File filling method and device, electronic equipment and storage medium

Abstract

The invention provides an archive filling method and device, electronic equipment and a storage medium. According to the method, after the entry set is obtained, the entries can be automatically classified to obtain the actual template entries and the actual information entries, the comprehensive matching degree of the actual information entries is calculated, the first target actual information entry serves as the first target archive field value to be filled in to obtain the archive, manual filling is not needed, and efficiency is improved; in a filling process, each entry is matched with each template entity library and each information entity library to determine an actual template entry and an actual information entry, so that the accuracy of the actual template entry and the actual information entry is relatively high; and then the comprehensive matching degree of each actual information entry and the matched archive field group is taken as a determination basis of the first target actual information entry, and the comprehensive matching degree is measured in combination with two aspects, so that the matching accuracy is relatively high, and the final filling accuracy is also relatively high. In other words, the file filling efficiency and quality are high.

Description

Technical Field

[0001] The present application relates to the field of logistics technology, and in particular to a file filling method, device, electronic device and storage medium. Background Art

[0002] With the development of standardization in the express delivery industry, more and more content needs to be filled in customer files, and customer file creation has become a tedious part of the entire business process. Currently, customer files are mostly filled in manually, which is inefficient and prone to irregular filling and errors, affecting the quality of file creation.

[0003] Therefore, the current customer file creation method has technical problems such as low filling efficiency and quality, and needs to be improved. Summary of the invention

[0004] The embodiments of the present application provide a file filling method, device, electronic device and storage medium to alleviate the technical problems of low filling efficiency and quality in the current customer file creation method.

[0005] To solve the above technical problems, the present application provides the following technical solutions:

[0006] This application provides a method for filling out a file, including:

[0007] Acquire a vocabulary entry set, wherein the vocabulary entry set includes a plurality of vocabulary entries;

[0008] Acquire a file template, the file template comprising a plurality of file field groups, each of the file field groups comprising a file field name and a file field value forming a corresponding relationship, each of the file field names in the file template respectively corresponds to a template entity library, and each of the file field values ​​in the file template respectively corresponds to an information entity library;

[0009] Match each of the terms with each of the template entity libraries to obtain multiple successfully matched target template entity libraries; match each of the terms with each of the information entity libraries to obtain multiple successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms;

[0010] According to the actual template terms, the actual information terms, and the correspondence between the file field names and the file field values, determining the matching relationship between each of the actual information terms and each of the actual template terms, and the matching relationship between each of the actual information terms and each of the file field groups;

[0011] Obtaining the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtaining the distance matching degree between each of the actual information terms and the matching actual template terms, and obtaining the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree;

[0012] The comprehensive matching degree of all actual information terms matched by each first target file field group is obtained, and the first target actual information term corresponding to the highest comprehensive matching degree is used as the first target file field value of each first target file field group to obtain the file of the target object.

[0013] At the same time, the embodiment of the present application also provides a file filling device, including:

[0014] A first acquisition module is used to acquire a vocabulary set, wherein the vocabulary set includes a plurality of vocabulary entries;

[0015] A second acquisition module is used to acquire a file template, wherein the file template includes a plurality of file field groups, each of which includes a file field name and a file field value that form a corresponding relationship, each of which file field names in the file template corresponds to a template entity library, and each of which file field values ​​in the file template corresponds to an information entity library;

[0016] A matching module is used to match each of the terms with each of the template entity libraries to obtain multiple successfully matched target template entity libraries; each of the terms is matched with each of the information entity libraries to obtain multiple successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms;

[0017] a determination module, for determining a matching relationship between each of the actual information terms and each of the actual template terms, and a matching relationship between each of the actual information terms and each of the file field groups, according to the actual template terms, the actual information terms, and the corresponding relationship between the file field name and the file field value;

[0018] The first obtaining module is used to obtain the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtain the distance matching degree between each of the actual information terms and the matching actual template terms, and obtain the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree;

[0019] The second obtaining module is used to obtain the comprehensive matching degree of all actual information entries matched by each first target file field group, and use the first target actual information entry corresponding to the highest comprehensive matching degree as the first target file field value of each first target file field group to obtain the file of the target object.

[0020] The present application also provides an electronic device, comprising a memory and a processor; the memory stores an application, and the processor is used to run the application in the memory to execute the steps in any of the above-mentioned file filling methods.

[0021] An embodiment of the present application provides a computer-readable storage medium, which stores a plurality of instructions suitable for loading by a processor to execute the steps in the above-mentioned file filling method.

[0022] Beneficial effects: The present application provides a method, device, electronic device and storage medium for filling in a file. After acquiring a set of entries, the method can automatically perform the steps of classifying actual template entries and actual information entries, calculating the comprehensive matching degree of actual information entries, and filling in the first target actual information entry as the first target file field value. The whole process does not require manual filling, so the efficiency of filling in the file is effectively improved; in addition, in the filling process of the present application, each entry in the entry set is first matched with each template entity library and each information entity library, and the actual template entry and actual information entry that are successfully matched are determined, so the accuracy of the actual template entry and the actual information entry is high, and then the comprehensive matching degree of each actual information entry and the matched first target file field group is used as the basis for determining the first target actual information entry. Since the comprehensive matching degree combines the similarity matching degree of the actual information entry and the matched target information entity library itself, and the distance matching degree of the actual information entry and the matched actual template entry, the matching accuracy is high, so that the accuracy of the first target file field value finally filled in is also high. That is, the filling efficiency and quality of the file in the present application are both high. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The technical solution and other beneficial effects of the present application will be made apparent by describing in detail the specific implementation methods of the present application in conjunction with the accompanying drawings.

[0024] Figure 1 A flowchart of the file filling method provided in the embodiment of the present application.

[0025] Figure 2 Schematic diagram of the implementation method of each stage of the file filling method provided in the embodiment of the present application.

[0026] Figure 3 This is a schematic diagram of the file content of the target object before review in an embodiment of the present application.

[0027] Figure 4 Schematic diagram of the file content of the target object in the review stage in the embodiment of the present application.

[0028] Figure 5 Schematic diagram of the file content of the target object after review in the embodiment of the present application.

[0029] Figure 6 It is a structural schematic diagram of the file filling device provided in the embodiment of the present application.

[0030] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0032] The embodiments of the present application provide a method and device for filling out a file, an electronic device, and a computer-readable storage medium, wherein the device for filling out a file can be integrated into an electronic device, and the electronic device can be a server, a terminal, or other device.

[0033] See also Figure 1 and Figure 2 , are respectively a flow chart of the file filling method provided in the embodiment of the present application and a schematic diagram of the implementation method of each stage, the method specifically includes:

[0034] S1: Obtain a term set, where the term set includes multiple terms.

[0035] The file that needs to be filled in this application is established in the customer management system, which can be displayed to the user through the application in the user terminal. When filling in the file, a set of entries containing multiple entries is first obtained through the application, and then each entry in the set of entries is used as the source of filling in the file content. After filling in is completed, the filled-in file content is displayed through the visual interface of the application.

[0036] In one embodiment, S1 specifically includes:

[0037] S11: Obtaining original voice data.

[0038] The various stages of the file filling method in this application are: start input - end input - review - generate file. In the first stage, the original voice data can be the voice data input in real time through the application, or the conversation recording data uploaded through the application. Therefore, when obtaining the original voice data, it can be obtained in real time during the normal conversation between the staff and the customer who needs to create a file, or the conversation between the staff and the customer who needs to create a file can be recorded first, and then the record can be obtained later. That is, the scenario applicable to the file filling method in this application is relatively flexible.

[0039] S12: Process the original speech data through the human voice model to obtain original text data.

[0040] Since the content that needs to be filled in the file is text content, the original voice data needs to be processed first in this step and converted into original text data. In addition to human voice, the original voice data usually contains some noise, which will interfere with the file filling process. Therefore, it is necessary to use the trained human voice model to identify and distinguish the human voice and noise in the original voice data, remove noise and extract the human voice, and then convert the human voice into original text data. In addition, before processing with the human voice model, the original voice data can also be pre-processed, such as noise reduction, filtering, normalization, etc., to reduce the interference caused by noise and make the recognition result more accurate.

[0041] S13: Perform word segmentation processing on the original text data to obtain multiple entries.

[0042] The original text data usually contains a text sequence, which is composed of multiple entries. Since some of the entries need to be referenced and filled in when filling out the file later, the original text data needs to be further processed in this step to obtain multiple entries, and all the entries constitute an entry set.

[0043] After processing the original speech data within a certain time period, the obtained original text data is the text sequence of the human voice within that time period. By performing word segmentation on these text sequences, they can be divided into multiple word entries. For example, the text sequence includes "The phone number of Zhang xx is 176xxxx7777". After word segmentation, the multiple word entries obtained are "Zhang xx", "of", "phone number", "is", "176xxxx7777" respectively. Various methods can be used for word segmentation. For example, a preset dictionary can be used. The dictionary contains various reference word entries. A single character or several consecutive character strings in the text sequence are matched with each reference word entry in the dictionary. If the character or string can be successfully matched with a certain reference word entry, then the character or string is segmented to obtain a word entry. Another example is that an N-gram language model can be used. This language model performs various forms of word entry segmentation on the input text sequence and calculates the probability of each form of word entry segmentation result, and selects the segmentation method with the highest probability as the final segmentation result to obtain multiple word entries. Another example is that a deep learning model can be used. The deep learning model is trained with a large number of labeled word entries, and then the text sequence is input into the trained deep learning model to obtain the word segmentation result.

[0044] In one embodiment, after S13, it further includes:

[0045] S14: Extract the attribute features of each word entry. The attribute features include part of speech, grammatical function, character type, and length.

[0046] After word segmentation, the attribute features of each word entry are extracted, and these attribute features are labeled to the corresponding word entries. All the labeled word entries form a word entry set. The attribute features can include the part of speech, grammatical function, character type, length, etc. of the word entry. Among them, the part of speech includes nouns, verbs, adjectives, auxiliary words, pronouns, numerals, etc., the grammatical function includes subject, predicate, object, etc., the character type includes characters, letters, numbers, etc., and the length is the total length of all characters that make up the word entry.

[0047] For example, the attribute features of "Zhang xx" are noun, character, length 3, the attribute features of "of" are auxiliary word, character, length 1, the attribute features of "phone number" are noun, subject, character, length 2, the attribute features of "is" are verb, predicate, character, length 2, and the attribute features of "176xxxx7777" are noun, object, number, length 11.

[0048] S15: Label each word entry according to the attribute features, and generate a word entry set based on the multiple labeled word entries.

[0049] After extracting the attribute features of these word entries, the attribute features are labeled to each word entry to obtain the labeled data of each word entry.

[0050] S2: Obtain an archive template, the archive template includes multiple archive field groups, each archive field group includes an archive field name and an archive field value that form a corresponding relationship, each archive field name in the archive template corresponds to a template entity library, and each archive field value in the archive template corresponds to an information entity library.

[0051] The archive template is a text frame with a fixed format and structure. Multiple archive field groups are pre-set in the archive template. Each archive field group includes an archive field name and an archive field value that form a corresponding relationship. The archive field name refers to the specific name of the archive field, and the archive field value refers to the text content that needs to be filled in. Before the archive template is filled in, the content of each archive field name in the archive template is fixed and already exists, while the archive field value corresponding to each archive field name has not been filled in and the content is empty. After filling in the archive field values ​​in the second stage, the archive template will be displayed to the user as a newly created archive.

[0052] For example, the file field name 1 of the first file field group in the file template is "customer type", and its corresponding file field value is 1; the file field name 2 of the second file field group is "customer name", and its corresponding file field value is 2; the file field name 3 of the third file field group is "primary industry", and its corresponding file field value is 3; the file field name 4 of the fourth file field group is "secondary industry", and its corresponding file field value is 4; the file field name 5 of the fifth file field group is "primary sender", and its corresponding file field value is 5; the file field name 6 of the sixth file field group is "contact person", and its corresponding file field value is 6; the file field name 7 of the seventh file field group is "contact information", and its corresponding file field value is 7. Before filling in, the file template contains the file field names 1 to 7 of the above 7 file field groups, and the corresponding file field values ​​1 to 7 are all empty. After filling in all or part of the file field values, the file template will be displayed to the user as a newly created file. For example, fill in the file field value 1 with "Enterprise", fill in the file field value 6 with "Zhang xx", and fill in the file field value 7 with "176xxxx7777". After filling in, the created and displayed file is as follows: Figure 3 shown.

[0053] Each file field name in the file template has a corresponding template entity library, which contains all possible entries with semantics similar to the file field name. For example, file field name 7 is "contact information", and its corresponding template entity library is the contact information entity library, which contains "contact information", "telephone number", "mobile phone number", "telephone", "number" and other entries with semantics similar to "contact information". These entries are reference template entries. Each file field value in the file template also has a corresponding information entity library, which contains all possible contents of the file field value. For example, file field value 7 is a specific contact method, and its corresponding information entity library is the mobile phone number entity library, which contains a large number of telephone numbers such as "173xxxx1211", "136xxxx51111", etc. These entries are reference information entries.

[0054] S3: Match each term with each template entity library respectively to obtain multiple successfully matched target template entity libraries; match each term with each information entity library respectively to obtain multiple successfully matched target information entity libraries; among them, the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms.

[0055] For each entry in the entry set, the entry is matched with each template entity library respectively. If a template entity library has a high degree of similarity matching with one or more entries, the template entity library is determined as a target template entity library, and the entries with a high degree of similarity matching with the target template entity library are determined as actual template entries that have successfully matched. Similarly, for each entry in the entry set, the entry is matched with each information entity library respectively. If a certain information entity library has a high degree of similarity matching with one or more entries, the information entity library is determined as a target information entity library, and the entries with a high degree of similarity matching with the target information entity library are determined as actual information entries that have successfully matched.

[0056] In one embodiment, S3 specifically includes:

[0057] S31: Match each term with a plurality of reference template terms in each template entity library to obtain each first similarity matching degree between each term and each template entity library.

[0058] When each entry is matched with each template entity library, the first similarity matching degree is used as the matching degree measurement standard, and the similarity matching degree of each entry is calculated with multiple reference template entries in the template entity library. Assuming that the archive template corresponds to n template entity libraries, n first similarity matching degrees can be obtained.

[0059] S32: When there is a target first similarity matching degree greater than a first preset threshold among the first similarity matching degrees, the template entity library with the target first similarity matching degree is determined as the successfully matched target template entity library, and the entry that successfully matches the target template entity library is determined as the actual template entry.

[0060] For the n first similarity matching degrees calculated for a certain term, if there is a target first similarity matching degree greater than the first preset threshold value among the n first similarity matching degrees, it means that there is a template entity library that successfully matches the term. The value of the first preset threshold value can be set as needed, such as 50%. Since the target first similarity matching degree is obtained by calculating the similarity matching degree between the term and a certain template entity library, the template entity library that obtains the target first similarity matching degree is determined as the target template entity library that successfully matches, and the term is determined as the actual template term that successfully matches the target template entity library.

[0061] S33: Match each term with each information entity library respectively to obtain each second similarity matching degree between each term and each information entity library.

[0062] When each term is matched with each information entity library, the second similarity matching degree is used as the matching degree measurement standard, and the similarity matching degree of each term is calculated with multiple reference information terms in the information entity library. Assuming that the archive template corresponds to n information entity libraries, n second similarity matching degrees can be obtained.

[0063] S34: When there is a target second similarity matching degree greater than the first preset threshold among the second similarity matching degrees, the information entity library with the target second similarity matching degree is determined as the successfully matched target information entity library, and the entry successfully matched with the target information entity library is determined as the actual information entry.

[0064] For the n second similarity matching degrees calculated for a certain term, if there is a target second similarity matching degree greater than the first preset threshold value among the n second similarity matching degrees, it means that there is a certain information entity library that successfully matches the term. Since the target second similarity matching degree is obtained by calculating the similarity matching degree between the term and a certain information entity library, the information entity library that obtains the target second similarity matching degree is determined as the target information entity library that successfully matches, and the term is determined as the actual information term that successfully matches the target information entity library.

[0065] It should be noted that, since the content of each archive field name is usually quite different, when a term successfully matches one of the template entity libraries, the probability of successfully matching with other template entity libraries is low, so an actual template term usually has only one successfully matched target template entity library. Similarly, since the content of each archive field value is usually quite different, when a term successfully matches one of the information entity libraries, the probability of successfully matching with other information entity libraries is low, so an actual information term usually has only one successfully matched target information entity library.

[0066] However, a target template entity library may have two or more successfully matched actual template entries, and a target information entity library may have two or more successfully matched actual information entries. For example, if two mobile phone numbers are mentioned in the original voice data, the entries corresponding to these two mobile phone numbers will both be successfully matched with the mobile phone number entity library. The mobile phone number entity library is determined as a target information entity library, and the target information entity library has two successfully matched actual information entries.

[0067] In addition, since the content of the archive field name and the content of the archive field value in the archive template are also quite different, if an entry has a high degree of match with a certain template entity library during matching, its match with all information entity libraries will be lower. Therefore, under normal circumstances, an entry will only be determined as an actual template entry, or will only be determined as an actual information entry, and will not be determined as both an actual template entry and an actual information entry at the same time.

[0068] For example, the term "telephone" is matched with n template entity libraries to obtain n first similarity matching degrees, among which there is one first similarity matching degree greater than the first preset threshold, and the other n-1 first similarity matching degrees are all less than the first preset threshold. Then the first similarity matching degree greater than the first preset threshold is used as the target first similarity matching degree, and the target first similarity matching degree is obtained by matching the term "telephone" with the contact entity library. The contact entity library is the target template entity library that matches successfully, and the term "telephone" is an actual template term that matches successfully with the contact entity library. At the same time, the term "telephone" is matched with n information entity libraries to obtain n second similarity matching degrees. Assuming that all second similarity matching degrees are less than the first preset threshold, there is no target second similarity matching degree, and the term "telephone" will not be an actual information term.

[0069] For another example, the term "176xxxx7777" is matched with each template entity library to obtain n first similarity matching degrees. Since all the first similarity matching degrees are less than the first preset threshold, there is no target first similarity matching degree, and the term "176xxxx7777" will not be an actual template term. At the same time, the term "176xxxx7777" is matched with each information entity library to obtain n second similarity matching degrees, of which there is 1 second similarity matching degree greater than the first preset threshold, and the other n-1 second similarity matching degrees are all less than the first preset threshold. The second similarity matching degree greater than the first preset threshold is used as the target second similarity matching degree. The target second similarity matching degree is obtained by matching the term "176xxxx7777" with the mobile phone number entity library. The mobile phone number entity library is the target information entity library that has been successfully matched, and the term "176xxxx7777" is an actual information term that has been successfully matched with the mobile phone number entity library.

[0070] In one embodiment, after step S32, the method further includes:

[0071] S32a: Obtain the first term attribute matching condition of each target template entity library, the first term attribute matching condition includes the first common attribute feature of all reference template terms in the target template entity library, the first common attribute feature includes part of speech, grammatical function, character type and length.

[0072] Each template entity library contains all possible entries that are semantically similar to a certain file field name, that is, reference template entries, and the first common attribute features of these reference template entries can be used as the first entry attribute matching conditions of the template entity library. For example, the contact information entity library contains various entries that are semantically similar to the file field name 7 "contact information", such as "telephone number", "mobile number", "telephone", "number", etc. The first common attribute features of these reference template entries are nouns, subjects, text, and character lengths not exceeding 4, etc. These first common attribute features form the first entry attribute matching conditions of the contact information entity library.

[0073] S32b: Based on the first term attribute matching condition of each target template entity library, the attribute characteristics of all actual template terms matched by each target template entity library are respectively checked, and the actual template terms whose attribute characteristics do not meet the first common attribute characteristics are filtered out from all actual template terms.

[0074] In the above steps, the attribute features of each term are extracted. If an actual template term has a high first similarity match with its matching target template entity library, but the attribute features of the actual template term do not meet the first term attribute matching condition of the target template entity library, then the actual template term and the target template entity library are still not considered to be matched successfully, and the actual template term may be misjudged as an actual template term, which does not actually belong to the actual template term. Therefore, when determining the final actual template term and the target template entity library, it is necessary to add a verification process to verify the attribute features of all matched actual template terms using the first term attribute matching condition of the target template entity library. If the verification result shows that any attribute feature of an actual template term does not meet the first common attribute feature in the first term attribute matching condition, then the actual template term is marked as discarded and removed from all actual template terms.

[0075] Through the above-mentioned method, the actual template terms determined according to the first similarity matching degree are further screened by means of attribute feature verification, and the actual template terms that are finally left have a higher accuracy.

[0076] In one embodiment, after step S34, the method further includes:

[0077] S34a: Obtain the second term attribute matching condition of each target information entity library, the second term attribute matching condition includes the second common attribute characteristics of all reference information terms in the target information entity library, the second common attribute characteristics include part of speech, grammatical function, character type and length.

[0078] Similarly, each information entity library contains all possible entries that are semantically similar to a certain archive field value, that is, reference information entries, and the second common attribute characteristics of these reference information entries can be used as the second entry attribute matching conditions of the information entity library. For example, the telephone number entity library contains a large number of telephone numbers such as "173xxxx1211" and "136xxxx51111". The second common attribute characteristics of these reference information entries are nouns, objects, numbers, and a length of 11. These second common attribute characteristics form the second entry attribute matching conditions of the telephone number entity library.

[0079] S34b: Based on the second term attribute matching condition of each target information entity library, the attribute characteristics of all actual information terms matched by each target information entity library are respectively checked, and the actual information terms whose attribute characteristics do not meet the second common attribute characteristics are filtered out from all actual information terms.

[0080] In the above steps, the attribute features of each term are extracted. If an actual information term has a high second similarity matching degree with the target information entity library it matches, but the attribute features of the actual information term do not meet the second term attribute matching condition of the target information entity library, then the actual information term and the target information entity library are still not considered to be matched successfully, and the actual information term may be misjudged as an actual information term, which does not actually belong to the actual information term. Therefore, when determining the final actual information term and the target information entity library, it is necessary to add a verification process to verify the attribute features of all matched actual information terms using the second term attribute matching condition of the target information entity library. If the verification result shows that any attribute feature of a certain actual information term does not meet the second common attribute feature in the second term attribute matching condition, then the actual information term is marked as discarded and removed from all actual information terms.

[0081] In the above manner, the actual information terms determined according to the second similarity matching degree are further screened by means of attribute feature verification, and the accuracy of the actual information terms finally remaining is relatively high.

[0082] In the above embodiment, the first similarity matching degree and the second similarity matching degree are calculated first, and then the final actual template entry and the actual information entry are screened based on the first preset threshold and the first and second entry attribute matching conditions. The above order can also be reversed, that is, multiple actual template entries and multiple actual information entries can be screened based on the first and second entry attribute matching conditions, and then the first similarity matching degree and the second similarity matching degree are calculated, and finally further screening is performed based on the first preset threshold to obtain the final actual template entry and the actual information entry. In this way, each entry can be screened once before matching. If a certain entry does not meet the entry attribute matching conditions of a certain entity library, the similarity matching degree of the entry and the entity library can be omitted in the subsequent calculation, thereby saving computing resources while ensuring a certain accuracy.

[0083] S4: According to the correspondence between the actual template terms, the actual information terms, and the file field names and file field values, the matching relationship between each actual information term and each actual template term, and the matching relationship between each actual information term and each file field group are determined.

[0084] In the previous step, the actual template entries that successfully match each target template entity library are obtained, and the actual information entries that successfully match each target information entity library are obtained, then a matching relationship is formed between each actual template entry and each target template entity library, and a matching relationship is formed between each actual information entry and each target information entity library. Each target template entity library has a corresponding relationship with a file field name, and each target information entity library has a corresponding relationship with a file field value. With each target template entity library as a reference, the file field name that matches each actual template entry can be determined, and with each target information entity library as a reference, the file field value that matches each actual information entry can be determined.

[0085] For each archive field group, there is also a corresponding relationship between the archive field name and the archive field value contained therein. Then, with the archive field name and the archive field value in each archive field group as reference, the matching relationship between each actual information entry and each actual template entry, and the matching relationship between each actual information entry and each archive field group can be determined. Specifically, when the archive field name and the archive field value in a certain archive field group match an actual template entry and an actual information entry respectively, the actual information entry has a matching relationship with the actual template entry, and the actual information entry also has a matching relationship with the archive field group. When there is a matching actual information entry for the archive field value in a certain archive field group, but there is no matching actual template entry for the archive field name, the actual information entry has a matching relationship with the archive field group, but the actual information entry does not have a matching actual template entry.

[0086] For example, after matching, the term "telephone" is determined to be an actual template term, and its successfully matched target template entity library is the contact information entity library. After matching, the term "176xxxx7777" is determined to be an actual information term, and its successfully matched target information entity library is the telephone number entity library. Since the term "telephone" matches the contact information entity library, and the contact information entity library corresponds to the file field name 7 "contact information", the term "telephone" is the actual template term that matches the file field name 7 "contact information". Since the term "176xxxx7777" matches the telephone number entity library, and the telephone number entity library corresponds to the file field value 7, the term "176xxxx7777" is the actual information term that matches the file field value 7.

[0087] Since the file field name 7 "Contact information" and the file field value 7 both belong to the 7th file field group, it can be determined that the actual template entry matched by the actual information entry "176xxxx7777" is "Telephone", and the file field group matched by the actual information entry "176xxxx7777" is the 7th file field group.

[0088] In one embodiment, S4 specifically includes:

[0089] S41: Obtain the first entity type of each target template entity library, the first entity type is used to represent the corresponding relationship between each target template entity library and each file field name, and obtain the first entry type of each actual template entry according to each first entity type, the first entry type is used to represent the matching relationship between each actual template entry and each file field name.

[0090] Each template entity library has a corresponding file field name. In order to distinguish which template entity library corresponds to which file field name, each template entity library can be pre-labeled with a first entity type. The first entity type is used to indicate the correspondence between each template entity library and each file field name. For example, if the first entity type of a template entity library is "Contact information", it means that the template entity library is the template entity library corresponding to the file field name 7 "Contact information".

[0091] Since the first entity type is used to represent the corresponding relationship between each target template entity library and each file field name, and there is also a matching relationship between each target template entity library and each actual template entry, each target template entity library is used as a reference to determine the file field name that each actual template entry matches, that is, to obtain the matching relationship between each actual template entry and each file field name. In order to clearly represent the matching relationship, each actual template entry can be marked with the first entry type, and the content of the first entry type is determined by the first entity type, specifically: when the first entity type of the target template entity library is "contact information", it means that the target template entity library is the template entity library corresponding to the file field name 7 "contact information", then the first entry type of the actual template entry that successfully matches the target template entity library can also be marked as "contact information", indicating that the actual template entry is the actual template entry that matches the file field name 7 "contact information".

[0092] S42: Obtain the second entity type of each target information entity library, the second entity type is used to represent the corresponding relationship between each target information entity library and each archive field value, and obtain the second entry type of each actual information entry according to each second entity type, the second entry type is used to represent the matching relationship between each actual information entry and each archive field value.

[0093] Each information entity library has a corresponding file field value. In order to distinguish which information entity library corresponds to which file field value, each information entity library can be marked with a second entity type in advance. The second entity type is used to indicate the correspondence between each information entity library and each file field value. For example, if the second entity type of an information entity library is "telephone number", it means that the information entity library is the information entity library corresponding to the file field value 7.

[0094] Since the second entity type is used to represent the corresponding relationship between each information entity library and each archive field value, and there is also a matching relationship between each target information entity library and each actual information entry, each target information entity library is used as a reference to determine the archive field value that matches each actual information entry, that is, to obtain the matching relationship between each actual information entry and each archive field value. In order to clearly represent the matching relationship, each actual information entry can be marked with a second entry type, and the content of the second entry type is determined by the second entity type, specifically: when the second entity type of the target information entity library is "telephone number", it means that the information entity library is the information entity library corresponding to the archive field value 7, then the second entry type of the actual information entry that successfully matches the target information entity library can also be marked as "telephone number", indicating that the actual information entry is the actual information entry that matches the archive field value 7.

[0095] S43: According to the first entry type, the second entry type, and the correspondence between the file field name and the file field value, determine the matching relationship between each actual information entry and each actual template entry, and the matching relationship between each actual information entry and each file field group.

[0096] Since the first entry type can represent the matching relationship between each actual template entry and each file field name, and the second entry type can represent the matching relationship between each actual information entry and each file field value, and for each file field group, there is also a corresponding relationship between the file field name and the file field value contained therein, specifically, the file field name 1 of the first file field group corresponds to the file field value 1, and the file field name 7 of the seventh file field group corresponds to the file field value 7. Therefore, according to the first entry type, the second entry type, and the corresponding relationship between the file field name and the file field value, it can be determined which file field group each actual template entry matches, which file field group each actual information entry matches, and if a certain file field group matches an actual template entry and an actual information entry at the same time, then the matching relationship between the actual template entry and the actual information entry. If a certain file field group only matches a certain actual information entry and does not match any actual template entry, then there is no matching actual template entry for the actual information entry.

[0097] Through the above method, a first entity type is set in advance for each template entity library, and a second entity type is set for each information entity library. After obtaining the matching relationship between each actual template entry and each archive field name, the matching relationship is annotated to each actual template entry in the form of the first entry type. After obtaining the matching relationship between each actual information entry and each archive field value, the matching relationship is annotated to each actual information entry in the form of the second entry type. Based on the annotated information, each actual template entry and the actual information entry can be accurately distinguished, and then the actual template entry that matches each actual information entry and the archive field group that matches each actual information entry can be quickly determined.

[0098] S5: Obtain the similarity matching degree between each actual information entry and the matching target information entity library, and obtain the distance matching degree between each actual information entry and the matching actual template entry. According to the similarity matching degree and the distance matching degree, obtain the comprehensive matching degree between each actual information entry and the matching archive field group.

[0099] In the above matching process, the similarity matching degree between each term and each template entity library and each information entity library is calculated. Then, for each actual information term, the similarity matching degree between the actual information term and the successfully matched target information entity library can be directly obtained, and the distance matching degree between each actual information term and the matched actual template term can be obtained. The distance matching degree is calculated by the Pearson correlation coefficient, and the specific formula is as follows:

[0100]

[0101] Among them, Q and D represent the vector representation of the actual template term and the actual information term respectively, and the calculated coefficient r(Q, D) converted into a percentage is the distance matching degree between the two. Under normal circumstances, if the actual information term and the actual template term that match the same archive field group are close in distance, their credibility and accuracy are higher. If the distance between the two is far, the credibility and accuracy will also decrease accordingly. Therefore, this application also uses the distance matching degree between the two as a factor to measure the accuracy of filling in. If there is no matching actual template term for a certain actual information term, the distance matching degree of the actual information term is recorded as 0, and then the comprehensive matching degree is calculated.

[0102] After obtaining the similarity matching degree and distance matching degree of each actual information term, the two are weighted and calculated to obtain the comprehensive matching degree of each actual information term and the matching archive field group. The comprehensive matching degree is used to measure the matching situation of each actual information term and the matching archive field group in the archive template. Specifically, the similarity matching degree has a first weight, and the distance matching degree has a second weight. The specific values ​​of the first weight and the second weight can be set as needed. For example, the first weight can be set to 80% and the second weight can be set to 20%. The calculation formula of the comprehensive matching degree is as follows:

[0103] Comprehensive matching degree = 80% similarity matching degree + 20% distance matching degree

[0104] In one embodiment, S5 specifically includes:

[0105] S51: Compare the distance matching degree of each actual information entry with a second preset threshold, and remove the actual information entries whose distance matching degree is not greater than the second preset threshold from all actual information entries, where the second preset threshold is less than the first preset threshold.

[0106] After calculating the distance matching degree of each actual information term, the distance matching degree is compared with the second preset threshold value, and the value of the second preset threshold value can be set as needed, for example, 30%. If the distance matching degree of an actual information term is greater than the second preset threshold value, it can be retained, and if the distance matching degree of an actual information term is not greater than the second preset threshold value, it needs to be marked as discarded and removed.

[0107] Since the first preset threshold is used to measure the similarity matching degree between each term and the template entity library, or to measure the similarity matching degree between each term and the information entity library, if the similarity matching degree of a certain term is low, then even if the term has a high distance matching degree and the calculated comprehensive matching degree is high, it is very likely that it will not be the correct actual template term or actual information term. In other words, if the first preset threshold is set too low, the final error may be large, so it is necessary to set the value of the second preset threshold slightly higher to indicate that the accuracy requirement for the similarity matching degree is higher. The second preset threshold is used to measure the distance between the actual information term and the matching actual template term. Considering that different people have different speaking habits, the distance matching degree measurement standard can be slightly lower than the similarity matching degree measurement standard. Taking the above considerations into account, the second preset threshold can be set to be smaller than the first preset threshold.

[0108] S52: Obtain a first weight of the similarity matching degree and a second weight of the distance matching degree.

[0109] The file filling method in the present application is completed by inputting the entries into the file filling model, and the first weight of the similarity matching degree and the second weight of the distance matching degree are parameters in the trained file filling model. Before training, the file filling model includes the first initial weight of the similarity matching degree and the second initial weight of the distance matching degree. After training the file filling model, the adjusted first initial weight is determined as the first weight of the similarity matching degree, and the adjusted second initial weight is determined as the second weight of the distance matching degree. In the embodiment of the present application, the first weight obtained after adjustment is 80%, and the second weight obtained after adjustment is 20%.

[0110] Specifically, the training data set includes multiple historical terms, each of which is input into the model, and the model performs the same steps as in S2 to S5 to obtain multiple historical information terms and multiple historical template terms. The model calculates the historical comprehensive matching degree of each historical information term and the matching archive field group based on the similarity matching degree and the first initial weight, the distance matching degree and the second initial weight of the historical information terms. Then, the model obtains the historical comprehensive matching degree of all historical information terms matched by each archive field group, takes the first historical information term corresponding to the highest historical comprehensive matching degree as the historical archive field value of each archive field group, and outputs the historical archive field value of each archive field group. The historical archive field value (historical information term) of each archive field group is marked in the test data set.

[0111] When the settings of the first initial weight and the second initial weight are reasonable, the output result of the model should be consistent with the annotation result, that is, the historical comprehensive matching degree calculated by the model is highly accurate, which can provide a more accurate reference for the determination of subsequent historical archive field values. If the settings of the first initial weight and the second initial weight are not reasonable, the historical comprehensive matching degree calculated by the model is quite different from the annotation result, which cannot provide a more accurate reference for the determination of subsequent historical archive field values. Therefore, the model output result is used as the predicted value, the annotation result of the test data set is used as the target value, the difference between the predicted value and the target value is measured by minimizing the loss function, and the values ​​of the first initial weight and the second initial weight are continuously adjusted until the difference between the predicted value and the target value after adjustment is reduced to the expected value. At this time, the calculated comprehensive matching degree is more reasonable.

[0112] S53: Based on the first weight and the second weight, weighted calculation is performed on the similarity matching degree and the distance matching degree of each remaining actual information term to obtain the comprehensive matching degree of each actual information term and the matching archive field group.

[0113] After all actual information entries are screened once, each remaining actual information entry is input into the trained profile filling model, and the comprehensive matching degree is calculated based on the adjusted first weight and second weight.

[0114] In the above embodiment, all entries, after various checks, whether they pass or fail, will be stored in the entry candidate library. For entries that fail to pass the check, they are marked as discarded, and no entries will be selected from this type of entries for filling in later. For entries that pass the check, their corresponding first entry type / second entry type, similarity matching degree, distance matching degree, comprehensive matching degree and other data will be marked, and these marked data will be used as the basis for selection later.

[0115] S6: Obtain the comprehensive matching degree of all actual information terms matched by each first target file field group, and use the first target actual information term corresponding to the highest comprehensive matching degree as the first target file field value of each first target file field group to obtain the file of the target object.

[0116] After obtaining the matching relationship between each actual information entry and each archive field group, since not every archive field group has a matching actual information entry, if the matching relationship indicates that a certain archive field group has a matching actual information entry, then the archive field group is determined as the first target archive field group; if the matching relationship indicates that a certain archive field group does not have a matching actual information entry, then the archive field group is not used as the first target archive field group.

[0117] In the second stage, each archive field group ultimately only needs to determine a unique actual information entry as the first target actual information entry. In the above steps, a first target archive field group may have a comprehensive match with one or more actual information entries. Therefore, it is necessary to further screen the comprehensive match and determine the actual information entry with the most satisfactory comprehensive match as the first target actual information entry. Specifically, for a first target archive field group, the comprehensive match of all the actual information entries it matches can be obtained, and then the highest comprehensive match can be selected from them, and the actual information entry corresponding to the highest comprehensive match can be determined as the first target actual information entry. Finally, the first target actual information entry is filled into the archive template as the first target archive field value of the first target archive field group. For example, for the 7th archive field group, the first target actual information entry it matches is "176xxxx7777", then "176xxxx7777" is filled into the archive template as the archive field value 7, and the following is obtained: Figure 3 The effect shown.

[0118] In one embodiment, S6 specifically includes:

[0119] S61: Obtaining the comprehensive matching degree of all actual information entries matched by each first target file field group.

[0120] For a first target file field group, the number of actual information entries that match it may be only one or more. In this step, all of these actual information entries may be found first, and the comprehensive matching degree of each actual information entry may be obtained. When finding these actual information entries, they may be searched according to the second entry type annotated by each actual information entry in the entry candidate library. For example, if the second entry type annotated by an actual information entry is "mobile phone number", then the actual information entry is the actual information entry matched by the seventh file field group.

[0121] S62: Determine the highest comprehensive matching degree from all comprehensive matching degrees of each first target profile field group, and judge whether the highest comprehensive matching degree is not less than a third preset threshold, and the third preset threshold is greater than the first preset threshold and the second preset threshold.

[0122] For the comprehensive matching degrees of all actual information terms that match the same first target file field group, the highest comprehensive matching degree is first determined in order of size, and then it is determined whether the highest matching degree is not less than a third preset threshold value, and the value of the third preset threshold value can be set as needed, such as 90%. If there is only one actual information term, the comprehensive matching degree of the actual information term is used as the highest comprehensive matching degree to participate in the judgment.

[0123] The third preset threshold is greater than the first preset threshold and greater than the second preset threshold. Since the first preset threshold is used to preliminarily screen out the range of actual template entries and the second preset threshold is used to preliminarily screen out actual information entries, the values ​​of these two can be slightly lower. Usually the first preset threshold is set to 50% and the second preset threshold is set to 30%. The third preset threshold is used to determine the only actual information entry that needs to be filled in in the end, and has a higher requirement for accuracy. Therefore, it needs to be greater than the first preset threshold and the second preset threshold. The third preset threshold is usually set to 90%.

[0124] S63: If yes, use the first target actual information entry corresponding to the highest comprehensive matching degree as the first target file field value of the first target file field group.

[0125] If the highest comprehensive matching degree of a first target archive field group is not less than the third preset threshold, the actual information entry corresponding to the highest comprehensive matching degree is determined as the first target actual information entry matched by the first target archive field group, and is filled into the archive template as the first target archive field value of the first target archive field group. If it does not exist, the archive field value of the first target archive field group is left blank and is not filled in.

[0126] In one embodiment, after S6, the method further includes:

[0127] S7: Determine whether the value of each file field in the file needs to be modified.

[0128] Since the above-mentioned file filling process is carried out automatically, there may be errors in the value of a file field, or a file field value may still be left blank. In order to ensure the correctness and completeness of the file filling, it is necessary to enter the third stage after filling in, that is, to review the file to determine whether the values ​​of each file field in the file are correct and whether they need to be modified.

[0129] S8: If there is a second target file field value in the second target file field group that needs to be modified, obtain the comprehensive matching degree of all actual information entries matched by the second target file field group, and determine the first N actual information entries with the highest comprehensive matching degree as the second target actual information entries, and generate modification prompt information based on the first N second target actual information entries, where N is a positive integer.

[0130] If it is found that there is a file field value of a certain file field group in the file that needs to be modified, the file field group is determined as the second target file field group, and the file field value is determined as the second target file field value. First, the comprehensive matching degree of all actual information entries matched by the second target file field group is obtained, and then these comprehensive matching degrees are arranged in order from large to small, and the actual information entries corresponding to the first N comprehensive matching degrees are taken as the second target actual information entries, and the contents of the first N second target actual information entries are taken as N optional items in the modification prompt information, and recommended to the user through the application. The value of N can be set as needed, such as 3. If the number of the second target actual information entries is less than N, all actual information entries are taken as optional items, and "None of the above" can also be provided as an optional item.

[0131] Specifically, Figure 3 The file field values ​​of the file field name "First-level Industry" of the third file field group in the file created in the , are all empty. In the review stage, the top three with the highest comprehensive matching degree are found from all the actual information entries matched by the third file field group. Since there are only two actual information entries, namely "Transportation" and "Logistics and Transportation Services", the modification prompt information finally displayed to the user is as follows: Figure 4 As shown, there are three options: "transportation", "logistics and transportation services" and "none of the above".

[0132] S9: receiving a modification operation performed based on the modification prompt information, determining a modified target file field value in response to the modification operation, and filling the modified target file field value into the file.

[0133] The user uses the modification prompt information as a reference and performs a modification operation on the second target file field value through voice input. The server responds to the modification operation, determines the modified content as the modified second target file field value, and fills it into the file. After performing the above steps once for all the second target file field values ​​that need to be modified, a complete file can be obtained. Of course, if there are some second file field values ​​that do not match the actual information entries, they can also be filled in with manual assistance.

[0134] Specifically, the user uses "transportation" and "logistics and transportation services" as references and determines that the content to be modified is "logistics services". Then the user inputs the term "logistics services" by voice and sends it to the server. After receiving the term, the server fills it into the file as the modified second file field value of "primary industry". After the review is completed, the generated full file is as follows Figure 5 shown.

[0135] During the first filling process, due to the high requirement for the comprehensive matching degree of the first target actual information entry, there may be a situation where some file field groups do not determine the first target actual information entry, resulting in the file field value being left blank. However, if there are other matching actual information entries in the file field group, and the comprehensive matching degree of these actual information entries with the file field group is still relatively high among all entries, then these actual information entries still have a certain reference value for filling in the file field value. Therefore, the above method can improve the efficiency of the review and improve the accuracy of the file filling after the review.

[0136] It can be seen from the above embodiments that the file filling method provided by the present application enables high efficiency and quality of file filling.

[0137] Based on the method described in the above embodiment, this embodiment will be further described from the perspective of the file filling device. Figure 6 , the file filling device may include:

[0138] A first acquisition module 10 is used to acquire a vocabulary entry set, wherein the vocabulary entry set includes a plurality of vocabulary entries;

[0139] A second acquisition module 20 is used to acquire a file template, wherein the file template includes a plurality of file field groups, each of which includes a file field name and a file field value that form a corresponding relationship, each of which file field names in the file template corresponds to a template entity library, and each of which file field values ​​in the file template corresponds to an information entity library;

[0140] The matching module 30 is used to match each of the terms with each of the template entity libraries to obtain a plurality of successfully matched target template entity libraries; and to match each of the terms with each of the information entity libraries to obtain a plurality of successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms;

[0141] A determination module 40, for determining a matching relationship between each of the actual information terms and each of the actual template terms, and a matching relationship between each of the actual information terms and each of the file field groups, according to the actual template terms, the actual information terms, and the corresponding relationship between the file field name and the file field value;

[0142] The first obtaining module 50 is used to obtain the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtain the distance matching degree between each of the actual information terms and the matching actual template terms, and obtain the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree;

[0143] The second obtaining module 60 is used to obtain the comprehensive matching degree of all actual information entries matched by each first target file field group, and use the first target actual information entry corresponding to the highest comprehensive matching degree as the first target file field value of each first target file field group to obtain the file of the target object.

[0144] In one embodiment, the first acquisition module 10 includes:

[0145] The first acquisition submodule is used to acquire original voice data;

[0146] A processing submodule, used for processing the original speech data through a human voice model to obtain original text data;

[0147] The word segmentation submodule is used to perform word segmentation processing on the original text data to obtain multiple entries.

[0148] In one embodiment, the first acquisition module 10 further includes:

[0149] An extraction submodule, for extracting attribute features of each of the entries, wherein the attribute features include part of speech, grammatical function, character type and length;

[0150] The first annotation submodule is used to annotate each of the terms according to the attribute feature, and generate a term set according to the annotated terms.

[0151] In one embodiment, the matching module 30 includes:

[0152] A first matching submodule, used for matching each of the terms with each of the template entity libraries respectively, to obtain first similarity matching degrees between each of the terms and each of the template entity libraries;

[0153] A first determination submodule is used for, when there is a target first similarity matching degree greater than a first preset threshold among the first similarity matching degrees, determining the template entity library that obtains the target first similarity matching degree as a successfully matched target template entity library, and determining the entry that successfully matches the target template entity library as an actual template entry;

[0154] A second matching submodule is used to match each of the terms with each of the information entity libraries respectively, to obtain each second similarity matching degree between each of the terms and each of the information entity libraries;

[0155] The second determination submodule is used to determine the information entity library that obtains the target second similarity matching degree as the successfully matched target information entity library when there is a target second similarity matching degree greater than the first preset threshold among each of the second similarity matching degrees, and determine the entry that successfully matches the target information entity library as the actual information entry.

[0156] In one embodiment, the matching module 30 further includes:

[0157] A second acquisition submodule is used to acquire a first term attribute matching condition of each target template entity library, wherein the first term attribute matching condition includes a first common attribute feature of all reference template terms in the target template entity library, wherein the first common attribute feature includes part of speech, grammatical function, character type and length;

[0158] A first screening submodule is used to check the attribute characteristics of all actual template terms matched by each target template entity library based on the first term attribute matching condition of each target template entity library, and screen out the actual template terms whose attribute characteristics do not meet the first common attribute characteristics from all the actual template terms;

[0159] A third acquisition submodule is used to acquire a second term attribute matching condition of each of the target information entity libraries, wherein the second term attribute matching condition includes a second common attribute feature of all reference information terms in the target information entity library, wherein the second common attribute feature includes part of speech, grammatical function, character type and length;

[0160] The second screening submodule is used to check the attribute characteristics of all actual information terms matched by each target information entity library based on the attribute matching condition of the second term of each target information entity library, and screen out the actual information terms whose attribute characteristics do not meet the first common attribute characteristics from all the actual information terms.

[0161] In one embodiment, the first obtaining module 50 includes:

[0162] a removal submodule, configured to compare the distance matching degree of each of the actual information terms with a second preset threshold, and remove the actual information terms whose distance matching degree is not greater than the second preset threshold from all the actual information terms, and the second preset threshold is less than the first preset threshold;

[0163] A sixth acquisition submodule, used to acquire a first weight of a similarity matching degree and a second weight of a distance matching degree;

[0164] The calculation submodule is used to perform weighted calculation on the similarity matching degree and distance matching degree of each remaining actual information term based on the first weight and the second weight to obtain the comprehensive matching degree between each actual information term and the matching first target archive field group.

[0165] In one embodiment, the second obtaining module 60 includes:

[0166] A seventh acquisition submodule, for acquiring the comprehensive matching degree of all actual information terms matched by each of the first target archive field groups;

[0167] a judgment submodule, configured to determine a highest comprehensive matching degree from all comprehensive matching degrees of each of the first target profile field groups, and to judge whether the highest comprehensive matching degree is not less than a third preset threshold, wherein the third preset threshold is greater than the first preset threshold and the second preset threshold;

[0168] The fourth determining submodule is configured to, if yes, use the first target actual information entry corresponding to the highest comprehensive matching degree as the first target file field value of the first target file field group to obtain the file of the target object.

[0169] In one embodiment, the file filling device further includes:

[0170] A judgment module, used to judge whether the value of each file field in the file needs to be modified;

[0171] a generating module, configured to obtain the comprehensive matching degree of all actual information entries matched by the second target file field group if there is a second target file field value that needs to be modified, and determine the first N actual information entries with the highest comprehensive matching degree as the second target actual information entries, and generate modification prompt information according to the first N second target actual information entries, where N is a positive integer;

[0172] The modification module is used to receive a modification operation performed based on the modification prompt information, determine a modified target file field value in response to the modification operation, and fill the modified target file field value into the file.

[0173] Different from the prior art, the file filling device provided by the present application enables higher efficiency and quality of file filling.

[0174] Accordingly, the present application also provides an electronic device, such as Figure 7 As shown, the electronic device may include a radio frequency (RF) circuit 101, a memory 102 including one or more computer-readable storage media, an input unit 103, a display unit 104, a sensor 105, an audio circuit 106, a WiFi module 107, a processor 108 including one or more processing cores, and a power supply 109. Those skilled in the art will appreciate that Figure 7 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0175] The radio frequency circuit 101 can be used for receiving and sending signals during information transmission or calls. In particular, after receiving the downlink information of the base station, it is handed over to one or more processors 108 for processing; in addition, the data related to the uplink is sent to the base station. The memory 102 can be used to store software programs and modules. The processor 108 executes various functional applications and file filling by running the software programs and modules stored in the memory 102. The input unit 103 can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal input related to customer settings and function control.

[0176] The display unit 104 may be used to display information input by a client or information provided to a client and various graphical client interfaces of the server. These graphical client interfaces may be composed of graphics, texts, icons, videos and any combination thereof.

[0177] The electronic device may also include at least one sensor 105, such as a light sensor, a motion sensor, and other sensors. The audio circuit 106 includes a speaker, which may provide an audio interface between the user and the electronic device.

[0178] WiFi is a short-range wireless transmission technology. Electronic devices can help customers send and receive emails, browse web pages, and follow streaming media through WiFi modules 107. It provides customers with wireless broadband Internet follow-up. Figure 7A WiFi module 107 is shown, but it is understandable that it is not an essential component of the electronic device and can be omitted as needed without changing the essence of the application.

[0179] The processor 108 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire mobile phone. It executes various functions of the electronic device and processes data by running or executing software programs and / or modules stored in the memory 102 and calling data stored in the memory 102, thereby monitoring the mobile phone as a whole.

[0180] The electronic device also includes a power source 109 (such as a battery) for supplying power to each component. Preferably, the power source can be logically connected to the processor 108 through a power management system, so that the power management system can manage charging, discharging, power consumption and other functions.

[0181] Although not shown, the electronic device may also include a camera, a Bluetooth module, etc., which will not be described in detail here. Specifically in this embodiment, the processor 108 in the server will load the executable files corresponding to the processes of one or more applications into the memory 102 according to the following instructions, and the processor 108 will run the application stored in the memory 102, thereby realizing the following functions:

[0182] Acquire a vocabulary entry set, wherein the vocabulary entry set includes a plurality of vocabulary entries;

[0183] Acquire a file template, the file template comprising a plurality of file field groups, each of the file field groups comprising a file field name and a file field value forming a corresponding relationship, each of the file field names in the file template respectively corresponds to a template entity library, and each of the file field values ​​in the file template respectively corresponds to an information entity library;

[0184] Match each of the terms with each of the template entity libraries to obtain multiple successfully matched target template entity libraries; match each of the terms with each of the information entity libraries to obtain multiple successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms;

[0185] According to the actual template terms, the actual information terms, and the correspondence between the file field names and the file field values, determining the matching relationship between each of the actual information terms and each of the actual template terms, and the matching relationship between each of the actual information terms and each of the file field groups;

[0186] Obtaining the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtaining the distance matching degree between each of the actual information terms and the matching actual template terms, and obtaining the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree;

[0187] The comprehensive matching degree of all actual information terms matched by each first target file field group is obtained, and the first target actual information term corresponding to the highest comprehensive matching degree is used as the first target file field value of each first target file field group to obtain the file of the target object.

[0188] In the above embodiments, the description of each embodiment has its own emphasis. For the part that is not described in detail in a certain embodiment, please refer to the detailed description above and will not be repeated here.

[0189] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0190] To this end, an embodiment of the present application provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions can be loaded by a processor to implement the following functions:

[0191] Acquire a vocabulary entry set, wherein the vocabulary entry set includes a plurality of vocabulary entries;

[0192] Acquire a file template, the file template comprising a plurality of file field groups, each of the file field groups comprising a file field name and a file field value forming a corresponding relationship, each of the file field names in the file template respectively corresponds to a template entity library, and each of the file field values ​​in the file template respectively corresponds to an information entity library;

[0193] Match each of the terms with each of the template entity libraries to obtain multiple successfully matched target template entity libraries; match each of the terms with each of the information entity libraries to obtain multiple successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms;

[0194] According to the actual template terms, the actual information terms, and the correspondence between the file field names and the file field values, determining the matching relationship between each of the actual information terms and each of the actual template terms, and the matching relationship between each of the actual information terms and each of the file field groups;

[0195] Obtaining the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtaining the distance matching degree between each of the actual information terms and the matching actual template terms, and obtaining the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree;

[0196] The comprehensive matching degree of all actual information terms matched by each first target file field group is obtained, and the first target actual information term corresponding to the highest comprehensive matching degree is used as the first target file field value of each first target file field group to obtain the file of the target object.

[0197] The above is a detailed introduction to a file filling method, device, electronic device and computer-readable storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the technical solution and its core idea of ​​the present application; ordinary technicians in this field should understand that: they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some of the technical features therein with equivalents; and these modifications or replacements do not make the essence of the corresponding technical solution deviate from the scope of the technical solution of the embodiments of the present application.

Claims

1. A method for filling out a file, characterized in that: include: Acquire a vocabulary entry set, wherein the vocabulary entry set includes a plurality of vocabulary entries; Acquire a file template, the file template comprising a plurality of file field groups, each of the file field groups comprising a file field name and a file field value forming a corresponding relationship, each of the file field names in the file template respectively corresponds to a template entity library, and each of the file field values ​​in the file template respectively corresponds to an information entity library; Matching each of the terms with each of the template entity libraries respectively to obtain a plurality of successfully matched target template entity libraries; Match each of the terms with each of the information entity libraries respectively to obtain a plurality of successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms; According to the actual template terms, the actual information terms, and the correspondence between the file field names and the file field values, determining the matching relationship between each of the actual information terms and each of the actual template terms, and the matching relationship between each of the actual information terms and each of the file field groups; Obtaining the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtaining the distance matching degree between each of the actual information terms and the matching actual template terms, and obtaining the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree; The comprehensive matching degree of all actual information terms matched by each first target file field group is obtained, and the first target actual information term corresponding to the highest comprehensive matching degree is used as the first target file field value of each first target file field group to obtain the file of the target object.

2. The file filling method according to claim 1, characterized in that: The steps to obtain the entry set include: Get the original voice data; Processing the original speech data through a human voice model to obtain original text data; The original text data is segmented to obtain a plurality of entries.

3. The file filling method according to claim 2, characterized in that: After the step of obtaining multiple entries, it also includes: Extracting attribute features of each of the entries, the attribute features including part of speech, grammatical function, character type and length; Each of the entries is annotated according to the attribute features, and an entry set is generated according to the annotated multiple entries.

4. The file filling method according to claim 3 is characterized in that: The steps of matching each of the terms with each of the template entity libraries to obtain a plurality of successfully matched target template entity libraries; matching each of the terms with each of the information entity libraries to obtain a plurality of successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms, include: Matching each of the terms with each of the template entity libraries respectively to obtain first similarity matching degrees between each of the terms and each of the template entity libraries; When there is a target first similarity matching degree greater than a first preset threshold value among the first similarity matching degrees, the template entity library that obtains the target first similarity matching degree is determined as the successfully matched target template entity library, and the entry that successfully matches the target template entity library is determined as the actual template entry; Matching each of the terms with each of the information entity libraries respectively to obtain each second similarity matching degree between each of the terms and each of the information entity libraries; When there is a target second similarity matching degree greater than the first preset threshold among each of the second similarity matching degrees, the information entity library that obtains the target second similarity matching degree is determined as the successfully matched target information entity library, and the entry that successfully matches the target information entity library is determined as the actual information entry.

5. The file filling method according to claim 4, characterized in that: After the step of determining the entry that successfully matches the target template entity library as the actual template entry, the method further includes: Acquire a first term attribute matching condition of each target template entity library, wherein the first term attribute matching condition includes a first common attribute feature of all reference template terms in the target template entity library, wherein the first common attribute feature includes part of speech, grammatical function, character type and length; Based on the first term attribute matching condition of each of the target template entity libraries, respectively checking the attribute characteristics of all actual template terms matched by each of the target template entity libraries, and filtering out the actual template terms whose attribute characteristics do not meet the first common attribute characteristics from all the actual template terms; After the step of determining the entry that successfully matches the target information entity library as the actual information entry, the method further includes: Acquire a second term attribute matching condition for each of the target information entity libraries, wherein the second term attribute matching condition includes a second common attribute feature of all reference information terms in the target information entity library, wherein the second common attribute feature includes part of speech, grammatical function, character type, and length; Based on the second term attribute matching condition of each target information entity library, the attribute characteristics of all actual information terms matched by each target information entity library are checked respectively, and the actual information terms whose attribute characteristics do not meet the second common attribute characteristics are filtered out from all the actual information terms.

6. The file filling method according to claim 4 or 5, characterized in that: The step of obtaining the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree comprises: Comparing the distance matching degree of each of the actual information terms with a second preset threshold, removing the actual information terms whose distance matching degree is not greater than the second preset threshold from all the actual information terms, and the second preset threshold is less than the first preset threshold; Obtain a first weight of a similarity matching degree and a second weight of a distance matching degree; Based on the first weight and the second weight, weighted calculation is performed on the similarity matching degree and the distance matching degree of each remaining actual information term to obtain a comprehensive matching degree between each actual information term and the matching archive field group.

7. The file filling method according to claim 6, characterized in that: The step of obtaining the comprehensive matching degree of all actual information terms matched by each first target file field group, and taking the first target actual information term corresponding to the highest comprehensive matching degree as the first target file field value of each first target file field group to obtain the file of the target object comprises: Obtaining the comprehensive matching degree of all actual information entries matched by each first target archive field group; Determining a highest comprehensive matching degree from all comprehensive matching degrees of each of the first target profile field groups, and judging whether the highest comprehensive matching degree is not less than a third preset threshold, wherein the third preset threshold is greater than the first preset threshold and the second preset threshold; If so, the first target actual information entry corresponding to the highest comprehensive matching degree is used as the first target file field value of the first target file field group to obtain the file of the target object.

8. The file filling method according to claim 1, characterized in that: After obtaining the target object's file, the following steps are also included: Determine whether the value of each file field in the file needs to be modified; If there is a second target file field value of the second target file field group that needs to be modified, obtain the comprehensive matching degree of all actual information entries matched by the second target file field group, and determine the first N actual information entries with the highest comprehensive matching degree as the second target actual information entries, and generate modification prompt information according to the first N second target actual information entries, where N is a positive integer; A modification operation performed based on the modification prompt information is received, and in response to the modification operation, a modified target file field value is determined, and the modified target file field value is filled into the file.

9. A file filling device, characterized in that: include: A first acquisition module is used to acquire a vocabulary set, wherein the vocabulary set includes a plurality of vocabulary entries; A second acquisition module is used to acquire a file template, wherein the file template includes a plurality of file field groups, each of which includes a file field name and a file field value that form a corresponding relationship, each of which file field names in the file template corresponds to a template entity library, and each of which file field values ​​in the file template corresponds to an information entity library; A matching module, used for matching each of the terms with each of the template entity libraries to obtain a plurality of successfully matched target template entity libraries; Match each of the terms with each of the information entity libraries respectively to obtain a plurality of successfully matched target information entity libraries; wherein the terms that successfully match the target template entity library are actual template terms, and the terms that successfully match the target information entity library are actual information terms; a determination module, for determining a matching relationship between each of the actual information terms and each of the actual template terms, and a matching relationship between each of the actual information terms and each of the file field groups, according to the actual template terms, the actual information terms, and the corresponding relationship between the file field name and the file field value; The first obtaining module is used to obtain the similarity matching degree between each of the actual information terms and the matching target information entity library, and obtain the distance matching degree between each of the actual information terms and the matching actual template terms, and obtain the comprehensive matching degree between each of the actual information terms and the matching archive field group according to the similarity matching degree and the distance matching degree; The second obtaining module is used to obtain the comprehensive matching degree of all actual information entries matched by each first target file field group, and use the first target actual information entry corresponding to the highest comprehensive matching degree as the first target file field value of each first target file field group to obtain the file of the target object.

10. An electronic device, characterized in that: It comprises a memory and a processor; the memory stores an application program, and the processor is used to run the application program in the memory to execute the steps in the file filling method described in any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps in the file filling method described in any one of claims 1 to 8.

Images