Disclosure of Invention
In order to solve the technical problems, the technical scheme solves the problems of low efficiency and insufficient accuracy in the prior art, and mainly solves the problems of long time consumption in understanding and analyzing the user consultation content, inaccurate problem identification and solution recommendation caused by human factors or incomplete information, thereby influencing the overall service quality and user satisfaction.
In order to achieve the above purpose, the invention adopts the following technical scheme:
a network technology service consultation method, comprising:
Based on the solution of the historical network technical problem and the known network technical problem, combining the historical network technical problem and the known network technical problem into a network technical knowledge database;
Based on the characteristic expression of the network technical problem, generating the characteristic attribute of the corresponding network technical problem, and constructing a characteristic matrix of the network technical problem;
acquiring network technology consultation content of a user;
According to the network technology consultation content of the user, keyword decomposition is carried out to obtain a plurality of problem characteristics of the user, and structural characteristic recombination is carried out aiming at the plurality of problem characteristics of the user to obtain the problem characteristic attribute of the user consultation;
giving decision weight according to the characteristic attribute of the user consultation problem, and generating a user consultation problem characteristic attribute reply decision by utilizing a priority reply algorithm;
according to each element in the user consultation problem feature matrix and the network technical problem feature matrix, carrying out sharing feature attribute matching to determine the network technical problem of the user;
and according to the network technical problem of the user, retrieving the corresponding solution from the network technical knowledge database to reply.
Preferably, based on the characteristic expression of the network technical problem, generating the characteristic attribute of the corresponding network technical problem, and constructing a characteristic matrix of the network technical problem;
determining a plurality of characteristics of the network technical problem based on historical experience of a network technical engineer;
according to a plurality of characteristics of the network technical problem, determining the characteristic performance of each characteristic, and determining the characteristic attribute;
carrying out standardization processing on feature attributes corresponding to a plurality of features of the network technical problem;
performing association mapping on feature attributes corresponding to a plurality of features of the standardized network technical problem to construct a network technical problem feature matrix A;
wherein,Is a characteristic matrix of the technical problem of the network,Is the firstProblem of personal network technologyThe attributes of the individual features are used to determine,As a total number of network technical problems,Is the total number of feature attributes.
Preferably, according to the network technology consultation content of the user, keyword decomposition is performed to obtain a plurality of network technology problem features of the user, and structural feature recombination is performed aiming at the plurality of network technology problem features of the user, so that the feature attributes of the user consultation problem are obtained specifically including:
constructing a ternary network technology consultation content identification model based on an N-gram natural language processing technology;
based on the network technology consultation content of the user, carrying out word preprocessing, deleting stop words and extracting word shapes;
Dividing the preprocessed text into three continuous text sequences;
Performing traversal statistics on text elements in three continuous text sequences to obtain text frequency of the text sequences;
According to the text frequency of the text sequences, calculating the occurrence probability of each text sequence through a conditional probability formula;
Substituting the network technology consultation content of the user into a ternary network technology consultation content identification model, taking the occurrence probability of a text sequence of the consultation content as input, evaluating the naturalness of the text sequence, screening a structured text sequence with optimal naturalness as output, obtaining a plurality of network technology problem characteristics of the user, and determining the characteristic attribute of the user consultation problem according to the characteristic expression of a plurality of problem characteristics of the network technology consultation content of the user;
the ternary network technology consultation content identification model specifically comprises the following steps:
In the method, in the process of the invention,The j-th characteristic attribute of the i-th network technology problem for the user network technology consultation content,To give the first two wordsAndIn the case of (a) wordsThe conditional probability of the occurrence of the event,For three consecutive text sequencesThe number of times of occurrence in the user network technology consultation content is V, which is the total number of words of the user network technology consultation content.
Preferably, the decision weight is given according to the attribute of the user consultation problem feature, and the generating the decision of the user consultation problem feature attribute reply by using a priority reply algorithm specifically comprises:
the severity level of the implanted network technical problem refers to a data table;
giving importance weight to each question feature attribute of the user's network technology consultation content according to the severity level of the network technology question with reference to the data table；
Giving time weight T to each problem characteristic attribute according to the time length of the network technology consultation content of the user；
Taking each problem characteristic attribute grade weight in the network technology consultation content time length of the user as a first element and taking the problem characteristic attribute time weight as a second element;
Calculating the priority level of each problem feature attribute of the network technology consultation content of each user by using a priority reply algorithm, judging whether the priority level of each problem feature attribute is the same, if not, carrying out sequencing reply according to the priority level, and if so, carrying out reply according to the comparison result of the importance weight of the problem feature attribute and the time weight of the problem feature attribute;
the priority reply algorithm is as follows:
In the method, in the process of the invention,Is the firstProblem of personal network technologyThe priority of the individual characteristic attributes,Is the firstProblem of personal network technologyThe importance weight of the individual feature attributes,Weighting functions for the severity level of the network technical problem with reference to the data table,Is a coefficient of the temporal priority,Is the firstProblem of personal network technologyThe temporal weight of the individual characteristic attributes,For the length of time the user is consulted,A time length threshold value, wherein the time length threshold value is dynamically allocated according to the current time length consulted by the user network technology,Is the firstProblem of personal network technologyThe priority of the individual characteristic attributes,Is the firstProblem of personal network technologyThe importance weight of the individual feature attributes,Is the firstProblem of personal network technologyTemporal weight of the individual feature attributes.
Preferably, the step of matching the shared characteristic attribute according to each element in the characteristic matrix of the user consultation problem and the network technical problem, and the step of determining the network technical problem of the user specifically includes:
determining the characteristic attribute of the user consultation problem;
converting the network technical problem feature matrix and the user consultation problem feature attribute into a user consultation problem feature attribute vector;
according to each consultation problem feature attribute vector in the network technical problem feature matrix and the user consultation problem feature attribute vector, calculating the feature attribute vector of each consultation problem of the user and the feature attribute vector of each network technical problem in the network technical problem feature matrix through a matching algorithm to perform sharing feature attribute matching, obtaining feature attribute matching degree of each network technical problem, and screening out the problem feature attribute with highest matching degree;
determining a final network technical problem according to the problem feature attribute with the highest matching degree;
Wherein the matching algorithm expression is as follows
In the method, in the process of the invention,For the problem feature attribute with the highest matching degree,First, theProblem of personal network technologyThe attributes of the individual features are used to determine,Is the first one in the characteristic matrix of network technical problemProblem of personal network technologyAnd a characteristic attribute.
Further, a network technology service consultation system is provided, which is configured to implement the network technology service consultation method based on the above, and includes:
The knowledge database is used for solving the problems based on the historical network technology and the known network technology, and is combined into a network technology knowledge database;
the problem feature matrix module is used for generating corresponding network technical problem feature attributes based on the feature expression of the network technical problems and constructing a network technical problem feature matrix;
The consultation content acquisition module is used for acquiring the network technology consultation content of the user;
The problem identification module is electrically connected with the consultation content acquisition module and is used for decomposing keywords according to the network technology consultation content of the user to obtain a plurality of problem characteristics of the user, and carrying out structural characteristic recombination aiming at the plurality of problem characteristics of the user to obtain the attribute of the user consultation problem characteristics;
The priority decision module is electrically connected with the problem identification module and is used for giving decision weight according to the characteristic attribute of the user consultation problem and generating a characteristic attribute reply decision of the user consultation problem by utilizing a priority reply algorithm;
the retrieval matching module is electrically connected with the problem feature matrix module and the priority decision module and is used for carrying out shared feature attribute matching according to each element in the user consultation problem feature attribute and the network technical problem feature matrix to determine the network technical problem of the user;
and the reply module is electrically connected with the knowledge database and the retrieval matching module and is used for retrieving a corresponding solution from the network technical knowledge database according to the network technical problem of the user to reply.
Optionally, the problem identification module includes:
The identification model construction unit is used for constructing a ternary network technology consultation content identification model based on the N-gram natural language processing technology;
the word preprocessing unit is used for carrying out word preprocessing, deleting stop words and extracting word shapes based on the network technology consultation content of the user;
A text segmentation unit for segmenting the preprocessed text into three continuous text sequences;
The text frequency unit is used for performing traversal statistics on text elements in three continuous text sequences to obtain text frequency of the text sequences;
the text probability unit calculates the probability of each text sequence according to the text frequency of the text sequence through a conditional probability formula;
The recognition analysis unit substitutes the network technology consultation content of the user into a ternary network technology consultation content recognition model, takes the occurrence probability of a text sequence of the consultation content as input, evaluates the naturalness of the text sequence, screens out a structured text sequence with optimal naturalness as output, obtains a plurality of network technology problem characteristics of the user, and determines the characteristic attribute of the user consultation problem according to the characteristic expression of a plurality of problem characteristics of the network technology consultation content of the user.
Optionally, the priority decision module internally includes:
The control unit is used for referencing a data table to the severity level of the technical problem of the implanted network;
an importance weighting unit for assigning importance weight to each problem feature attribute of the network technology consultation content of the user according to the severity level of the network technology problem and referring to the data table;
a time weighting unit for weighting time of each problem characteristic attribute according to the time length of the network technology consultation content of the user;
The condition unit takes the attribute grade weight of each problem characteristic in the time length of the network technical consultation content of the user as a first element and takes the attribute time weight of the problem characteristic as a second element;
The reply decision unit calculates the priority level of each problem feature attribute of the network technology consultation content of each user by using a priority reply algorithm, judges whether the priority level of each problem feature attribute is the same, if not, replies in sequence according to the priority level, and if so, replies according to the comparison result of the importance weight of the problem feature attribute and the time weight of the problem feature attribute;
optionally, the search matching module includes:
the feature attribute confirming unit is used for confirming the feature attribute of the user consultation problem;
the vector conversion unit converts the network technical problem feature matrix and the user consultation problem feature attribute into a user consultation problem feature attribute vector;
The matching unit calculates each consultation problem feature vector of the user and the feature attribute vector of each network technical problem in the network technical problem feature matrix through a matching algorithm according to each consultation problem feature vector in the network technical problem feature matrix and the user consultation problem feature vector, performs sharing feature attribute matching to obtain feature attribute matching degree of each network technical problem, and screens out problem feature attributes with highest matching degree;
And the network technical problem confirming unit is used for confirming the final network technical problem according to the problem characteristic attribute with the highest matching degree.
Compared with the prior art, the invention has the beneficial effects that:
The invention provides a network technical service consultation scheme, which automatically processes the network technical consultation content of a user by constructing a network technical knowledge database and a problem feature matrix, and rapidly and accurately identifies the problem and provides a solution by utilizing a keyword decomposition, feature recombination, decision weight and priority reply algorithm, thereby improving the efficiency and accuracy of solving the problem.
Detailed Description
The following description is presented to enable one of ordinary skill in the art to make and use the invention. The preferred embodiments in the following description are by way of example only and other obvious variations will occur to those skilled in the art.
Referring to fig. 1, a network technical service consultation method includes:
Based on the solution of the historical network technical problem and the known network technical problem, combining the historical network technical problem and the known network technical problem into a network technical knowledge database;
Based on the characteristic expression of the network technical problem, generating the characteristic attribute of the corresponding network technical problem, and constructing a characteristic matrix of the network technical problem;
acquiring network technology consultation content of a user;
According to the network technology consultation content of the user, keyword decomposition is carried out to obtain a plurality of problem characteristics of the user, and structural characteristic recombination is carried out aiming at the plurality of problem characteristics of the user to obtain the problem characteristic attribute of the user consultation;
giving decision weight according to the characteristic attribute of the user consultation problem, and generating a user consultation problem characteristic attribute reply decision by utilizing a priority reply algorithm;
according to each element in the user consultation problem feature matrix and the network technical problem feature matrix, carrying out sharing feature attribute matching to determine the network technical problem of the user;
and according to the network technical problem of the user, retrieving the corresponding solution from the network technical knowledge database to reply.
The scheme is based on natural language processing NLP and machine learning, and is a reply scheme for completing user consultation problems through automatic identification, and specifically comprises the following steps: constructing a knowledge database containing known network technical problems and solutions thereof, analyzing the consultation content of the user by using an NLP technology, extracting keywords, converting the keywords into structural characteristic attributes, and matching the characteristic attributes with a pre-constructed network technical problem characteristic matrix to identify specific problems facing the user. Finally, based on the identified problem, the corresponding solution is retrieved from the knowledge database.
Preferably, based on the characteristic expression of the network technical problem, generating the characteristic attribute of the corresponding network technical problem, and constructing a characteristic matrix of the network technical problem;
determining a plurality of characteristics of the network technical problem based on historical experience of a network technical engineer;
according to a plurality of characteristics of the network technical problem, determining the characteristic performance of each characteristic, and determining the characteristic attribute;
carrying out standardization processing on feature attributes corresponding to a plurality of features of the network technical problem;
performing association mapping on feature attributes corresponding to a plurality of features of the standardized network technical problem to construct a network technical problem feature matrix A;
wherein,Is a characteristic matrix of the technical problem of the network,Is the firstProblem of personal network technologyThe attributes of the individual features are used to determine,As a total number of network technical problems,Is the total number of feature attributes.
It should be noted that, the characteristic attribute of the network technical problem refers to specific information for feeding back the current network technical problem, and is exemplary: the WiFi router has the advantages that the network speed is reduced, the uploading speed is normal, the downloading speed is slow, the network technical problem is a WiFi bandwidth bottleneck problem, the characteristic is characterized in that the uploading is normal, the downloading is slow, and the data base is improved for subsequent user consultation by collecting special attributes corresponding to the characteristic of the network technical problem.
Referring to fig. 2, according to the network technology consultation content of the user, keyword decomposition is performed to obtain a plurality of network technology problem features of the user, and structural feature recombination is performed for the plurality of network technology problem features of the user, so that feature attributes of the user consultation problem are obtained specifically including:
constructing a ternary network technology consultation content identification model based on an N-gram natural language processing technology;
based on the network technology consultation content of the user, carrying out word preprocessing, deleting stop words and extracting word shapes;
Dividing the preprocessed text into three continuous text sequences;
Performing traversal statistics on text elements in three continuous text sequences to obtain text frequency of the text sequences;
According to the text frequency of the text sequences, calculating the occurrence probability of each text sequence through a conditional probability formula;
Substituting the network technology consultation content of the user into a ternary network technology consultation content identification model, taking the occurrence probability of a text sequence of the consultation content as input, evaluating the naturalness of the text sequence, screening a structured text sequence with optimal naturalness as output, obtaining a plurality of network technology problem characteristics of the user, and determining the characteristic attribute of the user consultation problem according to the characteristic expression of a plurality of problem characteristics of the network technology consultation content of the user;
the ternary network technology consultation content identification model specifically comprises the following steps:
In the method, in the process of the invention,The j-th characteristic attribute of the i-th network technology problem for the user network technology consultation content,To give the first two wordsAndIn the case of (a) wordsThe conditional probability of the occurrence of the event,For three consecutive text sequencesThe number of times of occurrence in the user network technology consultation content is V, which is the total number of words of the user network technology consultation content.
It should be noted that, the goal of performing structural feature reorganization on a plurality of problem features of a user is to perform structural feature reorganization on a plurality of problem features of a user to ensure that a system can identify a network technical problem of the user, which is a necessary condition for a subsequent reply solution, because the user generally has no technical causal concept when facing the network technical problem, and description consultation content is generally disordered.
Referring to fig. 3, according to the decision weight given by the user consultation problem feature attribute, the generating a user consultation problem feature attribute reply decision by using a priority reply algorithm specifically includes:
the severity level of the implanted network technical problem refers to a data table;
giving importance weight to each question feature attribute of the user's network technology consultation content according to the severity level of the network technology question with reference to the data table；
Giving time weight T to each problem characteristic attribute according to the time length of the network technology consultation content of the user；
Taking each problem characteristic attribute grade weight in the network technology consultation content time length of the user as a first element and taking the problem characteristic attribute time weight as a second element;
Calculating the priority level of each problem feature attribute of the network technology consultation content of each user by using a priority reply algorithm, judging whether the priority level of each problem feature attribute is the same, if not, carrying out sequencing reply according to the priority level, and if so, carrying out reply according to the comparison result of the importance weight of the problem feature attribute and the time weight of the problem feature attribute;
the priority reply algorithm is as follows:
In the method, in the process of the invention,Is the firstProblem of personal network technologyThe priority of the individual characteristic attributes,Is the firstProblem of personal network technologyThe importance weight of the individual feature attributes,Weighting functions for the severity level of the network technical problem with reference to the data table,Is a coefficient of the temporal priority,Is the firstProblem of personal network technologyThe temporal weight of the individual characteristic attributes,For the length of time the user is consulted,A time length threshold value, wherein the time length threshold value is dynamically allocated according to the current time length consulted by the user network technology,Is the firstProblem of personal network technologyThe priority of the individual characteristic attributes,Is the firstProblem of personal network technologyThe importance weight of the individual feature attributes,Is the firstProblem of personal network technologyTemporal weight of the individual feature attributes.
According to the scheme, the severity level reference data table is introduced, the importance weight is distributed to the characteristic attribute of the network technical problem, the time weight is given by considering the duration time of the problem, and the priority level recovery algorithm is used for integrating the factors to determine the priority level of the problem, so that the reasonable parameters are improved for responding to the consultation of the user in an automatic mode, and the key problem can be timely and effectively processed.
Referring to fig. 4, according to each element in the user consultation problem feature matrix and the network technical problem feature attribute, performing sharing feature attribute matching, and determining the network technical problem of the user specifically includes:
determining the characteristic attribute of the user consultation problem;
converting the network technical problem feature matrix and the user consultation problem feature attribute into a user consultation problem feature attribute vector;
according to each consultation problem feature attribute vector in the network technical problem feature matrix and the user consultation problem feature attribute vector, calculating the feature attribute vector of each consultation problem of the user and the feature attribute vector of each network technical problem in the network technical problem feature matrix through a matching algorithm to perform sharing feature attribute matching, obtaining feature attribute matching degree of each network technical problem, and screening out the problem feature attribute with highest matching degree;
determining a final network technical problem according to the problem feature attribute with the highest matching degree;
Wherein the matching algorithm expression is as follows
In the method, in the process of the invention,For the problem feature attribute with the highest matching degree,First, theProblem of personal network technologyThe attributes of the individual features are used to determine,Is the first one in the characteristic matrix of network technical problemProblem of personal network technologyAnd a characteristic attribute.
According to the scheme, the consultation content of the user is converted into the feature vector through a natural language processing technology and is matched with the pre-constructed problem feature matrix, so that the problems can be rapidly identified and classified, the technical support flow is optimized, and the problem recovery efficiency and the problem recovery accuracy are improved.
Referring to fig. 5, based on the same inventive concept of a network technical service consultation method, a network technical service consultation system is proposed, including:
The knowledge database is used for solving the problems based on the historical network technology and the known network technology, and is combined into a network technology knowledge database;
the problem feature matrix module is used for generating corresponding network technical problem feature attributes based on the feature expression of the network technical problems and constructing a network technical problem feature matrix;
The consultation content acquisition module is used for acquiring the network technology consultation content of the user;
The problem identification module is electrically connected with the consultation content acquisition module and is used for decomposing keywords according to the network technology consultation content of the user to obtain a plurality of problem characteristics of the user, and carrying out structural characteristic recombination aiming at the plurality of problem characteristics of the user to obtain the attribute of the user consultation problem characteristics;
The priority decision module is electrically connected with the problem identification module and is used for giving decision weight according to the characteristic attribute of the user consultation problem and generating a characteristic attribute reply decision of the user consultation problem by utilizing a priority reply algorithm;
the retrieval matching module is electrically connected with the problem feature matrix module and the priority decision module and is used for carrying out shared feature attribute matching according to each element in the user consultation problem feature attribute and the network technical problem feature matrix to determine the network technical problem of the user;
and the reply module is electrically connected with the knowledge database and the retrieval matching module and is used for retrieving a corresponding solution from the network technical knowledge database according to the network technical problem of the user to reply.
Wherein, inside the problem identification module includes:
The identification model construction unit is used for constructing a ternary network technology consultation content identification model based on the N-gram natural language processing technology;
the word preprocessing unit is used for carrying out word preprocessing, deleting stop words and extracting word shapes based on the network technology consultation content of the user;
A text segmentation unit for segmenting the preprocessed text into three continuous text sequences;
The text frequency unit is used for performing traversal statistics on text elements in three continuous text sequences to obtain text frequency of the text sequences;
the text probability unit calculates the probability of each text sequence according to the text frequency of the text sequence through a conditional probability formula;
The recognition analysis unit substitutes the network technology consultation content of the user into a ternary network technology consultation content recognition model, takes the occurrence probability of a text sequence of the consultation content as input, evaluates the naturalness of the text sequence, screens out a structured text sequence with optimal naturalness as output, obtains a plurality of network technology problem characteristics of the user, and determines the characteristic attribute of the user consultation problem according to the characteristic expression of a plurality of problem characteristics of the network technology consultation content of the user.
Wherein, the priority decision module internally comprises:
The control unit is used for referencing a data table to the severity level of the technical problem of the implanted network;
an importance weighting unit for assigning importance weight to each problem feature attribute of the network technology consultation content of the user according to the severity level of the network technology problem and referring to the data table;
a time weighting unit for weighting time of each problem characteristic attribute according to the time length of the network technology consultation content of the user;
The condition unit takes the attribute grade weight of each problem characteristic in the time length of the network technical consultation content of the user as a first element and takes the attribute time weight of the problem characteristic as a second element;
The reply decision unit calculates the priority level of each problem feature attribute of the network technology consultation content of each user by using a priority reply algorithm, judges whether the priority level of each problem feature attribute is the same, if not, replies in sequence according to the priority level, and if so, replies according to the comparison result of the importance weight of the problem feature attribute and the time weight of the problem feature attribute;
wherein, retrieve the matching module inside and include:
the feature attribute confirming unit is used for confirming the feature attribute of the user consultation problem;
the vector conversion unit converts the network technical problem feature matrix and the user consultation problem feature attribute into a user consultation problem feature attribute vector;
The matching unit calculates each consultation problem feature vector of the user and the feature attribute vector of each network technical problem in the network technical problem feature matrix through a matching algorithm according to each consultation problem feature vector in the network technical problem feature matrix and the user consultation problem feature vector, performs sharing feature attribute matching to obtain feature attribute matching degree of each network technical problem, and screens out problem feature attributes with highest matching degree;
And the network technical problem confirming unit is used for confirming the final network technical problem according to the problem characteristic attribute with the highest matching degree.
The application process of the network technical service consultation system comprises the following steps:
step 1: based on the solution of the historical network technical problem and the known network technical problem, combining the historical network technical problem and the known network technical problem into a network technical knowledge database;
step 2: based on the characteristic expression of the network technical problem, generating the characteristic attribute of the corresponding network technical problem, and constructing a characteristic matrix of the network technical problem;
step 3: acquiring network technology consultation content of a user;
step 4: constructing a ternary network technology consultation content identification model based on an N-gram natural language processing technology;
step 5: based on the network technology consultation content of the user, carrying out word preprocessing, deleting stop words and extracting word shapes;
Step 6: dividing the preprocessed text into three continuous text sequences;
step 7: performing traversal statistics on text elements in three continuous text sequences to obtain text frequency of the text sequences;
Step 8: according to the text frequency of the text sequences, calculating the occurrence probability of each text sequence through a conditional probability formula;
Step 9: substituting the network technology consultation content of the user into a ternary network technology consultation content identification model, taking the occurrence probability of a text sequence of the consultation content as input, evaluating the naturalness of the text sequence, screening a structured text sequence with optimal naturalness as output, obtaining a plurality of network technology problem characteristics of the user, and determining the characteristic attribute of the user consultation problem according to the characteristic expression of a plurality of problem characteristics of the network technology consultation content of the user;
step 10: the severity level of the implanted network technical problem refers to a data table;
Step 11: according to the severity level of the network technical problem, referring to a data table, giving importance weight to each problem characteristic attribute of the network technical consultation content of the user;
Step 12: giving time weight to each problem characteristic attribute according to the time length of the network technology consultation content of the user;
step 13: taking each problem characteristic attribute grade weight in the network technology consultation content time length of the user as a first element and taking the problem characteristic attribute time weight as a second element;
Step 14: calculating the priority level of each problem feature attribute of the network technology consultation content of each user by using a priority reply algorithm, judging whether the priority level of each problem feature attribute is the same, if not, carrying out sequencing reply according to the priority level, and if so, carrying out reply according to the comparison result of the importance weight of the problem feature attribute and the time weight of the problem feature attribute;
step 15: determining the characteristic attribute of the user consultation problem;
step 16: converting the network technical problem feature matrix and the user consultation problem feature attribute into a user consultation problem feature attribute vector;
Step 17: according to each consultation problem feature attribute vector in the network technical problem feature matrix and the user consultation problem feature attribute vector, calculating the feature attribute vector of each consultation problem of the user and the feature attribute vector of each network technical problem in the network technical problem feature matrix through a matching algorithm to perform sharing feature attribute matching, obtaining feature attribute matching degree of each network technical problem, and screening out the problem feature attribute with highest matching degree;
step 18: determining a final network technical problem according to the problem feature attribute with the highest matching degree;
Step 19: and according to the network technical problem of the user, retrieving the corresponding solution from the network technical knowledge database to reply.
In summary, the invention has the advantages that: by constructing a network technology knowledge database and a problem feature matrix, automatically processing the network technology consultation content of a user, and utilizing keyword decomposition, feature recombination, decision weight and priority reply algorithm, the problems are rapidly and accurately identified and solutions are provided, so that the problem solving efficiency and accuracy are improved.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, and that the above embodiments and descriptions are merely illustrative of the principles of the present invention, and various changes and modifications may be made therein without departing from the spirit and scope of the invention, which is defined by the appended claims. The scope of the invention is defined by the appended claims and equivalents thereof.