CN116136872A

CN116136872A - Big data analysis method and system for intelligent interaction response of meta-universe

Info

Publication number: CN116136872A
Application number: CN202310300532.6A
Authority: CN
Inventors: 白隆光
Original assignee: Binzhou Xinruo Network Technology Co ltd
Current assignee: Binzhou Xinruo Network Technology Co ltd
Priority date: 2023-03-27
Filing date: 2023-03-27
Publication date: 2023-05-19

Abstract

The invention provides a big data analysis method and a big data analysis system for a meta-universe intelligent interactive response, and relates to the technical field of artificial intelligence. According to the invention, through optimizing an interaction data analysis network, key information mining operation or interaction data semantic analysis operation is carried out on a first member of an interaction data set corresponding to interaction data to be responded; from the optimized interaction data set, candidate interaction data corresponding to interaction data integrators with key information matching parameters larger than a predetermined first preset matching parameter between the first members of the interaction data set are determined, or candidate interaction data corresponding to interaction data integrators with interaction data semantic matching parameters larger than a predetermined second preset information matching parameter between the first members of the interaction data set are determined; and feeding back the candidate interaction data as response interaction data corresponding to the interaction data to be responded to the target network user. Based on the above, the reliability of the intelligent interactive response can be improved.

Description

Big data analysis method and system for intelligent interaction response of meta-universe

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to a big data analysis method and a big data analysis system for intelligent interaction response of a meta space.

Background

Artificial intelligence (AI, artificial Intelligence) technology, mainly includes computer vision technology, speech processing technology, natural language processing technology, machine learning/deep learning, automatic driving, intelligent traffic, etc. Among them, machine Learning (ML) is a multi-domain interdisciplinary, and involves multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, etc. It is specially studied how a computer simulates or implements learning behavior of a human to acquire new knowledge or skills, and reorganizes existing knowledge structures to continuously improve own performance. Machine learning is the core of artificial intelligence, a fundamental approach to letting computers have intelligence, which is applied throughout various areas of artificial intelligence. Machine learning and deep learning typically include techniques such as artificial neural networks, confidence networks, reinforcement learning, transfer learning, induction learning, teaching learning, and the like.

In the prior art, intelligent interactive response is required in many fields based on various user demands, but the reliability of the intelligent interactive response is not high. For example, a simple comparison search based on keywords generally results in a relatively low degree of matching between the response data and the data to be responded, i.e., there is a problem of low reliability.

Disclosure of Invention

In view of the above, the present invention aims to provide a big data analysis method and system for intelligent meta-universe interaction response, so as to improve the reliability of intelligent interaction response.

In order to achieve the above purpose, the embodiment of the present invention adopts the following technical scheme:

a big data analysis method for a meta-universe intelligent interactive response, comprising:

forming a corresponding optimized interaction data analysis network through network optimization;

under the condition that interaction data to be responded of a target network user are received, performing key information mining operation or interaction data semantic analysis operation on a first member of an interaction data set corresponding to the interaction data to be responded through the optimized interaction data analysis network;

candidate interaction data corresponding to interaction data integrators with key information matching parameters larger than a predetermined first preset matching parameter among first members of the interaction data set are determined from an optimized interaction data set, or candidate interaction data corresponding to interaction data integrators with interaction data semantic matching parameters larger than a predetermined second preset information matching parameter among the first members of the interaction data set are determined from the optimized interaction data set;

And taking the determined candidate interaction data as response interaction data corresponding to the interaction data to be responded, and feeding back the response interaction data to the target network user.

In some preferred embodiments, in the above method for analyzing big data for a meta-universe intelligent interactive response, the step of forming a corresponding optimized interactive data analysis network through network optimization includes:

determining an interaction data set to be optimized, wherein the interaction data set to be optimized comprises a member description vector to be optimized and member semantic data to be optimized of each interaction data set member in a data set member network, the data set member network comprises at least one high-level interaction data set member, the data set member network further comprises at least one low-level interaction data set member, the high-level interaction data set member has corresponding member semantic data to be optimized, and the low-level interaction data set member does not have corresponding member semantic data to be optimized;

based on each member description vector to be optimized in the interaction data set to be optimized and the member semantic data to be optimized, performing a first optimization operation on the interaction data analysis network to be optimized to form an intermediate interaction data analysis network corresponding to the interaction data analysis network to be optimized;

Extracting relevant interaction data set members corresponding to each advanced interaction data set member in the data set member network from the at least one low-level interaction data set member, and performing interaction data semantic analysis operation on each relevant interaction data set member through an intermediate interaction data analysis network so as to output interaction data semantic analysis results corresponding to each relevant interaction data set member;

marking the interaction data semantic analysis result corresponding to each related interaction data set member to mark the member semantic data of each related interaction data set member, and combining the member semantic data of each related interaction data set member and the member semantic data to be optimized based on the member semantic data of each related interaction data set member to form a corresponding member semantic data set;

updating each related interaction data set member from the data set member network to update to a new advanced interaction data set member so as to form an updated data set member network;

determining an optimized member description vector of each interactive data set member mined by the interactive data analysis network to be optimized in the execution of the first optimizing operation to form a corresponding optimized interactive data set; the optimized interaction data set comprises an optimized member description vector of each interaction data set member in an updated data set member network and the member semantic data set, one interaction data set member corresponds to one interaction data, the updated data set member network comprises a first number of high-level interaction data set members and a second number of low-level interaction data set members, the member semantic data set comprises member semantic data of each high-level interaction data set member, and the member semantic data is used for reflecting semantic content of interaction data corresponding to the corresponding high-level interaction data set member;

Performing a second optimization operation on the intermediate interaction data analysis network based on each optimization member description vector in the optimization interaction data set and the member semantic data set to form an advanced interaction data analysis network corresponding to the intermediate interaction data analysis network;

performing interaction data semantic analysis operation on each low-level interaction data set member through the high-level interaction data analysis network so as to output interaction data semantic analysis results corresponding to each low-level interaction data set member and output reliability evaluation parameters corresponding to each interaction data semantic analysis result;

determining at least one complex interaction data set member from the second number of low-level interaction data set members according to the reliability evaluation parameters corresponding to each interaction data semantic analysis result, wherein the complex interaction data set member belongs to the low-level interaction data set member corresponding to the interaction data semantic analysis result of which the reliability evaluation parameters are smaller than the predetermined reference evaluation parameters;

and performing a third optimization operation on the advanced interaction data analysis network based on the at least one complex interaction data set member to form a corresponding optimized interaction data analysis network.

In some preferred embodiments, in the above big data analysis method for meta-universe intelligent interactive responses, the step of performing a second optimization operation on the intermediate interactive data analysis network based on each optimized member description vector in the optimized interactive data set and the member semantic data set to form a high-level interactive data analysis network corresponding to the intermediate interactive data analysis network includes:

polling each interactive data integration member in the updated data set member network to determine a first class related member and a second class related member of the currently polled polling interactive data set member in the updated data set member network, wherein the correlation degree between the first class related member and the polling interactive data set member is higher than that between the second class related member and the polling interactive data set member;

extracting an association description vector combination with association relation between the polling interaction data set members from the optimization interaction data set, wherein the association description vector combination comprises an optimization member description vector of the polling interaction data set members, an optimization member description vector of the first class related members and an optimization member description vector of the second class related members;

Updating each optimized member description vector in the association description vector combination through the intermediate interaction data analysis network to form an update description vector combination corresponding to the polling interaction data set member;

and based on the update description vector combination, analyzing a network optimization error index corresponding to the intermediate interaction data analysis network, and based on the network optimization error index corresponding to the intermediate interaction data analysis network, optimizing the intermediate interaction data analysis network to form an advanced interaction data analysis network corresponding to the intermediate interaction data analysis network.

In some preferred embodiments, in the above big data analysis method for a metauniverse intelligent interactive response, the update description vector combination includes an update member description vector of the polling interaction data set member, an update member description vector of the first type of related member, and an update member description vector of the second type of related member;

the step of analyzing the network optimization error index corresponding to the intermediate interaction data analysis network based on the update description vector combination, and optimizing the intermediate interaction data analysis network based on the network optimization error index corresponding to the intermediate interaction data analysis network to form an advanced interaction data analysis network corresponding to the intermediate interaction data analysis network includes:

Under the condition that the polling interaction data integrator belongs to the advanced interaction data set member, analyzing and processing each update member description vector in the update description vector combination based on a first analysis rule so as to analyze a corresponding error first local value;

determining semantic description vectors of member semantic data of the polling interaction data set members;

based on a second analysis rule, analyzing and processing the semantic description vector and the updated member description vector corresponding to the polling interaction data set member so as to analyze a corresponding error second local value;

based on the error first local value and the error second local value, fusing to obtain a network optimization error index corresponding to the intermediate interaction data analysis network;

and optimizing the intermediate interaction data analysis network based on the network optimization error index to form an advanced interaction data analysis network corresponding to the intermediate interaction data analysis network.

In some preferred embodiments, in the foregoing big data analysis method for a meta-universe intelligent interactive response, the step of analyzing each update member description vector in the update description vector combination to analyze a corresponding error first local value based on a first analysis rule in a case that the poll interaction data integrator belongs to an advanced interaction data set member includes:

Under the condition that the polling interaction data integrator belongs to the advanced interaction data set member, carrying out vector correlation analysis based on the updated member description vector of the polling interaction data set member and the updated member description vector of the first class correlation member so as to output corresponding first class vector correlation parameters;

performing vector correlation analysis based on the updated member description vector of the polling interaction data set member and the updated member description vector of the second class of related members, and outputting a second class of vector correlation parameters;

based on the vector length of the updated member description vector of the polling interaction data set member and the vector length of the updated member description vector of the first class related member, performing parameter interval mapping operation on the first class vector related parameter to output a mapped first class vector related parameter;

based on the vector length of the updated member description vector of the polling interaction data set member and the vector length of the updated member description vector of the second class of related members, performing parameter interval mapping operation on the second class of vector related parameters to output mapped second class of vector related parameters;

and analyzing an error first local value according to the mapped first class vector related parameters and the mapped second class vector related parameters.

In some preferred embodiments, in the above big data analysis method for a meta-universe intelligent interactive response, the step of analyzing the semantic description vector and the updated member description vector corresponding to the polling interactive data set member based on the second analysis rule to analyze a second local value of the corresponding error includes:

determining a reference semantic description vector of reference member semantic data in the intermediate interaction data analysis network, and performing vector correlation analysis on the reference semantic description vector and an updated member description vector of the polling interaction data set member to output a third class of vector correlation parameters;

and performing parameter interval mapping operation on the third class vector related parameters based on the vector length of the reference semantic description vector and the vector length of the updated member description vector of the polling interaction data set member to output mapped third class vector related parameters, and analyzing a corresponding error second local value based on the semantic description vector and the mapped third class vector related parameters.

In some preferred embodiments, in the above method for analyzing big data of a meta-universe intelligent interactive response, the step of fusing the first local value based on the error and the second local value based on the error to obtain a network optimization error index corresponding to the intermediate interactive data analysis network includes:

Determining an importance characterization parameter of the error second local value, and updating the error second local value based on the importance characterization parameter to output an updated error second local value;

and performing superposition operation on the error first local value and the updated error second local value to output a network optimization error index corresponding to the intermediate interaction data analysis network.

In some preferred embodiments, in the above-mentioned big data analysis method for meta-universe intelligent interaction response, the optimized interaction data set includes a plurality of optimized interaction data subsets having precedence relations, each of the optimized interaction data subsets has a corresponding precedence relation characterization parameter, and each of the optimized interaction data subsets sequentially performs a second optimization operation on the intermediate interaction data analysis network according to the precedence relation;

the step of determining the importance characterizing parameter of the error second local value comprises the following steps:

analyzing target precedence relation characterization parameters of a target optimization interaction data subset corresponding to the polling interaction data integrator;

and determining an importance characterization parameter of the error second partial value based on the target precedence relationship characterization parameter, wherein the target precedence relationship characterization parameter and the importance characterization parameter of the error second partial value have a corresponding relationship with negative correlation.

In some preferred embodiments, in the above large data analysis method for a meta-universe intelligent interactive response, the step of performing a third optimization operation on the advanced interactive data analysis network based on the at least one complex interactive data set member to form a corresponding optimized interactive data analysis network includes:

polling the at least one complex interactive data set member to determine polling first class related members and polling second class related members of the currently polled complex interactive data set member in the updated data set member network;

extracting an exemplary description vector combination with association relation with the polling complex interaction data set members from the optimized interaction data set, wherein the exemplary description vector combination comprises an optimized member description vector of the polling complex interaction data set members, an optimized member description vector of the polling first class related members and an optimized member description vector of the polling second class related members;

updating each optimized member description vector in the exemplary description vector combination through the advanced interactive data analysis network to form an updated exemplary description vector combination corresponding to the polling complex interactive data integrator;

And analyzing a network optimization error index corresponding to the advanced interactive data analysis network based on the updated exemplary description vector combination, and optimizing the advanced interactive data analysis network based on the network optimization error index corresponding to the advanced interactive data analysis network to form an optimized interactive data analysis network corresponding to the advanced interactive data analysis network.

The embodiment of the invention also provides a big data analysis system for the meta-universe intelligent interactive response, which comprises a processor and a memory, wherein the memory is used for storing a computer program, and the processor is used for executing the computer program so as to realize the big data analysis method for the meta-universe intelligent interactive response.

The big data analysis method and the system for the meta-universe intelligent interaction response form a corresponding optimized interaction data analysis network through network optimization; through optimizing an interaction data analysis network, carrying out key information mining operation or interaction data semantic analysis operation on a first member of an interaction data set corresponding to interaction data to be responded; candidate interaction data corresponding to the interaction data integrator with the key information matching parameters between the first members of the interaction data set being greater than the predetermined first preset matching parameters are determined from the optimized interaction data set, or candidate interaction data corresponding to the interaction data integrator with the interaction data semantic matching parameters of the first members of the interaction data set being greater than the predetermined second preset information matching parameters are determined from the optimized interaction data set; and feeding back the determined candidate interaction data as response interaction data corresponding to the interaction data to be responded to the target network user. Based on the above, since the determined response interaction data is the candidate interaction data matched with the interaction data to be responded based on the optimized interaction data set, the reliability of the determined response interaction data is relatively high, so that the reliability of intelligent interaction response can be improved.

In order to make the above objects, features and advantages of the present invention more comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

Fig. 1 is a schematic flow chart of steps included in a method for analyzing big data for a meta-universe intelligent interactive response according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of each module included in the big data analysis device for a meta-universe intelligent interactive response according to an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

The embodiment of the invention provides a big data analysis system for a meta-universe intelligent interactive response. Wherein the big data analysis system may comprise a memory and a processor.

In detail, the memory and the processor are electrically connected directly or indirectly to realize transmission or interaction of data. For example, electrical connection may be made to each other via one or more communication buses or signal lines. The memory may store at least one software functional module (computer program) that may exist in the form of software or firmware. The processor may be configured to execute the executable computer program stored in the memory, thereby implementing the big data analysis method for the meta-universe intelligent interaction response provided by the embodiment of the invention.

It should be appreciated that in some possible embodiments, the Memory may be, but is not limited to, random access Memory (Random Access Memory, RAM), read Only Memory (ROM), programmable Read Only Memory (Programmable Read-Only Memory, PROM), erasable Read Only Memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable Read Only Memory (Electric Erasable Programmable Read-Only Memory, EEPROM), and the like.

It should be appreciated that in some possible embodiments, the processor may be a general purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), a System on Chip (SoC), etc.; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

It should be appreciated that in some possible embodiments, the big data analysis system for a metauniverse intelligent interactive response may be a server with data processing capabilities.

With reference to fig. 1, the embodiment of the invention also provides a big data analysis method for the meta-universe intelligent interaction response, which can be applied to the big data analysis system for the meta-universe intelligent interaction response. The method steps defined by the flow related to the large data analysis method for the meta-universe intelligent interaction response can be realized by the large data analysis system for the meta-universe intelligent interaction response.

The specific flow shown in fig. 1 will be described in detail.

Step S110, forming a corresponding optimized interaction data analysis network through network optimization.

In the embodiment of the invention, the big data analysis system for the meta-universe intelligent interactive response can form a corresponding optimized interactive data analysis network through network optimization.

Step S120, under the condition that the interaction data to be responded of the target network user is received, performing a key information mining operation or an interaction data semantic analysis operation on a first member of the interaction data set corresponding to the interaction data to be responded through the optimized interaction data analysis network.

In the embodiment of the invention, the big data analysis system for the meta-universe intelligent interactive response can carry out key information mining operation or interactive data semantic analysis operation on the first member of the interactive data set corresponding to the interactive data to be responded through the optimized interactive data analysis network under the condition that the interactive data to be responded of the target network user is received. The interactive data to be answered may be, for example, interactive text words, interactive text sentences, interactive text paragraphs or the like.

Step S130, determining candidate interaction data corresponding to interaction data integrators with key information matching parameters larger than a predetermined first preset matching parameter between first members of the interaction data set from an optimized interaction data set, or determining candidate interaction data corresponding to interaction data integrators with interaction data semantic matching parameters larger than a predetermined second preset information matching parameter between first members of the interaction data set from the optimized interaction data set.

In the embodiment of the invention, the big data analysis system for the meta-universe intelligent interaction response can determine candidate interaction data corresponding to the interaction data integrator, wherein the key information matching parameters (such as the matching degree or the similarity between member description vectors) between the first members of the interaction data set are larger than the predetermined first preset matching parameters, from the optimized interaction data set, or determine candidate interaction data corresponding to the interaction data integrator, wherein the interaction data semantic matching parameters (such as the matching degree or the similarity between member semantic data) between the first members of the interaction data set are larger than the predetermined second preset information matching parameters, from the optimized interaction data set. The specific parameter values of the first preset matching parameter and the second preset matching parameter can be configured according to actual application requirements.

And step S140, taking the determined candidate interaction data as response interaction data corresponding to the interaction data to be responded, and feeding back the response interaction data to the target network user.

In the embodiment of the invention, the big data analysis system for the meta-universe intelligent interaction response can take the determined candidate interaction data as the response interaction data corresponding to the interaction data to be responded, and feed back the response interaction data to the target network user, for example, the response interaction data is transmitted to the target user terminal equipment corresponding to the target network user. The interactive data to be responded can also be formed and sent out based on the operation of the target network user on the target user terminal equipment.

Based on the above, since the determined response interaction data is the candidate interaction data matched with the interaction data to be responded based on the optimized interaction data set, the reliability of the determined response interaction data is relatively high, so that the reliability of intelligent interaction response can be improved.

It should be appreciated that in some possible embodiments, the step S110 described above, i.e. the optimization forming of the optimized interaction data analysis network, may further comprise the following sub-steps:

determining an interaction data set to be optimized, wherein the interaction data set to be optimized comprises a member description vector to be optimized and member semantic data to be optimized of each interaction data set member in a data set member network, the data set member network comprises at least one high-level interaction data set member, the data set member network further comprises at least one low-level interaction data set member, the high-level interaction data set member has corresponding member semantic data to be optimized, the low-level interaction data set member does not have corresponding member semantic data to be optimized, the data set member network comprises a text word, a text sentence, a text paragraph and the like in each interaction data set member, in addition, the member description vector to be optimized can be obtained by carrying out key information mining on the interaction data set member, the member semantic data to be optimized is obtained by analyzing based on the member description vector to be optimized, or the member semantic data to be optimized can be formed based on configuration operation of a user;

Based on each member description vector to be optimized in the interaction data set to be optimized and the member semantic data to be optimized, performing a first optimization operation on the interaction data analysis network to be optimized to form an intermediate interaction data analysis network corresponding to the interaction data analysis network to be optimized, wherein the specific processing process of the first optimization operation can refer to the processing process of the second optimization operation;

extracting relevant interaction data set members corresponding to each advanced interaction data set member in the data set member network from the at least one low-level interaction data set member, and performing interaction data semantic analysis operation on each relevant interaction data set member through an intermediate interaction data analysis network to output interaction data semantic analysis results corresponding to each relevant interaction data set member, wherein the interaction data semantic analysis results can be used for reflecting corresponding member semantic data; in addition, the related interaction data set member may refer to a low-level interaction data set member directly related to the high-level interaction data set member, and the specific representation of the direct relationship is not limited, for example, in the data set member network, a designated identifier may be provided between two directly related interaction data set members, and a designated identifier may not be provided between two indirectly related interaction data set members, in addition, the indirect relationship may refer to that two interaction data set members are not directly related, but may refer to that the two interaction data set members are not directly related, for example, the interaction data set member 1 is not directly related to the interaction data set member 2, but the interaction data set member 1 is directly related to the interaction data set member 3, and the interaction data set member 2 is directly related to the interaction data set member 3.

through the advanced interaction data analysis network, each low-level interaction data set member is subjected to interaction data semantic analysis operation so as to output an interaction data semantic analysis result corresponding to each low-level interaction data set member, and a reliability evaluation parameter corresponding to each interaction data semantic analysis result is output;

and performing third optimization operation on the advanced interactive data analysis network based on the at least one complex interactive data set member to form a corresponding optimized interactive data analysis network, and performing further more complex third optimization operation after the first optimization operation and the second optimization operation based on the third optimization operation, so that the accuracy of the obtained optimized interactive data analysis network can be higher, namely semantic analysis and the like can be performed on more complex interactive data, and the advanced interactive data analysis network can have relatively higher analysis capability through the pre-guarantee of the first optimization operation and the second optimization operation, thereby guaranteeing the effective performance of the third optimization operation.

It should be understood that, in some possible embodiments, the step of mining key information of the interaction dataset members to obtain corresponding member description vectors to be optimized may further include the following sub-steps:

Performing feature space mapping operation on the interactive data corresponding to the interactive data integrator to form a corresponding interactive data mapping vector, wherein the interactive data mapping vector is used as easy-to-represent and wide-range data of the interactive data corresponding to the interactive data integrator, and in addition, parameters for performing feature space mapping operation can be used as an object of network optimization to optimize formation;

performing depth mining operation on the interaction data mapping vector, such as performing further filtering processing and the like, to form a corresponding interaction data depth mining vector, wherein the interaction data depth mining vector is used as data which is not easy to describe but specific for interaction data corresponding to the interaction data integrator, and in addition, parameters for performing the mining operation can be used as an object of network optimization to optimize formation;

intercepting vector parameters of the interactive data mapping vector to form a correlated data mapping vector and an uncorrelated data mapping vector, wherein a vector distance between the correlated data mapping vector and the interactive data mapping vector is smaller than a vector distance between the uncorrelated data mapping vector and the interactive data mapping vector, for example, the vector distance between the correlated data mapping vector and the interactive data mapping vector is smaller than a first preset distance, and the vector distance between the uncorrelated data mapping vector and the interactive data mapping vector is larger than a second preset distance, and in addition, the second preset distance can be larger than the first preset distance, and the vector distance can be a cosine distance;

Processing the interactive data mapping vector, the related data mapping vector and the uncorrelated data mapping vector through up-sampling and down-sampling operations of vector parameters to obtain a sampled interactive data mapping vector, a sampled related data mapping vector and a sampled uncorrelated data mapping vector which are respectively corresponding, wherein the sampled interactive data mapping vector, the sampled related data mapping vector and the sampled uncorrelated data mapping vector all have the same vector size;

performing weighted superposition operation on the sample interaction data mapping vector, the sample correlation data mapping vector and the sample non-correlation data mapping vector to obtain an enhanced interaction data mapping vector, wherein a weighting coefficient corresponding to the sample correlation data mapping vector is larger than a weighting coefficient corresponding to the sample non-correlation data mapping vector, and the weighting coefficient corresponding to the sample interaction data mapping vector is larger than the weighting coefficient corresponding to the sample non-correlation data mapping vector;

and carrying out aggregation operation on the enhanced interaction data mapping vector and the interaction data depth mining vector to obtain a corresponding member description vector to be optimized, if the member description vector to be optimized is spliced, wherein the number of vector dimensions of the member description vector to be optimized is equal to the sum value of the number of vector dimensions of the enhanced interaction data mapping vector and the number of vector dimensions of the interaction data depth mining vector.

It should be appreciated that, in some possible embodiments, the step of performing the second optimization operation on the intermediate interaction data analysis network based on each optimized member description vector in the optimized interaction data set and the member semantic data set to form the advanced interaction data analysis network corresponding to the intermediate interaction data analysis network may further include the following sub-steps:

each interactive data integration member in the updated data set member network is polled, so that a first class related member and a second class related member of the currently polled interactive data set member are determined in the updated data set member network, the degree of correlation between the first class related member and the polled interactive data set member is higher than the degree of correlation between the second class related member and the polled interactive data set member, the degree of correlation can be related to the direct connection and the indirect connection, for example, the two interactive data set members which are directly connected can have the largest degree of correlation, and for the two interactive data set members which are indirectly connected, the degree of correlation can be inversely related to the number of the interactive data set members in the middle, namely, the larger the number of the interactive data set members is, the smaller the degree of correlation is, based on the first class related member and the interactive data integration member with the degree of correlation being higher than the preset degree of correlation can be marked as the first class related member, and the interactive data integration member with the degree of correlation being lower than or equal to the preset degree of correlation can be marked as the second class related member;

updating each optimized member description vector in the association description vector combination through the intermediate interaction data analysis network, for example, further convolution processing, filtering operation or coding processing can be performed on the optimized member description vector, the coding processing can also be attention-based coding, so as to form an updated description vector combination corresponding to the polling interaction data set member, and in addition, processing parameters of the convolution processing, filtering operation or coding processing can be used as an optimized object;

It should be appreciated that in some possible embodiments, the update description vector combination may include an update member description vector of the polling interaction data set member, an update member description vector of the first type of related member, and an update member description vector of the second type of related member, based on which the network optimization error index corresponding to the intermediate interaction data analysis network is analyzed based on the update description vector combination, and the step of optimizing the intermediate interaction data analysis network to form an advanced interaction data analysis network corresponding to the intermediate interaction data analysis network based on the network optimization error index corresponding to the intermediate interaction data analysis network may further include the following sub-steps:

determining semantic description vectors of member semantic data of the members of the polling interaction data set, for example, key information mining can be performed on the member semantic data to form corresponding semantic description vectors, and in addition, the key information mining can refer to mapping corresponding data to a feature space for representation;

and optimizing the intermediate interaction data analysis network based on the network optimization error index to form an advanced interaction data analysis network corresponding to the intermediate interaction data analysis network, wherein the network optimization error index needs to be reduced until convergence in the process of optimizing.

It should be appreciated that in some possible embodiments, the step of analyzing each update member description vector in the update description vector combination to analyze the corresponding error first local value based on the first analysis rule in the case that the poll interaction data integrator belongs to the advanced interaction data set member may further include the following sub-steps:

Based on the updated member description vector of the polling interaction data set member and the updated member description vector of the second class of related members, carrying out vector correlation analysis, outputting a second class of vector related parameters, and referring to the calculation process of the first class of vector related parameters;

based on the vector length of the updated member description vector of the polling interaction data set member and the vector length of the updated member description vector of the second class related member, performing a parameter interval mapping operation on the second class vector related parameter to output a mapped second class vector related parameter, which may refer to a later determination process of the mapped second class vector related parameter;

Wherein, it should be understood that, in some possible embodiments, in the case that the poll interaction data set member belongs to an advanced interaction data set member, the vector correlation analysis is performed based on the updated member description vector of the poll interaction data set member and the updated member description vector of the first class related member to output the corresponding first class vector correlation parameter, and the following sub steps may be further included:

performing a rank parameter permutation operation on vector parameters in the updated member description vector of the polling interaction data set member to form a new updated member description vector;

and multiplying the new updated member description vector and the updated member description vector of the first class related member to output a corresponding first class vector related parameter.

Wherein, it should be understood that, in some possible embodiments, the step of performing the parameter interval mapping operation on the first type vector-related parameter based on the vector length of the updated member description vector of the polling interaction dataset member and the vector length of the updated member description vector of the first type related member to output the mapped first type vector-related parameter may further include the following sub-steps:

Calculating the vector length of the updated member description vector of the polling interaction data set member to obtain a first vector length, and calculating the vector length of the updated member description vector of the first class related member to obtain a second vector length;

calculating a product of a predetermined reference coefficient, the first vector length and the second vector length to obtain a target product value, wherein the reference coefficient can belong to (1, 2).

Wherein it should be understood that, in some possible embodiments, the step of analyzing the error first local value according to the mapped first class vector-related parameter and the mapped second class vector-related parameter may further include the following sub-steps:

extracting a predetermined offset parameter, the product of which and the reference coefficient may be equal to 1, and may belong to an interval (0, 1);

calculating the sum of the first-class vector related parameters and the offset parameters, and then carrying out logarithmic operation on the sum to obtain a first logarithmic result;

calculating the difference value between the offset parameter and the second class vector related parameter, and then carrying out logarithmic operation on the difference value to obtain a second logarithmic result;

Subtracting the first logarithm result from a predetermined base parameter, which may be equal to a value of 0, and subtracting the second logarithm result from the first logarithm result, may result in a first local value of the error.

It should be appreciated that, in some possible embodiments, the step of analyzing the semantic description vector and the updated member description vector corresponding to the polling interaction data set member based on the second analysis rule to analyze the corresponding error second local value may further include the following sub-steps:

determining a reference semantic description vector of reference member semantic data in the intermediate interaction data analysis network, and performing vector correlation analysis on the reference semantic description vector and an updated member description vector of the polling interaction data set member to output a third type of vector correlation parameter, wherein the analysis mode of the third type of vector correlation parameter can refer to the correlation processing procedure, in addition, the reference semantic description vector can be formed by carrying out key information mining on the reference member semantic data, the reference member semantic data can be formed based on configuration operation of a user, or in other embodiments, the reference semantic description vector can be used as a network optimization object to be continuously optimized and formed under the condition of an initial vector, and the initial vector can be randomly generated;

And performing parameter interval mapping operation on the third class vector related parameters based on the vector length of the reference semantic description vector and the vector length of the updated member description vector of the polling interaction data set member to output mapped third class vector related parameters, and analyzing corresponding error second local values based on the semantic description vector and the mapped third class vector related parameters, wherein the formation process of the mapped third class vector related parameters is as described in the previous correlation.

Wherein it should be understood that, in some possible embodiments, the step of analyzing the corresponding error second local value based on the semantic description vector and the mapped third class vector related parameter may further include the following sub-steps:

performing normalization mapping on the third-class vector related parameters of the mapping to obtain a target processing result, for example, normalization mapping can be realized through a softmax function;

and carrying out logarithmic operation on the target processing result to obtain a third logarithmic result, then carrying out multiplication operation on the semantic description vector and the third logarithmic result, and finally calculating the difference between the basic parameter and the multiplication operation result to obtain an error second local value.

It should be understood that, in some possible embodiments, the step of merging, based on the first local error value and the second local error value, the network optimization error indicator corresponding to the intermediate interactive data analysis network may further include the following sub-steps:

determining an importance characterization parameter of the error second local value, and updating the error second local value based on the importance characterization parameter to output an updated error second local value, for example, multiplying the importance characterization parameter and the error second local value, so as to obtain a corresponding updated error second local value;

It should be appreciated that, in some possible embodiments, the optimized interaction data set may include a plurality of optimized interaction data subsets having precedence relationships, each of the optimized interaction data subsets having a corresponding precedence relationship characterizing parameter, and each of the optimized interaction data subsets sequentially performs a second optimizing operation on the intermediate interaction data analysis network according to the precedence relationships, based on which the step of determining the importance characterizing parameter of the error second local value may further include the following sub-steps:

Analyzing target precedence relation characterization parameters of target optimization interaction data subsets corresponding to the polling interaction data integrator, wherein the target optimization interaction data subsets are optimization interaction data subsets comprising polling interaction data set members;

Wherein, it should be understood that, in some possible embodiments, the step of determining the importance characterizing parameter of the error second local value based on the target precedence characterizing parameter, where the target precedence characterizing parameter and the importance characterizing parameter of the error second local value have a correspondence with negative correlation may further include the following sub-steps:

in the case that the polling interaction dataset member belongs to a low-level interaction dataset member, the corresponding importance characterizing parameter may be determined as a first value, such as a value of 0;

And under the condition that the polling interaction data integrator belongs to an advanced interaction data set member, carrying out index operation on the negative correlation coefficient of the target precedence relation representation parameter to obtain a target index operation result, then calculating a sum value between a second predetermined value and the target index operation result, calculating a difference value between the second value and the target index operation result, and finally calculating a ratio of the difference value and the sum value to obtain an importance representation parameter of the second partial value of the error, wherein the second value can belong to (0.5, 1.5), such as 1 and the like.

It should be appreciated that in some possible embodiments, the step of performing a third optimization operation on the advanced interaction data analysis network based on the at least one complex interaction data set member to form a corresponding optimized interaction data analysis network may include the following sub-steps:

polling the at least one complex interactive data set member to determine polling first class related members and polling second class related members of the currently polled complex interactive data set member in the updated data set member network, as previously described;

Extracting an exemplary description vector combination with association relation with the polling complex interaction data set members from the optimized interaction data set, wherein the exemplary description vector combination comprises an optimized member description vector of the polling complex interaction data set members, an optimized member description vector of the polling first class related members and an optimized member description vector of the polling second class related members, and the like related description;

updating each optimized member description vector in the exemplary description vector combination through the advanced interactive data analysis network to form an updated exemplary description vector combination corresponding to the polling complex interactive data integrator, as described in the previous related description;

and based on the updated exemplary description vector combination, analyzing a network optimization error index corresponding to the advanced interactive data analysis network, and based on the network optimization error index corresponding to the advanced interactive data analysis network, optimizing the advanced interactive data analysis network to form an optimized interactive data analysis network corresponding to the advanced interactive data analysis network, as described in the previous related description.

With reference to fig. 2, the embodiment of the invention further provides a big data analysis device for the meta-universe intelligent interaction response, which can be applied to the big data analysis system for the meta-universe intelligent interaction response. The big data analysis device for the meta-universe intelligent interaction response can comprise the following software modules:

the network optimization module is used for forming a corresponding optimized interaction data analysis network through network optimization;

the interaction data analysis module is used for carrying out key information mining operation or interaction data semantic analysis operation on a first member of the interaction data set corresponding to the interaction data to be responded through the optimized interaction data analysis network under the condition that the interaction data to be responded of the target network user is received;

the interactive data determining module is used for determining candidate interactive data corresponding to the interactive data integrator, wherein the key information matching parameters between the first members of the interactive data set are larger than the predetermined first preset matching parameters, from the optimized interactive data set, or determining candidate interactive data corresponding to the interactive data integrator, wherein the interactive data semantic matching parameters of the first members of the interactive data set are larger than the predetermined second preset information matching parameters, from the optimized interactive data set;

And the interaction data feedback module is used for taking the determined candidate interaction data as response interaction data corresponding to the interaction data to be responded and feeding back the response interaction data to the target network user.

In summary, the big data analysis method and the system for the meta-universe intelligent interaction response form a corresponding optimized interaction data analysis network through network optimization; through optimizing an interaction data analysis network, carrying out key information mining operation or interaction data semantic analysis operation on a first member of an interaction data set corresponding to interaction data to be responded; candidate interaction data corresponding to the interaction data integrator with the key information matching parameters between the first members of the interaction data set being greater than the predetermined first preset matching parameters are determined from the optimized interaction data set, or candidate interaction data corresponding to the interaction data integrator with the interaction data semantic matching parameters of the first members of the interaction data set being greater than the predetermined second preset information matching parameters are determined from the optimized interaction data set; and feeding back the determined candidate interaction data as response interaction data corresponding to the interaction data to be responded to the target network user. Based on the above, since the determined response interaction data is the candidate interaction data matched with the interaction data to be responded based on the optimized interaction data set, the reliability of the determined response interaction data is relatively high, so that the reliability of intelligent interaction response can be improved.

The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for analyzing big data for a meta-universe intelligent interactive response, comprising:

2. The big data analysis method for a meta-universe intelligent interactive response according to claim 1, wherein the step of forming a corresponding optimized interactive data analysis network through network optimization comprises:

3. The big data analysis method for a meta-universe intelligent interactive response according to claim 2, wherein the step of performing a second optimization operation on the intermediate interactive data analysis network based on each optimized member description vector in the optimized interactive data set and the member semantic data set to form a high-level interactive data analysis network corresponding to the intermediate interactive data analysis network comprises:

4. The big data analysis method for a metauniverse intelligent interactive response of claim 3, wherein the update description vector combination includes an update member description vector of the members of the polled interactive dataset, an update member description vector of the members of the first class of correlations, and an update member description vector of the members of the second class of correlations;

5. The method for big data analysis of a meta-universe intelligent interactive response according to claim 4, wherein the step of analyzing each update member description vector of the update description vector combination based on a first analysis rule to analyze a corresponding error first local value in a case that the poll interactive data integrator belongs to a high-level interactive data set member, comprises:

6. The method for big data analysis of a meta-universe intelligent interactive response according to claim 4, wherein the step of analyzing the semantic description vector and the updated member description vector corresponding to the polled interactive data set member based on a second analysis rule to analyze a second local value of the corresponding error comprises:

7. The method for analyzing big data of a meta-universe intelligent interactive response according to claim 4, wherein the step of fusing the first local value based on the error and the second local value based on the error to obtain a network optimization error index corresponding to the intermediate interactive data analysis network comprises the steps of:

8. The big data analysis method for meta-universe intelligent interactive response according to claim 7, wherein the optimized interactive data set comprises a plurality of optimized interactive data subsets with precedence relations, each optimized interactive data subset has a corresponding precedence relation characterization parameter, and each optimized interactive data subset sequentially carries out a second optimization operation on the intermediate interactive data analysis network according to the precedence relation;

9. The big data analysis method for a meta-universe intelligent interactive response according to claim 2, wherein the step of performing a third optimization operation on the advanced interactive data analysis network based on the at least one complex interactive data set member to form a corresponding optimized interactive data analysis network comprises:

10. A big data analysis system for a meta-universe intelligent interactive response, characterized by comprising a processor and a memory, the memory being for storing a computer program, the processor being for executing the computer program for implementing the method of any of claims 1-9.