the server 200 perform data interaction through the wired or wireless network.

[0036] Optionally, the client 100 can run on a terminal, and the terminal specifically includes but is not limited to a smart phone, a tablet computer, an e-book reader, an MP3 (Moving Picture Experts Group Audio Layer III) player, an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop computer, a vehicle-mounted computer, a desktop computer, a set-top box, a smart television, a wearable device, etc. The client 100 can correspond to a website or an APP (Application).

[0037] The service end of the embodiment of the present application can be a cloud service end (cloud end). The cloud end is a kind of computing service which is simple and efficient, safe and reliable, and has elastic processing capacity. The resource information of the cloud end is dynamic, so that the processing capacity thereof is elastic.

[0038] The embodiment of the present application can be applied to the recommendation scene of personalized entries to meet the fine input requirement of a user and reduce storage space.

[0039] The embodiment of the present application can be applied to an input method program of various input modes such as keyboard symbols, handwriting and speech, that is, a user can input characters through an encoded string. The input string can refer to the encoded string input by the user. In the field of input methods, for an input method program of Chinese, Japanese, Korean or other languages, the input string input by the user can be usually converted into a candidate of the corresponding language. Hereinafter, Chinese is mainly taken as an example for description, and other languages such as Japanese and Korean can be referred to each other. It can be understood that the Chinese input method can include, but is not limited to, full spelling, simple spelling, strokes and five strokes, and the embodiment of the present application does not limit the specific input method program corresponding to a certain language.

[0040] Taking the input of Chinese as an example, the type of the encoded string can include a pinyin string, a character string (such as a five-stroke string) and speech information (such as a speech input mode). Taking the input of English as an example, the type of the encoded string can include a letter string.

[0041] In actual application, for the input mode of keyboard symbols, the user can input the above input string through a physical keyboard or a virtual keyboard. For example, for a terminal with a touch screen, a virtual keyboard can be set in an input interface to input the input string by triggering a virtual key included in the virtual keyboard. Optionally, examples of the virtual keyboard can include a 9-key keyboard and a 26-key keyboard. It can be understood that in addition to the virtual keys corresponding to letters, the input interface can also be provided with symbol keys, number keys, function keys such as Chinese-English switching keys, or tool bar keys, and it can be understood that the embodiment of the present application does not limit the specific keys contained in the input interface.

[0042] According to some embodiments, the input string can include, but is not limited to, one key symbol or a combination of multiple key symbols input by the user through a key. The key symbol can specifically include pinyin, strokes and hiragana.

[0043] In the embodiment of the present application, the candidate can be used to represent one or more characters provided by the input method program to be selected by the user. The candidate can correspond to the context, or the candidate can correspond to the input string, or the candidate can correspond to the input string and the context. The candidate can be a character of Chinese, English, Japanese, etc. The candidate can also be a combination of symbols such as color text and pictures. The color text includes but is not limited to a picture composed of lines, symbols, and characters. For example, the color text can include ": P", ":-o", ":-)", etc.

[0044] In the embodiment of the present application, the input string can be used to search in the word library to obtain the candidate corresponding to the input string. The above-mentioned word library can be a traditional word library, and the traditional word library specifically includes a system word library, a user word library, a cell word library, a cloud word library, etc. The embodiment of the present application adds the first personalized word to the traditional word library to meet the real-time personalized needs of the user.

[0045] Method embodiment one

[0046] Reference Figure 2 Fig. 1 shows a step flowchart of a personalized word processing method embodiment one of the present application, which can specifically include the following steps:

[0047] Step 201, determining a user feature word from the input content of the user;

[0048] Step 202, determining a related word corresponding to the user feature word according to the vector relationship;

[0049] Step 203, determining a personalized word corresponding to the user according to the related word.

[0050] Figure 2 The method embodiment shown can be executed by the client and / or the server. It can be understood that the embodiment of the present application does not limit the specific execution subject of the method embodiment.

[0051] In step 201, the feature word can represent a word with exclusive properties. The input content of the user can include a word with exclusive properties, or a word without exclusive properties. The embodiment of the present application determines the user feature word from the input content of the user, which can make the user feature word represent the user feature or the user input demand.

[0052] Optionally, the above-mentioned determination of the user feature word from the input content of the user specifically includes matching the input content of the user with a feature word set to obtain the user feature word contained in the input content.

[0053] For example, the input content of the user A is a sentence A "Today we go to Dikaon to buy a jacket, and go skiing in the afternoon", and the user feature words such as "Dikaon", "jacket", and "skiing" can be determined from the sentence A.

[0054] Optionally, the feature word in the embodiment of the present application specifically includes at least one of a domain word, a self-made word, and a cloud word.

[0055] The domain can refer to a professional range, and all matters within the range can be referenced. The domain vocabulary can correspond to a domain keyword.

[0056] The embodiment of the present application can determine the domain vocabulary by the following determination method.

[0057] Determination method 1: determining the domain vocabulary of the personal feature category, and the personal feature category can correspond to a domain keyword.

[0058] The domain vocabulary of the personal feature category specifically includes a professional attribute category (doctor, express delivery personnel, …), a long-term interest category (animation, military, …), and the like.

[0059] Determination method 2: determining the domain vocabulary of the topic category, and the topic category can correspond to a domain keyword.

[0060] The domain vocabulary of the topic category specifically includes a temporary topic (fish raising, a certain TV series, …), a stage topic (such as buying insurance, buying a house, learning to drive, and decoration, …), and the like.

[0061] Determination method 3: extracting a knowledge entry, and after screening the extracted knowledge entry, a corresponding domain keyword is marked. For example, the knowledge entry and the cell entry can be fused, and after screening the fusion result, a domain word is obtained.

[0062] Determination method 4: clustering the high-frequency entries input by the user, and the category keyword corresponds to a domain keyword, and the high-frequency entries under the category correspond to domain words.

[0063] It can be understood that the determination method 1 to the determination method 4 are only examples of the determination method of the domain vocabulary, and the specific determination method of the domain vocabulary is not limited in the embodiment of the present application, and other determination methods can be used according to actual application requirements by those skilled in the art. For example, an entity word can be collected, and the entity word can be used as a domain vocabulary.

[0064] In the embodiments of the present application, an entity is a specific thing or concept. Entities are generally classified into types, such as a person type entity, a movie type entity, an animal type entity, a history type entity, and the like. One entity can correspond to multiple entity instances, and an entity instance can be a descriptive page (content) of an entity in a network (or other medium), such as a page in a encyclopedia that contains an entity instance corresponding to an entity.

[0065] Optionally, an entity can include a named entity, which can refer to a person name, an organization name, a place name, and all other entities identified by a name. More broadly, named entities can also include a book title, a song title, a film or television drama title, a product name, a brand name, a number, a date, a currency, an address, and the like.

[0066] Optionally, the entity categories can include at least one of the following categories: a person, a place, a fruit, a vegetable, an animal, a plant, a building, a piece of clothing, a food, a medicine, a vehicle, furniture, a musical instrument, an electrical appliance, and a natural phenomenon. The entity category can serve as a domain keyword.

[0067] A self-made entry can be a self-made entry of a user to supplement the deficiencies of a traditional entry library. In actual applications, a self-made entry interface can be provided to receive a mapping relationship between an input string set by the user and the self-made entry.

[0068] A cloud entry can be an entry provided by a cloud input method based on cloud computing. The cloud input method is an input method relying on cloud computing technology and has incomparable advantages over traditional input methods, such as a larger and more updated entry library and a more advanced language model, which significantly improves the accuracy of input. In actual applications, a cloud request can be sent to the cloud, and the cloud request can include input information of a user, such as an input string or voice information. The cloud entry returned by the cloud can be received.

[0069] In step 202, the word vector technology can convert a word into a dense vector. Moreover, for words with similar semantics and / or contexts, their corresponding word vectors are also similar. The dimensionally dense word vector can contain more information, and a single dimension of the word vector can have a specific meaning.

[0070] For example, the number of dimensions of a word vector can include 128, 256, 512, 1024, or the like. After training, the word vector can represent the matching relationship between words. For example, the distance between “banana” and “apple” is closer than the distance between “banana” and “eggplant”, and thus the matching degree between “banana” and “apple” is greater than the matching degree between “banana” and “eggplant”.

[0071] Multiple word vectors can be fused into a new word vector. For example, word vector "woman" + word vector "beautiful" = word vector "goddess". A word vector minus a word vector can result in a new vector. For example, word vector "king" - word vector "man" + word vector "woman" = word vector "queen", and the like.

[0072] The embodiment of the present application determines the relevant word entries corresponding to the user feature word according to the vector relationship, and the relevant word entries obtained can match the user feature word in terms of semantics and / or context, and thus the user input demand can be met.

[0073] The embodiment of the present application can provide the following technical solution for determining the relevant word entries corresponding to the user feature word:

[0074] Technical solution A1,

[0075] In technical solution A1, the determination of the relevant word entries corresponding to the user feature word specifically includes: determining a first word vector corresponding to the user feature word; and determining a second word vector matching the first word vector according to a vector retrieval method, and taking a word entry corresponding to the second word vector as a relevant word entry.

[0076] In the embodiment of the present application, the training data can be trained to obtain a generation model of word vectors, and the generation model can include a mapping relationship between word entries and word vectors. The training data can include input corpus in an input method environment, instant messaging environment chat corpus, microblog environment corpus, and question and answer environment corpus.

[0077] In an optional embodiment of the present application, the training data can include input corpus in a second time range. The length of the second time range can be determined by a person skilled in the art according to actual application requirements, for example, the length of the second time range is two years, and the second time range is the last two years. It can be understood that the embodiment of the present application does not limit the specific second time range.

[0078] The embodiment of the present application can use a vector tool to realize vector retrieval. The vector tool can provide an index of vectors and match the first word vector with the index of vectors to obtain a second word vector matching the first word vector. It can be understood that the embodiment of the present application does not limit the specific vector retrieval method.

[0079] For example, the user feature word A is "skiing", and the embodiment of the present application can obtain the relevant word entries corresponding to the user feature word A, such as "skiing board", "skiing board wax", "helmet", "windshield", and the like, based on vector retrieval.

[0080] In actual application, a corresponding relevant word table corresponding to the user feature word can be maintained, and the relevant word table can include relevant word entries corresponding to the user feature word. For example, n relevant word tables corresponding to n user feature words of a user, where n can be a natural number. Alternatively, one relevant word table corresponding to n user feature words of a user, where the relevant word table includes n data records, and each data record corresponds to one user feature word.

[0081] Technical solution A2,

[0082] In technical solution A2, the determining of the relevant word entry corresponding to the user feature word specifically includes: determining at least two relevant word sets corresponding to the user feature word according to at least two word vector types; and determining the relevant word entry corresponding to the user feature word and the matching degree between the user feature word and the corresponding relevant word entry according to the matching degrees of the user feature word and the at least two relevant word sets.

[0083] In the embodiment of the application, the word vector type can be used to represent different types of word vectors, which can include different models used to generate word vectors. The word vector type can include a word2vec (word to vector) type, a graph vector type, and the like. The graph vector type can include deepwalk, node2vec, and the like.

[0084] In actual application, for a single word vector type, there can be an accuracy deviation problem caused by insufficient training, which can make the matching degree between the relevant word entry and the user feature word lower. To solve the above problem, the embodiment of the application obtains at least two relevant word sets corresponding to the user feature word according to at least two word vector types, and fuses the at least two relevant word sets to obtain a fused relevant word entry and a corresponding matching degree. Since the at least two relevant word sets of the embodiment of the application are derived from at least two word vector types, the accuracy deviation problem caused by insufficient training of a single word vector type can be overcome to some extent, and thus the matching degree between the relevant word entry and the user feature word can be improved.

[0085] Suppose that the two relevant word sets include relevant word set 1 and relevant word set 2, then the relevant word entries in relevant word set 1 and relevant word set 2 are fused according to the weights corresponding to relevant word set 1 and relevant word set 2, respectively.

[0086] For example, if a related term i appears in multiple related word sets, and related word set 1 and related word set 2 correspond to weights a and b respectively, the matching degree between related term i and user feature term in related word set 1 is matching degree 1, and the matching degree between related term i and user feature term in related word set 2 is matching degree 2, then matching degree 1 and matching degree 2 can be weighted and averaged according to a and b to obtain the fused matching degree.

[0087] For example, if a related term j appears only in one set of related terms, then the matching degree between related term j and user feature terms in other sets of related terms can be considered to be 0, and thus the matching degree after fusion can be obtained.

[0088] It is understood that those skilled in the art can adopt any or a combination of technical solutions A1 to A2 according to actual application needs. For example, when technical solutions A1 and A2 are combined, technical solution A1 can be used first to retrieve at least two sets of related words from at least two word vector types using a vector retrieval method. Then, technical solution A2 can be used to determine the related terms corresponding to the user feature words and the matching degree between the user feature words and the corresponding related terms based on the matching degree between the user feature words and the at least two sets of related words.

[0089] In step 203, based on the aforementioned relevant terms, the personalized terms corresponding to the aforementioned users are determined, and input services corresponding to the personalized terms can be provided to the users.

[0090] In practical applications, after determining personalized terms, the server can push these terms to the user's corresponding client, enabling the client to provide input services for those personalized terms. Personalized terms can correspond to input information and word frequency, among other things. Input information can include input strings or voice messages.

[0091] During the user's input process, the user's input information can be matched with the input information corresponding to personalized terms. If the match is successful, the corresponding personalized terms can be used as candidate terms.

[0092] For example, user A wants to input the target term "snowboard wax" using the input string "huaxuebanla". However, "snowboard wax" does not exist in the traditional dictionary. But the first personalized term provided based on user A's user feature words includes "snowboard wax". Therefore, "snowboard wax" can be provided as a candidate based on the personalized term, thus improving input efficiency.

[0093] Or, even if there is "ski wax" in the traditional word library, but according to the matching relationship between the input information and the personalized word, the position of the personalized word corresponding candidate can be adjusted, and the position of the personalized word corresponding candidate can be advanced, for example, the personalized word corresponding candidate can be advanced to the first, second or third position. Since the personalized word corresponding candidate can be provided to the user in priority, the efficiency of the user selecting the candidate can be improved, and the input efficiency can be improved.

[0094] In practical applications, the above related words can be directly used as the personalized words corresponding to the user; or the related words can be screened, and the screened related words can be used as the personalized words corresponding to the user.

[0095] In the embodiment of the application, the determination of the personalized words corresponding to the user can specifically include: sorting the related words according to the matching degree between the user feature words and the related words and / or the weight parameters corresponding to the user feature words, and determining the personalized words from the related words according to the sorting result.

[0096] The matching degree between the user feature words and the related words can represent the matching degree between them, and thus can represent the matching degree between the related words and the user input requirements.

[0097] The weight parameter can represent the importance of the user feature word in the input corpus. According to the weight parameter, the importance of the related words corresponding to different user feature words can be determined in the embodiment of the application.

[0098] Optionally, the weight parameter can represent the importance of the user feature word in all feature words in the input corpus; or the weight parameter can represent the importance of the user feature word in the sentence input corpus containing the feature word. If the feature word is contained in the sentence input corpus k, it can be used in the determination process of the weight parameter. If the feature word is contained in the sentence input corpus p, it cannot be used in the determination process of the weight parameter.

[0099] Optionally, the weight parameter can be determined by using the TF-IDF (Term Frequency-Inverse Document Frequency) technology. TF-IDF is a statistical method used to evaluate the importance of a word in a document set or a document in a corpus. The importance of a word increases in direct proportion to the number of times it appears in a document, but decreases in inverse proportion to the frequency of its appearance in the corpus. The idea of TF-IDF is that if a word appears frequently in a document and rarely in other documents, the importance of the word is considered to be high.

[0100] Correspondingly, the process of determining the weight parameter specifically includes: determining a frequency parameter (TF) of the user characteristic word in input corpus in a first time range; determining a discrimination parameter (IDF) of the user characteristic word in input corpus in a second time range; the second time range is greater than the first time range; and determining the weight parameter according to the frequency parameter and the discrimination parameter.

[0101] The first time range can reflect the timeliness of the user input demand of the target user, and the input corpus of the target user in the first time range can be regarded as a document. Therefore, the first time range can correspond to a small time length, such as a time length within 1 month, a time length within 2 months, etc. The target user can refer to a specific user corresponding to the personalized term, that is, a user to be recommended, such as user A, etc. Therefore, the embodiment of the application can obtain the TF according to the frequency of occurrence of the user characteristic word in the input corpus of the target user in the first time range.

[0102] The input corpus in the second time range can be input corpus of multiple users, which can be used to provide multiple documents. The input corpus of a user in the second time range can be regarded as a document, so that multiple documents can be obtained to provide statistical data for the statistical method of the weight parameter.

[0103] The embodiment of the application can take the number of users corresponding to the input corpus in the second time range as the total number of documents T, and take the number of users inputting the user characteristic word as the number of documents D containing the user characteristic word. Optionally, the total number of documents T can be divided by a factor corresponding to the number of documents D, and then the quotient obtained is taken as the logarithm to obtain the IDF; wherein the factor corresponding to the number of documents D can be the sum of the number of documents D and a preset positive number, which can avoid the factor being zero in the case that the number of documents D is zero. It can be understood that the embodiment of the application does not limit the specific calculation method of the IDF.

[0104] The embodiment of the application can determine the score corresponding to the related term according to the matching degree between the user characteristic word and the related term, and / or the weight parameter corresponding to the user characteristic word, and then sort the related terms according to the order of the scores from large to small.

[0105] The above determining the personalized term from the related terms according to the sorting result specifically includes: selecting the first M related terms in the front as the personalized term; or, the related term whose score exceeds a score threshold can be selected as the personalized term. In this way, the quality of the personalized term can be improved, and the number of personalized terms can be controlled.

[0106] The embodiment of the application can provide any one of the following sorting methods:

[0107] The first sorting manner is to sort the relevant entries according to matching degrees between the user characteristic words and the relevant entries, and to determine the personalized entry from the relevant entries according to the first sorting result; or

[0108] The second sorting manner is to sort the relevant entries corresponding to the plurality of user characteristic words according to weight parameters corresponding to the user characteristic words, and to determine the personalized entry from the relevant entries corresponding to the plurality of user characteristic words according to a second sorting result; or

[0109] The third sorting manner is to sort the relevant entries corresponding to the plurality of user characteristic words according to matching degrees between the user characteristic words and the relevant entries and weight parameters corresponding to the user characteristic words, and to determine the personalized entry from the relevant entries corresponding to the plurality of user characteristic words according to a third sorting result.

[0110] For the third sorting manner, the matching degree and the weight parameter can be fused to obtain a score corresponding to the relevant entry. Correspondingly, the fusion manner can include multiplication, summation, or weighted average, etc.

[0111] In summary, the processing method of the personalized entry in the embodiment of the present application determines the relevant entry corresponding to the user characteristic word according to the vector relationship, and determines the personalized entry corresponding to the user according to the relevant entry. Since the user characteristic word represents the user characteristic and the user input requirement, the embodiment of the present application can recommend the personalized entry based on the word granularity corresponding to the user characteristic word, and thus the obtained personalized entry can meet the fine input requirement of the user.

[0112] In addition, the embodiment of the present application provides the corresponding personalized entry based on the word granularity corresponding to the user characteristic word. The number of the above-mentioned personalized entry is usually less than the number of the entries in the cell entry library, and thus the storage space can be reduced.

[0113] In addition, the current cell entry library does not sort different cell entries, so that the matching degree between the cell entry and the user input requirement cannot be measured. The present application sorts the relevant entry according to the matching degree and / or the weight parameter, and the sorting result can represent the matching degree between the relevant entry and the user input requirement. The higher the matching degree, the more suitable for recommendation. Therefore, the embodiment of the present application can recommend the relevant entry with a higher matching degree as the personalized entry, and thus the recommendation accuracy of the personalized entry can be improved.

[0114] Method embodiment two

[0115] Referring to Figure 3 FIG. 2 shows a step flowchart of a processing method embodiment two of a personalized entry of the present application, which can specifically include the following steps:

[0116] Step 301, according to the characteristic word set, determining the user characteristic word of the user from the input corpus of the user, and determining the weight parameter corresponding to the user characteristic word;

[0117] Step 302, determining at least two generation models of at least two word vector types according to the sentence input corpus of the user;

[0118] The sentence input corpus of the user can be used as training data of the generation model.

[0119] Step 303, determining the user characteristic word from the input content of the target user;

[0120] Step 304, determining at least two relevant word sets corresponding to the user characteristic word by using at least two generation models and vector retrieval methods;

[0121] Step 305, fusing the relevant word entries in the at least two relevant word sets to obtain the matching degree between the user characteristic word and the corresponding relevant word entry;

[0122] Step 306, sorting the relevant word entries corresponding to the multiple user characteristic words according to the weight parameter and the matching degree;

[0123] Step 307, selecting the relevant word entry meeting the preset condition from the sorting result as the personalized word entry for recommendation.

[0124] The embodiment of the application can recommend the personalized word entry for the user characteristic word contained in the input corpus of the target user in the first time range by using the concept of word pushing word.

[0125] The embodiment of the application can obtain at least two relevant word sets corresponding to the user characteristic word according to at least two generation models of at least two word vector types, and can overcome the precision deviation problem caused by insufficient training of a single word vector type to some extent, and thus can improve the matching degree between the relevant word entry and the user characteristic word.

[0126] Moreover, the embodiment of the application can sort the relevant word entries corresponding to the multiple user characteristic words according to the weight parameter and the matching degree, can adjust the positions of the relevant word entries corresponding to different user characteristic words according to the matching degree between the weight parameter and the user input demand, and can recommend the relevant word entry with higher matching degree as the personalized word entry, and thus can improve the recommendation precision of the personalized word entry.

[0127] It should be noted that, for the method embodiments, in order to simply describe, they are all expressed as a series of motion action combinations, but those skilled in the art should know that the embodiments of the present application are not limited by the described action sequence, because according to the embodiments of the present application, certain steps can be performed in other order or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification are all preferred embodiments, and the motion actions involved are not necessarily necessary for the embodiments of the present application.

[0128] Device embodiments

[0129] Referring to Figure 4 , a structure block diagram of an embodiment of a personalized entry processing device of the present application is shown, which can specifically include:

[0130] A user feature word determination module 401 is configured to determine user feature words from the user's input content;

[0131] A related entry determination module 402 is configured to determine related entries corresponding to the user feature words according to vector relationships; and

[0132] A personalized entry determination module 403 is configured to determine personalized entries corresponding to the user according to the related entries.

[0133] Optionally, the user feature word determination module 401 can include:

[0134] A matching module is configured to match the user's input content with a feature word set to obtain user feature words contained in the input content.

[0135] Optionally, the feature word set can include at least one of domain entries, self-made entries and cloud entries.

[0136] Optionally, the related entry determination module 402 can include:

[0137] A first word vector determination module is configured to determine a first word vector corresponding to the user feature words;

[0138] A second word vector determination module is configured to determine a second word vector matched with the first word vector according to a vector retrieval device, and the entry corresponding to the second word vector is taken as a related entry.

[0139] Optionally, the personalized entry determination module 403 can include:

[0140] A sorting module is configured to sort the related entries according to the matching degree between the user feature words and the related entries, and / or the weight parameter corresponding to the user feature words, and determine personalized entries from the related entries according to the sorting result.

[0141] Optionally, the apparatus can further include:

[0142] a frequency parameter determining module configured to determine a frequency parameter of the user characteristic word in input corpus in a first time range;

[0143] a distinguishing parameter determining module configured to determine a distinguishing parameter of the user characteristic word in input corpus in a second time range; the second time range is greater than the first time range;

[0144] a weight parameter determining module configured to determine the weight parameter according to the frequency parameter and the distinguishing parameter.

[0145] Optionally, the personalized entry determining module 403 can include:

[0146] a first sorting module configured to perform first sorting on the related entries according to the matching degree between the user characteristic word and the related entries, and determine the personalized entry from the related entries according to the first sorting result; or

[0147] a second sorting module configured to perform second sorting on the related entries corresponding to the plurality of user characteristic words respectively according to the weight parameter corresponding to the user characteristic word, and determine the personalized entry from the related entries corresponding to the plurality of user characteristic words respectively according to the second sorting result; or

[0148] a third sorting module configured to perform third sorting on the related entries corresponding to the plurality of user characteristic words respectively according to the matching degree between the user characteristic word and the related entries and the weight parameter corresponding to the user characteristic word, and determine the personalized entry from the related entries corresponding to the plurality of user characteristic words respectively according to the third sorting result.

[0149] Optionally, the related entry determining module 402 can include:

[0150] a first determining module configured to determine at least two related word sets corresponding to the user characteristic word according to at least two word vector types;

[0151] a second determining module configured to determine the related entry corresponding to the user characteristic word and the matching degree between the user characteristic word and the corresponding related entry according to the matching degree between the user characteristic word and the at least two related word sets.

[0152] For the apparatus embodiment, it is basically similar to the method embodiment, so the description is relatively simple, and the related parts are described in the part of the method embodiment.

[0153] The various embodiments described in this specification are intended to be exemplary only. Each embodiment was chosen for illustration only, and not as a limitation of the scope of the disclosure. Numerous specific implementation details are described to provide a thorough understanding of the embodiments. However, in some instances, well-known methods, procedures, components and circuits have not been described in detail in order to avoid unnecessarily obscuring the embodiments. Reference throughout this specification to "one embodiment", "an embodiment", or "a specific embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Thus, the appearances of the phrase "in one embodiment" or "an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment. Furthermore, the particular features, structures, or characteristics can be combined in any suitable manner in one or more embodiments.

[0154] With regard to the devices in the above-described embodiments, where a specific manner for carrying out the operations of each module has been described in detail in the embodiments of the methods, the detailed description will not be repeated here.

[0155] An apparatus for processing personalized terms is provided in embodiments of the application. The apparatus includes a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors to include instructions for determining a user feature term from input content of a user; determining a related term corresponding to the user feature term according to a vector relationship; and determining a personalized term corresponding to the user according to the related term.

[0156] Figure 5 FIG. 8 is a block diagram of an apparatus 800 for processing personalized terms according to an example embodiment. The apparatus 800 can be a mobile phone, a computer, a digital broadcast terminal, a message receiving device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and the like, for example.

[0157] Referring to Figure 5 The apparatus 800 can include one or more of the following components: a processing component 802, a memory component 804, a power supply component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, a sensor component 814, and a communication component 816.

[0158] The processing component 802 usually controls overall operations of the apparatus 800, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 802 can include one or more processors 820 to execute instructions to complete all or part of steps of the methods described above. In addition, the processing component 802 can include one or more modules to facilitate interaction between the processing component 802 and other components. For example, the processing component 802 can include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0159] The memory 804 is configured to store various types of data to support the operation of the device 800. Examples of such data include instructions for any application or method operating on the device 800, contact data, phonebook data, messages, pictures, videos, and the like. The memory 804 can be implemented by any type of volatile or nonvolatile storage devices 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.

[0160] The power supply component 806 supplies electrical power for the various components of the device 800. The power supply component 806 can include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing electrical power for the device 800.

[0161] The multimedia component 808 includes a screen providing an output interface between the device 800 and a user. In some embodiments, the screen can include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, swiping, and gestures on the touch panel. The touch sensor can not only sense a boundary of a touching or swiping action, but also detect duration and pressure related to the touching or swiping action. In some embodiments, the multimedia component 808 includes a front camera and / or a rear camera. The front and / or rear camera can receive external multimedia data when the device 800 is in an operation mode such as a shooting mode or a video mode. Each of the front and rear camera can be a fixed optical lens system or have a focal length and optical zoom capability.

[0162] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when the device 800 is in an operation mode such as a call mode, a recording mode, and a voice input mode. The received audio signals can be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.

[0163] The I / O interface 812 provides an interface between the processing component 802 and peripheral interface modules such as a keyboard, a click wheel, buttons, and the like. The buttons can include, but are not limited to, a home button, a volume button, a start button, and a lock button.

[0164] The sensor component 814 includes one or more sensors for providing status assessments for various aspects of the device 800. For example, the sensor component 814 can detect an open / closed position of the device 800, relative positioning of components, such as a display and keypad of the device 800, a change in position of the device 800 or a component of the device 800, the presence or absence of user contact with the device 800, the orientation or acceleration / deceleration / g-force and a temperature change of the device 800. The sensor component 814 can include proximity sensor(s) configured to detect the presence of objects in a proximity without any physical contact. The sensor component 814 can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 814 can also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor or a temperature sensor.

[0165] The communication component 816 is configured to facilitate wired or wireless communication between the device 800 and another device. The device 800 can access a wireless network based on a corresponding communication standard, such as WiFi, 2G, or 3G, or a combination thereof. In an example embodiment, the communication component 816 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In an example embodiment, the communication component 816 can further include a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technology.

[0166] In an example embodiment, the device 800 can be implemented using 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, micro-controllers, microprocessors or other electronic modules to perform the above-described methods.

[0167] In an example embodiment, a non-transitory computer-readable storage medium, such as the memory 804 including instructions, is also provided, which can be executed by the processor 820 of the device 800 to implement the above-described methods. 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 disc, and an optical data storage device, etc.

[0168] Figure 6Figure 19 is a schematic diagram of a server in some embodiments of the application. The server 1900 can vary greatly in configuration and performance, and can include one or more central processing units (CPU) 1922 (e.g., one or more processors) and memory 1932, one or more storage media 1930 (e.g., one or more mass storage devices) storing application programs 1942 and data 1944. The memory 1932 and the storage media 1930 can be of the same type or different types, and can be volatile or non-volatile storage. The programs stored on the storage media 1930 can include one or more modules (not shown), each of which can include a series of instructions for operating the server. Further, the CPU 1922 can be configured to communicate with the storage media 1930 to execute the series of instructions of the storage media 1930 on the server 1900.

[0169] The server 1900 can also include one or more power supplies 1926, one or more wired or wireless network interfaces 1950, one or more input / output interfaces 1958, one or more keyboards 1956, and / or one or more operating systems 1941, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0170] A non-transitory computer-readable storage medium having instructions stored therein, which when executed by a processor of an apparatus (a server or a terminal), cause the apparatus to perform a method of processing personalized terms as shown in the above. Figure 2 or Figure 3 A method of processing personalized terms as shown in the above.

[0171] A non-transitory computer-readable storage medium having instructions stored therein, which when executed by a processor of an apparatus (a server or a terminal), cause the apparatus to perform a method of processing personalized terms, the method comprising: determining a user feature term from input content of a user; determining a related term corresponding to the user feature term according to a vector relationship; and determining a personalized term corresponding to the user according to the related term.

[0172] An embodiment of the application discloses A1, a method of processing personalized terms, the method comprising:

[0173] determining a user feature term from input content of a user;

[0174] determining a related term corresponding to the user feature term according to a vector relationship;

[0175] determining a personalized term corresponding to the user according to the related term.

[0176] A2. The method of A1, wherein the determining the user feature word from the input content of the user comprises:

[0177] matching the input content of the user with a feature word set to obtain the user feature word contained in the input content.

[0178] A3. The method of A2, wherein the feature word set comprises at least one of a domain word, a self-made word, and a cloud word.

[0179] A4. The method of any one of A1 to A3, wherein the determining the relevant word corresponding to the user feature word comprises:

[0180] determining a first word vector corresponding to the user feature word;

[0181] determining a second word vector matching the first word vector according to a vector retrieval method, and taking a word corresponding to the second word vector as the relevant word.

[0182] A5. The method of any one of A1 to A3, wherein the determining the personalized word corresponding to the user comprises:

[0183] sorting the relevant word according to a matching degree between the user feature word and the relevant word, and / or a weight parameter corresponding to the user feature word, and determining the personalized word from the relevant word according to a sorting result.

[0184] A6. The method of A5, wherein the weight parameter is determined by:

[0185] determining a frequency parameter of the user feature word in an input corpus in a first time range;

[0186] determining a distinguishing parameter of the user feature word in an input corpus in a second time range; the second time range is greater than the first time range;

[0187] determining the weight parameter according to the frequency parameter and the distinguishing parameter.

[0188] A7. The method of any one of A1 to A3, wherein the determining the personalized word corresponding to the user comprises:

[0189] sorting the relevant word according to a matching degree between the user feature word and the relevant word, and determining the personalized word from the relevant word according to a first sorting result; or

[0190] According to the weight parameter corresponding to the user characteristic word, the related word items corresponding to the plurality of user characteristic words are secondly sorted, and according to the second sorting result, the personalized word item is determined from the related word items corresponding to the plurality of user characteristic words.

[0191] According to the matching degree between the user characteristic word and the related word item and the weight parameter corresponding to the user characteristic word, the related word items corresponding to the plurality of user characteristic words are thirdly sorted, and according to the third sorting result, the personalized word item is determined from the related word items corresponding to the plurality of user characteristic words.

[0192] A8. The method of any one of A1 to A3, wherein the determining the related word item corresponding to the user characteristic word comprises:

[0193] According to at least two word vector types, at least two related word sets corresponding to the user characteristic word are determined;

[0194] According to the matching degree between the user characteristic word and the at least two related word sets, the related word item corresponding to the user characteristic word and the matching degree between the user characteristic word and the corresponding related word item are determined.

[0195] Embodiments of the present application disclose B9, a personalized word item processing device, comprising:

[0196] A user characteristic word determination module is configured to determine a user characteristic word from input content of a user;

[0197] A related word item determination module is configured to determine a related word item corresponding to the user characteristic word according to a vector relationship; and

[0198] A personalized word item determination module is configured to determine a personalized word item corresponding to the user according to the related word item.

[0199] B10. The device of B9, wherein the user characteristic word determination module comprises:

[0200] A matching module is configured to match the input content of the user with a characteristic word set to obtain a user characteristic word contained in the input content.

[0201] B11. The device of B10, wherein the characteristic word set comprises at least one of a domain word item, a self-made word item and a cloud word item.

[0202] B12. The device of any one of B9 to B11, wherein the related word item determination module comprises:

[0203] A first word vector determination module is configured to determine a first word vector corresponding to the user characteristic word.

[0204] a second word vector determination module configured to determine, according to the vector retrieval device, a second word vector matched with the first word vector, the second word vector corresponding to a relevant word entry.

[0205] B13. The apparatus according to any one of B9 to B11, wherein the personalized word entry determination module comprises:

[0206] a ranking module configured to rank the relevant word entries according to a matching degree between the user characteristic word and the relevant word entries, and / or a weight parameter corresponding to the user characteristic word, and determine a personalized word entry from the relevant word entries according to a ranking result.

[0207] B14. The apparatus according to B13, further comprising:

[0208] a frequency parameter determination module configured to determine a frequency parameter of the user characteristic word in an input corpus in a first time range;

[0209] a distinction parameter determination module configured to determine a distinction parameter of the user characteristic word in an input corpus in a second time range; the second time range is greater than the first time range;

[0210] a weight parameter determination module configured to determine the weight parameter according to the frequency parameter and the distinction parameter.

[0211] B15. The apparatus according to any one of B9 to B11, wherein the personalized word entry determination module comprises:

[0212] a first ranking module configured to rank the relevant word entries according to a matching degree between the user characteristic word and the relevant word entries, and determine a personalized word entry from the relevant word entries according to a first ranking result; or

[0213] a second ranking module configured to rank the relevant word entries corresponding to a plurality of user characteristic words according to a weight parameter corresponding to the user characteristic word, and determine a personalized word entry from the relevant word entries corresponding to the plurality of user characteristic words according to a second ranking result; or

[0214] a third ranking module configured to rank the relevant word entries corresponding to a plurality of user characteristic words according to a matching degree between the user characteristic word and the relevant word entries, and a weight parameter corresponding to the user characteristic word, and determine a personalized word entry from the relevant word entries corresponding to the plurality of user characteristic words according to a third ranking result.

[0215] B16. The apparatus according to any one of B9 to B11, wherein the relevant word entry determination module comprises:

[0216] The first determining module is configured to determine at least two sets of relevant words corresponding to the user characteristic word according to at least two word vector types.

[0217] The second determining module is configured to determine a relevant word item corresponding to the user characteristic word and a matching degree between the user characteristic word and the corresponding relevant word item according to a matching degree of the user characteristic word and the at least two sets of relevant words.

[0218] Embodiments of the present application disclose C17, an apparatus for processing personalized word items, comprising a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, and the one or more programs comprise instructions for performing the following operations:

[0219] determining a user characteristic word from input content of a user;

[0220] determining a relevant word item corresponding to the user characteristic word according to a vector relationship;

[0221] determining a personalized word item corresponding to the user according to the relevant word item.

[0222] C18, the apparatus according to C17, wherein the user characteristic word is determined from the input content of the user, comprising:

[0223] matching the input content of the user with a characteristic word set to obtain the user characteristic word contained in the input content.

[0224] C19, the apparatus according to C18, wherein the characteristic word set comprises at least one of a domain word item, a self-made word item and a cloud word item.

[0225] C20, the apparatus according to any one of C17 to C19, wherein the relevant word item corresponding to the user characteristic word is determined, comprising:

[0226] determining a first word vector corresponding to the user characteristic word;

[0227] determining a second word vector matched with the first word vector according to a vector retrieval apparatus, and a word item corresponding to the second word vector is taken as a relevant word item.

[0228] C21, the apparatus according to any one of C17 to C19, wherein the personalized word item corresponding to the user is determined, comprising:

[0229] According to the matching degree between the user characteristic word and the related word, and / or the weight parameter corresponding to the user characteristic word, the related words are sorted, and according to the sorting result, the personalized word is determined from the related words.

[0230] C22. The apparatus of claim 21, the one or more programs, when executed by the one or more processors, to further cause the one or more processors to perform:

[0231] determine a frequency parameter of the user characteristic word in an input corpus in a first time range;

[0232] determine a distinctiveness parameter of the user characteristic word in an input corpus in a second time range; the second time range is greater than the first time range;

[0233] determine the weight parameter according to the frequency parameter and the distinctiveness parameter.

[0234] C23. The apparatus of any of claims 17-19, the determining the personalized word corresponding to the user comprises:

[0235] performing a first sorting of the related words according to the matching degree between the user characteristic word and the related word, and determining the personalized word from the related words according to the first sorting result; or

[0236] performing a second sorting of the related words corresponding to the plurality of user characteristic words respectively according to the weight parameter corresponding to the user characteristic word, and determining the personalized word from the related words corresponding to the plurality of user characteristic words respectively according to the second sorting result; or

[0237] performing a third sorting of the related words corresponding to the plurality of user characteristic words respectively according to the matching degree between the user characteristic word and the related word, and the weight parameter corresponding to the user characteristic word, and determining the personalized word from the related words corresponding to the plurality of user characteristic words respectively according to the third sorting result.

[0238] C24. The apparatus of any of claims 17-19, the determining the related word corresponding to the user characteristic word comprises:

[0239] determining at least two related word sets corresponding to the user characteristic word according to at least two word vector types;

[0240] determining the related word corresponding to the user characteristic word, and the matching degree between the user characteristic word and the corresponding related word according to the matching degree between the user characteristic word and the at least two related word sets.

[0241] The embodiment of the present application discloses a D25, a machine readable medium, which stores instructions, when executed by one or more processors, cause an apparatus to perform the processing method of personalized terms as one or more of A1 to A8.

[0242] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. It is intended that the specification and examples be considered as exemplary only, with the true scope and spirit of the application being indicated by the following claims.

[0243] It is to be understood that the application is not limited to the specific structures described herein and illustrated in the accompanying drawings, and that since modifications and changes varied to fit particular applications will be apparent to those skilled in the art from the general nature of the application. The application resides in each individual feature or combination of features described herein.

[0244] The above-described embodiments are merely illustrative of the present application and are not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the present application.

[0245] The above provides a detailed introduction of the processing method of personalized terms, the processing device of personalized terms and the device for processing personalized terms. The principles and implementation manners of the present application are described by using specific examples. The above embodiment description is only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, the specific implementation manners and application ranges can be changed according to the idea of the present application. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for processing personalized terms, characterized in that, The method includes: Identify user-specific keywords from the user's input; Based on vector relationships, determine the relevant terms corresponding to the user feature words; Based on the relevant terms, determine the personalized terms corresponding to the user; Provide the user with input services corresponding to the personalized terms; The step of determining the relevant terms corresponding to the user feature words based on vector relationships includes: Based on at least two word vector types, determine at least two sets of related words corresponding to the user feature words; The at least two sets of related words are fused.

2. The method according to claim 1, characterized in that, The process of determining user-specific keywords from the user's input includes: The user's input content is matched with a set of feature words to obtain the user feature words contained in the input content.

3. The method according to claim 2, characterized in that, The feature word set includes at least one of domain terms, user-created terms, and cloud terms.

4. The method according to any one of claims 1 to 3, characterized in that, The step of determining the relevant terms corresponding to the user feature words includes: Determine the first word vector corresponding to the user feature word; Based on the vector retrieval method, a second word vector that matches the first word vector is determined, and the word corresponding to the second word vector is taken as the relevant word.

5. The method according to any one of claims 1 to 3, characterized in that, The process of determining the personalized terms corresponding to the user includes: Based on the matching degree between the user feature words and the related terms, and / or the weight parameters corresponding to the user feature words, the related terms are sorted, and personalized terms are determined from the related terms based on the sorting results.

6. The method according to claim 5, characterized in that, The weight parameters are determined through the following steps: Determine the frequency parameters of user feature words in the input corpus within the first time range; Determine the distinguishing parameters of user feature words in the input corpus within a second time range; the second time range is greater than the first time range; The weighting parameter is determined based on the frequency parameter and the differentiation parameter.

7. The method according to any one of claims 1 to 3, characterized in that, The process of determining the personalized terms corresponding to the user includes: Based on the matching degree between the user feature words and the related terms, the related terms are first sorted, and personalized terms are determined from the related terms based on the first sorting result; or Based on the weight parameters corresponding to the user feature words, a second sort is performed on the relevant terms corresponding to the multiple user feature words, and personalized terms are determined from the relevant terms corresponding to the multiple user feature words based on the second sorting results; or Based on the matching degree between the user feature words and the related terms, and the weight parameters corresponding to the user feature words, the related terms corresponding to the multiple user feature words are sorted in a third order, and personalized terms are determined from the related terms corresponding to the multiple user feature words based on the third sorting results.

8. A device for processing personalized entries, characterized in that, include: The user feature word determination module is used to determine user feature words from the user's input content; The related term determination module is used to determine the related terms corresponding to the user feature words based on vector relationships; as well as The personalized term determination module is used to determine the personalized term corresponding to the user based on the relevant terms; The device is also used for: Provide the user with input services corresponding to the personalized terms; The related term determination module includes: Based on at least two word vector types, determine at least two sets of related words corresponding to the user feature words; The at least two sets of related words are fused.

9. An apparatus for processing personalized entries, characterized in that, It includes memory and one or more programs, wherein one or more programs are stored in memory and configured to be executed by one or more processors. The one or more programs contain instructions for performing the following operations: Identify user-specific keywords from the user's input; Based on vector relationships, determine the relevant terms corresponding to the user feature words; Based on the relevant terms, determine the personalized terms corresponding to the user; Provide the user with input services corresponding to the personalized terms; The step of determining the relevant terms corresponding to the user feature words based on vector relationships includes: Based on at least two word vector types, determine at least two sets of related words corresponding to the user feature words; The at least two sets of related words are fused.

10. A machine-readable medium having instructions stored thereon that, when executed by one or more processors, cause a device to perform the processing method for personalized terms as described in one or more of claims 1 to 7.

Images