Energy knowledge question and answer dialogue management method and system based on regular matching
By constructing a regular expression rule base and a knowledge database, and combining preprocessing and regular expression matching techniques, the problem of poor relevance of query information in the operation and maintenance of energy equipment in existing technologies has been solved, and a highly accurate and efficient energy knowledge question-and-answer system has been realized.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-17
- Publication Date
- 2026-03-31
AI Technical Summary
In the operation and maintenance of energy equipment, existing technologies struggle to accurately understand professional equipment codes, standard fault codes, and natural language questions in spoken language at the scene. This results in poor relevance of query information and fails to meet the needs of energy scenarios with high safety requirements.
We construct a regular expression rule base and knowledge database for the energy operation and maintenance field. By preprocessing and matching user input statements with regular expressions, we extract a list of slot values and query tags. Combining keyword scoring, we select the candidate answer with the highest comprehensive score to reply to the user and generate reminder instructions based on the user score.
It achieves highly accurate and efficient knowledge-based Q&A in energy operation and maintenance, improves the relevance and security of user queries, and provides accurate responses to professional equipment parameters, fault handling, and operation procedures.
Smart Images
Figure CN121764971A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, specifically to an energy knowledge question-and-answer dialogue management method and system based on regular expression matching. Background Technology
[0002] The stable operation of energy production and transmission facilities is a crucial guarantee for national energy security and the economic lifeline. With the large-scale construction of new energy power plants such as photovoltaic and wind power, and the advancement of intelligent transformation of traditional power plants, the operation and maintenance of energy equipment is becoming increasingly complex, placing higher demands on the professional knowledge and emergency response capabilities of on-site operation and maintenance personnel.
[0003] Currently, energy operation and maintenance personnel mainly rely on static knowledge bases and search engines to query equipment parameters, troubleshooting, and operation procedures, and even attempt to apply general deep learning dialogue models. However, these existing technical solutions all have significant drawbacks: static knowledge bases and general search engines have difficulty accurately understanding natural language questions that contain professional equipment codes, standard fault codes, and on-site spoken language; due to the scarcity of professional data and poor interpretability, general deep learning models produce answers that are poorly relevant to the user's query information in energy scenarios with extremely high safety requirements. Summary of the Invention
[0004] To address the shortcomings of existing technologies, this invention provides an energy knowledge question-and-answer dialogue management method and system based on regular expression matching, which has the advantages of high accuracy and high efficiency, and solves the aforementioned technical problems.
[0005] To achieve the above objectives, the present invention provides the following technical solution: an energy knowledge question-answering dialogue management method based on regular expression matching, comprising the following steps:
[0006] S1: Construct a regular rule base and knowledge database for the energy operation and maintenance field;
[0007] S2: Preprocess the user input statement to obtain a standardized input statement, extract keywords from the standardized input statement to obtain a keyword list of the standardized input statement. ;
[0008] S3: Based on the regular expression rule base for energy operation and maintenance built on S1, match the S2 standard input statement to extract the list of slot values for the standard input statement. Data table name and query tags ;
[0009] S4: Slot value list of the standardized input statement obtained from S3 Data table name and A search was conducted in the knowledge database of the energy operation and maintenance field to obtain a set of candidate answers corresponding to the standardized input statement. ;
[0010] S5: Keyword list for input statements based on S2 specification Slot value list for S4 specification input statements For the S4 candidate answer set Scoring was performed to obtain candidate answer sets. Keyword score for each candidate answer And slots are worth dividing Keyword scores based on candidate answers And slots are worth dividing Get the overall score Select the overall score The highest-ranking candidate answer will be sent to the user as the reply answer;
[0011] S6: Users rate the answers provided by S5 and receive reminder instructions based on the ratings.
[0012] As a preferred technical solution of the present invention, the regular expression rule base in the S1 energy operation and maintenance field includes multiple rules, and the expression of the regular expression rule base is as follows:
[0013]
[0014] in, This represents a regular expression rule base for the energy operation and maintenance field. These represent the first regularity rule in the energy operation and maintenance field. Rule number, ..., number Rule number, ..., number Rules;
[0015] The first Rules Includes regular expressions, query intent labels, slot templates, and table names. Rules The expression is as follows:
[0016] in, Indicates the first Rules; Indicates the first The regular expression for the rule; Indicates the first The query intent tags for each rule; Indicates the first Slot template with rules; Indicates the first The name of the corresponding data table in the knowledge database when retrieving a rule;
[0017] The first Query intent tags of the rules This includes equipment parameter query, fault handling query, and operation step query; the first Rules Includes multiple slots, specifically ,in, They represent the first Rule # Slot number, ..., number Rule # Slot number, ..., number Rule # One slot, For the first The rule corresponds to the total number of slots in the slot template.
[0018] As a preferred technical solution of the present invention, S2 preprocesses the user input statement. The preprocessing steps specifically include: text cleaning, word segmentation, and stop word filtering of the user input statement.
[0019] The S2 keyword list The specific expression is:
[0020]
[0021] in, Indicates the first A list of keywords for a standardized input statement; They represent the first The first standardized input statement The first keyword, ..., the first The first keyword..., the first One keyword.
[0022] As a preferred technical solution of the present invention, the S3 extracts the list of slot values for the standard input statement. Data table name and query tags This includes the following steps:
[0023] S3.1: Based on regular expressions To standardize the rules of the regular expression rule base for matching input statements ;
[0024] S3.2: According to the rules Corresponding slot template Extract the slot value for each slot in the standard input statement. The specific expression is as follows:
[0025]
[0026] in, Indicates the first A list of slot values for a standard input statement; They represent the first The first of the standardized input statements The slot value, ..., the first The slot value, ..., the first Each slot value;
[0027] S3.3: Rules based on regular expression rule base Extract the first one respectively The name of the data table for each standardized input statement. and .
[0028] As a preferred technical solution of the present invention, the candidate answer set corresponding to the candidate standardized input statement in S4 The expression is:
[0029]
[0030] in, This represents the set of candidate answers corresponding to the current standard input statement; These represent the first and second statements corresponding to the current standard input statement. The candidate answer, ..., the first The candidate answer, ..., the first The candidate answer; the first one corresponding to the current standard input statement. The expressions for the candidate answers are as follows:
[0031]
[0032] in, This indicates the first number corresponding to the current standard input statement. 10 candidate answers; These represent the first and second statements corresponding to the current standard input statement. The first candidate answer The first keyword, ..., the first The first keyword, ..., the first One keyword; The first corresponding to the current standard input statement The first candidate answer The slot value, ..., the first The slot value, ..., the first Each slot value; This indicates the first number corresponding to the current standard input statement. The search content for each candidate answer.
[0033] As a preferred technical solution of the present invention, the keyword score of the candidate answer in S5 The formula is:
[0034]
[0035] in, Indicates the first Keyword scores for each candidate answer; Indicates the first The first standardized input statement One keyword; This indicates the first number corresponding to the current standard input statement. The first candidate answer One keyword; This function represents the number of identical words. This function represents the total number of words. This indicates the first [number]th [statement] corresponding to the current standard input statement. Sum all the keywords of each candidate answer;
[0036] The slot values of the candidate answers in S5 are scored. The formula is:
[0037]
[0038] in, Indicates the first Each candidate answer slot is worth points; Indicates the first The first standardized input statement Each slot value; This indicates the first number corresponding to the current standard input statement. The first candidate answer Each slot value; This function represents the number of identical words. This function represents the total number of words. This indicates the first [number]th [statement] corresponding to the current standard input statement. The summation operation is performed on all slot values of each candidate answer.
[0039] As a preferred technical solution of the present invention, the keyword score based on the candidate answer in S5 And slots are worth dividing Get the overall score Overall score The formula is:
[0040] in, Indicates the first The scores of each candidate answer; Indicates the first Keyword scores for each candidate answer; Indicates the first Each candidate answer slot is worth points; and These represent the weighting coefficients.
[0041] As a preferred technical solution of the present invention, in step S6, the user rates the answer given in step S5, and a reminder instruction is obtained based on the rate. The specific steps are as follows:
[0042] S6.1: Based on the user's rating of the response to each user input statement in the current dialogue, obtain the score for the current dialogue. ;
[0043] S6.2: Based on the score of the current dialogue, a reminder instruction is generated. The specific reminder instruction is as follows:
[0044] If the current dialogue score If the threshold is ≥, no reminder will be issued;
[0045] If the current dialogue score If a threshold is set, staff will be prompted to make improvements.
[0046] As a preferred embodiment of the present invention, step S6.1 obtains the score of the current dialogue. The formula is:
[0047]
[0048] in, Indicates the score of the current conversation; Indicates the user's response to the current conversation. The score for the response to each standardized input statement; This indicates that all replies in the current dialogue are summed.
[0049] This invention provides an energy knowledge question-and-answer dialogue management system based on regular expression matching, used to execute the aforementioned energy knowledge question-and-answer dialogue management method based on regular expression matching.
[0050] Compared with existing technologies, this invention provides an energy knowledge question-answering dialogue management method and system based on regular expression matching, which has the following beneficial effects:
[0051] This invention constructs a regular expression rule base and knowledge database in the field of energy operation and maintenance. It preprocesses user input statements to obtain standardized input statements, performs regular expression matching on these standardized input statements to obtain a list of slot values, table names, and query intent tags. Based on these slot value lists, table names, and query intent tags, it searches the knowledge database in the energy operation and maintenance field to obtain a set of candidate answers corresponding to the standardized input statements. Each candidate answer in the set is scored to obtain a keyword score and a slot value score. Based on these scores, a comprehensive score is obtained for each candidate answer. The candidate answer with the highest comprehensive score is selected as the reply answer and sent to the user. The user scores the reply answer and receives a reminder instruction based on the score. Attached Figure Description
[0052] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation
[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0054] Please see Figure 1 A dialogue management method for energy knowledge question answering based on regular expression matching includes the following steps:
[0055] S1: Construct a regular rule base and knowledge database for the energy operation and maintenance field;
[0056] The S1 Energy Operations and Maintenance Regular Expression Rule Base includes multiple rules, and the expressions in the regular expression rule base are as follows:
[0057]
[0058] in, This represents a regular expression rule base for the energy operation and maintenance field. These represent the first regularity rule in the energy operation and maintenance field. Rule number, ..., number Rule number, ..., number Rules;
[0059] No. Rules Includes regular expressions, query intent tags, and slot templates, the first Rules The expression is as follows:
[0060] in, Indicates the first Rules; Indicates the first The regular expression for the rule; Indicates the first The query intent tags for each rule; Indicates the first Slot template with rules; Indicates the first The name of the corresponding data table in the knowledge database when retrieving a rule;
[0061] No. Query intent tags of the rules Includes equipment parameter query, fault handling query, and operation procedure query; Rules Includes multiple slots, specifically ,in, They represent the first Rule # Slot number, ..., number Rule # Slot number, ..., number Rule # One slot, For the first The total number of slots contained in the corresponding slot template for each rule;
[0062] The regular expression rule base in the field of energy operation and maintenance contains approximately a number of rules. These rules are summarized and written by experts in the field of energy operation and maintenance. Each rule has a different regular expression, and each regular expression corresponds to a fixed slot template and a data table name. The data table name and query intent tag are used to retrieve the corresponding data table in the knowledge database, and the slot template is used to retrieve the corresponding attribute in the data table. Each rule is assigned a query intent tag to facilitate retrieval in the knowledge database of the field of energy operation and maintenance.
[0063] The construction of a knowledge database in the energy operation and maintenance field begins with cleaning and labeling the collected equipment parameter tables, fault solutions, and SOP libraries. This includes removing invalid records, standardizing terminology, and extracting key entities, which are keywords. The cleaned and labeled data is then divided into multiple tables. The knowledge database comprises these tables, each named with a query intent tag, keyword, slot, and content. Query intent tags in the table are specifically named as equipment parameter query, fault handling query, and operation step query. Keywords are specifically named as equipment, cooling system, and handling, etc. Slots are specifically named as inverter, operating temperature, and range, etc. Content in the table includes specific information such as equipment parameter information, fault handling information, and fault handling steps.
[0064] S2: Preprocess the user input statement to obtain a standardized input statement, extract keywords from the standardized input statement to obtain a keyword list of the standardized input statement. ;
[0065] S2 preprocesses the user input statement. The preprocessing steps are: text cleaning, word segmentation, and stop word filtering of the user input statement.
[0066] Text cleaning is the process of removing extra spaces and special characters from user input statements; word segmentation is the process of dividing the cleaned user input statements into word sequences; stop word filtering is the process of removing stop words such as prepositions, conjunctions, articles, and pronouns from the segmented user input statements; and keyword extraction is the process of extracting core nouns, quantifiers, and verbs from the standardized input statements as keywords.
[0067] S2 Keyword List The specific expression is:
[0068]
[0069] in, Indicates the first A list of keywords for a standardized input statement; They represent the first The first standardized input statement The first keyword, ..., the first The first keyword..., the first One keyword;
[0070] S3: Based on the regular expression rule base for energy operation and maintenance built on S1, match the S2 standard input statement to extract the list of slot values for the standard input statement. Data table name and query tags ;
[0071] S3 Extracts Standard Input Statement Slot Value List Data table name and query tags This includes the following steps:
[0072] S3.1: Based on regular expressions To standardize the rules of the regular expression rule base for matching input statements ;
[0073] S3.2: According to the rules Corresponding slot template Extract the slot value for each slot in the standard input statement. The specific expression is as follows:
[0074]
[0075] in, Indicates the first A list of slot values for a standard input statement; They represent the first The first of the standardized input statements The slot value, ..., the first The slot value, ..., the first Each slot value;
[0076] S3.3: Rules based on regular expression rule base Extract the first one respectively The name of the data table for each standardized input statement. and ;
[0077] S4: Slot value list of the standardized input statement obtained from S3 Data table name and query tags A search was conducted in the knowledge database of the energy operation and maintenance field to obtain a set of candidate answers corresponding to the standardized input statement. ;
[0078] The candidate answer set corresponding to the candidate standardized input statement in S4 The expression is:
[0079]
[0080] in, This represents the set of candidate answers corresponding to the current standard input statement; These represent the first and second statements corresponding to the current standard input statement. The candidate answer, ..., the first The candidate answer, ..., the first The candidate answer; the first one corresponding to the current standard input statement. The expressions for the candidate answers are as follows:
[0081]
[0082] in, This indicates the first number corresponding to the current standard input statement. 10 candidate answers; These represent the first and second statements corresponding to the current standard input statement. The first candidate answer The first keyword, ..., the first The first keyword, ..., the first One keyword; The first corresponding to the current standard input statement The first candidate answer The slot value, ..., the first The slot value, ..., the first Each slot value; This indicates the first number corresponding to the current standard input statement. The search content for each candidate answer;
[0083] Each candidate answer is a response that combines keywords, slot values, and content retrieved from the knowledge database and is then sent to the user.
[0084] S5: Keyword list for input statements based on S2 specification Slot value list for S4 specification input statements For the S4 candidate answer set Scoring was performed to obtain candidate answer sets. Keyword score for each candidate answer and slot value Score, based on keyword scores of candidate answers And slots are worth dividing Get the overall score Select the overall score The highest-ranking candidate answer will be sent to the user as the reply answer;
[0085] Keyword scores of candidate answers in S5 The formula is:
[0086]
[0087] in, Indicates the first Keyword scores for each candidate answer; Indicates the first The first standardized input statement One keyword; This indicates the first number corresponding to the current standard input statement. The first candidate answer One keyword; This function represents the method to find the number of identical characters, obtained by iterating through the keywords of the current standard input statement and candidate answers. This function, which calculates the total number of words, is a commonly used programming function. This indicates the first [number]th [statement] corresponding to the current standard input statement. Sum all the keywords of each candidate answer;
[0088] Indicates the first The first standardized input statement keyword and its corresponding first The keyword matching degree of each candidate answer. Indicates the first The first standardized input statement The keyword and its corresponding first The first candidate answer The number of identical words in each keyword; Indicates the first The first standardized input statement The keyword and its corresponding first The first candidate answer The number of identical words in the keyword accounts for the [number]th [percentage]. The first of the standardized input statements The ratio of the number of characters in each keyword is considered a higher value, indicating that the input statement is more standardized. The first keyword and candidate answer The higher the keyword matching degree;
[0089] The overall formula represents the degree of matching between all keywords in the standardized input statement and all keywords in the corresponding candidate answers. The larger the value, the greater the keyword matching degree.
[0090] The slots for candidate answers in S5 are worth points. The formula is:
[0091]
[0092] in, Indicates the first Each candidate answer slot is worth points; Indicates the first The first standardized input statement Each slot value; This indicates the first number corresponding to the current standard input statement. The first candidate answer Each slot value; This function represents the method to find the number of identical characters, obtained by iterating through the keywords of the current standard input statement and candidate answers. This function, which calculates the total number of words, is a commonly used programming function. This indicates the first [number]th [statement] corresponding to the current standard input statement. Sum of all slot values for each candidate answer;
[0093] Indicates the first The first standardized input statement The value of the slot and its corresponding first slot value The first candidate answer Matching degree of each slot value Indicates the first The first standardized input statement The value of the slot and its corresponding first slot value The first candidate answer The number of words with the same value in each slot; Indicates the first The first standardized input statement The value of the slot and its corresponding first slot value The first candidate answer The number of slots with the same value accounts for the percentage of slots with the same character. The first of the standardized input statements The ratio of the number of characters in each slot value is considered higher if the value is larger, indicating that the input statement is in the correct order. The value of the slot and the candidate answer of the first slot. The greater the matching degree of the slot values;
[0094] The overall formula represents the matching degree between all slot values of the standardized input statement and all slot values of its corresponding candidate answers. The larger the value, the greater the matching degree of the slot values.
[0095] Keyword scoring based on candidate answers in S5 And slots are worth dividing Get the overall score Overall score The formula is:
[0096] in, Indicates the first The scores of each candidate answer; Indicates the first Keyword scores for each candidate answer; Indicates the first Each candidate answer slot is worth points; and These represent the weighting coefficients;
[0097] S6: Users rate the answers provided in S5 and receive reminder instructions based on the ratings;
[0098] In S6, users rate the answers provided in S5, and receive notification instructions based on these ratings. The specific steps are as follows:
[0099] S6.1: Based on the user's rating of the response to each user input statement in the current dialogue, obtain the score for the current dialogue. ;
[0100] Score of the current conversation The formula is:
[0101]
[0102] in, Indicates the score of the current conversation; Indicates the user's response to the current conversation. The score for the response to each standardized input statement; This indicates that the summation operation is performed on all responses in the current dialogue;
[0103] S6.2: Based on the score of the current dialogue, a reminder instruction is generated. The specific reminder instruction is as follows:
[0104] If the current dialogue score If the threshold is ≥, no reminder will be issued;
[0105] If the current dialogue score If a threshold is set, staff will be reminded to make improvements.
[0106] An energy knowledge question-and-answer dialogue management system based on regular expression matching is used to implement an energy knowledge question-and-answer dialogue management method based on regular expression matching.
[0107] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A regular matching based energy knowledge QnA dialogue management method, characterized in that: Comprise the following steps: S1: Constructing the regular rule base and knowledge database in the field of energy operation and maintenance; S2: pre-process the user input sentence to obtain a standard input sentence, extract keywords from the standard input sentence to obtain a keyword list of the standard input sentence ; S3: match the S2 specification input sentence based on the energy operation and maintenance field regular rule library constructed in S1, and extract the specification input sentence slot value list , data table name , and query label ; S4: obtaining a slot value list of the canonical input sentence based on the slot value list obtained in S3 , a data table name , and a query label , searching in a knowledge database in the energy operation and maintenance field to obtain a candidate answer set corresponding to the canonical input sentence ; S5: get the keyword list of the input sentence based on S2 and the slot value list of the input sentence based on S4 , score the S4 candidate answer set , get the keyword score and slot value score of each candidate answer of the candidate answer set , get the comprehensive score based on the keyword score and slot value score of the candidate answer , select the candidate answer with the highest comprehensive score as the reply answer to reply to the user; S6: The user scores the answer of S5, and gets the reminder instruction based on the score.
2. The regular matching based energy knowledge QnA dialogue management method according to claim 1, characterized in that: The S1 regular rule base in the field of energy operation and maintenance comprises a plurality of rules, and the expression of the regular rule base is as follows: ； wherein, represents a regular rule library in the field of energy operation and maintenance; respectively represent the first rule, the second rule, …, the nth rule in the regular rule library in the field of energy operation and maintenance; respectively represent the first rule, the second rule, …, the nth rule in the regular rule library in the field of energy operation and maintenance; respectively represent the first rule, the second rule, …, the nth rule in the regular rule library in the field of energy operation and maintenance; respectively represent the first rule, the second rule, …, The first Article rule including regular expressions, query intent labels, slot templates, and data table names, the first Article rule The expression is as follows: ; wherein, represents the rule of the th clause; represents a regular expression of the rule of the th clause; represents a query intent tag of the rule of the th clause; represents a slot template of the rule of the th clause; represents a data table name of the knowledge database corresponding to the rule retrieval of the th clause; The first Query intent label of the rule Including device parameter query, fault handling query and operation step query; the first Rule Including a plurality of slots, specifically , wherein Respectively represent the first Slot of the first rule, …, the first Slot of the first rule, …, the first Slot of the first rule, …, the first Slot of the first rule, …, the first Slot of the first rule, the first Slot of the first rule The total number of slots contained in the slot template corresponding to the first Rule 3. The regular matching based energy knowledge Q&A dialogue management method according to claim 2, characterized in that: The S2 pre-processes the user input sentence, and the pre-processing steps are specifically: text cleaning, word segmentation processing and stop word filtering on the user input sentence; The S2 keyword list The specific expression is: ； wherein, denotes a keyword list of the th canonical input sentence; denotes the th keyword, the th keyword,..., the th keyword,..., the th keyword of the th canonical input sentence, respectively.
4. The regular matching based energy knowledge Q&A dialogue management method according to claim 3, characterized in that: The S3 extraction specification input sentence slot value list , data table name , and query label , comprising the following steps: S3.1: Based on regular expressions To match the regular rules library rules for the normative input sentence ; S3.2: According to the rule Corresponding slot template The slot value of each slot of the specification input sentence is extracted, and the specific expression is as follows: ； wherein, represents a slot value list of the th canonical input sentence; respectively represents the th slot value, the th slot value, …, the th slot value, …, the th slot value of the th canonical input sentence. S3.3: Rules based on canonical rule base , respectively extract the data table name and of the canonical input sentence.
5. The regular matching based energy knowledge QnA dialogue management method according to claim 4, characterized in that: The candidate answer set corresponding to the candidate canonical input sentence in the S4 The expression is: ； wherein, denotes the candidate answer set corresponding to the current canonical input sentence; denote the first candidate answer, …, the th candidate answer, …, the th candidate answer, …, the th candidate answer corresponding to the current canonical input sentence, respectively; The expression of the th candidate answer corresponding to the current canonical input sentence is as follows: ； in, This indicates the first number corresponding to the current standard input statement. 10 candidate answers; These represent the first and second statements corresponding to the current standard input statement. The first candidate answer The first keyword, ..., the first The first keyword, ..., the first One keyword; The first corresponding to the current standard input statement The first candidate answer The slot value, ..., the first The slot value, ..., the first Each slot value; This indicates the first number corresponding to the current standard input statement. The search content for each candidate answer.
6. The regular matching based energy knowledge QnA dialogue management method according to claim 5, characterized in that: The keyword score of the candidate answer in S5 The formula is: ； wherein, denotes the keyword score of the th candidate answer; denotes the th keyword of the th canonical input sentence; denotes the th keyword of the th candidate answer corresponding to the current canonical input sentence; denotes a function that counts the same number of words; denotes a function that counts the total number of words; denotes a summation operation on all keywords of the th candidate answer corresponding to the current canonical input sentence; The slot value score of the candidate answer in S5 The formula is: ； wherein, denotes the slot value score of the th candidate answer; denotes the th slot value of the th canonical input sentence; denotes the th slot value of the th candidate answer corresponding to the current canonical input sentence; denotes a function to count the same number of words; denotes a function to count the total number of words; denotes a summation operation on all slot values of the th candidate answer corresponding to the current canonical input sentence.
7. The regular matching based energy knowledge Q&A dialogue management method according to claim 6, characterized in that: The keyword score based on the candidate answer in S5 And slot value score Get the comprehensive score The comprehensive score The formula is: wherein, denotes the score of the th candidate answer; denotes the score of the th candidate answer keyword; denotes the score of the th candidate answer slot value; and denote the weight coefficients, respectively.
8. The regular matching based energy knowledge QnA dialogue management method according to claim 7, characterized in that: In the S6, the user scores the answer of S5, and gets the reminder instruction based on the score, and the specific steps are: S6.1: obtaining a score of the current dialogue based on the user's reply answer score of each user input sentence of the current dialogue ; S6.2: Based on the score of the current dialogue, the reminder instruction is obtained, and the reminder instruction is specifically: If the score of the current dialogue ≥ the set threshold, no reminder is taken. If the score of the current conversation < Set threshold, then alert person to improve.
9. The regular matching based energy knowledge QnA dialogue management method according to claim 8, characterized in that: S6.1 is the score of the current dialogue The formula is: ； wherein, represents the score of the current dialogue; represents the score of the reply answer corresponding to the th canonical input sentence of the user to the current dialogue; represents summing all the reply answers of the current dialogue.
10. A regular matching based energy knowledge question and answer dialog management system, characterized in that: A regular matching-based energy knowledge question and answer dialogue management method for executing any one of claims 1-9.