CN117454986A - Business expert thinking digitization and dynamic evolution system capable of realizing interactive learning - Google Patents
Business expert thinking digitization and dynamic evolution system capable of realizing interactive learning Download PDFInfo
- Publication number
- CN117454986A CN117454986A CN202311749356.0A CN202311749356A CN117454986A CN 117454986 A CN117454986 A CN 117454986A CN 202311749356 A CN202311749356 A CN 202311749356A CN 117454986 A CN117454986 A CN 117454986A
- Authority
- CN
- China
- Prior art keywords
- thinking
- business
- expert
- business expert
- element information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 230000002452 interceptive effect Effects 0.000 title claims abstract description 16
- 238000000034 method Methods 0.000 claims abstract description 50
- 230000003542 behavioural effect Effects 0.000 claims abstract description 21
- 238000005065 mining Methods 0.000 claims abstract description 19
- 238000010276 construction Methods 0.000 claims abstract description 18
- 230000000877 morphologic effect Effects 0.000 claims abstract description 12
- 238000003062 neural network model Methods 0.000 claims abstract description 9
- 230000006399 behavior Effects 0.000 claims description 58
- 238000004458 analytical method Methods 0.000 claims description 53
- 238000004422 calculation algorithm Methods 0.000 claims description 33
- 230000008569 process Effects 0.000 claims description 30
- 238000012545 processing Methods 0.000 claims description 13
- 238000013528 artificial neural network Methods 0.000 claims description 12
- 230000003993 interaction Effects 0.000 claims description 12
- 239000013598 vector Substances 0.000 claims description 10
- 230000000694 effects Effects 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 7
- 238000007711 solidification Methods 0.000 claims description 6
- 230000008023 solidification Effects 0.000 claims description 6
- 238000009412 basement excavation Methods 0.000 claims description 5
- 238000010224 classification analysis Methods 0.000 claims description 5
- 238000004883 computer application Methods 0.000 claims description 4
- 238000013461 design Methods 0.000 claims description 4
- 238000010801 machine learning Methods 0.000 claims description 4
- 238000012544 monitoring process Methods 0.000 claims description 2
- 230000019771 cognition Effects 0.000 abstract description 2
- 230000006870 function Effects 0.000 description 8
- 238000003860 storage Methods 0.000 description 8
- 238000012549 training Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000008676 import Effects 0.000 description 4
- 238000005295 random walk Methods 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000008713 feedback mechanism Effects 0.000 description 1
- 230000006698 induction Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/02—Knowledge representation; Symbolic representation
- G06N5/022—Knowledge engineering; Knowledge acquisition
- G06N5/025—Extracting rules from data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/26—Discovering frequent patterns
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/047—Probabilistic or stochastic networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0475—Generative networks
Abstract
The invention discloses a business expert thinking digitizing and dynamic evolving system capable of interactive learning, which belongs to the intelligent cognition field, and comprises the following steps: the element information forming module is used for forming element information; the element information classification acquisition module is used for classifying element information and acquiring information by adopting different methods after classification; the business expert thinking element representation framework construction and updating module is used for constructing a business expert thinking element representation framework and fusing the multi-business expert thinking elements based on graph theory; extracting behavioral characteristics and non-behavioral characteristics of business expert thinking, mining network characteristics of the business expert thinking, and finding out corresponding thinking modes for recommendation; and adopting a deep confidence neural network model to carry out continuous iterative updating on the thought knowledge of the business expert. The invention learns and mines expert thinking mode, realizes the evolution of expert thinking, and provides support for morphological representation, construction, updating, recommended use and the like of the expert thinking elements of the follow-up business.
Description
Technical Field
The invention relates to the field of intelligent cognition, in particular to a business expert thinking digitizing and dynamic evolution system capable of interactive learning.
Background
The intelligent expert system is an important branch of artificial intelligence development, and the main objective is to complete the conversion of artificial intelligence from theory to practical application, and representative applications thereof include a domain expert system, a decision support system and the like. Through decades of development, expert systems and decision support systems have penetrated into various fields of engineering, agriculture, medical treatment, electric power, aerospace, construction, chemistry, military industry and the like, and have produced great economic benefits. At present, the intelligent expert system has the advantages of weak knowledge representation and low flexibility; the parallelism, mass and adaptability of knowledge reasoning are to be improved; the problems of self-learning ability, self-error correction ability, knowledge self-acquisition ability and the like are to be enhanced, and the problems need to be solved through continuous and intensive research.
In the field of business analysis, a large number of expert systems and decision support systems are applied at present, but on one hand, the expert systems are difficult to effectively abstract business analysis experience, and the knowledge width and basic principle understanding related to the business are relatively single; likewise, the decision support system can only assist the decision-making and decision-making of the service in a quantitative manner, and cannot well support the decision-making of the service task with uncertainty, so that a great development space is still provided.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, provides a business expert thinking digitizing and dynamic evolving system capable of interactively learning, analyzes and defines the business expert thinking element information, classifies and collects non-behavior element information and behavior element information according to characteristics, achieves descriptability of the business expert thinking, completes acquisition and digitizing of the business expert thinking information, learns and discovers expert thinking modes by man-machine interaction feedback of 'people in a loop', achieves evolution of expert thinking, and provides support for morphological representation, construction, updating, recommended use and the like of follow-up business expert thinking elements.
The invention aims at realizing the following scheme:
a business expert thought digitizing and dynamic evolving system capable of interactive learning, comprising:
the element information forming module is used for forming element information by analyzing and defining business expert thinking information;
the element information classification acquisition module is used for classifying element information into non-behavior element information and behavior element information, and acquiring information by adopting different methods aiming at different element information;
the business expert thinking element representation framework construction and updating module is used for constructing a business expert thinking element representation framework based on a graph network aiming at a polymorphic expert knowledge representation mode in the business analysis field and fusing multi-business expert thinking elements based on graph theory; extracting behavior features and non-behavior features of business expert thinking through frequent subgraphs and network representation learning respectively, mining network features of the business expert thinking in a clustering analysis and classification analysis mode, and finding out corresponding thinking modes in different business analysis tasks for recommendation; in the continuous learning process, the deep confidence neural network model is adopted to carry out continuous iterative updating on the thought knowledge of the business expert, so as to form a model which better accords with the thought of the business expert and support intelligent analysis and processing of the business.
Further, the element information includes knowledge data, rule template class data, algorithm model data, and execution flow data.
Further, in the element information classification module, non-behavior element information is collected through an import interface provided by an analysis platform, behavior element information is collected through a workflow management component of a design platform, and the collected behavior element information is respectively stored in a corresponding knowledge base of a business storage and analysis platform to finish the digitization of business expert thinking information; interaction between a user and a computer application tool in the execution process of various activities is realized through man-machine interaction, so that a computer learns business expert thinking and continuously and dynamically evolves.
Further, in the business expert thinking element representation framework construction and updating module, the construction of the business expert thinking element representation framework based on the graph network aiming at the polymorphic expert knowledge representation mode in the business intelligent analysis field specifically comprises the following steps:
aiming at polymorphic expert knowledge in the field of business analysis, a graph network is constructed by a flow method and a rule template thinking form in business expert thinking, and the construction of the business expert thinking network is finally completed by fusing a Petri network with a semantic network of knowledge thinking elements, wherein the Petri network is mathematical representation of a discrete parallel system.
Further, the characteristics of the business expert thinking constructed through the Petri net include behavioral characteristics and non-behavioral characteristics.
Further, the learning through frequent subgraphs and network representation extracts behavioral features and non-behavioral features of business expert thinking respectively, which specifically comprises: outputting a frequent item set by adopting an Apriori frequent subgraph mining algorithm to obtain thinking element behavior characteristics based on frequent subgraphs; the network is recoded by reserving the network topology structure, vertex content and side information of the business expert thinking graph, the network vertices are embedded into a low-dimensional vector space, the problem of non-behavior feature extraction of the business expert thinking is solved by using a machine learning method in a new vector space, and the Apriori frequent subgraph mining algorithm is an association rule mining algorithm.
Further, the continuous iterative updating of the business expert thinking knowledge by using the deep confidence neural network model specifically comprises the following steps: the deep confidence neural network is utilized to integrate the excavated knowledge of various business specialists into a business specialist thinking element library, so that the dynamic update of the business specialist thinking elements is realized, and the thinking elements with the expert knowledge in the field of business analysis are formed through continuous learning, excavation and solidification.
Further, the non-behavioral element information specifically includes: rule template type morphological element information, knowledge type morphological element information and algorithm model type morphological element information.
Further, the behavior element information specifically includes: the flow method includes form factor information.
Further, the workflow management component is used for creating and executing a workflow execution service formed by a plurality of workflow engines of the process instance, and monitoring and managing the execution state of the workflow process instance.
The beneficial effects of the invention include:
(1) According to the invention, the business expert thinking element information of a plurality of angles is acquired through the analysis platform import interface and the platform workflow management component, so that the problem that the traditional business expert thinking acquisition mode is single and the acquired information is incomplete is solved.
(2) The invention carries out the example description on the acquisition of the expert thinking element information, and shows the feasibility of the expert thinking element in the business analysis work and the connotation of the expert thinking element information.
(3) The invention provides accurate data support for inheritance and solidification of business expert thinking, morphological representation, construction, updating, recommended use and the like of the business expert thinking.
(4) The invention is based on the business expert thinking network space, so that the business expert thinking behavior characteristics and non-behavior characteristics can be extracted and represented.
(5) The invention acquires and solidifies expert thinking information by man-machine interaction, and can realize the mining of the business expert thinking mode based on the business expert thinking network characteristics, thereby realizing the solidification and use of expert knowledge and realizing the dynamic evolution of expert thinking.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
FIG. 1 is a diagram showing the construction of general element information for business specialists in the embodiment of the invention;
FIG. 2 is a diagram of the construction of business expert thinking element information in a specific scenario in an embodiment of the present invention;
FIG. 3 is a flow chart of business expert thinking element information collection in an embodiment of the invention;
FIG. 4 is a schematic diagram of an implementation of a workflow management component in an embodiment of the invention;
FIG. 5 is an overall architecture diagram of a workflow management assembly in an embodiment of the invention.
Detailed Description
All of the features disclosed in all of the embodiments of this specification, or all of the steps in any method or process disclosed implicitly, except for the mutually exclusive features and/or steps, may be combined and/or expanded and substituted in any way.
In view of the current situation in the background, the inventors of the present invention consider that: the intelligent expert system for carrying out business analysis has analysis capability approaching to human beings, such as induction, summarization, reasoning, deduction and the like, solidifies the thinking and processing capability of business experts, and realizes the true intelligence of the computer system for carrying out business analysis and processing. In order to make the service intelligent application better approach to the use requirement of users, an interactive feedback mechanism needs to be introduced into a service loop, so that the man-machine hybrid intelligence of 'man-in-loop' is realized. In the service production process, service structuring personnel, service duty personnel, service analysis personnel and the like can correct the system processing result through opinion feedback and online labeling, and the system processing result is stored as system training data, so that the system background algorithm is continuously trained, optimized and perfected, the service information processing system has learning growth capacity, active learning and incremental learning are formed, and the system robustness is enhanced. Aiming at the point, the invention designs a business expert thinking digitizing and dynamic evolution scheme capable of interactively learning, acquires and learns expert thinking information based on man-machine interaction, digitally constructs an expert thinking network, and finally realizes business expert thinking digitizing and dynamic evolution based on expert thinking network characteristics and by mining and utilizing an expert thinking mode.
As shown in FIG. 1, the method for collecting the essential information of the business expert thinking first needs to perform descriptive definition analysis, namely clear essential information, on the business expert thinking. By analyzing and summarizing the thinking mode of the expert at ordinary times and observing the behavior mode of the summarized expert at ordinary times, knowledge data in the relevant field, some algorithm models with excellent performance can be mainly used in the process of processing business information of the expert, meanwhile, a rule template relevant to business processing is designed, and meanwhile, the expert shows some business processing modes special to the expert and is embodied in details of business processing flows.
As shown in fig. 2, the invention demonstrates the meaning of the business expert thinking element information and clarifies the element information acquisition object. The acquisition object mainly comprises the following categories:
(1) knowledge data: the thematic writing scene has element information such as choice knowledge, thematic vocabulary, a method for remedying, a writing mode and the like; the high-value target behavior prediction scene has element information such as target basic parameters, target activity rules, remote sensing data, radar data and the like; the important crisis event auxiliary prediction scene comprises element information such as character encyclopedia knowledge, event encyclopedia knowledge, industry files, event libraries and the like.
(2) Rule template data: the thematic writing scene has element information such as a theme generation template, a manuscript comparison rule, a writing style template, a manuscript reporting rule and the like; the high-value target behavior prediction scene has element information such as target recognition rules, abnormal alarm rules, historical data matching rules, target portrait templates and the like; the critical event auxiliary prediction scene has element information such as event relation templates, critical event representation templates, conflict certificate false rules, critical event judgment rules, evaluation index templates and the like.
(3) Algorithm model data: the thematic writing scene has element information such as a convolutional neural network, a user attention network model, a similarity calculation algorithm, a Bayesian network model and the like; the high-value target behavior prediction has element information such as a cyclic neural network, a long-short-term memory network model, a convolutional neural network model, a deep confidence neural network and the like; the major crisis event auxiliary prediction scene comprises element information such as an incremental clustering algorithm, a transfer learning model, a probability map model, a Bayesian network and the like.
(4) Executing flow data: the three scenes are intersected and cover the flow steps of analysis and judgment, the calling behaviors of algorithms and models, the using behaviors of rule templates, the service consulting behaviors and other element information.
As shown in fig. 3, for the business expert thinking element information defined by the analysis, the invention classifies the element information into two types of non-behavior element information and behavior element information, and respectively collects the business expert thinking element information, and the specific steps are as follows:
step 1: and (5) acquiring non-behavior element information. The non-behavior element information in the thought of the business expert comprises three forms of rule templates, knowledge and algorithm models, and specifically refers to rule data used in analysis and judgment, algorithm data used in analysis and judgment, model data used in analysis and judgment and knowledge data used in analysis and judgment. Business expert thinking element collection interfaces fall into two categories: firstly, an analysis platform importing interface; and secondly, a workflow management component. The non-behavior element information is stored into a corresponding knowledge base of a business storage and analysis platform data warehouse through an analysis platform import interface. The rule template type data is input into the platform by a business analysis personnel or expert through an input template provided by the business storage and analysis platform, and is stored in a corresponding knowledge base; and the knowledge data and the algorithm model data are stored into the corresponding knowledge base through the importing interface of the business storage and analysis platform.
