CN115470905B - Big data analysis processing method and system - Google Patents
Big data analysis processing method and system Download PDFInfo
- Publication number
- CN115470905B CN115470905B CN202211182210.8A CN202211182210A CN115470905B CN 115470905 B CN115470905 B CN 115470905B CN 202211182210 A CN202211182210 A CN 202211182210A CN 115470905 B CN115470905 B CN 115470905B
- Authority
- CN
- China
- Prior art keywords
- big data
- knowledge
- preference
- business interaction
- analysis
- 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
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- Game Theory and Decision Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
According to the big data analysis processing method and system, the a priori type business interaction big data is used as a reference for service preference analysis, expert knowledge mining is conducted on the a priori type business interaction big data and the business interaction big data to be analyzed respectively, then feature association scores between the a priori type business interaction big data and the business interaction big data to be analyzed are determined through preference linkage analysis operation, service preference topic analysis of the business interaction big data to be analyzed is efficiently achieved, second digital user service preferences are obtained, analysis processing of the digital user service preferences is conducted through only one round of analysis processing by means of the reference of the service preference analysis, on one hand, preference positioning analysis can be conducted on the second digital user service preferences in time, on the other hand, the type of the first digital user service preferences can be accurately judged through the second digital user service preferences, and accordingly analysis processing quality of different digital user service preferences is improved.
Description
Technical Field
The invention relates to the technical field of big data processing, in particular to a big data analysis processing method and a big data analysis processing system.
Background
With the continuous proliferation of competitive pressures of various digital industries, proactive user survival schemes and demand prediction means are key to improving the competitiveness of the digital industries. Currently, big data mining analysis has been widely used in various digital industries as one of the main means of user analysis. However, the increasing number of users and number of services presents a number of challenges to user analysis, such as in terms of user preference analysis and mining, conventional big data mining techniques typically have the following problems: the timeliness is poor and the precision is difficult to meet the requirements.
Disclosure of Invention
In order to improve the technical problems in the related art, the invention provides a big data analysis processing method and a big data analysis processing system.
In a first aspect, an embodiment of the present invention provides a big data analysis processing method, where the method is applied to a big data analysis processing system, and the method at least includes: acquiring business interaction big data to be analyzed covering the first digital user service preference and prior business interaction big data covering the second digital user service preference; loading the prior-type business interaction big data and the business interaction big data to be analyzed into a service preference analysis strategy, and respectively carrying out expert knowledge mining on the prior-type business interaction big data and the business interaction big data to be analyzed by an expert knowledge mining module in the service preference analysis strategy to obtain first preference analysis decision knowledge and second preference analysis decision knowledge; and in a preference linkage analysis module of the service preference analysis strategy, carrying out preference linkage analysis operation on the second preference analysis decision knowledge through the first preference analysis decision knowledge, and judging whether the first digital user service preference covered by the business interaction big data to be analyzed is the second digital user service preference.
When the method is applied, the prior business interaction big data is used as a reference for analyzing the service preference, after expert knowledge mining is carried out on the prior business interaction big data and the business interaction big data to be analyzed respectively, the feature association scores between the prior business interaction big data and the business interaction big data to be analyzed are determined through preference linkage analysis operation, the service preference theme analysis of the business interaction big data to be analyzed is effectively realized, the second digital user service preference is obtained, the analysis processing of the digital user service preference is carried out by only executing one round of analysis processing through the reference of the service preference analysis, on one hand, the preference positioning analysis can be carried out on the second digital user service preference in time, on the other hand, the type of the first digital user service preference can be accurately judged by means of the second digital user service preference, and therefore the analysis processing quality of different digital user service preferences is improved.
In a possible embodiment, the performing, in the preference linkage analysis module in the service preference analysis policy, a preference linkage analysis operation on the second preference analysis decision knowledge by the first preference analysis decision knowledge includes:
Determining a sliding filter operator according to the knowledge scale of the first preference analysis decision knowledge;
performing sliding filtering operation on the second preference analysis decision knowledge through the sliding filtering operator to obtain a target knowledge description variable;
and determining whether the first digital user service preference is the second digital user service preference according to the service interaction data set corresponding to the target knowledge description variable in the service interaction big data to be analyzed.
In a possible embodiment, the determining the sliding filter operator according to the knowledge scale of the first preference analysis decision knowledge comprises: and carrying out knowledge redundancy elimination operation on the first preference analysis decision knowledge so as to adjust the knowledge scale of the first preference analysis decision knowledge, and taking the first preference analysis decision knowledge with the knowledge redundancy elimination operation as the sliding filter operator.
In a possible embodiment, the performing, by the sliding filter operator, a sliding filter operation on the second preference analysis decision knowledge, to obtain a target knowledge description variable includes:
performing sliding filtering operation on the second preference analysis decision knowledge through the sliding filtering operator to obtain a linkage analysis description variable;
And performing sliding filtering operation on the linkage analysis description variable through the sliding filtering operator to obtain a target knowledge description variable.
In a possible embodiment, the expert knowledge mining module is a weight sharing model, and the performing expert knowledge mining on the prior-type business interaction big data and the business interaction big data to be analyzed by the expert knowledge mining module to obtain first preference analysis decision knowledge and second preference analysis decision knowledge includes:
expert knowledge mining is carried out on the prior-type business interaction big data through a first knowledge mining unit of the weight sharing model, so that first preference analysis decision knowledge is obtained;
and carrying out expert knowledge mining on the business interaction big data to be analyzed through a second knowledge mining unit of the weight sharing model to obtain the second preference analysis decision knowledge.
In a possible embodiment, the first knowledge mining unit and the second knowledge mining unit respectively include a session habit mining node and a user interest mining node, and performing expert knowledge mining on the a priori service interaction big data and the service interaction big data to be analyzed through an expert knowledge mining module includes:
The prior-type business interaction big data and the business interaction big data to be analyzed are respectively loaded to the conversation habit mining node, and conversation habit visual vectors in the prior-type business interaction big data and the business interaction big data to be analyzed are obtained;
and loading session habit visual vectors in the prior-type business interaction big data and the business interaction big data to be analyzed to the user interest mining node, performing global sliding filtering operation on the session habit visual vectors, and determining session habit updating characteristics of the session habit visual vectors to obtain the first preference analysis decision knowledge and the second preference analysis decision knowledge.
In one possible embodiment, the session habit visual vector in the prior-type business interaction big data and the business interaction big data to be analyzed comprises a session habit visual vector of a GUI interaction event, and the session habit visual vector of the GUI interaction event comprises an activity weight relation network of business interaction links of the GUI interaction event.
In one possible embodiment, the tuning method of the service preference analysis policy includes:
acquiring a first business interaction big data example and a plurality of second business interaction big data examples, wherein at least one of the second business interaction big data example and the first business interaction big data example cover the same class of service preference topics;
Inputting the first business interaction big data example and the plurality of second business interaction big data examples into a general service preference analysis strategy, wherein the general service preference analysis strategy comprises a general expert knowledge mining module and a general preference linkage analysis module, and expert knowledge mining is respectively carried out on the first business interaction big data example and the plurality of second business interaction big data examples through the general expert knowledge mining module to obtain a first preference analysis decision knowledge example and a second preference analysis decision knowledge example;
in the general preference linkage analysis module, preference linkage analysis operation is carried out on the second preference analysis decision knowledge example through the first preference analysis decision knowledge example to obtain a knowledge description variable example, and whether the service preference theme is covered in the business interaction big data to be analyzed is judged through the knowledge description variable example;
and improving the strategy configuration parameters of the general service preference analysis strategy according to the knowledge description variable examples and the preference theme judgment result to obtain the service preference analysis strategy.
In one possible embodiment, the improving the policy configuration parameters of the generic expert knowledge mining module according to the knowledge descriptive variable example and the preference topic determination result includes:
Acquiring word vector commonality values among the plurality of second business interaction big data examples through the knowledge description variable examples, and determining characteristic detail quality cost;
comparing and analyzing the preference theme judging result and the first business interaction big data example to determine probability distribution accuracy cost;
and carrying out iterative optimization on strategy configuration parameters of the general service preference analysis strategy according to the characteristic detail quality cost and the probability distribution accuracy cost.
In a second aspect, the present invention also provides a big data analysis processing system, including a processor and a memory; the processor is in communication with the memory, and the processor is configured to read the computer program from the memory and execute the computer program to implement the method described above.
In a third aspect, the present invention also provides a readable storage medium having stored thereon a program which, when executed by a processor, implements the method described above.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
Fig. 1 is a flow chart of a big data analysis processing method according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a communication architecture of an application environment of a big data analysis processing method according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the accompanying claims.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present invention and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order.
The method embodiments provided by the embodiments of the present invention may be performed in a big data analysis processing system, a computer device, or a similar computing device. Taking the example of running on a big data analysis processing system, big data analysis processing system 10 may include one or more processors 102 (processor 102 may include, but is not limited to, a microprocessor MCU, a programmable logic device FPGA, etc. processing means) and a memory 104 for storing data, and optionally, a transmission means 106 for communication functions. It will be appreciated by those of ordinary skill in the art that the above-described structure is merely illustrative and is not intended to limit the structure of the big data analysis processing system. For example, big data analysis processing system 10 may also include more or fewer components than those shown above, or have a different configuration than those shown above.
The memory 104 may be used to store a computer program, for example, a software program of application software and a module, such as a computer program corresponding to a big data analysis processing method in an embodiment of the present invention, and the processor 102 executes the computer program stored in the memory 104 to perform various functional applications and data processing, that is, implement the above-mentioned method. Memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, memory 104 may further include memory located remotely from processor 102, which may be connected to big data analysis processing system 10 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The transmission means 106 is arranged to receive or transmit data via a network. Specific examples of the network described above may include a wireless network provided by a communication provider of big data analysis processing system 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, simply referred to as NIC) that can connect to other network devices through a base station to communicate with the internet. In one example, the transmission device 106 may be a Radio Frequency (RF) module, which is used to communicate with the internet wirelessly.
Based on this, referring to fig. 1, fig. 1 is a flow chart of a big data analysis processing method according to an embodiment of the present invention, where the method is applied to a big data analysis processing system, and further may include the following technical solutions described below.
Step 110, collecting business interaction big data to be analyzed covering the first digital user service preference and prior business interaction big data covering the second digital user service preference.
In the embodiment of the invention, at least one first digital user service preference category is covered in the business interaction big data to be analyzed, and before the big data is pushed, the first digital user service preference is needed to be analyzed and positioned from the business interaction big data to be analyzed, so as to guide the big data to be pushed, based on the first digital user service preference category, the prior business interaction big data covering the second digital user service preference is collected, and the prior business interaction big data can be used as a preference analysis reference of the business interaction big data to be analyzed.
Further, the first digital user service preference may be understood as a digital user service preference to be analyzed and the second digital user service preference may be understood as a target digital user service preference. In addition, the digital user service preferences may reflect interest interests, preference information, demand information, etc. of the user during the digital service interaction. Taking electronic commerce as an example, the digital user service preference may be a shopping preference of the user, taking government business as an example, the digital user service preference may be a business handling guidance preference of the user, taking meta-space service as an example, the digital user service preference may be a somatosensory interaction preference of the user, and the like. Furthermore, the business interaction big data to be analyzed and the prior business interaction big data both comprise business service activity records or interaction records of the digital users, and the interaction big data can be used as raw materials for big data analysis and mining, so that valuable information in the interaction big data can be mined and identified.
And 130, loading the prior-type business interaction big data and the business interaction big data to be analyzed into a service preference analysis strategy, and respectively carrying out expert knowledge mining on the prior-type business interaction big data and the business interaction big data to be analyzed through an expert knowledge mining module in the service preference analysis strategy to obtain first preference analysis decision knowledge and second preference analysis decision knowledge.
The embodiment of the invention can pre-configure the service preference analysis strategy, wherein the service preference analysis strategy comprises an expert knowledge mining module and a preference linkage analysis module. The expert knowledge mining module is used for mining interest description vectors or preference description vectors in the prior-type business interaction big data and the business interaction big data to be analyzed respectively, and is beneficial to further processing of the preference linkage analysis module. Illustratively, expert knowledge mining is carried out on the prior-type business interaction big data through an expert knowledge mining module to obtain first preference analysis decision knowledge; and carrying out expert knowledge mining on the business interaction big data to be analyzed through an expert knowledge mining module to obtain second preference analysis decision knowledge.
For example, the service preference analysis policy may be a neural network model built based on an artificial intelligence technology (such as an expert system), such as a convolutional neural network model, a deep learning neural network model, a cyclic neural network model, a decision tree model, a multi-layer perceptron model, a naive bayes model, and the like, which are not limited herein. Further, the expert knowledge mining module and the preference linkage analysis module can be respectively understood as a feature mining module and a feature association analysis module, wherein the expert knowledge mining module is used for performing feature mining, and the preference linkage analysis module is used for analyzing association among different preference analysis decision knowledge obtained by mining.
Adaptively, expert knowledge mining can also be understood as feature mining or feature extraction, and the determined preference analysis decision knowledge is the user preference feature of corresponding business interaction big data, and can be recorded in the forms of feature vectors, description arrays and the like.
And 150, performing preference linkage analysis operation on the second preference analysis decision knowledge through the first preference analysis decision knowledge in a preference linkage analysis module of the service preference analysis strategy, and identifying whether the first digital user service preference covered by the business interaction big data to be analyzed is the second digital user service preference.
By combining the related content, the preference linkage analysis module can acquire the association score of one preference analysis decision knowledge relative to the other preference analysis decision knowledge, the embodiment of the invention inputs the first preference analysis decision knowledge and the second preference analysis decision knowledge into the preference linkage analysis module, carries out preference linkage analysis operation on the second preference analysis decision knowledge based on the first preference analysis decision knowledge, compares the first preference analysis decision knowledge with the second preference analysis decision knowledge, acquires the association score of each knowledge element (such as knowledge member or knowledge vector in the second preference analysis decision knowledge) in the second preference analysis decision knowledge and the first preference analysis decision knowledge, identifies whether the first digital user service preference in the business interaction big data to be analyzed is the second digital user service preference, and selects the digital user service preference corresponding to the second preference analysis decision knowledge with high association score to acquire the second digital user service preference in the business interaction big data to be analyzed.
By using the technical scheme, the prior business interaction big data is used as a reference for analyzing the service preference, the prior business interaction big data and the business interaction big data to be analyzed are subjected to expert knowledge mining respectively, and then the feature association scores between the prior business interaction big data and the business interaction big data to be analyzed are determined through preference linkage analysis operation, so that the service preference topic analysis of the business interaction big data to be analyzed is efficiently realized, the second digital user service preference is obtained, the analysis processing of the digital user service preference is realized by only executing one round of analysis processing through the reference of one service preference analysis, on one hand, the preference positioning analysis can be timely carried out on the second digital user service preference, and on the other hand, the type of the first digital user service preference can be accurately judged by means of the second digital user service preference, and the analysis processing quality of different digital user service preferences is improved.
The prior-type business interaction big data and the business interaction big data to be analyzed in the embodiment of the invention can be interaction records of GUI interaction events or streaming text interaction events, for example, can be interaction records of GUI interaction events.
The expert knowledge mining module of the embodiment of the invention is used for respectively carrying out consistent expert knowledge mining processing on the prior-type business interaction big data and the business interaction big data to be analyzed, and in some possible examples, the expert knowledge mining module is any one of the AI models, and can respectively carry out processing on the prior-type business interaction big data and the business interaction big data to be analyzed through the same model twice.
In other possible examples, the expert knowledge mining module is a weight sharing model (such as a gemini model), and performs expert knowledge mining on the prior-type service interaction big data and the service interaction big data to be analyzed through two sub-models of the weight sharing model. Expert knowledge mining is carried out on the prior-type business interaction big data through a first knowledge mining unit of the weight sharing model, and first preference analysis decision knowledge is obtained. Expert knowledge mining is carried out on the business interaction big data to be analyzed through a second knowledge mining unit of the weight sharing model, and second preference analysis decision knowledge is obtained.
For example, the following is one implementation of the expert knowledge mining technical solution according to the embodiment of the present invention, and may include the following.
Step 210, inputting the prior-type business interaction big data into a conversation habit mining node of the first knowledge mining unit, and obtaining conversation habit visual vectors in the prior-type business interaction big data.
In the embodiment of the invention, the weight sharing model comprises two sub-models, a first knowledge mining unit and a second knowledge mining unit, each sub-model comprises a session habit mining node and a user interest mining node, the session habit mining node is a GUI interaction event operation habit analysis sub-model, and the session habit mining node selectively performs feature mining on user session data to acquire content most relevant to GUI interaction event change features.
In some possible examples, the session habit mining node encompasses a depth residual sub-model and a reverse sliding filter sub-model (such as deconvolution) for determining an activity weight relationship net for GUI interactivity events business interactions links.
For example, the activity weight relation network in the embodiment of the present invention refers to a probability distribution map generated by setting a feature value with a probability statistics index of 1 based on a distribution variable of each business interaction link of the GUI interaction event. In view of too much noise in the business interaction big data, the operation cost is too high, and the information covered by the business interaction links is too little, so that the anti-interference performance is not strong. And the correlation degree between the business interaction links is very high, if regression analysis is performed mechanically, the correlation degree between the business interaction links is difficult to mine and analyze. Based on the above, the embodiment of the invention uses the activity weight relation network of the GUI interaction event as the session habit vision vector of the GUI interaction event. The session habit mining node determines a feature relation network set with a constant window size, and the feature relation network set can contain four-dimensional or more descriptive contents.
Step 230, inputting the conversation habit visual vector in the prior-type business interaction big data into a user interest mining node of the second knowledge mining unit, performing global sliding filtering operation on the conversation habit visual vector, extracting conversation habit updating characteristics of the conversation habit visual vector, and obtaining first preference analysis decision knowledge.
The user interest mining node is used for mining user preference change knowledge of the GUI interaction event on the basis of the conversation habit visual vector, and after the user preference change knowledge passes through the conversation habit mining node, first preference analysis decision knowledge of each group of business interaction information in prior business interaction big data is obtained, wherein the first preference analysis decision knowledge is data under a service scene level and comprises knowledge number, time interval, first relation network size and second relation network size. And loading the conversation habit visual vectors into an user interest mining node, adding a global sliding filtering operation (such as multi-dimensional convolution processing) of a time sequence layer to the conversation habit visual vectors, and obtaining conversation habit updating characteristic information of the conversation habit visual vectors in time to obtain first preference analysis decision knowledge.
Under some possible examples, an exemplary introduction to the structure of a user interest mining node is as follows. The method comprises the steps of obtaining a session habit vision vector with knowledge number C of 34, time interval T of 64, first relation network size H of 128 and second relation network size W of 96, performing global sliding filtering operation on the session habit vision vector with knowledge number C of 34 x 64 x 128 x 96, obtaining a session habit vision vector with knowledge number C of 64, time interval T of 64, first relation network size H of 128 and second relation network size W of 96, and performing global sliding filtering operation again on the session habit vision vector with knowledge number C of 64 x 128 x 96. In an actual implementation, the number of global sliding filter operations is not limited.
And 250, inputting the business interaction big data to be analyzed into a conversation habit mining node of the second knowledge mining unit, and obtaining conversation habit visual vectors of GUI interaction events in the business interaction big data to be analyzed.
Step 270: and inputting the conversation habit visual vector in the business interaction big data to be analyzed into a user interest mining node of the second knowledge mining unit, performing global sliding filtering operation on the conversation habit visual vector, extracting conversation habit updating characteristics of the conversation habit visual vector, and obtaining second preference analysis decision knowledge.
Similarly, the second knowledge mining unit also encompasses session habit mining nodes and user interest mining nodes. Because the first knowledge mining unit and the second knowledge mining unit are weight sharing models, the processing thought of the second knowledge mining unit to the interactive big data of the service to be analyzed is consistent with the processing thought of the first knowledge mining unit to the interactive big data of the prior service.
It will be appreciated that, for the above step 210, the execution of step 230 and step 250 is not limited. For example, steps 250 and 270 may be advanced before steps 210 and 230.
For the embodiment of the invention, the weight sharing model is used as an expert knowledge mining module, and the digital user service preference is determined through the activity weight relation network state of the GUI interaction event, so that the characteristics of interest change and user session habit of massive GUI interaction events are maintained as much as possible, and meanwhile, part of noise characteristics are effectively avoided. And then, further expert knowledge mining is carried out on the mined session habit visual vector on a time sequence level through global sliding filtering operation, session habit updating characteristics are obtained, the accuracy of expert knowledge mining can be improved, and an accurate and reliable data base is provided for the subsequent preference linkage analysis operation.
Another embodiment of the big data analysis processing method according to the embodiments of the present invention may further include the following matters under some independent design considerations.
Step 510, collecting business interaction big data to be analyzed covering the first digital user service preference and prior business interaction big data covering the second digital user service preference.
And 530, loading the prior-type business interaction big data and the business interaction big data to be analyzed into a service preference analysis strategy, and respectively carrying out expert knowledge mining on the prior-type business interaction big data and the business interaction big data to be analyzed through an expert knowledge mining module in the service preference analysis strategy to obtain first preference analysis decision knowledge and second preference analysis decision knowledge.
The following is an exemplary introduction to a service preference analysis policy according to an embodiment of the present invention: the service preference analysis strategy comprises a weight sharing model and a preference linkage analysis module. The prior-type business interaction big data and the business interaction big data to be analyzed are respectively loaded into two modules of a weight sharing model of a service preference analysis strategy, expert knowledge mining is carried out on the prior-type business interaction big data and the business interaction big data to be analyzed respectively through the expert knowledge mining method of the expert knowledge mining module, and first preference analysis decision knowledge and second preference analysis decision knowledge are obtained.
Step 550, determining a sliding filter operator based on the knowledge scale of the first preference analysis decision knowledge.
In view of the small scale difference between the first preference analysis decision knowledge and the second preference analysis decision knowledge generated by the same neural network, in order to ensure the accuracy of the preference linkage analysis operation, the knowledge redundancy elimination operation (such as downsampling or pooling) is performed on the first preference analysis decision knowledge to adjust the knowledge scale of the first preference analysis decision knowledge, and the first preference analysis decision knowledge after the knowledge redundancy elimination operation is used as a sliding filter operator (such as a convolution kernel).
Step 570: and performing sliding filtering operation on the second preference analysis decision knowledge through a sliding filtering operator to obtain a target knowledge description variable.
And performing preference linkage analysis operation on the second preference analysis decision knowledge through the first preference analysis decision knowledge, and in some possible examples, performing sliding filtering operation on the second preference analysis decision knowledge by using the first preference analysis decision knowledge as a sliding filtering operator. First, taking first preference analysis decision knowledge with knowledge redundancy elimination operation as a sliding filter operator, and performing global sliding filter operation on second preference analysis decision knowledge to adjust and learn the second preference analysis decision knowledge so as to obtain target knowledge description variables. The global sliding filtering operation in the preference linkage analysis module is different from the global sliding filtering operation in the user interest mining node, and in the global sliding filtering operation in the preference linkage analysis module, a sliding filtering operator is unchanged, in other words, is the first preference analysis decision knowledge subjected to the completion of the knowledge redundancy removal operation. And performing sliding filtering operation on second preference analysis decision knowledge of the business interaction big data to be analyzed through first preference analysis decision knowledge of the prior business interaction big data so as to update strategy configuration parameters of the business interaction big data to be analyzed, and if the strategy configuration parameters are consistent with the user preference of the prior business interaction big data, reducing the description variable of the data target knowledge. For example, the operation of the global sliding filter operation in the preference linkage analysis module may determine an algorithm formula according to actual requirements, which is not limited herein. It can be understood that the preference linkage analysis module adjusts the undetermined preference analysis decision knowledge through the known preference analysis decision knowledge, and processes the undetermined preference analysis decision knowledge through the advantage that the word vector commonality values of the same preference analysis decision knowledge are more similar, so that feature analysis and preference positioning are performed at one time.
In some possible examples, the global sliding-filter operation may be performed once, or multiple times. Such as in some examples, a global sliding filter operation is performed twice. And taking the first preference analysis decision knowledge with the knowledge redundancy elimination operation as a sliding filter operator, and performing global sliding filter operation on the second preference analysis decision knowledge to adjust and learn the second preference analysis decision knowledge, so as to obtain linkage analysis description variables with the knowledge number C of 128, the time interval T of 64, the first relation network size H of 128 and the second relation network size W of 96. And continuously taking the first preference analysis decision knowledge with the knowledge redundancy elimination operation as a sliding filter operator, and performing global sliding filter operation on the linkage analysis description variable to perform adjustment learning on the linkage analysis description variable to obtain a target knowledge description variable with the knowledge number C of 256, the time interval T of 64, the first relationship network size H of 128 and the second relationship network size W of 96.
Step 590, determining whether the first digital user service preference is the second digital user service preference based on the service interaction data set corresponding to the target knowledge description variable in the service interaction big data to be analyzed.
In some embodiments, redundancy elimination and standardization processing can be performed on the target knowledge description variable, wherein the standardization processing is implemented by adopting a softmax function, relevant business interaction data sets are matched according to the scale of the target knowledge description variable, and the business interaction data set with smaller target knowledge description variable is screened to determine the second digital user service preference in the business interaction big data to be analyzed. Wherein knowledge descriptive variables can be understood as eigenvalues.
In some possible examples, a priori type business interaction big data is used as a reference for service preference analysis, expert knowledge mining is carried out on the priori type business interaction big data and business interaction big data to be analyzed through a weight sharing model, then feature association scores between the priori type business interaction big data and the business interaction big data to be analyzed are determined through preference linkage analysis operation, service preference topic analysis of the business interaction big data to be analyzed is effectively achieved, second digital user service preference is obtained, pending preference analysis decision knowledge is updated through determined preference analysis decision knowledge, the pending preference analysis decision knowledge is selected through the advantage that word vector commonality values of the same preference analysis decision knowledge are closer, and preference analysis efficiency is improved. In addition, the sliding filter operators are obtained through learning, and the sliding filter operators corresponding to different preferences have differences, so that the intelligent degree and the flexibility of preference analysis can be ensured.
In some independent embodiments, the service preference analysis policy tuning method of the embodiments of the present invention may include the following.
Step 810, acquiring a first business interaction big data example and a plurality of second business interaction big data examples.
For example, a first business interaction big data example and a plurality of second business interaction big data examples are obtained, wherein the first business interaction big data example covers service preference topics to be positioned, each topic is provided with one business interaction big data example, the second business interaction big data example is correspondingly provided with a plurality of service preference topics, and each topic is provided with a plurality of business interaction big data examples. Not less than one second business interaction big data example and the first business interaction big data example cover the same class of service preference topics.
Step 830, inputting the first service interaction big data example and the plurality of second service interaction big data examples into a general service preference analysis strategy, where the general service preference analysis strategy includes a general expert knowledge mining module and a general preference linkage analysis module, and performing expert knowledge mining on the first service interaction big data example and the plurality of second service interaction big data examples through the general expert knowledge mining module to obtain a first preference analysis decision knowledge example and a second preference analysis decision knowledge example.
Wherein a generic service preference analysis policy can be created that encompasses a generic expert knowledge mining module and a generic preference linkage analysis module. Loading the first business interaction big data example and a plurality of second business interaction big data examples into a general service preference analysis strategy, and carrying out expert knowledge mining through a general expert knowledge mining module to respectively obtain a first preference analysis decision knowledge example and a second preference analysis decision knowledge example.
Step 850: and in the general preference linkage analysis module, performing preference linkage analysis operation on the second preference analysis decision knowledge example through the first preference analysis decision knowledge example to obtain a knowledge description variable example, and identifying whether service preference topics are covered in the business interaction big data to be analyzed through the knowledge description variable example.
Illustratively, a series of sliding filtering processes are performed through a general preference linkage analysis module, a knowledge description variable example is obtained, and a standardization operation is performed on the obtained knowledge description variable example. The standardization is realized based on a softmax function, and whether the second digital user service preference is covered in the business interaction big data to be analyzed is identified through a knowledge description variable example to obtain a preference theme judging result.
And 870, improving policy configuration parameters of the general service preference analysis policy based on the knowledge description variable examples and the preference theme judgment result to obtain the service preference analysis policy.
For the embodiment of the invention, two cost functions cost (loss function loss) of characteristic detail quality cost and probability distribution accuracy cost are used for improving the strategy configuration parameters of the general service preference analysis strategy. The feature detail quality cost is a triplet cost function. Illustratively, before a softmax link, outputting a knowledge description variable example of a strategy for determining a characteristic detail quality cost function, and illustratively, acquiring word vector commonality values among a plurality of second business interaction big data examples through the knowledge description variable example to determine the characteristic detail quality cost; after the softmax link, a probability distribution accuracy cost function (such as cross entropy loss) is accessed, and the preference subject judgment result and the first business interaction big data example are compared and analyzed to determine the probability distribution accuracy cost. Combining the characteristic detail quality cost and the probability distribution accuracy cost as the global cost (overall loss) of the general service preference analysis strategy, and performing iterative optimization on strategy configuration parameters of the general service preference analysis strategy through the global cost.
In some possible examples, the quality cost of the feature details (such as a triplet loss function) is used, so that the knowledge vector difference of big data examples of the same theme in the business interaction big data to be analyzed is as small as possible, and the knowledge vector difference of big data examples of different themes is as large as possible, thereby improving the processing precision and the reliability of the AI strategy in the processing process. Further, a general policy, a general module may be understood as a policy or module that has not been optimized, or may be understood as an original policy or module.
Under some independent design ideas, on the premise that the first digital user service preference covered by the business interaction big data to be analyzed is the second digital user service preference, extracting a digital user text session from the business interaction big data to be analyzed based on the first digital user service preference and the second digital user service preference; pushing demand mining is carried out on the digital user text session, and a pushing demand mining report is obtained; and pushing big data to the digital user terminal corresponding to the first digital user service preference based on the pushing demand mining report.
In the embodiment of the invention, if the first digital user service preference is the second digital user service preference, the digital user text conversation can be extracted by combining the first digital user service preference and the second digital user service preference, so that text big data which can be used for carrying out user push demand analysis can be obtained, and thus, a push demand mining report can be completely and accurately determined through the digital user text conversation, so that targeted and intelligent big data push can be carried out according to the push demand mining report.
Under some independent design ideas, pushing demand mining is carried out on the digital user text session to obtain a pushing demand mining report, and the pushing demand mining report can be realized through the following technical scheme which can be independently implemented: performing natural language processing (such as pushing demand feature mining operations of different levels) on the digital user text session to obtain a pushing demand field matrix (pushing demand feature distribution) set of the digital user text session; the push requirement field matrix set comprises at least two push requirement field matrices, and the sizes of inconsistent push requirement field matrices in the push requirement field matrix set are inconsistent; determining influence factors corresponding to all push requirement field matrixes in the push requirement field matrix set; the influence factor characterizes the determination of the association weight (importance) of the corresponding push demand field matrix relative to the rest push demand field matrix, wherein the rest push demand field matrix comprises push demand field matrices except the corresponding push demand field matrix in the push demand field matrix set; changing each push demand field matrix by using influence factors corresponding to each push demand field matrix to obtain a plurality of target push demand field matrixes of the digital user text session; and determining a push demand mining report through the plurality of target push demand field matrixes.
By means of the design, accurate and complete push demand mining reports can be obtained by fusing the target push demand field matrixes.
Under some independent design ideas, the determining the influence factors corresponding to each push requirement field matrix in the push requirement field matrix set includes: acquiring a description array of each push demand field matrix in the push demand field matrix set; acquiring an overall description array based on the description arrays of each push demand field matrix in the push demand field matrix set; and determining influence factors of all the push requirement field matrixes in the push requirement field matrix set based on the integral description array.
Under some independent design considerations, the influence factors include bias coefficients; the determining, based on the overall description array, an influence factor of each push requirement field matrix in the push requirement field matrix set includes: inputting the integral description array into a preset vector fusion model, and obtaining a spliced description array output by the preset vector fusion model, wherein the dimension of the spliced description array is the number of push demand field matrixes in the push demand field matrix set; and performing dimension simplification (standardization) processing on the spliced description array to obtain dynamic bias coefficients (global attention coefficients) of each push demand field matrix in the push demand field matrix set.
Under some independent design ideas, the obtaining the description array of each push requirement field matrix in the push requirement field matrix set includes: carrying out overall pooling on each push demand field matrix in the push demand field matrix set to obtain a description array of each push demand field matrix; the obtaining the overall description array based on the description array of each push demand field matrix in the push demand field matrix set includes: and combining the description arrays of the push requirement field matrixes to obtain the overall description array.
Under some independent design ideas, the changing the push requirement field matrix by using the influence factors corresponding to the push requirement field matrices to obtain a plurality of target push requirement field matrices of the digital user text session includes: and weighting each push demand field matrix based on the influence factors corresponding to each push demand field matrix to obtain the target push demand field matrix corresponding to each push demand field matrix.
Based on the same or similar inventive concept, please refer to fig. 2, the present invention further provides a schematic architecture of an application environment 30 of the big data analysis processing method, which includes a big data analysis processing system 10 and a digital client 20 that communicate with each other, where the big data analysis processing system 10 and the digital client 20 implement or partially implement the technical solutions described in the above method embodiments during operation.
Further, there is also provided a readable storage medium having stored thereon a program which when executed by a processor implements the above-described method.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other manners. The apparatus and method embodiments described above are merely illustrative, for example, flow diagrams and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, functional modules in the embodiments of the present invention may be integrated together to form a single part, or each module may exist alone, or two or more modules may be integrated to form a single part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a media service server 10, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes. It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (9)
1. A big data analysis processing method, wherein the method is applied to a big data analysis processing system, and the method at least comprises the following steps:
acquiring business interaction big data to be analyzed covering the first digital user service preference and prior business interaction big data covering the second digital user service preference;
loading the prior-type business interaction big data and the business interaction big data to be analyzed into a service preference analysis strategy, and respectively carrying out expert knowledge mining on the prior-type business interaction big data and the business interaction big data to be analyzed by an expert knowledge mining module in the service preference analysis strategy to obtain first preference analysis decision knowledge and second preference analysis decision knowledge;
performing preference linkage analysis operation on the second preference analysis decision knowledge through the first preference analysis decision knowledge in a preference linkage analysis module of the service preference analysis strategy, and judging whether the first digital user service preference covered by the business interaction big data to be analyzed is the second digital user service preference;
The expert knowledge mining module is a weight sharing model, and the expert knowledge mining module is used for respectively mining the prior-type business interaction big data and the business interaction big data to be analyzed to obtain first preference analysis decision knowledge and second preference analysis decision knowledge, wherein the steps of:
expert knowledge mining is carried out on the prior-type business interaction big data through a first knowledge mining unit of the weight sharing model, so that first preference analysis decision knowledge is obtained;
expert knowledge mining is carried out on the business interaction big data to be analyzed through a second knowledge mining unit of the weight sharing model, so that second preference analysis decision knowledge is obtained;
the tuning method of the service preference analysis strategy comprises the following steps:
acquiring a first business interaction big data example and a plurality of second business interaction big data examples, wherein at least one of the second business interaction big data example and the first business interaction big data example cover the same class of service preference topics;
inputting the first business interaction big data example and the plurality of second business interaction big data examples into a general service preference analysis strategy, wherein the general service preference analysis strategy comprises a general expert knowledge mining module and a general preference linkage analysis module, and expert knowledge mining is respectively carried out on the first business interaction big data example and the plurality of second business interaction big data examples through the general expert knowledge mining module to obtain a first preference analysis decision knowledge example and a second preference analysis decision knowledge example;
In the general preference linkage analysis module, preference linkage analysis operation is carried out on the second preference analysis decision knowledge example through the first preference analysis decision knowledge example to obtain a knowledge description variable example, and whether the service preference theme is covered in the business interaction big data to be analyzed is judged through the knowledge description variable example;
and improving the strategy configuration parameters of the general service preference analysis strategy according to the knowledge description variable examples and the preference theme judgment result to obtain the service preference analysis strategy.
2. The method of claim 1, wherein the performing, in the preference linkage analysis module in the service preference analysis policy, a preference linkage analysis operation on the second preference analysis decision knowledge by the first preference analysis decision knowledge comprises:
determining a sliding filter operator according to the knowledge scale of the first preference analysis decision knowledge;
performing sliding filtering operation on the second preference analysis decision knowledge through the sliding filtering operator to obtain a target knowledge description variable;
and determining whether the first digital user service preference is the second digital user service preference according to the service interaction data set corresponding to the target knowledge description variable in the service interaction big data to be analyzed.
3. The method of claim 2, wherein determining the sliding filter operator based on the knowledge scale of the first preference analysis decision knowledge comprises: and carrying out knowledge redundancy elimination operation on the first preference analysis decision knowledge so as to adjust the knowledge scale of the first preference analysis decision knowledge, and taking the first preference analysis decision knowledge with the knowledge redundancy elimination operation as the sliding filter operator.
4. The method of claim 2, wherein performing a sliding filter operation on the second preference analysis decision knowledge by the sliding filter operator to obtain a target knowledge descriptive variable comprises:
performing sliding filtering operation on the second preference analysis decision knowledge through the sliding filtering operator to obtain a linkage analysis description variable;
and performing sliding filtering operation on the linkage analysis description variable through the sliding filtering operator to obtain a target knowledge description variable.
5. The method of claim 1, wherein the first knowledge mining unit and the second knowledge mining unit respectively include a session habit mining node and a user interest mining node, and wherein performing expert knowledge mining on the prior-type business interaction big data and the business interaction big data to be analyzed respectively by the expert knowledge mining module comprises:
The prior-type business interaction big data and the business interaction big data to be analyzed are respectively loaded to the conversation habit mining node, and conversation habit visual vectors in the prior-type business interaction big data and the business interaction big data to be analyzed are obtained;
and loading session habit visual vectors in the prior-type business interaction big data and the business interaction big data to be analyzed to the user interest mining node, performing global sliding filtering operation on the session habit visual vectors, and determining session habit updating characteristics of the session habit visual vectors to obtain the first preference analysis decision knowledge and the second preference analysis decision knowledge.
6. The method of claim 5, wherein the session habit visual vector in the prior-type business interaction big data and the business interaction big data to be analyzed comprises a session habit visual vector of a GUI interaction event, and the session habit visual vector of the GUI interaction event comprises an activity weight relation network of business interaction links of the GUI interaction event.
7. The method of claim 6, wherein said refining the policy configuration parameters of the generic service preference analysis policy based on the knowledge descriptive variable instance and the preference topic determination result comprises:
Acquiring word vector commonality values among the plurality of second business interaction big data examples through the knowledge description variable examples, and determining characteristic detail quality cost;
comparing and analyzing the preference theme judging result and the first business interaction big data example to determine probability distribution accuracy cost;
and carrying out iterative optimization on strategy configuration parameters of the general service preference analysis strategy according to the characteristic detail quality cost and the probability distribution accuracy cost.
8. A big data analysis processing system is characterized by comprising a processor and a memory; the processor being communicatively connected to the memory, the processor being adapted to read a computer program from the memory and execute it to carry out the method of any of the preceding claims 1-7.
9. A readable storage medium, characterized in that it has stored thereon a program which, when executed by a processor, implements the method of any of the preceding claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211182210.8A CN115470905B (en) | 2022-09-27 | 2022-09-27 | Big data analysis processing method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211182210.8A CN115470905B (en) | 2022-09-27 | 2022-09-27 | Big data analysis processing method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115470905A CN115470905A (en) | 2022-12-13 |
| CN115470905B true CN115470905B (en) | 2023-07-25 |
Family
ID=84335169
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211182210.8A Active CN115470905B (en) | 2022-09-27 | 2022-09-27 | Big data analysis processing method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115470905B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115829421B (en) * | 2023-02-20 | 2023-05-05 | 杭银消费金融股份有限公司 | Intelligent analysis method and system for business requirements |
| CN116881571B (en) * | 2023-07-27 | 2024-01-12 | 广东麦视互联网科技有限公司 | Data mining method and system based on micro-service |
| CN117149986B (en) * | 2023-10-31 | 2024-02-09 | 杭州海兴泽科信息技术有限公司 | Real-time big data processing method and system based on multi-stage data channel |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140149441A1 (en) * | 2012-11-29 | 2014-05-29 | Fujitsu Limited | System and method for matching persons in an open learning system |
| CN114499911A (en) * | 2020-11-13 | 2022-05-13 | 三六零数字安全科技集团有限公司 | Attack user identification method, equipment, storage medium and device based on test machine |
| CN112685478B (en) * | 2021-02-23 | 2022-01-14 | 爱云保(上海)科技有限公司 | Information processing method for cloud service and user portrait mining and cloud server |
| CN114661886A (en) * | 2022-03-14 | 2022-06-24 | 李翠萍 | Online business topic processing method and server applied to big data intelligent analysis |
| CN114625804B (en) * | 2022-03-30 | 2022-11-08 | 深圳唯爱智云科技有限公司 | Big data-based user behavior data processing method and system and cloud platform |
-
2022
- 2022-09-27 CN CN202211182210.8A patent/CN115470905B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115470905A (en) | 2022-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115470905B (en) | Big data analysis processing method and system | |
| CN107423442B (en) | Application recommendation method and system based on user portrait behavior analysis, storage medium and computer equipment | |
| CN115757745B (en) | Business scene control method and system based on artificial intelligence and cloud platform | |
| CN111582341B (en) | User abnormal operation prediction method and device | |
| CN111611488A (en) | Information recommendation method and device based on artificial intelligence and electronic equipment | |
| EP4339843A1 (en) | Neural network optimization method and apparatus | |
| CN113435509A (en) | Small sample scene classification and identification method and system based on meta-learning | |
| EP3767549A1 (en) | Delivery of compressed neural networks | |
| CN113449011A (en) | Big data prediction-based information push updating method and big data prediction system | |
| CN113987236B (en) | Unsupervised training method and unsupervised training device for visual retrieval model based on graph convolution network | |
| WO2023279674A1 (en) | Memory-augmented graph convolutional neural networks | |
| CN115272242B (en) | YOLOv 5-based optical remote sensing image target detection method | |
| CN112817563B (en) | Target attribute configuration information determining method, computer device, and storage medium | |
| CN110956277A (en) | Interactive iterative modeling system and method | |
| CN115222051A (en) | Big data mining method and system based on digital Internet | |
| CN110427870B (en) | Eye picture recognition method, target recognition model training method and device | |
| EP3803712A1 (en) | An apparatus, a method and a computer program for selecting a neural network | |
| CN112906883B (en) | Hybrid precision quantization strategy determination method and system for deep neural network | |
| CN115906927B (en) | Data access analysis method and system based on artificial intelligence and cloud platform | |
| CN117194790A (en) | Big data mining method and system for analyzing pushing preference of digital user information | |
| CN116089727A (en) | Information analysis method and system based on user portrait mining | |
| CN114385876B (en) | Model search space generation method, device and system | |
| CN114625606A (en) | Big data intelligent service information optimization method based on digitization and server | |
| CN113762505A (en) | Clustering pruning method of convolutional neural network according to norm of channel L2 | |
| CN112532692A (en) | Information pushing method and device and storage medium |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230705 Address after: 430000, No.7, Longyang Avenue, Hanyang District, Wuhan City, Hubei Province Applicant after: Hubei Huatong Engineering Consulting Co.,Ltd. Address before: No. 59, Qingnian Road, Tiexi District, Siping, Jilin Province, 136000 Applicant before: Gao Qiang |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |