CN110059315B - Scientific and technological resource perception fusion decision method - Google Patents
Scientific and technological resource perception fusion decision method Download PDFInfo
- Publication number
- CN110059315B CN110059315B CN201910304666.9A CN201910304666A CN110059315B CN 110059315 B CN110059315 B CN 110059315B CN 201910304666 A CN201910304666 A CN 201910304666A CN 110059315 B CN110059315 B CN 110059315B
- Authority
- CN
- China
- Prior art keywords
- resource
- data
- behavior
- user
- tree
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to a scientific and technological resource perception fusion decision method, which comprises the following steps: the resource data is stored in a distributed mode in a node mode; collecting external corpus input by a user, and generating a service request according to the external corpus; generating a clone tree based on the nodes according to the service request; sensing data dynamic change and/or data interaction state through an induction link based on a clone tree; reconstructing or fusing the clone tree to obtain a fusion tree aiming at the data state change and/or the data interaction state; and releasing the resource data to the user according to the fusion tree. The invention can sense the resource demand change of the user in real time, dynamically adjust the released resources to the user, and enable the information between the resource supply and demand parties to be more matched.
Description
Technical Field
The invention relates to the technical field of data interaction management and information mining, in particular to a scientific and technological resource perception fusion decision method.
Background
In the field of technological resource innovation management, because the attribute of the technological resource is more various, the ownership is more complex, the application field or the application scene is specific and cross and coexistent, the management and the innovation fusion of the technological resource are more difficult, especially in the technological resource butt joint, the technological achievement of a provider and the technological requirement of an acquirer are huge, but in most of time, the two cannot be matched. The main reason is that the information is not equivalent to the information owned by the supply and demand parties. Currently, for the supplier, the scientific and technological resource data is detailed and huge, and is clearly shown as a scientific and technological result, but is debilitated as a scientific and technological result output, which is determined by the way of supplying and demanding scientific and technological resources. Unbalanced information or knowledge intake, passive big data acquisition or interaction, and inefficient network resource management for the scientific resource platform operators and resource management decision makers also make the connectors as both supply and demand sides passive. At present, a passive technological resource management mode or platform becomes an obstacle for the current technological resource fusion innovation.
Disclosure of Invention
The invention aims to overcome at least one defect (deficiency) of the prior art, and provides a scientific and technological resource perception fusion decision method which can be used for perceiving the resource demand change of a user in real time and dynamically adjusting the released resources to the user so as to enable the information between the resource supply and demand parties to be more matched.
The technical scheme adopted by the invention is as follows:
a technological resource perception fusion decision method comprises the following steps:
the resource data is stored in a distributed mode in a node mode;
collecting external corpus input by a user, and generating a service request according to the external corpus;
generating a clone tree based on the nodes according to the service request;
sensing data dynamic change and/or data interaction state through an induction link based on a clone tree;
reconstructing or fusing the clone tree to obtain a fusion tree aiming at the data state change and/or the data interaction state;
and releasing the resource data to the user according to the fusion tree.
Further, the method further comprises:
recording user behavior data in a layered mode according to behavior logic by using a user behavior growth tree, and calculating behavior complexity according to the user behavior data, wherein the user behavior data comprises one or more of access resource sample size, sample dispersion and interaction sample characteristics; and proportional adjustment is carried out on the resources released to the user according to the behavior complexity, more matched characteristic resources based on the fusion tree are released when the behavior complexity is larger, and more dissimilar characteristic resources based on the fusion tree are released when the behavior complexity is smaller.
Further, the method further comprises:
searching similar users according to the identity characteristics of the users;
and generating a user demand semantic vector according to the behavior data of the similar users, releasing the matched characteristic resources and the dissimilar characteristic resources based on the user demand semantic vector, and setting induction buried points.
Further, the method further comprises:
if the behavior complexity is monitored to be increased, an alarm signal is sent;
and if the behavior complexity is monitored to be reduced, adding the behavior semantic maximum frequent item set in the user behavior data into the resource data corresponding to the user.
Further, the clone tree-based inductive link is established by the following steps:
performing feature analysis on the scientific and technological resource feature library, and taking a feature analysis result as a first party of the induction link;
and carrying out real-time behavior analysis on the user behavior, and taking the real-time behavior analysis result as the second party of the induction link.
Further, the performing feature analysis on the scientific and technological resource feature library, and taking the feature analysis result as the first party of the induction link specifically includes:
performing feature analysis on the scientific and technological resource feature library to obtain a resource semantic tree classified by cluster feature types;
carrying out association rule analysis on the resource semantic tree to obtain a frequent item set;
and constructing a class resource decision classification tree according to the frequent item set, wherein the class resource decision classification tree and corresponding resource data are used as the first party of the induction link.
Further, the performing real-time behavior analysis on the user behavior, and taking the real-time behavior analysis result as the second party of the inductive link specifically includes:
setting a characteristic index for monitoring;
and extracting the strong correlation characteristic resources according to the characteristic indexes, releasing the strong correlation characteristic resources to users, and setting induction feedback buried points at the strong correlation characteristic resources as the second party of the induction link.
Further, the characteristic index comprises one or more of a maximum frequent item set proportion, a resource characteristic aggregation degree and a user resource semantic entropy value.
Further, the reconstructing the clone tree or fusing the data to obtain a fusion tree specifically includes:
and reconstructing the clone tree or fusing the data according to the association rule and/or the similarity calculation result to obtain a fusion tree.
Further, the resource data comprises user behavior semantics and/or technological association corpus, wherein the technological association corpus is obtained by directionally crawling corpus to an external data network of the system according to semantic analysis demand characteristics.
Further, the step of directionally crawling the corpus to the external data network of the system according to the semantic analysis demand features specifically comprises the following steps: generating a corpus demand according to feedback of semantic analysis, and transmitting the corpus demand to a corpus matcher in the form of a corpus demand tree or a corpus demand vector;
the corpus matcher matches or searches the corpus according to the characteristic relation or weight distribution presented in the corpus demand tree or the corpus demand vector.
Further, the service request includes a cluster index of service-associated resources, the method further comprising:
and calculating the resource cluster complexity according to the cluster index, and distributing calculation resources and/or network resources for the service request according to the resource cluster complexity.
Further, the service request also comprises a public key, a private key and a random request code; the service request is based on
The node generation clone tree specifically comprises the following steps:
analyzing the public key and the random request code to confirm whether the identity and the service request type of the input external corpus meet the authority requirements;
if the authority requirements are met, generating a matched private key according to the private key, analyzing the matched private key to obtain corresponding request data, and generating a clone tree based on the nodes according to the request data.
Further, the step of performing distributed storage on the resource data in the form of nodes specifically includes:
performing dependency syntactic analysis on the resource data to obtain labeled words;
creating a semantic tree, and carrying out distributed storage on resource data based on nodes of the semantic tree.
Further, the method further comprises: learning user behaviors and monitoring abnormal behaviors;
if abnormal behavior of the writing operation is monitored, temporarily blocking current behavior of the user through a permission prompt and/or a prompt box of a user grade, cloning current demand data of the user, adding a section of pseudo-data random code, and covering a page;
when the pseudo-data random code is monitored, a warning is sent to a true user, the true user is treated, if the warning is cancelled, the treatment feedback and the abnormal behavior mode are combined to be used as training data, and the internal attack mode is trained according to the training data, so that the internal abnormal behavior discrimination capability is improved;
if abnormal behavior of the non-writing operation is monitored, calculating the complexity of the behavior and dynamically adjusting the interactive data so that the complexity becomes smaller or tends to be stable, and simultaneously recording the behavior semantics and adding the behavior semantics into the stored resource data.
Compared with the prior art, the invention has the beneficial effects that:
(1) The change of the resource demand of the user is perceived in real time through the induction link based on the clone tree, and the released resource is dynamically adjusted, so that the information of the resource supply and demand double transmission is more matched;
(2) Recording user behavior data by using a user behavior growth tree, calculating behavior complexity, and determining the proportion of the matched characteristic resources and the different characteristic resources released based on the clone tree according to the behavior complexity, so that the resource requirements of users with unknown intention and wider resource requirement range can be met;
(3) Searching similar users according to the identity characteristics of the users, and releasing matching characteristic resources and dissimilar characteristic resources of the similar users for the users according to the identity characteristics of the users, so that the resource requirements of the users can be met as far as possible when the analysis depth of keywords input by the users is insufficient;
(4) The behaviors of a system visitor user, a login user and an administrator user are learned, and the recognition capability of monitoring abnormal behaviors is improved;
(5) Setting the coping mechanism when abnormal behavior of writing and non-writing operation occurs can improve the security of the system.
Detailed Description
The present embodiments are to be considered in all respects as illustrative and not restrictive. It will be appreciated by those skilled in the art that some well known techniques in the embodiments and descriptions thereof may be omitted.
Examples
The embodiment provides a scientific and technological resource perception fusion decision method, which comprises the following steps:
the resource data is stored in a distributed mode in a node mode;
collecting external corpus input by a user, and generating a service request according to the external corpus;
generating a clone tree based on the nodes according to the service request;
sensing data dynamic change and/or data interaction state through an induction link based on a clone tree;
reconstructing or fusing the clone tree to obtain a fusion tree aiming at the data state change and/or the data interaction state; and releasing the resource data to the user according to the fusion tree.
Sensing data dynamic change and/or data interaction state through the sensing links based on the clone tree means that sensing links are established between services (such as between user behavior analysis and resource semantic analysis), and/or between services and data (such as between resource semantic analysis and corpus references), and/or between data and data (such as between corpus references and semantic associations) for sensing data dynamic change or data interaction state in real time.
The data is stored in a distributed manner according to the cluster distribution, so that the Availability and partition tolerance of the distributed system are ensured for the system to operate normally, but in the practical requirement of the embodiment, if the user behavior semantics are perceived in real time and the user resources are matched in real time, the call of the data also ensures the real-time Consistency of the semantics association tree, the corpus or the resource semantics, namely, the CAP theory-Consistency, availability and partition fault tolerance (Partition tolerance) are required to be satisfied, so that the data is stored in a distributed manner based on nodes.
The clone tree is established by taking cluster resource association or resource semantic tree as a reference during data interaction in order to ensure data monitoring instantaneity, data feature fusion feasibility and safety and cluster distribution modification feasibility and safety during service processing, especially during data real-time perception, data feature fusion or data cluster distribution modification. The clone tree comprises node coordinates, node semantic codes and weights, and the dynamic change or the service condition of the data can be perceived in real time according to the node state information of the clone tree.
In practice, the behavior of a user is perceived, and as a great acceptable dissimilarity exists between the behavior semantics of a user or a plurality of users in the same family and the history semantics, the related resource semantics can change, for example, in the system operation, the current time domain has the behavior of A, B, C three login users at the same time, and when the behavior is perceived, if the user A originally belongs to the medical field, the user B belongs to the mechanical field and the user C belongs to the computer field; in the behavior monitoring, an A user generates acceptable abnormal behaviors in the same time stamp, if the A user has the requirement of acquiring a numerical control lathe or has the requirement of a certain computer technology, the requirement of the A user is that the actual problem of the A user is solved by using domain association resources of the B user or the C user, and in the domain of the B user or the C user, a scheme or a supplier wish which is spontaneously applied to the domain of the A user is difficult to generate, because resource data are stored in a distributed mode according to a distribution tree, the nodes cannot form association due to the lack of semantic analysis, at the moment, the system needs to feed back the behavior of the A user to other users in real time according to the perception condition of the A user, and node association or node modification is performed, so that the B user or the C user can timely perceive the change resources, namely the perception link is multi-end asynchronous perception, synchronous integrated release is not guaranteed due to the delay of the distribution network, the data are stored in a distributed mode, and the data are perceived in a clone tree mode, and the state of data dynamic change and/or data interaction state is perceived through an induction link based on the clone tree.
If the behavior of the user A causes the semantic change of the associated resources, namely, the nodes in the clone tree need to be changed, for example, medical application or medical resource associated nodes are added in the clone tree of the resource node of a certain computing technology, when the node change can be detected through an induction link based on the clone tree, the clone tree is reconstructed or data fusion is carried out to obtain a fusion tree, the front end can be directly modified through analysis of the newly added node codes, and the user can be directly fed back without a sub-network.
The clone tree comprises a resource semantic tree and/or a resource characteristic tree, the clone tree is a real tree reflecting real service requests in the system, and the fusion tree is a prediction tree obtained by reconstruction or information fusion based on the clone tree. The node state information of the clone tree can be used for sensing the fusion requirement in the real service request, and then the clone tree is reconstructed or data fused according to the sensed fusion requirement.
In the specific implementation process, after a service request is generated according to the collected external corpus, a clone tree is generated based on resource data nodes, the resource request related in the service request is transmitted in the form of tree nodes, the nodes are associated with the clone tree, the processing state of the service request is monitored according to the state information change of the nodes, corpus analysis or behavior analysis is carried out in the processing process, and if new resource semantics or resource characteristics are generated after analysis, the dynamic change of the data is perceived.
In this embodiment, the method further includes:
recording user behavior data in a layered mode according to behavior logic by using a user behavior growth tree, and calculating behavior complexity according to the user behavior data, wherein the user behavior data comprises one or more of access resource sample size, sample dispersion and interaction sample characteristics;
and proportional adjustment is carried out on the resources released to the user according to the behavior complexity, more matched characteristic resources based on the fusion tree are released when the behavior complexity is larger, and more dissimilar characteristic resources based on the fusion tree are released when the behavior complexity is smaller.
In practice, the large behavior complexity indicates that the user behavior has strong randomness and weak purpose, and may be caused by insufficient analysis depth, wider matching range, unclear user intention and the like of user input, for this, according to the higher weight resource characteristics associated with the user behavior in the fusion tree, more similar resources are released to reduce the access resource dispersion, the whole process is strong in real time, the monitoring resource characteristic indexes are all access resource weight curves, specifically, the characteristic indexes record the characteristic resource weight conditions by taking tens of seconds or minutes as one time stamp, after a few time stamps, corresponding scale resources are released in layers according to the weight curves, if the user input is acquired as a microscope, various types of microscopes are randomly displayed in a page, and supposing that one page displays 10 sample data, according to the type characteristics, various types of microscopes are needed to be balanced according to the proportion of the previous page, in the time stamp of statistics, 30 sample data are displayed in total, and the polarization factors are displayed in a total, and if the three or more than one page is displayed, and if the three sample data are displayed in a detail, the polarization factors are displayed by clicking the detail, and the user has a small scale factor is recorded according to the detailed characteristics of the previous time stamp, and the like.
In this embodiment, the method further includes:
searching similar users according to the identity characteristics of the users;
and generating a user demand semantic vector according to the behavior data of the similar users, releasing the matched characteristic resources and the dissimilar characteristic resources based on the user demand semantic vector, and setting induction buried points.
In the specific implementation process, the IP address of the user is matched with the preset IP address in the matching library, and the identity characteristic of the user is extracted according to the matching result and the attribution area of the IP address of the user.
The user ID can be allocated according to the identity characteristics, and a user database is established to manage and control the user. The management and control of the user database and the internal data processing are isolated from each other.
In this embodiment, the method further includes:
if the behavior complexity is monitored to be increased, an alarm signal is sent;
and if the behavior complexity is monitored to be reduced, adding the behavior semantic maximum frequent item set in the user behavior data into the resource data corresponding to the user.
In this embodiment, the clone tree-based inductive link is established by the following steps:
performing feature analysis on the scientific and technological resource feature library, and taking a feature analysis result as a first party of the induction link;
and carrying out real-time behavior analysis on the user behavior, and taking the real-time behavior analysis result as the second party of the induction link.
In this embodiment, the performing feature analysis on the scientific and technological resource feature library, and taking the feature analysis result as the first party of the induction link specifically includes:
performing feature analysis on the scientific and technological resource feature library to obtain a resource semantic tree classified by cluster feature types;
carrying out association rule analysis on the resource semantic tree to obtain a frequent item set;
and constructing a class resource decision classification tree according to the frequent item set, wherein the class resource decision classification tree and corresponding resource data are used as the first party of the induction link.
In this embodiment, the performing real-time behavior analysis on the user behavior, and taking the real-time behavior analysis result as the second party of the inductive link specifically includes:
setting a characteristic index for monitoring;
and extracting the strong correlation characteristic resources according to the characteristic indexes, releasing the strong correlation characteristic resources to users, and setting induction feedback buried points at the strong correlation characteristic resources as the second party of the induction link.
In this embodiment, the feature index includes one or more of a very frequent item set proportion, a resource feature aggregation degree, and a user resource semantic entropy value.
In a specific implementation process, the reconstruction or data fusion of the clone tree to obtain a fusion tree specifically includes:
and reconstructing the clone tree or fusing the data according to the association rule and/or the similarity calculation result to obtain a fusion tree.
The similarity calculation may employ a cross entropy value. And performing cross entropy calculation before reconstruction or information fusion, and performing decision fusion according to the cross entropy. The data server records the change of the characteristics of the master node and the slave node in the clone tree in real time so as to dynamically adjust the fusion decision.
In this embodiment, the resource data includes user behavior semantics and/or technological association corpus.
In the implementation process, the technological association corpus is obtained by directionally crawling corpus to an external data network of the system according to semantic analysis demand characteristics.
In this embodiment, the crawling corpus to the external data network of the system according to the semantic analysis demand features specifically includes:
generating a corpus demand according to feedback of semantic analysis, and transmitting the corpus demand to a corpus matcher in the form of a corpus demand tree or a corpus demand vector;
the corpus matcher matches or searches the corpus according to the characteristic relation or weight distribution presented in the corpus demand tree or the corpus demand vector.
In the specific implementation process, the crawled science and technology association corpus is classified in a hierarchical mode according to corpus types, demand types and demand feature types, and then syntactic analysis and word sense association analysis are conducted in a directed mode according to requirements.
In this embodiment, the service request includes a cluster index of service-related resources, and the method further includes:
and calculating the resource cluster complexity according to the cluster index, and distributing calculation resources and/or network resources for the service request according to the resource cluster complexity.
The task difficulty can be quantitatively evaluated according to the complexity of the associated resource cluster, and the computing processing resources can be distributed according to the current computing resource occupation condition and the data occupation condition.
In the specific implementation process, the resource cluster complexity is calculated according to the cluster index, specifically, the following formula is adopted for calculation:
wherein L is the complexity of the resource cluster, N is the number of clusters, N is the sample capacity, e is the sample entropy value calculated by taking the cluster characteristics as indexes, k is the maximum length of the frequent item sets of the association tree, C is the number of the extremely-large frequent item sets, and C is the total number of the item sets.
Preferably, the method may further comprise:
and allocating the computing resources and/or the network resources according to the current computing resources and/or the network resources occupation situation.
Preferably, the method may further comprise:
and distributing computing resources and/or network resources according to the corresponding task attributes in the service request.
In this embodiment, the service request further includes a public key and a private key; the generation of the clone tree based on the nodes according to the service request specifically comprises the following steps:
analyzing the public key and the random request code to confirm whether the identity and the service request type of the input external corpus meet the authority requirements;
if the authority requirements are met, generating a matched private key according to the private key, analyzing the matched private key to obtain corresponding request data, and according to the private key
The request data generates a clone tree based on the nodes.
In a specific implementation process, the step of analyzing the matching private key to obtain the corresponding request data may specifically include: and analyzing the received matching private key to obtain a corresponding data index, and obtaining corresponding request data according to the data index.
In this embodiment, the storing the resource data in a distributed manner in the form of nodes specifically includes:
performing dependency syntactic analysis on the resource data to obtain labeled words;
and carrying out belonging field semantic association analysis and/or belonging resource semantic association analysis on the resource data according to the labeling words to construct a semantic tree, and carrying out distributed storage on the resource data based on nodes of the semantic tree.
And processing data distribution storage, and carrying out distributed deployment through a training model.
For a segment of science and technology resource text corpus, for example: in a narrow sense, aerospace-like professions include subject matter professions such as aircraft design and engineering, aircraft power engineering, aircraft manufacturing engineering, aircraft environment and life support engineering, probe guidance and control technology, and the like. However, both airplanes and aerospace craft are crystals of comprehensive science and technology, and relate to aspects of materials, electronic communication equipment, instruments, remote control and telemetry, navigation, remote sensing and the like. Therefore, in a broad sense, material science and engineering, electronic information engineering, automation, computers, transportation, quality and reliability engineering and the like are all essential disciplines of aerospace technology. With the rapid development of aerospace industry, emerging professions such as aerospace transportation and control, remote sensing science and technology and the like are promoted in recent years "
The corpus is derived from aerospace resources and belongs to a paragraph in a report or paper under the resources, the characteristic of distributed storage shows that on one hand, the corpus is related to aerospace, on the other hand, three strategies are simultaneously used during corpus acquisition and analysis based on corpus attributes on the assumption that the paragraph belongs to a part of a feasibility report, the strategy A is semantic association analysis in the aerospace field, the strategy B is semantic association analysis of the resources, such as association analysis of aerospace and 'aircraft design and engineering, aircraft power engineering, aircraft manufacturing engineering, aircraft environment and life guarantee engineering, detection guidance and control technology', and the strategy C is acquisition of feasibility report labeling word materials.
For the strategies A and B, the set response key is a cluster resource feature, if the cluster resource feature is an aerospace tag code, the stored association corpus joint analysis can be called during analysis, meanwhile, the original corpus is optimized according to analysis feedback and the resource nodes are modified and optimized, the strategy C is classified as a text analysis strategy, words which indicate sentence meaning are obtained according to dependency syntax analysis, such as generalized, generalized speaking, subject, emerging specialized, and the like, and the words which indicate sentence meaning are marked, the annotation word corpus obtained through analysis is called by the strategies A and B to obtain target semantic words, then a semantic tree is constructed, and if sentence meaning feature words are utilized, the association tree of the main feature and the sentence meaning words can be constructed.
In this embodiment, the releasing the resource data to the user according to the fusion tree specifically includes:
and releasing the resource data to the user according to the fusion tree by a decision algorithm, wherein the decision algorithm is obtained by adopting one or more of a resource data clustering analysis method, a user resource matching analysis method, a semantic association rule analysis method and a user real-time behavior analysis method.
In a specific implementation process, the embodiment can perform concurrent computation for a plurality of users, and performs data communication and protocol frame cluster management, security protocol management and heterogeneous data fusion decision, and parallel processing interactive protocol management. The concurrency is mainly embodied in clone tree management, real-time data state management, real-time dynamic resource data release and the like.
In this embodiment, the method further includes: learning user behaviors and monitoring abnormal behaviors;
if abnormal behavior of the writing operation is monitored, temporarily blocking current behavior of the user through a permission prompt and/or a prompt box of a user grade, cloning current demand data of the user, adding a section of pseudo-data random code, and covering a page;
when the pseudo-data random code is monitored, a warning is sent to a true user, the true user is treated, if the warning is cancelled, the treatment feedback and the abnormal behavior mode are combined to be used as training data, and the internal attack mode is trained according to the training data, so that the internal abnormal behavior discrimination capability is improved;
if abnormal behavior of the non-writing operation is monitored, calculating the complexity of the behavior and dynamically adjusting the interactive data so that the complexity becomes smaller or tends to be stable, and simultaneously recording the behavior semantics and adding the behavior semantics into the stored resource data.
In the implementation process, the behavior of the visitor user generating the alarm signal can be learned, the alarm behavior processing mode or result is monitored, and the behavior mode of the visitor user is trained according to the complexity index, the behavior structure and the alarm behavior processing feedback, so that the behavior identification capability of the visitor user is improved.
In the implementation process, the behavior of the login user or the manager user can be learned, portrait construction is carried out on the login user or the manager user according to the historical behavior and the similar user behaviors, and accordingly buried points are set for the user dissimilarity behavior or the related resources so as to monitor abnormal behaviors.
It should be understood that the foregoing examples of the present invention are merely illustrative of the present invention and are not intended to limit the present invention to the specific embodiments thereof. Any modification, equivalent replacement, improvement, etc. that comes within the spirit and principle of the claims of the present invention should be included in the protection scope of the claims of the present invention.
Claims (9)
1. The technological resource perception fusion decision method is characterized by comprising the following steps of:
the resource data is stored in a distributed mode in a node mode;
collecting external corpus input by a user, and generating a service request according to the external corpus;
generating a clone tree based on the nodes according to the service request;
sensing data dynamic change and/or data interaction state through an induction link based on a clone tree;
reconstructing or fusing the clone tree to obtain a fusion tree aiming at the data state change and/or the data interaction state;
releasing the resource data to the user according to the fusion tree;
further comprises:
recording user behavior data in a layered mode according to behavior logic by using a user behavior growth tree, and calculating behavior complexity according to the user behavior data, wherein the user behavior data comprises one or more of access resource sample size, sample dispersion and interaction sample characteristics;
and proportional adjustment is carried out on the resources released to the user according to the behavior complexity, more matched characteristic resources based on the fusion tree are released when the behavior complexity is larger, and more dissimilar characteristic resources based on the fusion tree are released when the behavior complexity is smaller.
2. The technology resource aware fusion decision method of claim 1, further comprising:
searching similar users according to the identity characteristics of the users;
and generating a user demand semantic vector according to the behavior data of the similar users, releasing the matched characteristic resources and the dissimilar characteristic resources based on the user demand semantic vector, and setting induction buried points.
3. The scientific and technological resource perception fusion decision method according to claim 1, wherein the clone tree-based inductive link is established by the following steps:
performing feature analysis on the scientific and technological resource feature library, and taking a feature analysis result as a first party of the induction link;
and carrying out real-time behavior analysis on the user behavior, and taking the real-time behavior analysis result as the second party of the induction link.
4. The method for determining a fusion of technical resources according to claim 3, wherein the feature analysis is performed on the technical resource feature library, and the feature analysis result is taken as the first party of the induction link, and specifically comprises:
performing feature analysis on the scientific and technological resource feature library to obtain a resource semantic tree classified by cluster feature types;
carrying out association rule analysis on the resource semantic tree to obtain a frequent item set;
and constructing a class resource decision classification tree according to the frequent item set, wherein the class resource decision classification tree and corresponding resource data are used as the first party of the induction link.
5. The method for making a fusion decision by sensing technological resources according to claim 3, wherein the real-time behavior analysis is performed on the user behavior, and the real-time behavior analysis result is taken as the second party of the sensing link, specifically comprising:
setting a characteristic index for monitoring;
and extracting the strong correlation characteristic resources according to the characteristic indexes, releasing the strong correlation characteristic resources to users, and setting induction feedback buried points at the strong correlation characteristic resources as the second party of the induction link.
6. The method for determining a fusion of technical and scientific resource awareness according to claim 1, wherein the resource data comprises user behavior semantics and/or technical association corpus, and the technical association corpus is obtained by directionally crawling corpus to an external data network of a system according to semantic analysis demand characteristics.
7. The method for determining a fusion of technical resources according to claim 6, wherein the step of directionally crawling corpus to an external data network of the system according to semantic analysis demand features specifically comprises:
generating a corpus demand according to feedback of semantic analysis, and transmitting the corpus demand to a corpus matcher in the form of a corpus demand tree or a corpus demand vector;
the corpus matcher matches or searches the corpus according to the characteristic relation or weight distribution presented in the corpus demand tree or the corpus demand vector.
8. The scientific and technological resource perception fusion decision method according to claim 1, wherein the step of performing distributed storage on resource data in the form of nodes specifically comprises the following steps:
performing dependency syntactic analysis on the resource data to obtain labeled words;
and carrying out belonging field semantic association analysis and/or belonging resource semantic association analysis on the resource data according to the labeling words to construct a semantic tree, and carrying out distributed storage on the resource data based on nodes of the semantic tree.
9. The method for determining a fusion of technical resources according to any one of claims 1 to 8, further comprising: learning user behaviors and monitoring abnormal behaviors;
if abnormal behavior of the writing operation is monitored, temporarily blocking current behavior of the user through a permission prompt and/or a prompt box of a user grade, cloning current demand data of the user, adding a section of pseudo-data random code, and covering a page;
when the pseudo-data random code is monitored, a warning is sent to a true user, the true user is treated, if the warning is cancelled, the treatment feedback and the abnormal behavior mode are combined to be used as training data, and the internal attack mode is trained according to the training data, so that the internal abnormal behavior discrimination capability is improved;
if abnormal behavior of the non-writing operation is monitored, calculating the complexity of the behavior and dynamically adjusting the interactive data so that the complexity becomes smaller or tends to be stable, and simultaneously recording the behavior semantics and adding the behavior semantics into the stored resource data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910304666.9A CN110059315B (en) | 2019-04-16 | 2019-04-16 | Scientific and technological resource perception fusion decision method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910304666.9A CN110059315B (en) | 2019-04-16 | 2019-04-16 | Scientific and technological resource perception fusion decision method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110059315A CN110059315A (en) | 2019-07-26 |
| CN110059315B true CN110059315B (en) | 2023-07-07 |
Family
ID=67317682
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910304666.9A Active CN110059315B (en) | 2019-04-16 | 2019-04-16 | Scientific and technological resource perception fusion decision method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110059315B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114741197B (en) * | 2022-04-18 | 2024-06-28 | 中国电信股份有限公司 | Resource release method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102812486A (en) * | 2010-03-25 | 2012-12-05 | 诺基亚公司 | Method and apparatus for providing personalized information resource recommendation based on group behaviors |
| CN107704575A (en) * | 2017-09-30 | 2018-02-16 | 郑州轻工业学院 | User behavior analysis method and user behavior analysis device based on data mining |
| CN109190051A (en) * | 2018-11-29 | 2019-01-11 | 广东省科技基础条件平台中心 | A kind of user behavior analysis method and the resource recommendation method based on the analysis method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090199090A1 (en) * | 2007-11-23 | 2009-08-06 | Timothy Poston | Method and system for digital file flow management |
-
2019
- 2019-04-16 CN CN201910304666.9A patent/CN110059315B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102812486A (en) * | 2010-03-25 | 2012-12-05 | 诺基亚公司 | Method and apparatus for providing personalized information resource recommendation based on group behaviors |
| CN107704575A (en) * | 2017-09-30 | 2018-02-16 | 郑州轻工业学院 | User behavior analysis method and user behavior analysis device based on data mining |
| CN109190051A (en) * | 2018-11-29 | 2019-01-11 | 广东省科技基础条件平台中心 | A kind of user behavior analysis method and the resource recommendation method based on the analysis method |
Non-Patent Citations (1)
| Title |
|---|
| 面向知识服务的科技资源协同创新服务机制;饶元 等;《计算机科学》;20180731;第45卷(第7期);第286-292页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110059315A (en) | 2019-07-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR20210038860A (en) | Intent recommendation method, apparatus, device and storage medium | |
| Liang et al. | Search engine for the internet of things: Lessons from web search, vision, and opportunities | |
| US9230030B2 (en) | Method and system for mining websites | |
| CN112165462A (en) | Attack prediction method and device based on portrait, electronic equipment and storage medium | |
| CN101610190A (en) | Data network and system | |
| US9123006B2 (en) | Techniques for parallel business intelligence evaluation and management | |
| CN110855648B (en) | Early warning control method and device for network attack | |
| CN113094385A (en) | Data sharing fusion platform and method based on software definition open toolset | |
| CN110955801B (en) | Knowledge graph analysis method and system for cognos report indexes | |
| CN110851288A (en) | Message processing method and device | |
| KR101637504B1 (en) | System and method for processing informal data | |
| Liang et al. | Financial big data analysis and early warning platform: a case study | |
| CN106961441B (en) | User dynamic access control method for Hadoop cloud platform | |
| Wang et al. | A Dynamic multi-sensor data fusion approach based on evidence theory and WOWA operator | |
| EP4120096A1 (en) | Method and device for data retrieval, electronic device, and storage medium | |
| CN110059315B (en) | Scientific and technological resource perception fusion decision method | |
| Sun et al. | Expert system for automatic microservices identification using API similarity graph | |
| Elayni et al. | Using MongoDB databases for training and combining intrusion detection datasets | |
| US20130117245A1 (en) | Method and system for identification of asset records in a version managed datastore | |
| KR101546696B1 (en) | Information fusion method and system | |
| CN111797422A (en) | Data privacy protection query method and device, storage medium and electronic equipment | |
| Leclercq et al. | Snfreezer: a platform for harvesting and storing tweets in a big data context | |
| CN115510116A (en) | Data directory construction method, device, medium and equipment | |
| Moura et al. | Integration of linked data sources for gazetteer expansion | |
| Li | Construction of intelligent service system of university library based on internet of things in artificial intelligence environment |
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 |