Step 2: and (5) collecting behavior element information. The behavior element information in the thought of the business expert comprises a flow method type form, and concretely refers to the steps of the analysis and judgment flow of the expert, the adoption of rule template behaviors in the expert analysis and judgment, the invoking of algorithms and model behaviors in the expert analysis and judgment, and the consulting of knowledge behaviors in the expert analysis and judgment. The behavior data are collected through a workflow management component of a business expert thinking element collection interface and stored in a knowledge base corresponding to a business storage and analysis platform.
As shown in fig. 4, the workflow management component collects behavior element information, and in the implementation process, a process definition tool for analyzing and defining an actual business process, which is also called a workflow modeling tool, needs to be created, is responsible for creating and executing a workflow execution service formed by a plurality of workflow engines of a process instance, monitors and manages the execution state of the workflow process instance, is used for processing a workflow client application requiring a manual intervention task in the execution process of the workflow process instance, and is used for processing data.
The workflow management component for collecting behavioral element information in the present invention is similar to a software system, as shown in fig. 5, running on a platform with one or more workflow engines (also called workflow machines) that can interpret process definitions, interact with participants (including people or software) of the workflow, and invoke other application tools as needed, advance execution of workflow instances, and monitor the running state of the workflow. The specific implementation process is divided into an establishment stage and an operation stage, wherein the operation stage comprises two functions of process control and man-machine interaction, and the specific functions are as follows:
(1) function of the setup phase: the method mainly completes the definition and modeling functions of the workflow process and related activities thereof, and completes the conversion from the actual business process to formalized definition which can be processed by a computer.
(2) Control function of the operation phase: under a certain running environment, the definition of the business process is analyzed, the running instance of the business process is created and controlled, and the execution sequence arrangement and scheduling functions of activities in each execution process are completed.
(3) Man-machine interaction function in operation phase: interactions between the user and the computer application tool during execution of various activities are achieved.
The business expert thinking network construction and updating process comprises the steps of firstly constructing a model of polymorphic expert knowledge by workflow thinking and rule template thinking based on a Petri network, and fusing the multi-business expert thinking network to form business expert thinking elements; aiming at behavioral and non-behavioral characteristics in the business expert thinking network, the extraction of the business expert thinking network characteristics is realized by adopting a frequent subgraph and network representation learning mode; performing mode mining on the thought network characteristics of the business specialists through clustering and classification analysis, and finding out typical thought modes of the business specialists; the deep confidence neural network is adopted to complete the knowledge updating and solidification of the business expert thinking, and the specialized, intelligent and digital business expert thinking in the business analysis field is formed.
After the business expert thinking element is constructed by the Petri network, the feature extraction can be carried out on the business expert thinking element, so that the follow-up relevant excavation of the thinking element is facilitated. The business expert thinking element features constructed through the Petri network comprise two aspects of behavioral features and non-behavioral features:
(1) And (5) extracting the behavior characteristics of the thinking elements based on the frequent subgraphs. In order to extract the behavior characteristics of the thought elements of the business specialist, the thought elements of the specialist are firstly represented by a graph, and the representation of the graph relates to isomorphism, relative support degree, absolute support degree, label graph, sub graph isomorphism, connection graph and non-connection graph, graph set core and similar functions of the graph. After the graph representation of the thinking element is carried out, then frequent subgraphs are mined, firstly, the vertex and edge marks are ordered, then the thinking element graph information is stored as an adjacent matrix, the adjacent matrix is normalized, which coding value in the obtained normal shape matrix is the minimum, then the judgment of candidate subgraphs is carried out, and frequent item sets are output, so that the thinking element behavior characteristics based on the frequent subgraphs are obtained.
(2) A non-behavioral thinking element feature representation learned based on the network representation. In order to realize extraction of non-behavioral characteristics of thought elements of business specialists. The present invention contemplates the use of network representation learning (Network Representation Learning), also known as network embedding (Network Embedding), for feature extraction. Network vertices are embedded into the low-dimensional vector space by recoding the network by preserving business expert thought element graph network topology, vertex content, and other side information. In the new vector space, the complex analysis and mining problems of the business expert thinking network can be easily solved by using a machine learning method.
The original business expert thought element graph network generates a node sequence training library by using a Random Walk (Random-Walk) algorithm in Deep-Walk algorithm, and learns and updates parameters by using a Skip-Gram model. By means of such a walk algorithm, a series of training examples or sets of dimension reduction features can be obtained. There are two specific training modes in Skip-Gram model: hierarchical Softmax and Negative Sampling. Given a networkWhere V denotes all nodes in the network and E denotes all edges in the network, then it is obvious that there is +.>. The purpose of the network embedding algorithm is to find a mapping +.>Mapping each node to a +.>Space of dimensions. In this low-dimensional space, the structural information of the network, including the local network structure and the global network structure, is still preserved. The node vectorization model based on Deep-Walk is adopted to realize that the network vertex is embedded into a low-dimensional vector space, so that a foundation is laid for the follow-up development of classification analysis and excavation based on the neural network. The invention is thatThe algorithm to be built consists of two main parts: first, a training library is generated using Random Walk (Random Walk); second, the Skip-Gram model is used to learn and update parameters.
The invention provides a thought map network clustering method (Kfk-means) based on an improved K-means algorithm in cluster analysis and mining of business expert thought elements. The whole algorithm steps in two steps: firstly, obtaining a feature set, namely obtaining clusters, through frequent subgraph mining, random sampling and a K-means algorithm; and secondly, taking the cluster obtained in the first step as input, and then performing condensation to obtain the best k cluster centers.
After the graph network of the business expert thinking elements is trained into the representation of the hidden space, the nodes and edges can be converted into high-dimensional vectors which can be received by the deep neural network, and then the business expert thinking elements can be classified and mined by adopting the neural network models such as CNN, RNN and the like.
Business expert thinking element update based on deep belief neural network. The deep belief neural network is one type of deep neural network. The method can be used for unsupervised learning and is similar to a self-encoder; the method can also be used for supervised learning and used as a classifier. The invention adopts the deep confidence neural network model to classify and process the network characteristic data of the business expert thinking elements, thereby realizing the dynamic update of the knowledge of the business expert thinking elements and forming more specialized business expert thinking elements. The general network structure of the deep belief neural network can be expressed as: several layers of constrained boltzmann machines were used as feature extractors, followed by a softmax classifier in series.
It should be noted that, within the scope of protection defined in the claims of the present invention, the following embodiments may be combined and/or expanded, and replaced in any manner that is logical from the above specific embodiments, such as the disclosed technical principles, the disclosed technical features or the implicitly disclosed technical features, etc.
Example 1
A business expert thought digitizing and dynamic evolving system capable of interactive learning, comprising:
the element information forming module is used for forming element information by analyzing and defining business expert thinking information;
the element information classification acquisition module is used for classifying element information into non-behavior element information and behavior element information, and acquiring information by adopting different methods aiming at different element information;
the business expert thinking element representation framework construction and updating module is used for constructing a business expert thinking element representation framework based on a graph network aiming at a polymorphic expert knowledge representation mode in the business analysis field and fusing multi-business expert thinking elements based on graph theory; extracting behavior features and non-behavior features of business expert thinking through frequent subgraphs and network representation learning respectively, mining network features of the business expert thinking in a clustering analysis and classification analysis mode, and finding out corresponding thinking modes in different business analysis tasks for recommendation; in the continuous learning process, the deep confidence neural network model is adopted to carry out continuous iterative updating on the thought knowledge of the business expert, so as to form a model which better accords with the thought of the business expert and support intelligent analysis and processing of the business.
Example 2
On the basis of embodiment 1, the element information includes knowledge data, rule template class data, algorithm model data, and execution flow data.
Example 3
On the basis of embodiment 1, in the element information classification module, acquiring non-behavior element information through an import interface provided by an analysis platform, acquiring behavior element information through a design platform workflow management component, and respectively storing the acquired behavior element information into a corresponding knowledge base of a business storage and analysis platform to finish the digitization of business expert thinking information; interaction between a user and a computer application tool in the execution process of various activities is realized through man-machine interaction, so that a computer learns business expert thinking and continuously and dynamically evolves.
Example 4
On the basis of embodiment 1, in the business expert thinking element representation framework construction and updating module, the multi-form expert knowledge representation mode in the business intelligent analysis field is constructed, and the business expert thinking element representation framework based on the graph network specifically comprises:
aiming at polymorphic expert knowledge in the field of business analysis, a flow method and a regular template thinking form in business expert thinking are subjected to graph network construction, basic units in thinking elements are described by using a Petri network, the basic units are fused with a semantic network of knowledge thinking elements, and finally, the construction of the business expert thinking network is completed, wherein the Petri network is mathematical representation of a discrete parallel system.
Example 5
On the basis of embodiment 4, the characteristics of business expert thinking constructed through the Petri net include behavioral characteristics and non-behavioral characteristics.
Example 6
On the basis of embodiment 5, the learning through frequent subgraphs and network representation extracts behavioral characteristics and non-behavioral characteristics of business expert thinking respectively, which specifically comprises: outputting a frequent item set by adopting a frequent subgraph mining algorithm of the Apriori idea to obtain thinking element behavior characteristics based on the frequent subgraph; the network is recoded by reserving the network topology structure, vertex content and side information of the business expert thinking graph, the network vertices are embedded into a low-dimensional vector space, the problem of non-behavior feature extraction of the business expert thinking is solved by using a machine learning method in a new vector space, and the Apriori frequent subgraph mining algorithm is an association rule mining algorithm.
Example 7
Based on the embodiment 1, the continuous iterative updating of the business expert thinking knowledge by using the deep confidence neural network model specifically includes: the deep confidence neural network is utilized to integrate the excavated knowledge of various business specialists into a business specialist thinking element library, so that the dynamic update of the business specialist thinking elements is realized, and the thinking elements with the expert knowledge in the field of business analysis are formed through continuous learning, excavation and solidification.
Example 8
On the basis of embodiment 3, the non-behavioral element information specifically includes: rule template type morphological element information, knowledge type morphological element information and algorithm model type morphological element information.
Example 9
On the basis of embodiment 3, the behavior element information specifically includes: the flow method includes form factor information.
Example 10
On the basis of embodiment 3, the workflow management component is configured to create and execute a workflow execution service formed by a plurality of workflow engines of the process instance, and monitor and manage an execution state of the workflow process instance.
The units involved in the embodiments of the present invention may be implemented by software, or may be implemented by hardware, and the described units may also be provided in a processor. Wherein the names of the units do not constitute a limitation of the units themselves in some cases.
According to an aspect of embodiments of the present invention, there is provided a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The computer instructions are read from the computer-readable storage medium by a processor of a computer device, and executed by the processor, cause the computer device to perform the methods provided in the various alternative implementations described above.
As another aspect, the embodiment of the present invention also provides a computer-readable medium that may be contained in the electronic device described in the above embodiment; or may exist alone without being incorporated into the electronic device. The computer-readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to implement the methods described in the above embodiments.
Claims (10)
1. A business expert thinking digitizing and dynamic evolving system capable of interactive learning is characterized by comprising:
the element information forming module is used for forming element information by analyzing and defining business expert thinking information;
the element information classification acquisition module is used for classifying element information into non-behavior element information and behavior element information, and acquiring information by adopting different methods aiming at different element information;
the business expert thinking element representation framework construction and updating module is used for constructing a business expert thinking element representation framework based on a graph network aiming at a polymorphic expert knowledge representation mode in the business analysis field and fusing multi-business expert thinking elements based on graph theory; extracting behavior features and non-behavior features of business expert thinking through frequent subgraphs and network representation learning respectively, mining network features of the business expert thinking in a clustering analysis and classification analysis mode, and finding out corresponding thinking modes in different business analysis tasks for recommendation; in the continuous learning process, the deep confidence neural network model is adopted to carry out continuous iterative updating on the thought knowledge of the business expert, so as to form a model which better accords with the thought of the business expert and support intelligent analysis and processing of the business.
2. The business expert thought digitizing and dynamic evolving system for interactive learning according to claim 1, wherein the element information includes knowledge data, rule template class data, algorithm model data and execution flow data.
3. The interactive learning business expert thinking digitizing and dynamic evolving system according to claim 1, characterized in that in the element information classifying module, the non-behavior element information is collected through the importing interface provided by the analyzing platform, the behavior element information is collected through the design platform workflow managing component, and is stored in the corresponding knowledge base of the business storing and analyzing platform to complete the digitizing of the business expert thinking information; interaction between a user and a computer application tool in the execution process of various activities is realized through man-machine interaction, so that a computer learns business expert thinking and continuously and dynamically evolves.
4. The business expert thinking digitizing and dynamic evolving system capable of interactive learning according to claim 1, wherein in the business expert thinking element representing frame constructing and updating module, for the polymorphic expert knowledge representing mode in the business intelligent analysis field, the business expert thinking element representing frame based on the graph network is constructed, specifically comprising:
aiming at polymorphic expert knowledge in the field of business analysis, a graph network is constructed by a flow method and a rule template thinking form in business expert thinking, and the construction of the business expert thinking network is finally completed by fusing a Petri network with a semantic network of knowledge thinking elements, wherein the Petri network is mathematical representation of a discrete parallel system.
5. The business expert thinking digitizing and dynamically evolving system for interactive learning according to claim 4, wherein the features of the business expert thinking built through the Petri net include behavioral features and non-behavioral features.
6. The business expert thinking digitizing and dynamic evolving system capable of interactive learning according to claim 5, wherein the behavior feature and non-behavior feature of the business expert thinking are extracted by frequent sub-graph and network representation learning, respectively, specifically comprising: outputting a frequent item set by adopting an Apriori frequent subgraph mining algorithm to obtain thinking element behavior characteristics based on frequent subgraphs; the network is recoded by reserving the network topology structure, vertex content and side information of the business expert thinking graph, the network vertices are embedded into a low-dimensional vector space, the problem of non-behavior feature extraction of the business expert thinking is solved by using a machine learning method in a new vector space, and the Apriori frequent subgraph mining algorithm is an association rule mining algorithm.
7. The business expert thinking digitizing and dynamic evolving system capable of interactive learning according to claim 1, wherein the continuous iterative updating of the business expert thinking knowledge by the deep confidence neural network model comprises: the deep confidence neural network is utilized to integrate the excavated knowledge of various business specialists into a business specialist thinking element library, so that the dynamic update of the business specialist thinking elements is realized, and the thinking elements with the expert knowledge in the field of business analysis are formed through continuous learning, excavation and solidification.
8. The business expert thinking digitizing and dynamic evolving system for interactive learning according to claim 3, wherein the non-behavior element information specifically comprises: rule template type morphological element information, knowledge type morphological element information and algorithm model type morphological element information.
9. The business expert thinking digitizing and dynamic evolving system for interactive learning according to claim 3, wherein the behavior element information specifically comprises: the flow method includes form factor information.
10. The business expert thinking digitizing and dynamic evolving system of claim 3, wherein the workflow managing component is used for creating and executing the workflow executing service composed of a plurality of workflow engines of the process instance, and monitoring and managing the executing state of the workflow process instance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311749356.0A CN117454986B (en) | 2023-12-19 | 2023-12-19 | Business expert thinking digitization and dynamic evolution system capable of realizing interactive learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311749356.0A CN117454986B (en) | 2023-12-19 | 2023-12-19 | Business expert thinking digitization and dynamic evolution system capable of realizing interactive learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117454986A true CN117454986A (en) | 2024-01-26 |
| CN117454986B CN117454986B (en) | 2024-03-19 |
Family
ID=89585801
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311749356.0A Active CN117454986B (en) | 2023-12-19 | 2023-12-19 | Business expert thinking digitization and dynamic evolution system capable of realizing interactive learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117454986B (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110173189A1 (en) * | 2006-02-27 | 2011-07-14 | The Regents Of The University Of California | Graph querying, graph motif mining and the discovery of clusters |
| CN103400227A (en) * | 2013-08-05 | 2013-11-20 | 浙江大学 | Graph mining and graph distance-based flow recommendation method |
| CN106126577A (en) * | 2016-06-17 | 2016-11-16 | 北京理工大学 | A kind of weighted association rules method for digging based on data source Matrix dividing |
| CN107341611A (en) * | 2017-07-06 | 2017-11-10 | 浙江大学 | A kind of operation flow based on convolutional neural networks recommends method |
| US10311442B1 (en) * | 2007-01-22 | 2019-06-04 | Hydrojoule, LLC | Business methods and systems for offering and obtaining research services |
| CN112288091A (en) * | 2020-10-30 | 2021-01-29 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Knowledge inference method based on multi-mode knowledge graph |
| CN112328801A (en) * | 2020-09-28 | 2021-02-05 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Method for predicting group events by event knowledge graph |
| CN112904805A (en) * | 2021-01-15 | 2021-06-04 | 天津商业大学 | Digital twin modeling and virtual-real synchronization method for multi-axis numerical control machine tool |
| CN113095489A (en) * | 2021-05-10 | 2021-07-09 | 电子科技大学 | Inductive multi-level graph network representation learning method |
| CN113869609A (en) * | 2021-10-29 | 2021-12-31 | 北京宝兰德软件股份有限公司 | Method and system for predicting confidence of frequent subgraph of root cause analysis |
| CN114297404A (en) * | 2021-12-29 | 2022-04-08 | 北京信息科技大学 | Knowledge graph construction method for field evaluation expert behavior track |
| CN115187014A (en) * | 2022-06-24 | 2022-10-14 | 山东科技大学 | Business process prediction method based on Petri network reachability graph interpretation |
| CN115238197A (en) * | 2022-09-26 | 2022-10-25 | 中国电子科技集团公司第十研究所 | Expert thinking model-based field business auxiliary analysis method |
| CN115496570A (en) * | 2022-10-25 | 2022-12-20 | 西安交通大学 | Medium area identification method for complex tax data system |
| CN115602273A (en) * | 2022-10-19 | 2023-01-13 | 广东省中医院（广州中医药大学第二附属医院、广州中医药大学第二临床医学院、广东省中医药科学院）（Cn） | Traditional Chinese medicine auxiliary decision method and system combining clinical knowledge map and multi-mode information fusion |
-
2023
- 2023-12-19 CN CN202311749356.0A patent/CN117454986B/en active Active
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110173189A1 (en) * | 2006-02-27 | 2011-07-14 | The Regents Of The University Of California | Graph querying, graph motif mining and the discovery of clusters |
| US10311442B1 (en) * | 2007-01-22 | 2019-06-04 | Hydrojoule, LLC | Business methods and systems for offering and obtaining research services |
| CN103400227A (en) * | 2013-08-05 | 2013-11-20 | 浙江大学 | Graph mining and graph distance-based flow recommendation method |
| CN106126577A (en) * | 2016-06-17 | 2016-11-16 | 北京理工大学 | A kind of weighted association rules method for digging based on data source Matrix dividing |
| CN107341611A (en) * | 2017-07-06 | 2017-11-10 | 浙江大学 | A kind of operation flow based on convolutional neural networks recommends method |
| CN112328801A (en) * | 2020-09-28 | 2021-02-05 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Method for predicting group events by event knowledge graph |
| CN112288091A (en) * | 2020-10-30 | 2021-01-29 | 西南电子技术研究所（中国电子科技集团公司第十研究所） | Knowledge inference method based on multi-mode knowledge graph |
| CN112904805A (en) * | 2021-01-15 | 2021-06-04 | 天津商业大学 | Digital twin modeling and virtual-real synchronization method for multi-axis numerical control machine tool |
| CN113095489A (en) * | 2021-05-10 | 2021-07-09 | 电子科技大学 | Inductive multi-level graph network representation learning method |
| CN113869609A (en) * | 2021-10-29 | 2021-12-31 | 北京宝兰德软件股份有限公司 | Method and system for predicting confidence of frequent subgraph of root cause analysis |
| CN114297404A (en) * | 2021-12-29 | 2022-04-08 | 北京信息科技大学 | Knowledge graph construction method for field evaluation expert behavior track |
| CN115187014A (en) * | 2022-06-24 | 2022-10-14 | 山东科技大学 | Business process prediction method based on Petri network reachability graph interpretation |
| CN115238197A (en) * | 2022-09-26 | 2022-10-25 | 中国电子科技集团公司第十研究所 | Expert thinking model-based field business auxiliary analysis method |
| CN115602273A (en) * | 2022-10-19 | 2023-01-13 | 广东省中医院（广州中医药大学第二附属医院、广州中医药大学第二临床医学院、广东省中医药科学院）（Cn） | Traditional Chinese medicine auxiliary decision method and system combining clinical knowledge map and multi-mode information fusion |
| CN115496570A (en) * | 2022-10-25 | 2022-12-20 | 西安交通大学 | Medium area identification method for complex tax data system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117454986B (en) | 2024-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | Intelligent computing system based on pattern recognition and data mining algorithms | |
| Chiroma et al. | Progress on artificial neural networks for big data analytics: a survey | |
| CN104318340B (en) | Information visualization methods and intelligent visible analysis system based on text resume information | |
| Leng et al. | Mining and matching relationships from interaction contexts in a social manufacturing paradigm | |
| CN110914836A (en) | System and method for implementing continuous memory bounded learning in artificial intelligence and deep learning for continuously running applications across networked computing edges | |
| Gil et al. | Review of the complexity of managing big data of the internet of things | |
| Chefrour | Incremental supervised learning: algorithms and applications in pattern recognition | |
| Forestier et al. | Coastal image interpretation using background knowledge and semantics | |
| Huitzil et al. | Datil: Learning fuzzy ontology datatypes | |
| CN115238197A (en) | Expert thinking model-based field business auxiliary analysis method | |
| Pięta et al. | Applications of rough sets in big data analysis: an overview | |
| Wang et al. | An ontology-based framework for geospatial clustering | |
| Wang et al. | Research on evaluation model of music education informatization system based on machine learning | |
| CN113902223A (en) | Key target reasoning prediction system based on situation data | |
| CN117454986B (en) | Business expert thinking digitization and dynamic evolution system capable of realizing interactive learning | |
| Perner | Case‐based reasoning and the statistical challenges | |
| Nagaraj et al. | Enhancement of fraternal K-median algorithm with CNN for high dropout probabilities to evolve optimal time-complexity | |
| Garg et al. | A comparative study of clustering algorithms using mapreduce in hadoop | |
| CN112836718B (en) | Fuzzy knowledge neural network-based image emotion recognition method | |
| Perner | Case-based reasoning and the statistical challenges II | |
| Zhang et al. | Towards Urban General Intelligence: A Review and Outlook of Urban Foundation Models | |
| Wu et al. | Review of unsupervised learning techniques | |
| De Oliveira et al. | Visual content representation for cognitive systems: Towards augmented intelligence | |
| Gamal et al. | A fuzzy rough rule based system enhanced by fuzzy cellular automata | |
| Chahdi et al. | On the Use of Ontology as a priori Knowledge into Constrained Clustering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |