WO2019080419A1

WO2019080419A1 - Method for building standard knowledge base, electronic device, and storage medium

Info

Publication number: WO2019080419A1
Application number: PCT/CN2018/076484
Authority: WO
Inventors: 卢川; 高祎璠
Original assignee: 平安科技（深圳）有限公司
Priority date: 2017-10-26
Filing date: 2018-02-12
Publication date: 2019-05-02
Also published as: CN107832374A

Abstract

A method for building a standard knowledge base, relating to the field of database maintenance. The method for building a standard knowledge base comprises the following steps: building an answer file (S1); building a question template (S2); setting constant items and question words (S3); and presetting a question-answer pair generating rule, and forming question-answer pairs (S4). By using the method to build a standard knowledge base, batch import of data can be achieved. Moreover, by automatically generating question-answer pairs according to a rule, the maintenance workload of a basic database is reduced, and the working efficiency is greatly improved.

Description

Method for constructing standard knowledge base, electronic device and storage medium

The present application claims the priority of the Chinese patent application filed on October 26, 2017, with the application number of 201711031785.9, entitled "Construction Method, Electronic Device and Storage Medium of Standard Knowledge Base", the entire contents of which are incorporated by reference. The way is combined in this application.

Technical field

The present application relates to the field of database maintenance, and relates to a method for constructing a standard knowledge base, an electronic device, and a storage medium.

Background technique

With the rapid development of the Internet, online customer service has spread to all walks of life and goes deep into all aspects of daily business services. At present, common network customers are usually composed of intelligent customer service robots and manual customer service. Intelligent customer service can provide round-the-clock and day-to-day service, and divert the burden of manual customer service.

However, the answers of the intelligent customer service robots are all set in advance, and are usually paired and saved in the basic database according to the way one question corresponds to one answer. Therefore, when constructing the basic database, it is necessary to maintain the problem as much as possible - the answer pair In order to realize the intelligent answer of the intelligent customer service robot, the maintenance of the basic database is huge and consumes a lot of labor costs.

Summary of the invention

The purpose of the present application is to propose a method for constructing a standard knowledge base, an electronic device, and a computer readable storage medium. By setting a composition rule of a question and an answer, the system can perform the problem of the received content according to the set rule. And the automatic generation of the answer.

The present application solves the above technical problems by the following technical solutions:

A method for constructing a standard knowledge base, comprising the steps of: S1, constructing an answer file: collecting an answer, and parsing the answer into a same file in a uniform format, the file including a form or a text. S2, build problem template: determine the question template according to the answer in the answer file, the question template is a sequence of words including a number of change items, a number of constant items and question words; S3, set constant items and question words: according to the answer file The meaning of the answer in the answer, determining the constant item and the question word in the question template; S4, forming a question-answer pair: according to the generation rule of the question-answer pair, acquiring the content of the corresponding position in the answer file is embedded into the A question is generated in the corresponding change item in the question template, and the content of the corresponding position in the answer file is obtained to generate an answer, and the generated question and the answer link are saved as a question-answer pair.

An electronic device includes a memory and a processor, wherein the memory stores a build system of a standard knowledge base executable by the processor, the build system of the standard knowledge base includes: a file receiving module, configured to receive a good answer file, the answer file includes at least one answer, the answer is split into a uniform file and then organized into a same file, the file includes a form or text; a template setting module is used according to the answer file The answer setting problem template is a word sequence including a plurality of variation items, a plurality of constant items and question words; an input module for receiving the contents of the constant item and the question word; and a question-answer pair generation module for Generating a question according to a generation rule of the question-answer pair by embedding the content of the corresponding position in the received answer file into a corresponding change item in the question template, and acquiring the content of the corresponding position in the answer file to generate an answer, which will be generated The question and the answer are associated as a question-answer pair.

A computer readable storage medium having a built-in system of a standard knowledge base stored therein, the built-in system of the standard knowledge base being executable by at least one processor to implement the following steps: S1, constructing Answer file: collecting the answers, sorting the answers into a unified file in a uniform format, the files include a form or text; S2, constructing a question template: determining a question template according to an answer in the answer file, the question The template is a sequence of words including a plurality of variation items, a plurality of constant items, and an interrogative word; S3, setting a constant term and an interrogative word: determining a constant term and an interrogative word in the question template according to the meaning expressed by the answer in the answer file; S4 Forming a question-answer pair: according to the generation rule of the question-answer pair, acquiring the content of the corresponding position in the answer file is embedded into the corresponding change item in the question template to generate a question, and acquiring the content of the corresponding position in the answer file To generate an answer, save the generated question and the answer association as a question-answer pair.

The positive progress of the application is that the standard knowledge base is built by the application, the data is imported in batches, and the problem-answer pair is automatically generated according to the rules, and the maintenance workload of the basic database is reduced.

DRAWINGS

1 is a schematic diagram showing the hardware architecture of an embodiment of an electronic device of the present application;

2 is a schematic diagram showing a program module of an embodiment of a system for building a standard knowledge base in an electronic device of the present application;

FIG. 3 is a flowchart showing Embodiment 1 of a method for constructing a standard knowledge base of the present application;

4 is a flow chart showing the construction of an answer file in Embodiment 2 of the method of the present application;

FIG. 5 is a flowchart of a problem template in Embodiment 2 of the method of the present application;

6 is a flow chart showing setting constant items and interrogative words in Embodiment 2 of the method of the present application;

FIG. 7 is a flow chart showing a problem-answer pair formed in Embodiment 2 of the method of the present application;

Figure 8 is a schematic diagram showing the answer file in the form of a table in the method of the present application;

FIG. 9 is a flowchart of constructing an answer file in Embodiment 3 of the method of the present application;

FIG. 10 is a flowchart of constructing a problem template in Embodiment 3 of the method of the present application;

Figure 11 is a flow chart showing the formation of a question-answer pair in the third embodiment of the method of the present application;

Figure 12 is a diagram showing the answer file in the form of text in the method of the present application.

Detailed ways

The present application is further illustrated by the following examples, but is not intended to limit the scope of the embodiments.

First of all, the present application proposes an electronic device.

1 is a schematic diagram of a hardware architecture of an embodiment of an electronic device of the present application. In the present embodiment, the electronic device 2 is an apparatus capable of automatically performing numerical calculation and/or information processing in accordance with an instruction set or stored in advance. For example, it can be a smartphone, a tablet, a laptop, a desktop computer, a rack server, a blade server, a tower server, or a rack server (including a stand-alone server, or a server cluster composed of multiple servers). As shown, the electronic device 2 includes at least, but not limited to, a built-in system 20 that can communicate with each other via a system bus to the memory 21, the processor 22, the network interface 23, and a standard knowledge base. among them:

The memory 21 includes at least one type of computer readable storage medium including a flash memory, a hard disk, a multimedia card, a card type memory (eg, SD or DX memory, etc.), a random access memory (RAM), Static Random Access Memory (SRAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Programmable Read Only Memory (PROM), magnetic memory, magnetic disk, optical disk, and the like. In some embodiments, the memory 21 may be an internal storage unit of the electronic device 2, such as a hard disk or a memory of the electronic device 2. In other embodiments, the memory 21 may also be an external storage device of the electronic device 2, such as a plug-in hard disk equipped on the electronic device 2, a smart memory card (SMC), and a secure digital device. (Secure Digital, SD) card, flash card, etc. Of course, the memory 21 can also include both the internal storage unit of the electronic device 2 and its external storage device. In this embodiment, the memory 21 is generally used to store an operating system installed in the electronic device 2 and various types of application software, such as program code of the build system 20 of the standard knowledge base. Further, the memory 21 can also be used to temporarily store various types of data that have been output or are to be output.

The processor 22 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 22 is typically used to control the overall operation of the electronic device 2, such as performing control and processing associated with data interaction or communication with the electronic device 2. In this embodiment, the processor 22 is configured to run program code or process data stored in the memory 21, such as the build system 20 or the like that runs the standard knowledge base.

The network interface 23 may comprise a wireless network interface or a wired network interface, which is typically used to establish a communication connection between the electronic device 2 and other electronic devices. For example, the network interface 23 is configured to connect the electronic device 2 to an external terminal through a network, establish a data transmission channel, a communication connection, and the like between the electronic device 2 and an external terminal. The network may be an intranet, an Internet, a Global System of Mobile communication (GSM), a Wideband Code Division Multiple Access (WCDMA), a 4G network, or a 5G network. Wireless or wired networks such as network, Bluetooth, Wi-Fi, etc.

It is pointed out that FIG. 1 only shows the electronic device 2 with the components 21-23, but it should be understood that not all illustrated components are required to be implemented, and more or fewer components may be implemented instead.

In the present embodiment, the construction system 20 of the standard knowledge base stored in the memory 21 can be divided into one or more program modules, the one or more program modules are stored in the memory 21, and can be Or multiple processors (this embodiment is processor 22) are executed to complete the application.

For example, FIG. 2 shows a schematic diagram of a program module of an embodiment of the standard knowledge base construction system 20. In this embodiment, the standard knowledge base construction system 20 can be divided into a file receiving module 201 and a template setting module. 202. An input module 203 and a question-answer pair generation module 204. The following description will specifically describe the specific functions of the program modules 201-204.

The file receiving module 201 is configured to receive the prepared answer file, where the answer file includes at least one answer, and the answers are split into a unified file in a unified format, and the file includes a form or a text; The template setting module 202 is configured to set a question template according to an answer in the answer file, the question template is a word sequence including a plurality of change items, a plurality of constant items, and an interrogative word; the input module 203 is configured to receive the constant item and The content of the question word; the question-answer pair generating module 204 is configured to generate a question by embedding the content of the corresponding position in the received answer file into the corresponding change item in the question template according to the generation rule of the question-answer pair. At the same time, the content of the corresponding position in the answer file is obtained to generate an answer, and the generated question and the answer association are saved as a question-answer pair.

In this embodiment, the content in the answer file needs to be organized in a unified format in advance. It should be noted that the format compiled here needs to match the generation rule of the question-answer pair. For example, the answer file is organized in text format, then the generation rule of the question-answer pair is also for the file in text format, and another example: The answer file is organized in a tabular format, so the generation rule for the question-answer pair is for the file in tabular format.

Secondly, the present application proposes a method of constructing a standard knowledge base.

In the first embodiment, as shown in FIG. 3, the method for constructing the standard knowledge base includes the following steps: S1: constructing an answer file: collecting an answer, splitting the answer into a uniform format, and then collating the same file into the same file. The file includes a form or a text; S2, constructing a question template: determining a question template according to an answer in the answer file, the question template being a word sequence including a plurality of change items, a plurality of constant items, and an interrogative word; S3, setting a constant Item and question word: determine the constant item and the question word in the question template according to the meaning expressed by the answer in the answer file; S4, form a question-answer pair: according to the generation rule of the question-answer pair, obtain the answer file The content of the corresponding location is embedded into the corresponding change item in the question template to generate a question, and the content of the corresponding position in the answer file is obtained to generate an answer, and the generated question and the answer association are saved as a question-answer pair.

In this embodiment, the answer file may be a text or a table, and the steps in the method are further described in detail below by using the answer file as a text and a form.

In the second embodiment, on the basis of the first embodiment, the steps in the first embodiment are further illustrated by using the answer file as a table format, as follows:

First, build an answer file (as shown in Figure 4)

S11, collecting an answer; S12, splitting each answer into a sequence of words consisting of a plurality of keywords; S13, obtaining two meaning keywords representing the meaning of the answer in each word sequence; S14, de-duplicating the meaning keywords Classification; S15, one type of meaning keyword is used as the first row of the table, another type of meaning keyword is used as the first column of the table, and the intersecting cells of the first row and the first column are blank; S16, the value of the answer is represented in the sequence of acquired words. The numerical keyword; S17, the numeric keyword is filled in the cell in which the two meaning keywords in the sequence of the word of the numerical keyword are located and the column intersects.

Second, build a problem template (as shown in Figure 5)

S21. The type of the meaning keyword in the corresponding answer file determines the number of the change items in the question template. S22. Add a function word between the change item, the constant item, and the question word according to the grammar to form a grammatical problem template.

Third, set constant items and question words (as shown in Figure 6)

S31, splitting each answer into a sequence of words consisting of a plurality of keywords; S32, obtaining an abstract noun in the meaning keyword representing the meaning of the answer in the sequence of words as a constant term; S33, setting a suitable one for the query after the constant term Question words for numerical keywords in the answer file.

Fourth, the formation of the problem - the answer pair (as shown in Figure 7)

S40: Obtain the first non-blank cell in the first row and the first column of the table as the current position of the two change items; S41, according to the generation rule of the question-answer pair, obtain the meaning keyword in the current location and embed the corresponding keyword into the problem template. The change position generation problem is temporarily stored; S42, obtaining the two meaning keywords of the generated problem, the numerical keyword of the row and the column intersecting the cell is temporarily stored as an answer; S43, the temporarily stored question and the answer are associated with each other; S44 And determining whether the meaning keyword in the current position corresponding to the first change item is the last word in the first row or the first column of the meaning keyword, and if yes, executing step S46, if otherwise, performing step S45; S45, first The current position corresponding to the change item is sequentially shifted one by one along the first line or the first column of the meaning keyword in the current position, and the current position corresponding to the first change item is reset, and step S41 is performed; S46, determining the second The change item corresponds to whether the meaning keyword in the current position is the last word in the first example or the first line of the meaning keyword, and if yes, step S48 is performed, if no Then, step S47 is executed; S47, the current position corresponding to the second change item is sequentially shifted one by one along the first column or the first line of the meaning keyword in the current position, and step S41 is performed; S48;

The following is an example of building a standard knowledge base on the income of each type of insurance in each quarter:

1. Collect the answers to the income data of each type of insurance in each quarter, and organize the income data of each type of insurance in each quarter into an answer file as shown in Figure 8. The first row and the first column are two types of meaning keywords. , respectively, time and insurance (also can be the first behavior insurance, the first column is time), the specific insurance category and the specific time in the row of cells intersecting the income of the insurance at that time. 2. According to the two types of meaning keywords in the foregoing table, the change item in the problem template is determined as two items. According to the grammar, the question template is determined as “two variable items + one constant item + yes + question word”. 3. According to the collected answers, determine that the constant term is “income” and the question word is “how much”. The more clear question template that applies to the above answer file is “two changes + + income + yes + how much” . 4. The second cell in the first row of the table and the two meaning keywords in the second cell of the first column are respectively embedded in the position of the two variables in the aforementioned problem template, and the problem is "first quarter". What is the income of the accident insurance? The corresponding answer is the value “2560000” in the cell where the second column and the second row intersect, and then the generated question and answer are saved as a pair of questions-answer pairs to the standard knowledge. In the library; then, the second cell in the first row of the table and the two meaning keywords in the third cell of the first column are respectively embedded in the position of the two variables in the aforementioned problem template, and the problem is generated as " What is the income of health insurance in the first quarter? The corresponding answer is the value "5246286" in the cell intersecting the second and third rows; until the second cell and the first column of the first row in the table are obtained The meaning of the keyword in the last cell; then take the meaning keyword in the third cell of the first row in the table, and then get the meaning keywords in each cell in the first column, in order The problem-answer pair is saved to the standard knowledge base.

In the third embodiment, on the basis of the first embodiment, the steps in the first embodiment are further described by using the answer file as a text format, as follows:

First, build an answer file (as shown in Figure 9)

S11', collecting answers; S12', splitting each answer into a sequence of words consisting of several keywords; S13', a meaning keyword representing the meaning of the answer in the sequence of reserved words, and a numerical keyword representing the value of the answer; S14' The meaning keywords and numerical keywords in the same word sequence are arranged in order, and are separated by a uniform keyword separator between the keywords; S15', different keywords are used between different word sequences The uniform word sequence separators of the separators are separated.

Second, build a problem template (as shown in Figure 10)

S21', the number of keyword separators in the same word sequence in the corresponding answer file determines the number of change items in the question template; S22', according to the grammar, adds a function word between the change item, the constant item and the question word to form a match A grammar question template.

Third, set the constant term and the question word (the same as the second embodiment, no more details here.)

Fourth, the formation of the problem - the answer pair (as shown in Figure 11)

S40', obtaining the position of the first word sequence separator in the text as the position of the current word sequence separator, and the position of each keyword separator before the first word sequence separator as the position of each current keyword separator; S41' According to the generation rule of the question-answer pair, each meaning keyword before each current keyword separator is obtained, and the problem item generated in the problem template is temporarily stored in the problem; S42', the numerical keyword before the current word sequence separator is obtained. S43', the associated question and answer are saved; S44', determining whether the current word sequence separator is the last word sequence separator in the answer file, and if yes, executing step S47', if otherwise, performing step S45'; S45', the position of the current word sequence separator is sequentially shifted and the position of the current word sequence separator is reset; S46', the position of each current keyword separator is reset to the key before the current word sequence separator The position of the word separator is executed in step S41'; S47' and ended.

The following is a detailed description of building a standard knowledge base on the income of each type of insurance in each quarter:

1. Collect the answers to the income data of each type of insurance in each quarter (the same example above), and organize the income data of each type of insurance in each quarter into an answer file as shown in Figure 12, with a semicolon between each word sequence. Separate, separating each keyword in the same word sequence with a comma. (It can also be separated by other symbols according to the habit.) 2. According to the number of meaning keywords in each word sequence in the aforementioned text, the change item in the question template is determined to be two items, and according to the grammar, the question template is determined as " The +variation + of the two variables + is + question word. 3. According to the answers collected above, it can be determined that the constant term is “income” and the question word is “how much”, so a more specific question template applicable to the aforementioned answer file can be obtained as “two changes + + income + Yes + how much." 4. Get the position of the first colon and the position of each comma before the colon. According to the order of the comma, the meaning keyword before the comma is inserted into the position of the change item of the question template, and the problem is “the income of the first quarter accident insurance”. What is it?" Then get the numeric keyword before the colon, generate the answer as "2560000", save the above generated question and answer as the question-answer pair in the standard knowledge base; then get the position of the second colon And the position of each comma before the colon, the same rules as above will generate questions and answers, and the generated questions and answers will be saved as a question-answer pair in the standard knowledge base; according to this rule, the questions are generated in turn - the answer pairs are associated Save in the standard knowledge base until the last colon.

Furthermore, the present application is a computer readable storage medium having a built-in knowledge base of a built-in knowledge base 20 that can be implemented by one or more processors to implement the above The construction method of the standard knowledge base or the operation of the electronic device.

While the embodiments of the present invention have been described above, it will be understood by those skilled in the art that the scope of the invention is defined by the appended claims. A person skilled in the art can make various changes or modifications to these embodiments without departing from the spirit and scope of the present application, and such changes and modifications fall within the scope of the present application.

Claims

A method for constructing a standard knowledge base, comprising the steps of:

S1. Build an answer file: collect the answer, and split the answer into a unified file in a uniform format, the file including a form or a text;

S2. Build a problem template: determining a question template according to an answer in the answer file, the question template being a sequence of words including a plurality of change items, a plurality of constant items, and question words;

S3. Setting a constant term and a question word: determining a constant term and a question word in the question template according to the meaning expressed by the answer in the answer file;

S4. Forming a question-answer pair: according to a generation rule of the question-answer pair, acquiring a content of a corresponding position in the answer file is embedded in a corresponding change item in the question template to generate a question, and obtaining a corresponding position in the answer file The content is generated to generate an answer, and the generated question and the answer are associated as a question-answer pair.
The method for constructing a standard knowledge base according to claim 1, wherein when the file in step S1 is a table, the following sub-steps are included:

S11, collecting answers;

S12. Split each answer into a sequence of words consisting of several keywords;

S13. Obtain two meaning keywords in the sequence of each word that represent the meaning of the answer;

S14, de-duplicating and classifying the meaning keywords;

S15. A class of meaning keywords is used as the first row of the table, another type of meaning keyword is used as the first column of the table, and the intersecting cells of the first row and the first column are blank;

S16. Obtain a numerical keyword representing a value of the answer in the sequence of words;

S17. Fill in the numerical keyword in the cell where the two meaning keywords in the sequence of words of the numerical keyword are located and the cells intersecting the column.
The method for constructing a standard knowledge base according to claim 2, wherein the step S2 comprises the following sub-steps:

S21. The type of the meaning keyword in the corresponding answer file determines the number of the change items in the problem template;

S22. Add a function word between the variable item, the constant item, and the question word according to the grammar to form a grammatical problem template.
The method for constructing a standard knowledge base according to claim 2, wherein the step S4 specifically comprises the following sub-steps:

S40. Obtain the first non-blank cell in the first row and the first column of the table as the current position of the two variables;

S41. According to the generation rule of the question-answer pair, obtaining the meaning keyword in the current location and temporarily generating the problem corresponding to the location of the change item embedded in the problem template;

S42. Obtaining two meaning keywords of the generated problem: the numerical keyword in the cell where the row and the column intersecting are temporarily stored as an answer;

S43. Associate the temporarily stored question and the answer;

S44, determining whether the meaning keyword in the current position corresponding to the first change item is the last word in the first row or the first column of the meaning keyword, if yes, executing step S46, if otherwise, executing step S45;

S45, the current position corresponding to the first change item is sequentially shifted along the first row or the first column of the meaning keyword in the current position, and the current position corresponding to the first change item is reset, and step S41 is performed;

S46, determining whether the second change item corresponds to the meaning keyword in the current position is the first word in the first instance or the first line of the meaning keyword, if yes, step S48 is performed, otherwise step S47 is performed;

S47, the current position corresponding to the second change item is sequentially shifted along the first column or the first line of the meaning keyword in the current position, and step S41 is performed;

S48, the end.
The method for constructing a standard knowledge base according to claim 1, wherein when the file in step S1 is text, the following sub-steps are included:

S11’, collecting answers;

S12', splitting each answer into a sequence of words consisting of several keywords;

S13', a meaning keyword indicating a meaning of the answer in the sequence of reserved words, and a numerical keyword representing the value of the answer;

S14', the meaning keywords and the numerical keywords in the same word sequence are arranged in order, and are separated by a uniform keyword separator between the keywords;

S15' is separated by a uniform word sequence separator different from the keyword separator between different word sequences.
The method for constructing a standard knowledge base according to claim 5, wherein the step S2 comprises the following sub-steps:

S21', the number of keyword separators in the same word sequence in the corresponding answer file determines the number of change items in the question template;

S22', according to the grammar, adding a virtual word between the variable item, the constant item, and the question word to form a grammatical problem template.
The method for constructing a standard knowledge base according to claim 5, wherein the step S4 comprises the following substeps:

S40', obtaining the position of the first word sequence separator in the text as the position of the current word sequence separator, and the position of each keyword separator before the first word sequence separator as the position of each current keyword separator;

S41', according to the generation rule of the question-answer pair, obtaining each of the meaning keywords before each current keyword separator, and temporarily storing the change item position problem embedded in the problem template;

S42', before the current word sequence separator is obtained, the numerical keyword is temporarily stored as an answer;

S43’, saving the temporarily stored question and answer;

S44', determining whether the current word sequence separator is the last word sequence separator in the answer file, and if so, executing step S47', if otherwise performing step S45';

S45', moving the position of the current word sequence separator back and resetting the position of the current word sequence separator;

S46', the position of each current keyword separator is reset to the position of each keyword separator before the current word sequence separator, and step S41' is performed;

S47', the end.
The method for constructing a standard knowledge base according to any one of claims 1 to 7, wherein the step S3 specifically comprises the following sub-steps:

S31. Split each answer into a sequence of words consisting of several keywords;

S32. Obtain an abstract noun in the meaning keyword that represents the meaning of the answer in the sequence of words as a constant term;

S33. After the constant item, set a question word suitable for asking a numerical keyword in the answer file.
An electronic device includes a memory and a processor, wherein the memory stores a built-in system of a standard knowledge base executable by the processor, and the built-in system of the standard knowledge base includes:

a file receiving module, configured to receive the prepared answer file, where the answer file includes at least one answer, and the answers are split into a unified file in a unified format, and the file includes a form or a text;

a template setting module, configured to set a question template according to an answer in the answer file, where the question template is a word sequence including a plurality of change items, a plurality of constant items, and question words;

An input module for receiving the contents of the constant item and the question word;

a question-answer pair generation module, configured to embed a content of a corresponding position in the received answer file into a corresponding change item in the question template according to a generation rule of the question-answer pair to generate a question, and obtain a corresponding position in the answer file The content is generated to generate an answer, and the generated question and the answer are associated as a question-answer pair.
A computer readable storage medium, wherein the computer readable storage medium stores a build system of a standard knowledge base, and the build system of the standard knowledge base can be executed by at least one processor to implement the following steps :

S1. Build an answer file: collect the answer, and split the answer into a unified file in a uniform format, the file including a form or a text;

S2. Build a problem template: determining a question template according to an answer in the answer file, the question template being a sequence of words including a plurality of change items, a plurality of constant items, and question words;

S3. Setting a constant term and a question word: determining a constant term and a question word in the question template according to the meaning expressed by the answer in the answer file;

S4. Forming a question-answer pair: according to a generation rule of the question-answer pair, acquiring a content of a corresponding position in the answer file is embedded in a corresponding change item in the question template to generate a question, and obtaining a corresponding position in the answer file The content is generated to generate an answer, and the generated question and the answer are associated as a question-answer pair.
The computer readable storage medium according to claim 10, wherein when the file in step S1 is a table, the following substeps are included:

S11, collecting answers;

S12. Split each answer into a sequence of words consisting of several keywords;

S13. Obtain two meaning keywords in the sequence of each word that represent the meaning of the answer;

S14, de-duplicating and classifying the meaning keywords;

S15. A class of meaning keywords is used as the first row of the table, another type of meaning keyword is used as the first column of the table, and the intersecting cells of the first row and the first column are blank;

S16. Obtain a numerical keyword representing a value of the answer in the sequence of words;

S17. Fill in the numerical keyword in the cell where the two meaning keywords in the sequence of words of the numerical keyword are located and the cells intersecting the column.
The computer readable storage medium according to claim 11, wherein the step S2 comprises the following substeps:

S21. The type of the meaning keyword in the corresponding answer file determines the number of the change items in the problem template;

S22. Add a function word between the variable item, the constant item, and the question word according to the grammar to form a problem template that conforms to the grammar.
The computer readable storage medium according to claim 11, wherein the step S4 comprises the following substeps:

S40. Obtain the first non-blank cell in the first row and the first column of the table as the current position of the two variables;

S41. According to the generation rule of the question-answer pair, obtaining the meaning keyword in the current location and temporarily generating the problem corresponding to the location of the change item embedded in the problem template;

S42. Obtaining two meaning keywords of the generated problem: the numerical keyword in the cell where the row and the column intersecting are temporarily stored as an answer;

S43. Associate the temporarily stored question and the answer;

S44, determining whether the meaning keyword in the current position corresponding to the first change item is the last word in the first row or the first column of the meaning keyword, if yes, executing step S46, if otherwise, executing step S45;

S45, the current position corresponding to the first change item is sequentially shifted along the first row or the first column of the meaning keyword in the current position, and the current position corresponding to the first change item is reset, and step S41 is performed;

S46, determining whether the second change item corresponds to the meaning keyword in the current position is the first word in the first instance or the first line of the meaning keyword, if yes, step S48 is performed, otherwise step S47 is performed;

S47, the current position corresponding to the second change item is sequentially shifted along the first column or the first line of the meaning keyword in the current position, and step S41 is performed;

S48, the end.
The computer readable storage medium according to claim 10, wherein when the file in step S1 is text, the following substeps are included:

S11’, collecting answers;

S12', splitting each answer into a sequence of words consisting of several keywords;

S13', a meaning keyword indicating a meaning of the answer in the sequence of reserved words, and a numerical keyword representing the value of the answer;

S14', the meaning keywords and the numerical keywords in the same word sequence are arranged in order, and are separated by a uniform keyword separator between the keywords;

S15' is separated by a uniform word sequence separator different from the keyword separator between different word sequences.
The computer readable storage medium according to claim 14, wherein the step S2 comprises the following substeps:

S21', the number of keyword separators in the same word sequence in the corresponding answer file determines the number of change items in the question template;

S22', according to the grammar, adding a virtual word between the variable item, the constant item, and the question word to form a grammatical problem template.
The computer readable storage medium of claim 14, wherein step S4 comprises the following substeps:

S40', obtaining the position of the first word sequence separator in the text as the position of the current word sequence separator, and the position of each keyword separator before the first word sequence separator as the position of each current keyword separator;

S41', according to the generation rule of the question-answer pair, obtaining each of the meaning keywords before each current keyword separator, and temporarily storing the change item position problem embedded in the problem template;

S42', before the current word sequence separator is obtained, the numerical keyword is temporarily stored as an answer;

S43’, saving the temporarily stored question and answer;

S44', determining whether the current word sequence separator is the last word sequence separator in the answer file, and if so, executing step S47', if otherwise performing step S45';

S45', moving the position of the current word sequence separator back and resetting the position of the current word sequence separator;

S46', the position of each current keyword separator is reset to the position of each key word separator before the current word sequence separator, and step S41' is performed;

S47', the end.
The computer readable storage medium according to any one of claims 10 to 16, wherein the step S3 comprises the following substeps:

S31. Split each answer into a sequence of words consisting of several keywords;

S32. Obtain an abstract noun in the meaning keyword that represents the meaning of the answer in the sequence of words as a constant term;

S33. After the constant item, set a question word suitable for asking a numerical keyword in the answer file.