CN110083663B

CN110083663B - Classification optimization method and device for data display

Info

Publication number: CN110083663B
Application number: CN201910280277.7A
Authority: CN
Inventors: 黄浩
Original assignee: Beijing Zhongke Zhiying Technology Development Co ltd
Current assignee: Beijing Zhongke Zhiying Technology Development Co ltd
Priority date: 2019-04-09
Filing date: 2019-04-09
Publication date: 2021-08-17
Anticipated expiration: 2039-04-09
Also published as: CN110083663A

Abstract

The invention provides a classification optimization method and a classification optimization device for data display, which solve the technical problem that the field of the existing data classification process is greatly limited, and operators are not beneficial to quickly forming a classification display effect according to the requirements of business rules. The method comprises the following steps: establishing a classification data structure for storing classification categories; establishing the classification category through the classification data structure storage; storing classification rules for the classification categories by the classification data structure. The classification data structure with universality is used for planning and designing the storage of classification categories, the classification categories formed by deep learning and the classification categories formed by supervision classification are kept consistent with a data frame of a storage structure, and the classification category data has application and inherited data processing advantages. The classification data structure is used for storing the parameters of the classification categories, so that the display consistency of the same characteristics or parameters of different classification categories is formed, the classification categories and the corresponding parameters or characteristics are determined, and the application difficulty of data users is reduced.

Description

Classification optimization method and device for data display

Technical Field

The invention relates to the technical field of data display, in particular to a classification optimization method and device for data display.

Background

In the prior art, a mature classification system is biased to customization and pertinence and is often limited to a narrow data field, information needing to be classified is generally listed by using a table, a tree structure is formed by using a graphic technology to list classification information, and a classification effect is displayed by using a legend. The classification effect forming process needs to depend on the combination of professional graphic programming and classification information, and is not beneficial for operators to quickly form a classification display effect according to actual requirements of business rules, knowledge maps, data searching and the like. This directly results in the lack of effective addressing and operation ability of the corresponding data mathematical model in the memory environment for the model composition data and the model addressing architecture, so that the data and model addressing can not be effectively reused, and various costs of model reconstruction are greatly increased.

Disclosure of Invention

In view of the above problems, embodiments of the present invention provide a method and an apparatus for classifying and optimizing data display, which solve the technical problem that the field of the existing data classification process is limited greatly, and it is not favorable for an operator to quickly form a classification display effect according to actual needs such as business rules, knowledge maps, data search, and the like.

The classification optimization method for data display in the embodiment of the invention comprises the following steps:

establishing a classification data structure for storing classification categories;

establishing the classification category through the classification data structure storage;

storing classification rules for the classification categories by the classification data structure.

In an embodiment of the present invention, the establishing a classification data structure for storing classification categories includes:

forming a classification category base framework;

forming a set of classification categories in the base framework;

forming a classification category grade in the classification category set;

forming a classification category topology in a classification category level in the set of classification categories;

forming a storage field of a classification category in the classification category level;

and storing the associated data of the classification category according to the topological structure and the storage field.

In an embodiment of the present invention, the establishing the classification category through the classification data structure storage includes:

registering a classification identification through the classification category base framework;

establishing a determined classification subject set through the classification category set;

establishing a classification and grading theme through the classification category grade;

and establishing entity parameters for determining the subject through the storage field.

In an embodiment of the present invention, the storing the classification rule of the classification category by the classification data structure includes:

determining a classification identification and a corresponding classification subject set through the classification data structure;

obtaining classification categories of the determined topics through the classification topic set;

extracting a rule set for determining a theme through a data structure of the rule;

adding the rule set to a classification category association field of a determined topic according to the classification data structure.

The classification optimization device for data display of the embodiment of the invention comprises:

the memory is used for storing program codes corresponding to the processing procedures of the classification optimization method shown in the embodiment data of the claim;

a processor for executing the program code.

the classification structure forming module is used for establishing a classification data structure for storing classification categories;

the classified content storage module is used for establishing the classified category through the classified data structure storage;

and the classification rule association module is used for storing the classification rules of the classification categories through the classification data structure.

In an embodiment of the present invention, the classification structure forming module includes:

a frame forming unit for forming a classification category base frame;

the set forming unit is used for forming a classification category set in the basic framework;

a grade forming unit, which is used for forming the grade of the classification class in the classification class set;

an integral topology unit for forming a classification category topology in a classification category level in the classification category set;

a field forming unit, configured to form a storage field of a classification category in the classification category level;

and the parameter storage unit is used for storing the associated data of the classification categories according to the topological structure and the storage fields.

In an embodiment of the present invention, the classified content storage module includes:

the frame registration unit is used for registering classification identification through the classification category base frame;

the theme registration unit is used for establishing and determining a classification theme set through the classification category set;

the classification registration unit is used for establishing a classification topic through the classification category grade;

and the theme determining unit is used for establishing entity parameters for determining the theme through the storage field.

In an embodiment of the present invention, the classification rule association module includes:

the theme set structure determining unit is used for determining a classification identifier and a corresponding classification theme set through the classification data structure;

the classification type determining unit is used for acquiring the classification type of the determined theme through the classification theme set;

the rule determining unit is used for extracting a rule set of the determined subject through a data structure of the rule;

and the data structure mapping unit is used for adding the rule set into a classification category association field of a determined subject according to the classification data structure.

The classification optimization method and device for data display in the embodiment of the invention utilize the classification data structure with universality to plan and design the storage of classification categories, and keep the consistency of data frames of the storage structure of the classification categories formed by deep learning and the classification categories formed by supervision and classification, so that the classification category data has the advantages of application and inherited data processing. The classification data structure is used for storing the parameters of the classification categories, so that the display consistency of the same characteristics or parameters of different classification categories is favorably formed, the classification categories and the corresponding parameters or characteristics are favorably determined, an interaction means for reusing or updating the classification categories can be further formed, and the application difficulty of data users is reduced. Especially, the method has positive effect on vertical data classification in a single data field.

Drawings

Fig. 1 is a schematic flow chart illustrating a big data classification rule optimization method according to an embodiment of the present invention.

Fig. 2 is a schematic flow chart illustrating the establishment of a data structure in the big data classification rule optimization method according to an embodiment of the present invention.

Fig. 3 is a schematic flow chart illustrating the establishment of a rule set in the big data classification rule optimization method according to an embodiment of the present invention.

Fig. 4 is a schematic flow chart illustrating scene classification performed in the big data classification rule optimization method according to an embodiment of the present invention.

Fig. 5 is a schematic flow chart illustrating data processing performed on rules in the big data classification rule optimization method according to an embodiment of the present invention.

Fig. 6 is a schematic diagram illustrating an architecture of a big data classification rule optimization apparatus according to an embodiment of the present invention.

Fig. 7 is a flowchart illustrating a classification optimization method according to an embodiment of the present invention.

Fig. 8 is a schematic flow chart illustrating the processes of forming a classification data structure, establishing classification categories and associating related data in the classification optimization method for data presentation according to an embodiment of the present invention.

Fig. 9 is a schematic diagram illustrating an architecture of a classification optimization apparatus according to an embodiment of the present invention.

Fig. 10 is a schematic main flow chart of a data classification optimization method according to an embodiment of the present invention.

Fig. 11 is a detailed flowchart illustrating a data classification optimization method according to an embodiment of the present invention.

Fig. 12 is a schematic diagram illustrating an architecture of a data classification optimizing apparatus according to an embodiment of the invention.

Fig. 13 is a schematic main flow chart illustrating a data processing method of a classification interactive interface according to an embodiment of the present invention.

Fig. 14 is a detailed flowchart illustrating a data processing method for a classification interactive interface according to an embodiment of the present invention.

FIG. 15 is a block diagram of a data processing apparatus for classifying an interactive interface according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer and more obvious, the present invention is further described below with reference to the accompanying drawings and the detailed description. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention. The numbering of the steps is independent of the process of data processing.

Fig. 1 shows a method for optimizing a big data classification rule according to an embodiment of the present invention. In fig. 1, the present embodiment includes:

step 110: a data structure storing rules is established.

The data structure is used for storing specific parameters of the rules, determining classification standard of each rule on the big data source data according to the specific parameters, and forming a classification process.

The rule can be a numerical value, a character string, a characteristic vector value, a data reference position, data judgment logic or rule check data and the like of a specific parameter or an associated parameter, the data structure and the specific parameter or the associated parameter of the rule have adaptation universality, and the adaptation universality ensures that the data structure is adapted to the type and the magnitude of the specific parameter or the associated parameter.

Step 120: a rule set that defines a topic is formed by the data structure.

The data evolution process is matched with the process change formed by the source data in the objective environment, the change of the classification scenes can respond to the data evolution process of the source data of the classification subjects, the set of the classification scenes determines the classification characteristics of the classification subjects, and the classification subjects comprise the further detailed classification scenes. And storing specific parameters or associated parameters corresponding to the classification scenes by using a data structure, and forming rules corresponding to the classification scenes into a rule set with a single theme through the data structure.

Step 130: and carrying out scene classification on the source data according to the rule set.

And performing source data topic classification according to the rule set, and forming classification of different scenes in the topic classification on the basis of forming the data topic classification.

According to the type of the rule parameters in the rule set, the parameter meanings and the judgment logic formed among the parameters, the rule parameters in the rule set can also be used in data processing processes such as retrieval, sequencing or exclusion when scene classification is carried out.

According to the big data classification rule optimization method, a complete topological structure of topic classification and scene classification is formed by using a data structure, parameters of different classification processes are configured in order, a structured classification system is formed by a complex parameter reference process of a big data classification model, classification iteration parameter data, key enumeration data and a classification process threshold value adjustment interval in the classification model can be effectively managed and reasonably configured, the big data classification model has a proper adjustment technical means for source data characteristic change, training and learning processes are prevented from being adjusted frequently for evolution of an objective scene of the source data formation process, time cost and labor cost of big data classification are reduced, and application universality and model tolerance of the big data classification model are improved.

As shown in fig. 1, in an embodiment of the present invention, the method for optimizing big data classification rules further includes:

step 140: and analyzing the rule set and updating the storage rule.

The analysis and the update are realized by reading the data structure of the rule set, acquiring the parameters or related parameters of the determined rule according to the data structure, and modifying the content of the parameters to form the update of the determined rule, so as to update the rule set of the determined subject.

The big data classification rule optimization method of the embodiment of the invention provides an updating process of positioning, correcting and replacing local rules of a big data classification model. The integrity of the existing big data classification model is ensured, and meanwhile, a technical means for performing source data adaptability correction on classification parameters is provided. The data type or trend change of the source data can be properly adjusted in the big data classification model in time, and meanwhile, the general trend and the precision of big data classification are stable.

step 150: and analyzing the rule set, acquiring a local data structure and forwarding the rule.

The analysis and acquisition is to obtain the data topological structure and the stored rule parameters aiming at the local classification and the scene classification of the specific type data in the big data classification model by reading the data structure of the rule set, and form the rule subset for determining the theme.

The big data classification rule optimization method of the embodiment of the invention provides an extraction process of a big data classification model local rule. The classification rules of the existing big data classification model aiming at specific classification types or classification scenes can be accurately extracted from the rules for determining the classification, and the effective reuse of the reliable rule subsets in the mature big data classification model is realized. The multiplexing of the rule subset has a positive effect on the perfection of the classification rules of other initially formed big data classification models, so that the big data classification models can be adapted timely according to the business field or the application field of the source data, and the pertinence and the accuracy of big data classification in a business scene are met. The method has a good data response effect on the combination of business data, business knowledge maps and search structures in a business environment.

The process of establishing the data structure in the big data classification rule optimization method according to the embodiment of the invention is shown in fig. 2. In fig. 2, the process of building the data structure includes:

step 111: the structure field of a single rule is set.

And setting the data type matched with the specific parameters or the associated parameters of the rule by using the structure field to meet the form requirement of data storage. Inclusion logic exists between the configuration fields.

Step 112: setting field keywords and compound field keywords of a single rule.

The readability of the field with a single structure is improved by referring to the field type through the field keywords, the readability of the field with a composite structure is improved through the composite field keywords, and the reading comprehension of people is met.

Step 113: the data structure of the rule is formed according to the field topology of the single rule.

The data structure may adopt a standard structure in the prior art, such as an XML standard, a JSON standard, and the like. Taking an XML (extensible markup language) data structure standard as an example, a structure field is formed by using the XML standard, a keyword is set for the structure field, and a determined field topological structure is formed by using the keyword to be registered in the XML standard, so that a data structure of each level of rules is formed. The basic form of the data structure may be as follows:

< subject rule keywords > <

Rule 1 ((parameter type parameter data), (parameter type parameter data))

Rule 2 ((parameter type parameter data), (parameter type parameter data))

……

< keyword of subject rule >

Example 1

[1.0] < escrue > ("e-commerce", "micro-commerce")// use "e-commerce", "micro-commerce" as the classification keyword, the weight coefficient in the classification result is 1.0

[0.5] < escrlue > ("shop", "B2B", "O2O")// use "shop", "B2B", "O2O" as classification keywords with a weight coefficient of 0.5 in the classification results >

Example 2

[1.0] "electronic commerce", [1.0] "Weekly", [0.5] "shop", [0.5] "B2B", [0.5] "O2O")// the classification keyword "electronic commerce", "Weekly", "shop", "B2B", and "O2O" have corresponding weight coefficients of [1.0], [0.5] and [0.5] > "

According to the big data classification rule optimization method, the basic data structure is established according to the keywords and the topological structure of the keywords, the data structure is matched with the types of the rule parameters, and the readability of the data structure is improved through the keywords.

The process of establishing a rule set in the big data classification rule optimization method according to an embodiment of the present invention is shown in fig. 3. In fig. 3, the process of establishing a rule set includes:

step 121: and establishing a rule set keyword.

Step 122: and establishing a rule set subset of the rule set and a corresponding subset keyword.

Step 123: and establishing a scene set of the rule set subset and a corresponding scene keyword.

Step 124: and establishing a keyword topological structure according to the rule set keywords, the subset keywords or the scene keywords.

Step 125: and adding a rule data structure in the keyword topological structure, and adding rule parameters or related parameters in the rule data structure.

Taking an XML (extensible markup language) data structure standard as an example, the following rule set is formed by combining the data structure addition theme formed in the above embodiment and the rule parameters corresponding to the theme:

< rule set name >

< rule name >

< rule scene name > <

Rule 1

Rule 2

…>

Example 1

< topic name >

{ the political theory },

{ business },

{ society },

)]>

< topic name >

{ an hour review },

{ macro-economy },

{ political characters },

)]>

< topic name ═ time commentary "> < [ cdata >

[1.0] ("economic reform", "political reform", "educational reform", "education fairness", "college entrance examination and immigration", "household registration system")

[0.5] ("keny coefficient", "network public opinion", "liberality", "political inspection", "political construction", "political culture", "virtual nonsense")

)]>

</topic>

The big data classification rule optimization method provided by the embodiment of the invention establishes the topological structures of the classification rules and the scene rules according to the keywords of all levels, and fills in specific rule parameter data. The readability of the data structure and the data content is met on the basis of ensuring the parameter storage efficiency.

The process of performing scene classification in the big data classification rule optimization method according to an embodiment of the present invention is shown in fig. 4. In fig. 4, the process of scene classification includes:

step 131: and determining corresponding preliminary source data according to the preliminary classification.

The preliminary classification may be a basic classification or a classification formed by a basic data classification model. The preliminary source data is source data corresponding to a certain determined classification after the basic classification.

Step 132: a rule set corresponding to the preliminary classification is determined in the rule set.

The rule set may comprise (a) set of rules for different classifications, the (a) set of rules comprising (a) set of rules for different domain or class of scene classification classes.

Step 133: scene classification categories and scene classification parameter data are extracted from the rule set.

The scene classification category may have default parameters or may have customized parameter data.

Step 134: and classifying the preliminary source data by using the scene classification parameter data to form classification source data corresponding to the scene classification category determined under the preliminary classification.

The scene classification parameter data is used as further data classification, the preliminary source data is further classified into scene classification categories, threshold judgment of characteristics exists in the classification process, and the possibility that the source data is excluded from the categories to form classification source data exists.

Step 135: and forming result classification data of the preliminary classification according to the classification source data.

And the preliminary source data screening is formed by utilizing the set of the classified source data, so that the data noise after the evolution of the source data is favorably eliminated.

According to the big data classification rule optimization method, the initial source data which are formed by the basic big data classification model and are used for determining the initial classification are subjected to further scene classification by utilizing the rule set, the deviation data in the initial source data are filtered through the scene classification, more accurate classification scenes and corresponding classification data are formed, and further classification features and classification precision formed according to scene features in the basic classification model are improved.

Fig. 5 shows a process of performing data processing on a rule in the big data classification rule optimization method according to an embodiment of the present invention. In fig. 5, the process of performing update data processing on the rule includes:

step 141: obtaining the updating requirement of determining the classification and determining the scene.

Step 142: and determining a rule data structure and rule contents according to the updating requirement, and displaying through an interactive frame.

Step 143: and updating the data structure and the rule content of the corresponding rule by the updating content through the interactive framework.

In fig. 5, the process of locally acquiring the rule and processing the forwarding data includes:

step 151: and acquiring the associated topological structure of the data structure of the determined classification and determined scene according to the forwarding requirement.

Step 152: a temporary rule data structure and rule content are formed according to the associated topology.

Step 153: and forming the temporary rule data structure and the rule content into an independent data object to provide data linkage.

According to the big data classification rule optimization method, part of parameters of the complete big data classification model are extracted and updated according to the data structure, and damage to the integrity of the big data classification model and interference of the data classification process are avoided. The updating of the classification rules can have better timeliness and can basically meet the gradual evolution of classification source data types and domain characteristics. The local extraction of the classification rules and the formation of the response data objects can exist as independent rule data sources, so that the reliable rules of data classification in related fields and scenes are improved for other big data classification models, the reusability of the reliable data rules is improved, and the big data classification cost is reduced.

The big data classification rule optimization device in one embodiment of the invention comprises:

the memory is used for storing the program codes corresponding to the processing procedures of the big data classification rule optimization method in the embodiment;

a processor for executing the program code of the big data classification rule optimization method processing procedure of the embodiment

Fig. 6 shows a structure of a big data classification rule optimization apparatus according to an embodiment of the present invention. In fig. 6, the present embodiment includes:

a rule setting module 1110 for establishing a data structure for storing rules;

a rule forming module 1120 for forming a rule set for determining a topic by a data structure;

a rule application module 1130, configured to perform scene classification on the source data according to the rule set.

As shown in fig. 6, in an embodiment of the present invention, the method further includes:

a rule update module 1140 for parsing the rule set and updating the stored rules.

and a rule extraction module 1150, configured to parse the rule set and obtain a local data structure and a rule for forwarding.

As shown in fig. 6, in an embodiment of the present invention, the rule setting module 1110 includes:

a rule field forming unit 1111 configured to set a structure field of a single rule;

a rule key forming unit 1112 configured to set a field key and a composite field key of a single rule;

a rule structure forming unit 1113, configured to form a data structure of a rule according to a field topology of a single rule.

As shown in fig. 6, in an embodiment of the present invention, the rule forming module 1120 includes:

a main keyword setting unit 1121 configured to establish a rule set keyword;

a rule subset setting unit 1122 for establishing a rule set subset of the rule set and a corresponding subset keyword;

a rule scene setting unit 1123, configured to establish a scene set of the rule set subset and a corresponding scene keyword;

a keyword association setting unit 1124 configured to establish a keyword topology structure according to the rule set keyword, the subset keyword, or the scene keyword;

the rule parameter filling unit 1125 is used to add a rule data structure to the keyword topology structure and add a rule parameter or a related parameter to the rule data structure.

As shown in fig. 6, in an embodiment of the present invention, the rule application module 1130 includes:

a preliminary data obtaining unit 1131, configured to determine corresponding preliminary source data according to the preliminary classification;

a rule determining unit 1132 for determining a rule set corresponding to the preliminary classification in the rule set;

a rule configuration unit 1133, configured to extract scene classification categories and scene classification parameter data from the rule set;

a classification execution unit 1134, configured to classify the preliminary source data by using the scene classification parameter data, and form classification source data corresponding to the scene classification category determined under the preliminary classification;

a data synthesis unit 1135, configured to form result classification data of the preliminary classification according to the classification source data.

As shown in fig. 6, in an embodiment of the present invention, the rule updating module 1140 includes:

an update receiving unit 1141, configured to obtain an update requirement for determining a classification and determining a scene;

an update interaction unit 1142, configured to determine a rule data structure and rule contents according to an update requirement, and display the rule data structure and the rule contents through an interaction frame;

and the update receiving unit 1143 is configured to update the data structure and the rule content of the corresponding rule with the update content through the interaction framework.

As shown in fig. 6, in an embodiment of the present invention, the rule extracting module 1150 includes:

a forwarding requirement receiving unit 1151, configured to obtain, according to a forwarding requirement, an associated topology structure of a data structure of a determined classification and a determined scene;

a requirement determining unit 1152, configured to form a temporary rule data structure and rule contents according to the associated topology structure;

a requirement independent unit 1153 for forming the temporary rule data structure and the rule content into independent data objects, providing data links.

The optimization method of the big data classification rule enables the classification rule to be structured, has integral frame performance on the purposes of reuse, expansion, updating and the like of the classification rule, can meet the universality of the classification rule of data in related fields, and has a supporting function on forming a classification frame with the universality among similar fields.

And the data structure optimization is carried out in the data classification process aiming at the big data, and a classification optimization method beneficial to data display can be formed.

The classification optimization method of data presentation according to an embodiment of the present invention is shown in fig. 7. In fig. 7, the present embodiment includes:

step 210: a classification data structure storing classification categories is established.

The classification data structure is used for forming a topological structure of classification categories and forming a framework basis of the whole classification categories. So that the framework basis can adapt to classification category changes in complex application fields by combined adaptation.

The classification data structure is used to store basic parameters of the classification category including, but not limited to, citation, identifier or name information, and the like.

The classification data structure is also used to provide structural features such as control indexes for classification categories.

Step 220: the classification categories are established by classification data structure storage.

And forming a topological structure of classification categories through the classification data structure according to the classification characteristics of the field data in the application field.

The topology of the classification categories is formed into a user-readable classification category description format by the classification data structure.

Step 230: the classification rules for classifying the categories are stored by a classification data structure.

The classification data structure is utilized to establish a classification system, the universality of the classification system is utilized to form the compatibility with the data structure of the rules, and the classification rules of classification categories are formed through the rule processing process of the big data classification rule optimization method in the embodiment of the invention. The rule processing process of the big data classification rule optimization method in one embodiment of the invention is combined with the embodiment, and the local and overall citation or reuse of the data structure is established through the data structure establishing data mapping.

The classification optimization method for data display in one embodiment of the invention utilizes the classification data structure with universality to plan and design the storage of classification categories, and keeps the consistency of data frames of the storage structure of the classification categories formed by deep learning and the classification categories formed by supervision and classification, so that the classification category data has the advantages of application and inherited data processing. The classification data structure is used for storing the parameters of the classification categories, so that the display consistency of the same characteristics or parameters of different classification categories is favorably formed, the classification categories and the corresponding parameters or characteristics are favorably determined, an interaction means for reusing or updating the classification categories can be further formed, and the application difficulty of data users is reduced. Especially, the method has positive effect on vertical data classification in a single data field.

Fig. 8 shows a method for forming a classification data structure in the classification optimization method according to an embodiment of the present invention. In fig. 8, the classification data structure forming process includes:

step 211: forming a classification category base framework.

The basic framework includes the primary elements that form the complete framework, such as the basic data types, basic data structure standards, etc. that construct the taxonomic data structure.

Step 212: a set of classification categories in the base framework is formed.

A classification base of classification classes is formed using the set of classification classes. The category basis serves as an identification of a data port or a set of classification categories that form a data exchange with other data systems.

Step 213: forming a classification category level in the set of classification categories.

The class classification level identifies the association logic of the class classification and determines the inclusion characteristics of two adjacent class classifications. The classification category classes form an intrinsic association of adjacent classification categories.

Step 214: a classification category topology is formed that is in a classification category hierarchy of the set of classification categories.

The classification category topology is used to establish the overall membership characteristics of the classification category set. The classification category topology forms an intrinsic association of the overall classification category.

Step 215: a storage field for a classification category in the classification category hierarchy is formed.

Storage fields include, but are not limited to, numeric values, strings, feature vector values, data reference locations, data judgment logic or rule check data, and the like. The formed storage fields have adaptability, can be formed according to the preset conditions of the adaptation rules or according to the data input types, the number of the storage fields is not specifically specified, and the memory allocation mode by utilizing the queue or array structure can be adapted.

example 3:

step 216: and storing the associated data of the classification category according to the storage topology and the storage field.

The associated data may be description data of the classification category, such as feature data of the classification category, or associated data of the classification category, such as rule data of the classification category.

The classification optimization method for data display in the embodiment of the invention establishes a classification data structure newcastle and a field data classification category framework with better universality aiming at similar data fields, so that the classification category has a general data processing basis of multiplexing, reusing and quick updating.

As shown in fig. 8, in an embodiment of the present invention, the process of establishing the classification category includes:

step 221: and registering the classification identification through a classification base framework.

And the registration enables the classification category basic framework and the processing process of the existing data classification model to form data connection and form an addressable data port for data calling.

Step 222: and establishing a determined classification subject set through the classification category set.

The classification topic set determines a classification basis in one data domain and establishes a data domain or data addressing range that is relatively independent of other data domains. The data fields or data addressing scopes may be individual data objects, such as linked data sources, linked individual data files, or linked categorized subject sets.

Step 223: and establishing a classification grading theme through classification category grades.

The classification category level has classification category level logic description of two adjacent classification categories, completes the level or containing logic description of the two adjacent classification categories, and manages the construction of the classification categories through classification subjects.

Step 224: and establishing entity parameters for determining the subject through the storage field.

The classification category includes storage fields and the determined classification category may include the same or different storage fields.

The classification optimization method for data display in the embodiment of the invention utilizes a classification data structure to form a classification subject set, a classification subject level and a structured storage for determining classification and corresponding parameters. A complex classification hierarchy can be implemented while satisfying the expansion, reuse, and association of classification categories.

As shown in fig. 8, in an embodiment of the present invention, the process of associating the related data includes:

step 231: and determining a classification identification and a corresponding classification subject set through the classification data structure.

And obtaining a data interaction port of the classification subject set by using the classification data structure, and starting an interaction process of a data basic frame and data contents.

Step 232: and obtaining the classification category of the determined topic through the classification topic set.

Determining the classification category of the subject includes determining a topology and a specific field structure of each classification category of the subject.

Step 233: and extracting a rule set for determining the theme through a data structure of the rules.

And forming mapping of the data structure of the rule and the classification data structure, completing compatible conversion of the data structure, and correspondingly connecting the rule set of the determined theme with the classification category of the determined theme.

Step 234: the rule set is added to the classification category association field of the determined topic according to the classification data structure.

The data structure or rule set of the rule is added to the classification data structure using the mapping of the data structure of the rule to the classification data structure and forms a transmission of the corresponding field and the field content.

The data display classification optimization method provided by the embodiment of the invention utilizes the mapping formation structure compatibility of the classification data structure and the rule data structure, so that the rules and the classifications can be independently stored, independently adjusted and respectively associated, the classification process in the adjacent data fields can be formed by distributed classification categories and classification rules, and the universality of a classification framework and the adaptability to data types are effectively expanded.

The classification optimization device for data display in one embodiment of the invention comprises:

the memory is used for storing program codes corresponding to the processing procedures of the classification optimization method shown by the data of the embodiment;

and the processor is used for executing the program codes of the processing procedures of the classification optimization method shown by the data of the embodiment.

Fig. 9 shows a classification optimization apparatus according to data of an embodiment of the present invention. In fig. 9, the present embodiment includes:

a classification structure forming module 2210 for establishing a classification data structure for storing the classification categories;

a classified content storage module 2220, configured to establish a classification category through classified data structure storage;

a classification rule association module 2230 for storing the classification rules for the classification categories via the classification data structure.

As shown in fig. 9, in an embodiment of the invention, the classification structure forming module 2210 includes:

a frame forming unit 2211 for forming a classification category base frame;

a set forming unit 2212 for forming a set of classification categories in the base framework;

a rank forming unit 2213 for forming a classification category rank in the classification category set;

an overall topology unit 2214 for forming a classification category topology in a classification category level in the set of classification categories;

a field forming unit 2215 for forming a storage field of the classification category in the classification category level;

a parameter storage unit 2216, configured to store the associated data of the classification category according to the storage topology and the storage field.

As shown in fig. 9, in an embodiment of the present invention, the categorized content storage module 2220 includes:

a framework registration unit 2221, configured to register a classification identifier through a classification category base framework;

a topic registration unit 2222, configured to establish a classification topic set through the classification category set;

a hierarchical registration unit 2223 for establishing a classification hierarchical topic by classifying the category levels;

the topic determination unit 2224 is configured to establish an entity parameter for determining a topic through the storage field.

As shown in fig. 9, in an embodiment of the present invention, the classification rule association module 2230 includes:

a topic set structure determining unit 2231, configured to determine a classification identifier and a corresponding classification topic set according to the classification data structure;

a classification category determining unit 2232, configured to obtain a classification category of the determined topic from the classification topic set;

a rule determining unit 2233, configured to extract a rule set for determining a topic from a data structure of the rule;

a data structure mapping unit 2234, configured to add the rule set to the classification category association field of the determined topic according to the classification data structure.

By utilizing the optimization method of the big data classification rule and the data classification system formed by the classification optimization method of the data display in the embodiment, a generalized structure covering data classification and classification rules can be well formed. The combination of the generalization structure can form effective improvement on the existing data classification method.

Fig. 10 shows a data classification optimization method according to an embodiment of the present invention. In fig. 10, an embodiment of the present invention includes:

step 310: and forming an industry customized classification model by utilizing the classification data structure.

Based on the universality, expansibility and data reusability of a classification system formed by the classification data structure, partial classification categories of data in a specific industry or a specific field and corresponding classification rules form an ordered industry customized classification model by utilizing the classification data structure.

Step 320: and classifying the source data through an industry customized classification model to form a data label.

The source data classification has an explicit classification and an implicit classification, the explicit classification features can be directly defined and described in an industry customized classification model, and the data implicit feature expression of the implicit classification is not suitable for direct definition and description. And performing explicit classification by using an industry customized classification model, and performing data marking on the classification characteristic of each source data to form a corresponding data label.

Step 330: and dividing the source data into data training sets according to the data labels.

And the data labels are utilized to combine the source data with the same label type into a data training set, so that the consistency of the dominant characteristics is ensured.

Step 340: and optimizing an intelligent classification model according to the data training set and completing classification of the source data.

And training and shaping the intelligent classification model by combining the invisible features of the data training set obtained by utilizing supervised learning or semi-supervised learning.

The source data may be unclassified source data, which facilitates data classification of the delta data. The source data may also be the entire source data, which is beneficial to the accuracy of the data classification.

The data classification optimization method of the embodiment of the invention utilizes the classification data structure to form reusable and administrable artificial classification rules to be combined with a classification model based on a computer intelligent algorithm. An accurate large-scale data training set is formed by using the artificial classification rules, an intelligent classification model is optimized by using the large-scale data training set, and the accuracy and the classification efficiency of the invisible classification are realized by using the intelligent classification model.

Fig. 11 shows a data classification optimization method according to an embodiment of the present invention. In fig. 11, the process of forming the industry customized classification model in the present embodiment includes:

step 311: and establishing classification categories and topological structures of the classification categories through the classification data structure according to the service requirements.

Forming classification classes and topology of classification classes using normalization and setup logic of classification data structures

Step 312: and establishing a corresponding classification rule according to the classification category.

The classification rule may be a diversity data parameter corresponding to a classification category, such as a keyword, and may be used as a classification basis or a search basis.

Step 313: and setting a classification threshold value according to a classification rule.

The classification threshold may be a value corresponding to a relevance score of each piece of data in a search result, and a classification set is determined for data having a score higher than the predetermined classification threshold.

Step 314: and forming an industry customized classification model according to the topological structure, the classification rules and the classification threshold value.

The data structure may be a standard structure in the above embodiments or in the prior art, such as an XML standard, a JSON standard, or the like. Taking an XML (extensible markup language) data structure standard as an example, a structure field is formed by using the XML standard, a keyword is set for the structure field, and a determined field topological structure is formed by using the keyword to be registered in the XML standard, so that a data structure of each level of rules is formed. The basic form of the data structure may be as follows:

example 4

Establishing a classification system according to a hierarchical structure provided by an enterprise, wherein the classification system refers to the hierarchical structure of classification and specific names of the classification, and the classification system comprises the following steps: representing the first class as "business", the business includes the second class "business person"

The data classification optimization method of the embodiment of the invention forms an ordered classification system of the dominant classification features by utilizing the classification data structure, expresses the topological structure among the classification categories and the respective classification rule parameters and classification weight factors of the classification categories by utilizing the classification system, and orderly configures the dominant features of the industry service requirements through manual adjustment, thereby providing a reliable technical adjustment basis for the effective adjustment, multiplexing, reusing or updating of the dominant features. The industry business demand dominant feature classification can be independently completed by operators, and the whole data classification efficiency is improved.

As shown in fig. 11, in an embodiment of the present invention, a process of forming a data tag in this embodiment includes:

step 321: phased source data is acquired.

The staged source data refers to concurrent or separate continuous state data within a unit time length or a period time length. It may be continuous static data or dynamic data corresponding to traffic status.

Step 322: and filtering the staged source data through an industry customized classification model to obtain the dominant classification data.

The filtered data or the retrieved data formed by using the keyword parameters of the classification category in the model belongs to the dominant classification data.

Step 323: and adjusting the classification data by using the classification threshold value to form preliminary classification class data.

The classification threshold corresponds to the classification rule, and the classification threshold acts on data such as retrieval and filtration formed by the classification rule, adjusts the data migration degree, and adjusts the data classification.

Step 324: and performing the same data identification on each type of preliminary classification category data to form a data label.

The data labels take the classification category as a data feature of an individual dimension, so that the linear classification features of the preliminary classification category data are quantified.

The data classification optimization method provided by the embodiment of the invention filters and classifies the source data by using the rule keywords and the rule weights of the classification classes in the industry customized classification model and identifies the source data, can fully utilize the processing resources and the storage resources of a computer system, efficiently finishes dominant feature classification and forms data labels, and improves the whole data classification efficiency.

As shown in fig. 11, in an embodiment of the present invention, the process of forming the data training set in this embodiment includes:

step 331: and taking the data label as the characteristic data of each preliminary classification category data.

Step 332: and forming the feature dimension and the feature vector of each preliminary classification category data according to each type of feature data.

Step 333: and forming a data training set according to the feature dimension and the feature vector.

According to the data classification optimization method, the corresponding training set is formed according to the requirements of the intelligent classification model, the training set comprises necessary data characteristic dimensions and quantization vectors, the data labels are used as the basis of implicit characteristic calculation formed by explicit characteristics, and the explicit characteristics formed by processing resources and storage resources of a computer system are used for replacing training data characteristics of supervised learning, so that the formation of required grades is efficient and high-quality.

As shown in fig. 11, in an embodiment of the present invention, the process of optimizing and classifying in this embodiment includes:

step 341: and carrying out iterative optimization on the intelligent classification model by changing the data scale of the training set.

The training set data scale can be a continuous local data training set, a continuous complete data training set and a random data training set, and the training sets are applied to the same intelligent classification model one by one.

The intelligent classification model in one embodiment of the invention adopts a naive Bayes model.

Step 342: and classifying all source data through an intelligent classification model.

The intelligent classification model can form the integral classification of all source data, and the analysis and the classification of the data with determined service scale are met.

Step 343: and classifying the incremental source data through an intelligent classification model.

The intelligent classification model can form continuous classification of incremental data, and the analysis and classification of the determined business production data are met.

According to the data classification optimization method, the associated sub-training sets are formed through splitting of the training sets, iterative optimization is carried out on the intelligent classification model through data differences of the sub-training sets, and the classification efficiency and quality of the intelligent classification model on the implicit classification features are improved. And meanwhile, incremental and full data classification is carried out according to the generation time of the source data to obtain further detailed classification of the existing classified data.

The data classification optimizing device of one embodiment of the invention comprises:

the memory is used for storing the program codes corresponding to the processing procedures of the data classification optimization method in the embodiment;

and the processor is used for executing the program codes of the processing procedures of the data classification optimization method of the embodiment.

Fig. 12 shows an architecture of a data classification optimizing apparatus according to an embodiment of the present invention. In fig. 12, the present embodiment includes:

a classification model forming module 3310 for forming an industry customized classification model using the classification data structure;

the tag marking module 3320 is used for classifying the source data through an industry customized classification model to form a data tag;

a training set forming module 3330, configured to partition the source data into data training sets according to the data labels;

and the classification forming module 3340 is used for optimizing the intelligent classification model according to the data training set and completing classification of the source data.

As shown in fig. 12, in an embodiment of the invention, the classification model forming module 3310 further includes:

a topology structure forming unit 3311, configured to establish a classification category and a topology structure of the classification category according to the service requirement through the classification data structure;

a classification rule forming unit 3312 for establishing a corresponding classification rule according to the classification category;

a classification threshold value forming unit 3313 for setting a classification threshold value according to a classification rule;

a model forming unit 3314 for forming an industry customized classification model based on the topology, classification rules and classification thresholds.

As shown in fig. 12, in an embodiment of the present invention, the tag marking module 3320 further includes:

a source data obtaining unit 3321, configured to obtain the periodic source data;

the dominant classification unit 3322 is configured to filter the staged source data through an industry customized classification model to obtain dominant classification data;

a classification adjusting unit 3323, configured to adjust the classification data by using a classification threshold to form preliminary classification category data;

and the label marking unit 3324 is used for marking the same data on each type of the preliminary classification category data to form a data label.

As shown in fig. 12, in an embodiment of the present invention, the training set forming module 3330 further includes:

a tag feature forming unit 3331 configured to use the data tag as feature data of each preliminary classification category data;

the characteristic quantization unit 3332 is configured to form a characteristic dimension and a characteristic vector of each preliminary classification category data according to each type of characteristic data;

a training set composing unit 3333 for forming a data training set according to the feature dimensions and the feature vectors.

As shown in fig. 12, in an embodiment of the present invention, the classification forming module 3340 further includes:

the iterative optimization unit 3341 is used for performing iterative optimization on the intelligent classification model by changing the data scale of the training set;

the whole classification unit 3342 is used for classifying all the source data through an intelligent classification model;

and an increment classification unit 3343, configured to classify the increment source data through an intelligent classification model.

By using the optimization method of the big data classification rule of the embodiment, the data classification system formed by the classification optimization method of the data display of the embodiment, and the data classification system and the data classification result formed by the data classification optimization method of the embodiment have good data demand response characteristics. And a high-efficiency data interaction display technical scheme can be formed by combining the data structure of the data classification system.

Fig. 13 shows a data processing method for classifying an interactive interface according to an embodiment of the present invention. In fig. 13, an embodiment of the present invention includes:

step 410: and retrieving classification category data according to the first interactive data to form a classification category first topology of the relevant classification categories.

The first interactive data comprises search keywords which can be classification category keywords or words with similar meanings or characters to the classification category keywords, the closest related classification categories are obtained through similarity matching of the search keywords and the classification category keywords, and a classification category first topological structure is formed according to the inclusion logic of classification category data below the related classification categories.

Step 420: and forming ordered display of the keywords of the classification categories according to the first topological structure of the classification categories.

The taxonomy-class first topology can form an ordered display of trees, tiers, or groupings within the display framework through sophisticated data presentation techniques.

Step 430: and forming a first classification result data set adaptive to the retrieval result according to the classification type first topological structure, and displaying the classification result data in order.

The classification classes (including dominant and stealth attribute classification classes) retained in the classification class data structure system have corresponding classification data, and the classification data is formed according to a data classification model or a keyword classification system formed by corresponding training.

Step 440: and forming classification category combinational logic according to the second interactive data, forming a classification category second topological structure according to the classification category combinational logic, and displaying the classification category keywords in order according to the classification category second topological structure.

The second interactive data comprises selection of classification categories of topology nodes (classification categories of different node positions in the data structure) in the first topology structure, selection of combinations or choices of nodes, and selection of keywords of the classification categories of the topology nodes in the first topology structure to embody classification category combination logic.

And the determined classification category combination logic forms the combination or the rejection of the classification categories, and further forms a second topological structure of the classification categories.

The taxonomy class second topology can form an ordered adjustment of trees, tiers, or groupings within the display framework through sophisticated data presentation techniques.

Step 450: and forming a second classification result data set adaptive to the retrieval result according to the classification type second topological structure, and displaying the classification result data in order.

The adaptation includes sorting, deduplication, or indexing of data in different classification data.

The classification data structure and classification category formed in the data classification process of the data processing method of the classification interactive interface and the classification data formed by the data classification source data are combined with the display technology to form timely classification, classification combination and data matching display of mass retrieval data, so that the defects that the data processing process in the existing classification interactive interface is influenced by the interactive process, the matching degree of the search process and the data classification dimension is limited, and the data positioning and data matching combination cannot be quickly formed are overcome. The retrieval classification process in the interactive interface can be orderly combined with a classification system formed by a classification data structure, so that data information formed by data classification can be fully displayed in the interactive process. Fig. 14 shows a data processing method for classifying an interactive interface according to an embodiment of the present invention. In fig. 14, the process of forming the first topology of the classification category according to the embodiment of the present invention includes:

step 411: similar keywords are determined in the classification category data according to the first interaction data.

The first interaction data includes a query keyword, a fuzzy keyword, or a text paragraph.

Step 412: and determining related classification categories according to the similar keywords.

And comparing the similarity by using a fuzzy matching algorithm of characters and the weight parameters of the rules in the classification category data to determine the related classification category.

Step 413: and establishing a classification category first topological structure according to the related classification category and the classification category data structure.

Relative node positions and associated lower or upper node positions in the classification category data structure are obtained by determining the relevant classification categories, thereby forming a topology structure of the associated classification categories.

Fig. 14 shows a data processing method for classifying an interactive interface according to an embodiment of the present invention. In fig. 14, the process of forming the ordered display of the keywords of the classification category according to the embodiment of the present invention includes:

step 421: and forming first display data of an optimized topological structure by using the classification category keywords in the classification category first topological structure according to the first topological structure.

According to the classification category data structure in the data classification embodiment, the classification category first topology structure includes category information such as keywords, classification rules, and classification weights corresponding to the classification category.

Step 422: and displaying the first display data according to the data display strategy of the display frame.

The first display data includes a topological mapping structure between the keywords of the classification category, and has a tree data structure between words and phrases.

Fig. 14 shows a data processing method for classifying an interactive interface according to an embodiment of the present invention. In fig. 14, the process of forming the first classification result data set according to the embodiment of the present invention includes:

step 431: classification classes in the first topology are determined.

And according to the classification category, each data node of the first topological structure can obtain a determined classification category.

Step 432: and determining corresponding classification data according to the classification categories.

The classification category data structure according to the data classification embodiment described above indicates that the source data may form classification data.

Step 433: the classification data is de-registered and a first classification result data set is formed.

Redundant data exists in the classification data under the influence of the diversity of the classification features.

Step 434: and displaying the first classification result data set according to the data display strategy of the display frame.

The first classification result data set is used as query or retrieval result data to form display contents.

Fig. 14 shows a data processing method for classifying an interactive interface according to an embodiment of the present invention. In fig. 14, the process of forming the second topology of the classification category according to the embodiment of the present invention includes:

step 441: an interactive selection of a category-classified keyword is received.

Interactive selection includes selection of keywords, including addition or exclusion, for the category of classification.

Step 442: and determining classification category combinational logic according to the interactive selection to form a classification category second topological structure.

And performing attribution logic judgment on the corresponding data structure node on a selection result formed after the keyword is selected, and further forming a topological structure.

Step 443: and forming second display data of the optimized topological structure according to the classified second topological structure and displaying according to the data display strategy of the display frame.

The second display data includes a topological mapping structure between the classified category keywords and has a tree data structure between words.

Fig. 14 shows a data processing method for classifying an interactive interface according to an embodiment of the present invention. In fig. 14, the process of forming the second classification result data set according to the embodiment of the present invention includes:

step 451: and determining corresponding classification data according to the change of the second topological structure of the classification category.

Step 452: the classification data is de-registered and a second classification result data set is formed.

Step 453: and displaying the second classification result data set according to the data display strategy of the display frame along with the display of the second display data.

The data processing device for classifying the interactive interface in one embodiment of the invention comprises:

the memory is used for storing program codes corresponding to the processing procedures of the data processing method of the classified interactive interface in the embodiment;

and the processor is used for executing the program codes of the processing procedures of the data processing method of the classified interactive interface of the embodiment.

FIG. 15 shows a data processing apparatus for classifying an interactive interface according to an embodiment of the present invention. In fig. 15, the present embodiment includes:

a first structure forming module 4410 for retrieving classification category data from the first interaction data to form a classification category first topology of related classification categories;

a first structure display module 4420, configured to form a sorted category keyword ordered display according to the sorted category and the first topology structure;

a first data presentation module 4430, configured to form a first classification result data set adapted to the search result according to the classification type first topology structure, and perform ordered presentation of classification result data;

a second structure forming module 4440, configured to form a classification category combinational logic according to the second interactive data, form a classification category second topological structure according to the classification category combinational logic, and display the classification category keywords in order according to the classification category second topological structure;

the second data displaying module 4450 is configured to form a second classification result data set adapted to the search result according to the classification category and the second topology structure, and perform ordered display on the classification result data.

As shown in fig. 15, in one embodiment of the present invention, the first structure forming module 4410 comprises:

a similar vocabulary determining unit 4411 for determining similar keywords in the classification category data according to the first interaction data;

a classification determining unit 4412 for determining a related classification category according to the similar keyword;

a first topology establishing unit 4413, configured to establish a classification category first topology according to the relevant classification category and classification category data structure.

As shown in fig. 15, in an embodiment of the invention, the first structure representation module 4420 includes:

a first display planning unit 4421, configured to form a first display data of an optimized topology from the classification category keywords in the first topology according to the first topology;

the first display transmitting unit 4422 is configured to display the first display data according to a data display policy of the display frame.

As shown in fig. 15, in an embodiment of the invention, the first data presentation module 4430 includes:

a category determining unit 4431 for determining a category in the category-first topology;

a classification data determining unit 4432 for determining corresponding classification data according to the classification category;

a classification data integration unit 4433, configured to de-superpose the classification data and form a first classification result data set;

a classified data transmission unit 4434, configured to display the first classified result data set according to the data display policy of the display frame.

As shown in fig. 15, in an embodiment of the invention, the second structure forming module 4440 includes:

an interaction establishing unit 4441, configured to receive an interaction selection of a category-classified keyword;

a second topology determining unit 4442, configured to determine a classification category combinational logic to form a classification category second topology structure according to the interaction selection;

a second data transmission unit 4443, configured to form second display data of the optimized topology according to the classified class second topology and display according to the data display policy of the display frame.

As shown in fig. 15, in an embodiment of the invention, the second data presentation module 4450 includes:

a second classification determining unit 4451, configured to determine corresponding classification data according to the change of the second topology of the classification category;

a second data integration unit 4452, configured to de-superpose the classified data and form a second classification result data set;

a data set transmission unit 4453 for displaying the second classification result data set according to the data display policy of the display frame following the display of the second display data.

In an embodiment of the present invention, the processor may adopt a dsp (digital Signal processing) digital Signal processor, an FPGA (Field-Programmable Gate Array), an mcu (microcontroller unit) system board, an soc (system on a chip) system board, or a plc (Programmable Logic controller) minimum system including I/O.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A classification optimization method for data presentation is characterized by comprising the following steps:

storing classification rules for the classification categories by the classification data structure;

the establishing a classification data structure for storing classification categories comprises:

forming a classification category base framework;

forming a set of classification categories in the base framework;

forming a classification category grade in the classification category set;

2. The method for classification optimization of data presentation according to claim 1, wherein said establishing the classification category through the classification data structure storage comprises:

3. The method of classification optimization of data presentation of claim 2, wherein said storing classification rules for said classification categories by said classification data structure comprises:

4. A data presentation classification optimization apparatus, comprising:

a memory for storing program codes corresponding to the processing procedures of the data presentation classification optimization method according to any one of claims 1 to 3;

a processor for executing the program code.

5. A data presentation classification optimization apparatus, comprising:

a classification rule association module for storing the classification rules of the classification categories through the classification data structure;

the classification structure formation module includes:

a frame forming unit for forming a classification category base frame;

6. The data presentation classification optimization device of claim 5, wherein the classification content storage module comprises:

7. The data presentation classification optimization device of claim 6, wherein the classification rule association module comprises: