CN111858850A - Method for realizing accurate and rapid scoring of question and answer on intelligent customer service - Google Patents
Method for realizing accurate and rapid scoring of question and answer on intelligent customer service Download PDFInfo
- Publication number
- CN111858850A CN111858850A CN202010609405.0A CN202010609405A CN111858850A CN 111858850 A CN111858850 A CN 111858850A CN 202010609405 A CN202010609405 A CN 202010609405A CN 111858850 A CN111858850 A CN 111858850A
- Authority
- CN
- China
- Prior art keywords
- score
- matching
- scores
- word segmentation
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides a method for realizing accurate and quick scoring of questions and answers on an intelligent customer service, which comprises the following steps: establishing a knowledge route associated with a knowledge base, and matching the corresponding knowledge route according to the characteristics of the problem, wherein the knowledge route comprises a question and answer template, a search engine and semantic analysis; performing word segmentation on the problem through a word segmentation device hand to obtain words after word segmentation, and enabling the words after word segmentation to be matched with routing keywords of a target knowledge base according to corresponding knowledge routing; and acquiring a target answer corresponding to the question from a target knowledge base based on a final score obtained by full-text index score, semantic similarity score, part of speech weighting score and matching frequency score, and returning the target answer to the user. According to the embodiment of the invention, the final score is obtained through full-text index score, semantic similarity score, part of speech weighting score and matching frequency score, so that the most appropriate answer is matched, and accurate answer and quick score are realized.
Description
Technical Field
The invention relates to the technical field of intelligent customer service, in particular to a method for realizing accurate and quick grading of questions and answers on the intelligent customer service.
Background
With the continuous development of internet financial technology, more and more technologies are applied to the financial field, wherein intelligent customer service in the financial field relates to numerous technical applications.
At present, the question-answer function of the intelligent customer service is realized in three modes, one mode is to find a correspondingly matched question-answer template according to the format of a question of a questioner and then return the question, the mode is high in efficiency, the question format template needs to be maintained all the time, and once a sentence of the question is not in the template range, the question cannot be matched with the answer. The second is a search engine mode, which is used for segmenting words of a sentence, then storing keywords, segmenting words of a user question when the user asks the question, and then matching the words through the search engine. The third is to use semantic analysis to ask and answer, which is very flexible and can match answers to questions through semantic analysis and word stock training, but this method has high performance requirement, low efficiency and easy 'answer not all questions', and it is more important for intelligent customer service to match answers required by users. It is most important for intelligent customer service to accurately and quickly return answers required by users by means of matching scores.
Each of the three question-answering modes has its own disadvantages, and if the three question-answering modes are combined, the combination will be more humanized, so that how to realize accurate and rapid scoring of the question-answering on the intelligent customer service is needed on the basis of the prior art, and the trend of modernization of the internet of things is.
Summary of the invention
In order to overcome the defects of the prior art, the invention provides a method for realizing the accurate and quick grading of questions and answers on an intelligent customer service, which is used for improving the problem that the questions put forward by a user in the intelligent customer service can be quickly matched with answers and accurately matched with answers.
The technical scheme adopted by the invention for solving the technical problems is as follows: a method for realizing accurate and quick scoring of questions and answers on an intelligent customer service comprises the following steps: establishing a knowledge route associated with a knowledge base, and matching the corresponding knowledge route according to the characteristics of the problem, wherein the knowledge route comprises a question and answer template, a search engine and semantic analysis; performing word segmentation on the problem through a word segmentation device hand to obtain words after word segmentation, and enabling the words after word segmentation to be matched with routing keywords of a target knowledge base according to corresponding knowledge routing; and acquiring a target answer corresponding to the question from a target knowledge base based on a final score obtained by full-text index score, semantic similarity score, part of speech weighting score and matching frequency score, and returning the target answer to the user.
Preferably, the step of obtaining a target answer corresponding to the question from a target knowledge base based on a final score obtained by the full-text index score, the semantic similarity score, the part-of-speech weighted score and the matching frequency score includes:
matching the words after word segmentation with 6 matching results with the highest scores through the BM25 algorithm of the ElastcSearch full-text index;
acquiring full-text index scores (full score of 20) corresponding to the highest 6 matching results;
Score(Q,d)＝SUM(Wi*R(qi,d)) Wi＝IDF(qi)＝log((N-n(qi)+0.5)/(n(qi)+0.5))
R(qi,d)＝fi(k1+1)/(fi+K) K＝k1*(1-b+b*(dl/avg(dl)))
the character string after word segmentation is q1, q2, q3, …, qn, N is the total document number in the index, N (qi) is the document number containing the word segmentation qi, D is the search result, Wi is the correlation weight for matching the word segmentation of qi and the index document, k1 and b are algorithm adjustable parameters, dl is the length of the index document D, and avgdl is the average length of all the texts in the index text set D.
Preferably, the step of obtaining a target answer corresponding to the question from a target knowledge base based on a final score obtained by the full-text index score, the semantic similarity score, the part-of-speech weighted score and the matching frequency score includes:
calculating 6 matching results with the highest matching scores of the words after word segmentation by using a Python synnyms frame through semantic similarity;
Obtaining semantic similarity matching scores (full score of 20) corresponding to the highest 6 matching results;
d1＝max(compare(a1,b1),compare(a1,b2),…compare(a1,bm))；
d2＝max(compare(a2,b1),compare(a2,b2),…compare(a2,bm))；
…
dn＝max(compare(an,b1),compare(an,b2),…compare(an,bm))；
the semantic similarity matching score ═ avg (d1, d2, …, dn);
wherein: and the word set after word segmentation comprises Wi ═ a1, a2, … and an, the word set of the result matched with the search engine comprises Wj ═ { b1, b2, … and bm }, match (a and b) represents the distance between the word a and the word b, the value range is [0,1], and d represents the distance between the words.
Preferably, the step of obtaining a target answer corresponding to the question from a target knowledge base based on a final score obtained by the full-text index score, the semantic similarity score, the part-of-speech weighted score and the matching frequency score includes:
calculating 6 matching results with the highest matching scores of the segmented words through a special (a) function;
obtaining part-of-speech weighted scores of the top 6 matching results;
the semantic similarity matching score ═ avg (s1, s2, …, sn);
wherein: s is a part-of-speech weighted score, a word set after word segmentation is Wi ═ a1, a2, …, an, and special (a) is a special noun score function, the special noun score function is managed by a background management system, a score is set for a special noun, and the score is obtained through the special (a) function.
Preferably, the step of obtaining a target answer corresponding to the question from a target knowledge base based on a final score obtained by the full-text index score, the semantic similarity score, the part-of-speech weighted score and the matching frequency score includes:
Obtaining 6 matching results with highest matching frequency scores according to 20 (matching frequency of the problem in the database/highest problem matching frequency of the current database);
the match frequency score of the top 6 match results is obtained.
Preferably, the final score calculation method:
a final score of w1 full text index score + w2 semantic similarity score + w3 parts of speech weighted score + w4 match frequency score;
wherein: w1, w2, w3 and w4 are weighted values of four types of scores, the initialization values are all 0.5, the value range is [0,1], the values belong to a configurable value, and a system administrator adjusts the values according to actual question-answer matching results.
Preferably, the obtaining of the target answer corresponding to the question from the target knowledge base based on the final score obtained by the full-text index score, the semantic similarity score, the part-of-speech weighted score and the matching frequency score comprises:
when a user proposes a problem, if the user matches a small speech library and the small speech library has no matching result, obtaining the matching result through a third party API interface; or
When a user proposes a problem, if the user matches the small speech library and the small speech library has no matching result, the matching result is obtained through calculation of a neural network algorithm.
Preferably, the matching the corresponding knowledge route according to the features of the question includes:
The number of the same elements in the routing keyword set Ki ═ { k1, k2, … kn } and the word set W ═ a1, a2, …, am } is determined as di.
Preferably, the matching the corresponding knowledge route according to the features of the question includes:
and acquiring a maximum value max (di) and determining the knowledge base route with the highest matching degree.
Preferably, the routing according to the corresponding knowledge so that matching the word after word segmentation with the routing keyword of the target knowledge base comprises:
and setting a corresponding word segmentation strategy according to the corresponding knowledge route, so that the word after word segmentation is matched with the route keyword of the target knowledge base.
The invention has the beneficial effects that: the final score is obtained through the corresponding full-text index score, the semantic similarity score, the part of speech weighting score and the matching frequency score, and the accuracy and the efficiency of the question matching answer are improved.
Drawings
FIG. 1 is a flow diagram of a method for implementing accurate and fast scoring of questions and answers on an intelligent customer service.
Fig. 2 is a scoring schematic diagram of a method for implementing accurate and rapid scoring of questions and answers on an intelligent customer service.
Detailed Description
The invention is further illustrated with reference to the following figures and examples.
The conception, the specific structure, and the technical effects produced by the present invention will be clearly and completely described below in conjunction with the embodiments and the accompanying drawings to fully understand the objects, the features, and the effects of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments, and those skilled in the art can obtain other embodiments without inventive effort based on the embodiments of the present invention, and all embodiments are within the protection scope of the present invention. In addition, all the connection/connection relations referred to in the patent do not mean that the components are directly connected, but mean that a better connection structure can be formed by adding or reducing connection auxiliary components according to specific implementation conditions. All technical characteristics in the invention can be interactively combined on the premise of not conflicting with each other.
With reference to FIG. 1
S101, establishing a knowledge route associated with a knowledge base, and matching the corresponding knowledge route according to the characteristics of a question, wherein the knowledge route comprises a question and answer template, a search engine and semantic analysis;
establishing a knowledge route associated with a knowledge base, wherein the knowledge base comprises a cold conversation base, a question and answer base, a service base and the like, matching the corresponding knowledge route according to the characteristics of the question, and the knowledge route comprises a question and answer template, a search engine and semantic analysis, wherein when a user puts forward the question, for example, the question is: what will the smart POS fail? Matching answers through the knowledge routing path if the question exists in the question-answer template; if the question is retrieved by the search engine, matching answers according to a strategy needing word segmentation; if the question is searched through semantic analysis, the question needs to be subjected to semantic analysis, answers corresponding to semantics which are similar to or identical to the semantics of the question are searched, the answers are matched with the knowledge base, multiple ways for searching corresponding answers are provided, and the searching efficiency is improved.
S102, performing word segmentation on the problem through a word segmentation device hand to obtain words after word segmentation, and matching the words after word segmentation with the routing keywords of the target knowledge base according to the corresponding knowledge routing;
When the user proposes "what are the smart POS opportunities that have failed? When answers cannot be retrieved by a question-answer template and a semantic analysis knowledge route, the question is retrieved in a search engine mode, then the question is subjected to word segmentation processing through a word segmentation device hand to obtain words after word segmentation, the question is divided into words of various parts of speech, and a corresponding knowledge base is searched and matched. The system administrator collects the service questions and answers, collects the service questions and answers into a question-answer mapping relation, constructs a knowledge base in advance, trains the knowledge base to obtain the trained knowledge base, and matches the corresponding knowledge base according to the attributes of the questions so as to quickly match the corresponding knowledge base and improve the efficiency of searching the knowledge base.
And S103, acquiring a target answer corresponding to the question from a target knowledge base based on a final score obtained by full-text index score, semantic similarity score, part of speech weighting score and matching frequency score, and returning the target answer to the user.
Scoring is carried out through full-text index, semantic similarity, part of speech weighting and matching frequency to obtain final scoring, target answers corresponding to the problems are obtained from a target knowledge base and returned to the user, and when the user raises the problems, if the matching is carried out to a small speech database and no matching result exists in the small speech database, the matching result is obtained through a third party API interface; or when the user proposes a problem, if the user matches the small speech library and the small speech library has no matching result, the matching result is obtained through calculation of a neural network algorithm, so that the accuracy and the efficiency of matching scoring are realized.
Refer to FIG. 2
S201, scoring of full-text index
Matching the words after word segmentation with 6 matching results with the highest scores through the BM25 algorithm of the ElastcSearch full-text index;
obtaining the matching score (full score 20) of the search engine corresponding to the highest 6 matching results;
Score(Q,d)＝SUM(Wi*R(qi,d)) Wi＝IDF(qi)＝log((N-n(qi)+0.5)/(n(qi)+0.5))
R(qi,d)＝fi(k1+1)/(fi+K) K＝k1*(1-b+b*(dl/avg(dl)))
the character string after word segmentation is set to be qi-q 1, q2, q3, …, qn, N is the number of all documents in the index, N (qi) is the number of documents containing the word segmentation qi, D is a search result, Wi is a correlation weight for matching the qi segmentation and the index document, k1 and b are algorithm adjustable parameters, dl, and avgdl are the length of the index document D and the average length of all the texts in the index text set D respectively. The scoring mode of the search engine eliminates the problem that most semantics are basically not related, and improves the efficiency of searching the words after word segmentation.
S202, semantic similarity scoring
Calculating 6 matching results with the highest matching scores of the words after word segmentation by using a Python synnyms frame through semantic similarity;
and acquiring semantic similarity matching scores (full score of 20) corresponding to the top 6 matching results.
d1＝max(compare(a1,b1),compare(a1,b2),…compare(a1,bm))；
d2＝max(compare(a2,b1),compare(a2,b2),…compare(a2,bm))；
…
dn＝max(compare(an,b1),compare(an,b2),…compare(an,bm))；
Semantic similarity matching score ═ avg (d1, d2, …, dn);
wherein: and the word set after word segmentation comprises Wi ═ a1, a2, … and an, the word set of the result matched with the search engine comprises Wj ═ { b1, b2, … and bm }, match (a and b) represents the distance between the word a and the word b, the value range is [0,1], and d represents the distance between the words.
S203, weighted grading of parts of speech
Calculating 6 matching results with the highest matching scores of the segmented words through a special (a) function;
and acquiring the part-of-speech weighted scores of the top 6 matching results.
Obtaining word segmentation results of the 6 matching results and word segmentation results of the problems, performing word class weighted scoring according to specific nouns (16-20 points) > common nouns (15 points) > verbs (10 points) > other words (5 points), finally obtaining an average word class weighted scoring, and calculating the distance between Wi and Wj through a near word framework synnym, wherein the word class weighted scoring calculation method comprises the following steps:
semantic similarity matching score ═ avg (s1, s2, …, sn);
wherein: s is a part-of-speech weighted score, a word set after word segmentation is Wi ═ a1, a2, …, an, and special (a) is a special noun score function, the special noun score function is managed by a background management system, a score is set for a special noun, and the score is obtained through the special (a) function.
S204, matching frequency scoring
The matching frequency of 6 matching results in the database is obtained, the matching frequency score is obtained according to 20 (the matching frequency of the question in the database/the highest matching frequency of the question in the current database), the higher the matched frequency of the corresponding question is, the higher the score is, and the probability of matching the question to the answer required by the user is improved in such a way.
S205, final grading
Final score w1 full text index score + w2 semantic similarity score + w3 part of speech weighted score + w4 match frequency score.
Wherein: w1, w2, w3 and w4 are weighted values of four types of scores, the initialization values are all 0.5, the value range is [0,1], the values belong to a configurable value, and a system administrator adjusts the values according to actual question-answer matching results, for example, if the system administrator feels that the matching score of a search engine is more important to answer matching, the value of w1 is adjusted, or the values of w2, w3 and w4 are correspondingly reduced; or the system administrator feels the semantic similarity score to be more important for answer matching, the value of w2 is adjusted up or the values of w1, w3 and w4 are correspondingly adjusted down, and the like.
In the embodiment of the application, the final score is obtained through full-text index score, semantic similarity score, part of speech weighting score and matching frequency score, so that the most appropriate answer is matched, and the accuracy and efficiency of the answer to question matching are improved.
For the principle of knowledge base routing and word segmentation:
and (3) setting the routing keyword set of the knowledge base i as Ki { k1, k2, … kn }, setting the word set after word segmentation as W { a1, a2, …, am }, solving the same element number of the routing keyword set Ki and the word set W as di, obtaining the maximum value max (di), determining the knowledge base route with the highest matching degree, and if the plurality of knowledge base routes have the same score, determining the finally selected knowledge base route according to the weighted value of the knowledge base routes configured in the background. Generally, partial keyword matching is firstly carried out on the problems, then the problems are graded for the first time, and the knowledge base with the highest grade is selected for subsequent matching, so that on one hand, the subsequent matching times are reduced, the retrieval efficiency is improved, on the other hand, the matching of knowledge irrelevant to the problems is reduced, and the matching accuracy is improved.
And adopting different word segmentation strategies for routing according to corresponding knowledge.
Like cold-talk library strategy: the word segmentation range is widest, other words such as word meaning, such as nouns, verbs, and directional words, also participate in word segmentation, and word groups, articles, and punctuation marks do not participate in word segmentation, for example: does hello be; the question-answering strategy comprises the following steps: just nouns and verbs, for example: the intelligent POS machine has a fault; and (3) service strategies: just the nouns, for example: an intelligent POS machine; the system adopts different word segmentation strategies according to different types of the knowledge base, wherein one part is part of speech, the other part is a weighted word, and the different word segmentation strategies can provide greater flexibility and accuracy for subsequent scoring.
While the preferred embodiments of the present invention have been illustrated and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (10)
1. A method for realizing accurate and rapid scoring of questions and answers on an intelligent customer service is characterized by comprising the following steps:
establishing a knowledge route associated with a knowledge base, and matching the corresponding knowledge route according to the characteristics of the problem, wherein the knowledge route comprises a question and answer template, a search engine and semantic analysis;
Performing word segmentation on the problem through a word segmentation device hand to obtain words after word segmentation, and enabling the words after word segmentation to be matched with routing keywords of a target knowledge base according to corresponding knowledge routing;
and acquiring a target answer corresponding to the question from a target knowledge base based on a final score obtained by full-text index score, semantic similarity score, part of speech weighting score and matching frequency score, and returning the target answer to the user.
2. The method for achieving precise and rapid scoring of questions and answers on an intelligent customer service according to claim 1, wherein said step of obtaining target answers corresponding to questions from a target knowledge base based on final scores obtained from full-text index scores, semantic similarity scores, part-of-speech weighted scores and matching frequency scores comprises:
matching the words after word segmentation with 6 matching results with the highest scores through the BM25 algorithm of the ElastcSearch full-text index;
obtaining full-text index scores corresponding to the highest 6 matching results, wherein the full score is 20 scores;
Score(Q,d)＝SUM(Wi*R(qi,d))Wi＝IDF(qi)＝log((N-n(qi)+0.5)/(n(qi)+0.5))
R(qi,d)＝fi(k1+1)/(fi+K)K＝k1*(1-b+b*(dl/avg(dl)))
the character string after word segmentation is q1, q2, q3, …, qn, N is the total document number in the index, N (qi) is the document number containing the word segmentation qi, D is the search result, Wi is the correlation weight for matching the word segmentation of qi and the index document, k1 and b are algorithm adjustable parameters, dl is the length of the index document D, and avgdl is the average length of all the texts in the index text set D.
3. The method for achieving precise and rapid scoring of questions and answers on an intelligent customer service according to claim 1, wherein said step of obtaining target answers corresponding to questions from a target knowledge base based on final scores obtained from full-text index scores, semantic similarity scores, part-of-speech weighted scores and matching frequency scores comprises:
calculating 6 matching results with the highest matching scores of the words after word segmentation by using a Python synnyms frame through semantic similarity;
obtaining semantic similarity matching scores corresponding to the highest 6 matching results, wherein the full score is 20;
d1＝max(compare(a1,b1),compare(a1,b2),…compare(a1,bm))；
d2＝max(compare(a2,b1),compare(a2,b2),…compare(a2,bm))；
…
dn＝max(compare(an,b1),compare(an,b2),…compare(an,bm))；
the semantic similarity matching score ═ avg (d1, d2, …, dn);
wherein: and the word set after word segmentation comprises Wi ═ a1, a2, … and an, the word set of the result matched with the search engine comprises Wj ═ { b1, b2, … and bm }, match (a and b) represents the distance between the word a and the word b, the value range is [0,1], and d represents the distance between the words.
4. The method for achieving precise and rapid scoring of questions and answers on an intelligent customer service according to claim 1, wherein said step of obtaining target answers corresponding to questions from a target knowledge base based on final scores obtained from full-text index scores, semantic similarity scores, part-of-speech weighted scores and matching frequency scores comprises:
Calculating 6 matching results with the highest matching scores of the segmented words through a special (a) function;
obtaining part-of-speech weighted scores of the top 6 matching results;
the semantic similarity matching score ═ avg (s1, s2, …, sn);
wherein: s is a part-of-speech weighted score, a word set after word segmentation is Wi ═ a1, a2, …, an, and special (a) is a special noun score function, the special noun score function is managed by a background management system, a score is set for a special noun, and the score is obtained through the special (a) function.
5. The method for achieving precise and rapid scoring of questions and answers on an intelligent customer service according to claim 1, wherein said step of obtaining target answers corresponding to questions from a target knowledge base based on final scores obtained from full-text index scores, semantic similarity scores, part-of-speech weighted scores and matching frequency scores comprises:
obtaining 6 matching results with highest matching frequency scores according to 20 (matching frequency of the problem in the database/highest problem matching frequency of the current database);
the match frequency score of the top 6 match results is obtained.
6. The method for realizing the accurate and fast scoring of the questions and answers on the intelligent customer service according to any one of claims 2 to 5, wherein the final scoring calculation method comprises the following steps:
A final score of w1 full text index score + w2 semantic similarity score + w3 parts of speech weighted score + w4 match frequency score;
wherein: w1, w2, w3 and w4 are weighted values of four types of scores, the initialization values are all 0.5, the value range is [0,1], the values belong to a configurable value, and a system administrator adjusts the values according to actual question-answer matching results.
7. The method for achieving precise and rapid scoring of questions and answers on the intelligent customer service as claimed in claim 1, wherein the obtaining of the target answers corresponding to the questions from the target knowledge base based on the final scores obtained from the full-text index score, the semantic similarity score, the part-of-speech weighted score and the matching frequency score comprises:
when a user proposes a problem, if the user matches a small speech library and the small speech library has no matching result, obtaining the matching result through a third party API interface; or
When a user proposes a problem, if the user matches the small speech library and the small speech library has no matching result, the matching result is obtained through calculation of a neural network algorithm.
8. The method of claim 1, wherein said matching the corresponding knowledge routing according to the characteristics of the question comprises:
The number of the same elements in the routing keyword set Ki ═ { k1, k2, … kn } and the word set W ═ a1, a2, …, am } is determined as di.
9. The method of claim 1, wherein said matching the corresponding knowledge routing according to the characteristics of the question comprises:
and acquiring a maximum value max (di) and determining the knowledge base route with the highest matching degree.
10. The method for achieving precise and rapid question and answer scoring on an intelligent customer service according to claim 1, wherein the step of matching the participled words with the routing keywords of the target knowledge base according to the corresponding knowledge routing comprises the following steps:
and setting a corresponding word segmentation strategy according to the corresponding knowledge route, so that the word after word segmentation is matched with the route keyword of the target knowledge base.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010609405.0A CN111858850A (en) | 2020-06-29 | 2020-06-29 | Method for realizing accurate and rapid scoring of question and answer on intelligent customer service |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010609405.0A CN111858850A (en) | 2020-06-29 | 2020-06-29 | Method for realizing accurate and rapid scoring of question and answer on intelligent customer service |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111858850A true CN111858850A (en) | 2020-10-30 |
Family
ID=72988188
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010609405.0A Pending CN111858850A (en) | 2020-06-29 | 2020-06-29 | Method for realizing accurate and rapid scoring of question and answer on intelligent customer service |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111858850A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112988704A (en) * | 2021-03-05 | 2021-06-18 | 无锡星凝互动科技有限公司 | AI consultation database cluster building method and system |
| CN113157868A (en) * | 2021-04-29 | 2021-07-23 | 青岛海信网络科技股份有限公司 | Method and device for matching answers to questions based on structured database |
-
2020
- 2020-06-29 CN CN202010609405.0A patent/CN111858850A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112988704A (en) * | 2021-03-05 | 2021-06-18 | 无锡星凝互动科技有限公司 | AI consultation database cluster building method and system |
| CN113157868A (en) * | 2021-04-29 | 2021-07-23 | 青岛海信网络科技股份有限公司 | Method and device for matching answers to questions based on structured database |
| CN113157868B (en) * | 2021-04-29 | 2022-11-11 | 青岛海信网络科技股份有限公司 | Method and device for matching answers to questions based on structured database |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1474759B1 (en) | System, method, and software for automatic hyperlinking of persons' names in documents to professional directories | |
| US7376634B2 (en) | Method and apparatus for implementing Q&A function and computer-aided authoring | |
| US20160140187A1 (en) | System and method for answering natural language question | |
| US9390161B2 (en) | Methods and systems for extracting keyphrases from natural text for search engine indexing | |
| US20060230005A1 (en) | Empirical validation of suggested alternative queries | |
| CN106202153A (en) | The spelling error correction method of a kind of ES search engine and system | |
| CN106874441A (en) | Intelligent answer method and apparatus | |
| CN110705247B (en) | Based on x2-C text similarity calculation method | |
| WO2019218527A1 (en) | Multi-system combined natural language processing method and apparatus | |
| CN112307182B (en) | Question-answering system-based pseudo-correlation feedback extended query method | |
| CN111858850A (en) | Method for realizing accurate and rapid scoring of question and answer on intelligent customer service | |
| CN107341188A (en) | Efficient data screening technique based on semantic analysis | |
| Li et al. | National University of Singapore at the TREC-13 question answering main task | |
| Yusuf et al. | Query expansion method for quran search using semantic search and lucene ranking | |
| Aroussi et al. | Improving question answering systems by using the explicit semantic analysis method | |
| Kaur et al. | Effective question answering techniques and their evaluation metrics | |
| Almeida et al. | Universal Passage Weighting Mecanism (UPWM) in BioASQ 9b. | |
| Pizzato et al. | Extracting exact answers using a meta question answering system | |
| Pradhan et al. | Building a Foundation System for Producing Short Answers to Factual Questions. | |
| Liu et al. | Computing semantic relatedness using a word-text mutual guidance model | |
| CN117688319B (en) | Method for analyzing database structure by using AI | |
| CN112380830B (en) | Matching method, system and computer readable storage medium for related sentences in different documents | |
| Lin et al. | Query expansion for document retrieval by mining additional query terms | |
| CN110458658A (en) | A kind of book retrieval recommended method based on semantic information | |
| Doan-Nguyen et al. | Using semantic information to improve the performance of a restricted-domain question-answering system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |