CN115660738A - E-commerce transaction method and system based on big data - Google Patents

Abstract

According to the E-commerce transaction method and system based on the big data, under the condition that the digital E-commerce user keywords with different interaction behavior activeness are analyzed, the E-commerce service interaction big data are searched and matched through the service behavior preference knowledge mined by the updated user behavior preference mining model, the matching precision of the E-commerce service interaction big data with different interaction behavior activeness can be improved, and therefore the precision of determining the personalized E-commerce service requirements of the digital E-commerce user keywords to be analyzed based on the personalized E-commerce service requirements of the matched E-commerce service interaction big data is improved.

Description

E-commerce transaction method and system based on big data

Technical Field

The invention relates to the technical field of big data, in particular to an E-commerce transaction method and system based on big data.

Background

The term Electronic commerce is derived from Electronic Business, that is, business transactions conducted electronically. By using electronic tools such as the internet and the like, the electronic business is used for sharing information in companies, suppliers, clients and partners, the electronization of business processes among enterprises is realized, and the efficiency of each link of production, inventory, circulation and the like of the enterprises is improved by matching with an electronized production management system in the enterprises. At present, the e-commerce transaction is more focused on 'thousands of people and thousands of faces', namely personalized and differentiated services are provided for different e-commerce users, so that the purpose of accurately mining the requirements of different e-commerce users is very important, but the related technology is difficult to implement.

Disclosure of Invention

In order to solve the technical problems in the related art, the invention provides an E-commerce transaction method and system based on big data.

In a first aspect, an embodiment of the present invention provides a method for debugging a user behavior preference mining model, which is applied to a big data e-commerce platform system, and the method includes: obtaining example type e-commerce business interaction big data of behavior activity reference information of example type digital e-commerce user keywords of which annotation addition is completed; determining an error evaluation index of a to-be-debugged user behavior preference mining model by combining the behavior liveness reference information of the exemplary digital electric power business user keywords and the corresponding categories of the exemplary digital electric power business user keywords; updating a first commonality score between the exemplary digital electric power business user keyword and a category corresponding to the exemplary digital electric power business user keyword by combining the error evaluation index to obtain a model cost of the behavior preference mining model of the user to be debugged; and updating the model variables of the user behavior preference mining model to be debugged by combining the model cost so as to enable the model cost generated by the updated user behavior preference mining model to be in a stable state.

In some optional designs, the determining, by combining the behavior liveness reference information of the example-type digital electric appliance user keyword and the category corresponding to the example-type digital electric appliance user keyword, an error evaluation index of the user behavior preference mining model to be debugged includes: determining a first interactive behavior liveness of the example digital electric appliance user keyword in combination with the behavior liveness reference information of the example digital electric appliance user keyword; determining a second interactive behavior activity degree of a category corresponding to the example type digital television user keyword; determining a first comparison result of the exemplary digital electric appliance user keyword under the corresponding category by combining the first interactive behavior liveness of the exemplary digital electric appliance user keyword and the second interactive behavior liveness of the category corresponding to the exemplary digital electric appliance user keyword; and determining an error evaluation index corresponding to the example type digital electric power business user keyword by combining the first comparison result of the example type digital electric power business user keyword.

According to the design, a first interactive behavior activity determined by means of behavior activity reference information of example type electricity business interaction big data and a second interactive behavior activity of a type corresponding to example type digital electricity business user keywords are determined, an error evaluation index is determined according to a first comparison result, so that a model cost is obtained by means of updating a first common value through the error evaluation index, and then business behavior preference knowledge which is more beneficial to analysis of different interactive behavior activities is mined out through a user behavior preference mining model through the model cost.

In some optional designs, the determining the second interactive activity level of the category corresponding to the exemplary digital television user keyword includes: in an exemplary cloud sharing server, determining that the data content contains at least one group of target setting electric business interaction big data samples of alternative exemplary digital electric business user keywords which are consistent with the types corresponding to the exemplary digital electric business user keywords; the exemplary cloud sharing server is used for recording a plurality of sets of set electric business interaction big data examples of digital electric business user keywords of which the data contents comprise behavior liveness reference information of which the annotation addition is completed; and determining a second interactive activity degree of the type corresponding to the example type digital electric power business user keyword by combining the reference information of the activity degree of the alternative example type digital electric power business user keyword in the at least one group of target setting electric power business interaction big data sample.

By means of the design, the second interactive activity of the type can be determined according to the activity reference information matched with the example type digital electric power business user keywords to the same type of example type digital electric power business user keywords, and therefore the corresponding error evaluation index can be determined through the errors of the first interactive activity of each example type digital electric power business user keyword and the second interactive activity of the type.

In some optional designs, the determining, in combination with the reference information of the activity level of the alternative example type digital electric utility user keyword in the at least one set of target setting electric utility interaction big data sample, a second interaction activity level of a category corresponding to the example type digital electric utility user keyword includes: determining an interactive behavior activity mean value of a category corresponding to the example type digital electric power business user keyword by combining with the behavior activity reference information of the alternative example type digital electric power business user keyword in the at least one group of target setting electric power business interaction big data sample; and taking the average value of the interactive behavior liveness of the type corresponding to the example type digital electric appliance user keyword as the second interactive behavior liveness.

By means of the design, the average value of the interactive behavior liveness of the exemplary digital electric power business user keywords of the same type is determined according to the behavior liveness reference information matched with the exemplary digital electric power business user keywords of the same type with the exemplary digital electric power business user keywords, the average value of the interactive behavior liveness is used as the second interactive behavior liveness of the type, and therefore the corresponding error evaluation index is determined according to the error between the interactive behavior liveness of each exemplary digital electric power business user keyword and the average value of the interactive behavior liveness of the type, and the knowledge extraction of the exemplary digital electric power business user keywords with large errors of the average value of the interactive behavior liveness is strengthened by the user behavior preference mining model.

In some optional designs, the determining, in combination with the reference information of the activity level of the alternative example type digital electric utility user keyword in the at least one set of target setting electric utility interaction big data sample, a second interaction activity level of a category corresponding to the example type digital electric utility user keyword includes: determining target interaction behavior activity of a type corresponding to the example type digital electric appliance user keyword by combining with the behavior activity reference information of the alternative example type digital electric appliance user keyword in the at least one group of target setting electric appliance business interaction big data sample; and taking the target interactive behavior activity of the type corresponding to the example type digital television business user keyword as the second interactive behavior activity.

By means of the design, the target interactive behavior liveness of the exemplary digital electric power business user keywords of the same type is determined according to the behavior liveness reference information matched with the exemplary digital electric power business user keywords of the same type with the exemplary digital electric power business user keywords, the target interactive behavior liveness is used as the second interactive behavior liveness of the type, and therefore the corresponding error evaluation indexes are determined by means of the errors of the interactive behavior liveness of each exemplary digital electric power business user keyword and the target interactive behavior liveness of the type, and the knowledge extraction of the exemplary digital electric power business user keywords with large errors of the target interactive behavior liveness is strengthened by the user behavior preference mining model.

In some optional designs, the determining, by combining the first interactive activity level of the example digital tv provider user keyword and the second interactive activity level of the category corresponding to the example digital tv provider user keyword, a first comparison result of the example digital tv provider user keyword under the corresponding category includes: determining a first operation result of a first interactive behavior activity degree of the example type digital electric power business user keyword and a second interactive behavior activity degree of a category corresponding to the example type digital electric power business user keyword; and taking the mapping value of the first operation result as a first comparison result of the exemplary digital electric power business user keyword under the corresponding category.

By designing in this way, according to the mapping value of the first operation result between the first interactive behavior activity degree of the example type digital electric business user keyword of each example type electric business interaction big data and the second interactive behavior activity degree of the affiliated category, the first comparison result of the example type digital electric business user keyword under the affiliated category can be determined.

In some optional designs, the determining, by combining the first interactive activity level of the example digital tv provider user keyword and the second interactive activity level of the category corresponding to the example digital tv provider user keyword, a first comparison result of the example digital tv provider user keyword under the corresponding category includes: determining proportion data between a first interactive behavior activity of the example type digital electric power business user keyword and a second interactive behavior activity of a category corresponding to the example type digital electric power business user keyword; and taking the mapping value of the first operation result of the proportion data and the selected value as a first comparison result of the exemplary digital television user keyword under the corresponding category.

By designing in this way, according to the proportion data between the first interactive activity degree of the example type digital electric power business user keyword of each example type electric power business interaction big data and the second interactive activity degree of the affiliated category, the first comparison result of the example type digital electric power business user keyword under the affiliated category can be determined.

In some optional designs, the determining, in combination with the first comparison result of the example digital electric utility user keyword, an error rating index corresponding to the example digital electric utility user keyword includes: determining a maximum first comparison result from the first comparison results of the plurality of digital electric company user keywords of the same category as a second comparison result; wherein a category of the plurality of digital electric power consumer keywords of the same category is identical to a category of the exemplary digital electric power consumer keyword; determining a quantized difference between the first comparison result and the second comparison result of the exemplary digital television user keyword; and updating the quantization difference value by adopting a set updating instruction to obtain an error evaluation index corresponding to the keyword of the exemplary digital electric power business user.

By means of the design, the quantitative difference between the first comparison result of the example-type digital electric appliance user keyword and the second comparison result of the type corresponding to the example-type digital electric appliance user keyword is determined, the quantitative difference is updated by means of the set variable serving as the updating variable, the error evaluation index of the example-type electric appliance business interaction big data is determined, the first common score is updated by means of the error evaluation index to obtain the model cost, and then the business behavior preference mining model is mined out through the model cost to obtain business behavior preference knowledge which is more beneficial to activity analysis of different interaction behaviors.

In some optional design solutions, before the updating, in combination with the error evaluation index, a first commonality score between the example-type digital electric power business user keyword and a category corresponding to the example-type digital electric power business user keyword to obtain a model cost of the mining model of the user behavior preference to be debugged, the method further includes: performing behavior preference mining on the exemplary digital electric power business user keywords by combining the to-be-debugged user behavior preference mining model to obtain business behavior preference knowledge of the exemplary digital electric power business user keywords; performing behavior preference mining on alternative example type digital electric appliance user keywords which are consistent with the types corresponding to the example type digital electric appliance user keywords by combining the to-be-debugged user behavior preference mining model to obtain business behavior preference knowledge of the types corresponding to the example type digital electric appliance user keywords; determining the first common score between the business behavior preference knowledge of the example digital electric appliance user keyword and the business behavior preference knowledge of the category corresponding to the example digital electric appliance user keyword.

The design is that based on a to-be-debugged user behavior preference mining model, behavior preference mining is respectively carried out on an example type digital electric appliance user keyword and an alternative example type digital electric appliance user keyword of a type corresponding to the example type digital electric appliance user keyword to obtain business behavior preference knowledge of the example type digital electric appliance user keyword and business behavior preference knowledge of a type corresponding to the example type digital electric appliance user keyword, a first common score between the business behavior preference knowledge of the example type digital electric appliance user keyword and the business behavior preference knowledge of the type corresponding to the example type digital electric appliance user keyword can be determined, and therefore the first common score can be updated by means of an error evaluation index to obtain model cost; in view of more rigorous requirements on multivariate regression analysis, the service behavior preference knowledge dispersity of the electric business interaction big data mapped in the knowledge characteristic coordinate system is as low as possible, so that the user behavior preference mining model can strengthen the knowledge refinement of the exemplary digital electric business user keywords for setting the interaction behavior liveness, and the updated user behavior preference mining model can mine the service behavior preference knowledge more beneficial to recognition of different interaction behavior liveness.

In some optional designs, on the basis that the example digital electric utility user keyword is a registered member keyword, the alternative example digital electric utility user keyword is the registered member keyword with different interactive activity liveness, and in the example cloud sharing server, determining that the data content contains at least one set of target-setting electric utility interaction big data sample of the alternative example digital electric utility user keyword of which the category corresponds to the example digital electric utility user keyword includes: in the exemplary electric business interaction big database, determining that the data content comprises not less than one group of target setting electric business interaction big data examples of the registered member keywords with different interaction behavior liveness; the determining, by combining the reference information of the activity levels of the alternative exemplary digital electric appliance user keywords in the at least one set of target setting electric business interaction big data samples, a second interaction activity level of a category corresponding to the exemplary digital electric appliance user keyword includes: and determining a second interactive activity degree of the registered member keywords by combining different interactive activity degrees of the registered member keywords annotated in the at least one group of target setting e-commerce interaction big data samples.

By the design, under the condition of carrying out requirement mining on e-commerce member business interaction big data, at least one behavior activity reference information of an e-commerce business interaction big data sample is set at least one group of targets according to the same registered member keyword, and a second interaction behavior activity degree taking the registered member keyword as the keyword type of the digital e-commerce user can be determined; setting a first interactive behavior activity of the e-commerce business interaction big data sample in each group of targets by means of the registered member keywords, and determining a corresponding error evaluation index according to the error of the corresponding second interactive behavior activity; the user behavior preference mining model can further conduct reinforced mining on knowledge vectors of the interactive big data with different interactive behavior activeness, and the updated user behavior preference mining model can mine the interactive big data service behavior preference knowledge which is more beneficial to recognition of different interactive behavior activeness; in other words, when the E-commerce operator business interaction big data matching with different interaction behavior liveness degrees is carried out by means of the user behavior preference mining model, the precision of the E-commerce business interaction big data matching with different interaction behavior liveness degrees can be improved.

In a second aspect, an embodiment of the present invention further provides an e-commerce transaction method based on big data, which is applied to a big data e-commerce platform system, and the method includes: obtaining first e-commerce business interaction big data containing a digital e-commerce user keyword to be analyzed; performing behavior preference mining on the digital electric appliance user keywords to be analyzed by combining a user behavior preference mining model to obtain service behavior preference knowledge of the digital electric appliance user keywords to be analyzed; the user behavior preference mining model is obtained by debugging based on the debugging method of the user behavior preference mining model; retrieving at least one group of second e-commerce interaction big data matched with the first e-commerce interaction big data in a set cloud sharing server in combination with the business behavior preference knowledge; the cloud sharing server is set to comprise at least one group of E-commerce service interaction big data of personalized E-commerce service requirements of the added annotation-finished prior user keywords; the at least one group of second e-commerce interaction big data comprises prior user keywords with different interaction behavior liveness; and determining the personalized e-commerce service requirements of the digital e-commerce user keywords to be analyzed by combining the personalized e-commerce service requirements of the prior user keywords with different interactive behavior liveness degrees in the set cloud sharing server.

In a third aspect, the invention further provides a big data e-commerce platform system, which comprises a processor and a memory; the processor is connected with the memory in communication, and the processor is used for reading the computer program from the memory and executing the computer program to realize the method.

In a fourth aspect, the invention also provides a computer-readable storage medium, on which a program is stored, which program, when executed by a processor, performs the method described above.

In the embodiment of the invention, firstly, obtaining the example type electric business interaction big data added with the behavior liveness reference information of the example type digital electric business user keyword; thus, the example type E-commerce interaction big data of the example type digital E-commerce user keywords with different interaction activity degrees in one category can be obtained; secondly, determining an error evaluation index for updating the behavior preference mining model cost of the user to be debugged when the exemplary digital electric power business user keyword is used as a debugging exemplary type based on the behavior liveness reference information of the exemplary digital electric power business user keyword and the type corresponding to the exemplary digital electric power business user keyword; thus, the error evaluation index corresponding to each example type digital electric power business user keyword is determined by means of the relation between the interactive behavior activity of the example type digital electric power business user keyword and the interactive behavior activity corresponding to the belonged type; thirdly, updating a first commonality score between the example type digital electric power business user keyword and the category corresponding to the example type digital electric power business user keyword by means of the error evaluation index to obtain a model cost of the behavior preference mining model of the user to be debugged; in this way, the first common score is updated through the error evaluation index of each exemplary digital electric power business user keyword to obtain the model cost corresponding to each exemplary digital electric power business user keyword; finally, debugging the to-be-debugged user behavior preference mining model based on the model cost, and updating the model variables of the user behavior preference mining model so as to enable the model cost generated by the updated user behavior preference mining model to be in a stable state; therefore, the user behavior preference mining model can conduct reinforced mining on knowledge vectors of the example type digital electric power business user keywords with different interactive behavior liveness degrees by means of model costs corresponding to the example type digital electric power business user keywords with different interactive behavior liveness degrees, and business behavior preference knowledge which is more beneficial to recognition of different interactive behavior liveness degrees can be mined by the updated user behavior preference mining model.

It can be understood that model cost is updated for each exemplary digital electric power business user keyword by means of the behavior activity reference information, the user behavior preference mining model is debugged based on the model cost, the user behavior preference mining model can conduct reinforced mining on knowledge vectors of the exemplary digital electric power business user keywords with different interaction behavior activities, the updated user behavior preference mining model can mine business behavior preference knowledge which is more beneficial to recognition of different interaction behavior activities, and therefore under the condition of analyzing the digital electric power business user keywords with different interaction behavior activities, the business behavior preference knowledge mined by the updated user behavior preference mining model is used for retrieving and matching electric power business interaction big data, and accuracy of matching the electric business interaction big data with different interaction behavior activities can be improved.

Drawings

Fig. 1 is a schematic flow chart of an e-commerce transaction method based on big data according to an embodiment of the present invention.

Fig. 2 is a communication architecture diagram of an application environment of a big data-based e-commerce transaction method according to an embodiment of the present invention.

Detailed Description

The method provided by the embodiment of the invention can be executed in a big data e-commerce platform system, a computer device or a similar operation device. Taking the example of the large data e-commerce platform system running on the large data e-commerce platform system, the large data e-commerce platform system 10 may include one or more processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), and a memory 104 for storing data, and optionally, the large data e-commerce platform system may further include a transmission device 106 for communication function.

Based on this, please refer to fig. 1, fig. 1 is a schematic flowchart of a debugging method for a user behavior preference mining model according to an embodiment of the present invention, where the method is applied to a big data e-commerce platform system, and further includes the technical solutions described in STEP101-STEP 104.

STEP101, obtaining example e-commerce business interaction big data of activity reference information of the example digital e-commerce user keyword with completed annotation adding.

For some example embodiments, the big data e-commerce platform system obtains not less than one group of example type e-commerce interaction big data of the activity reference information of the example type digital e-commerce user keyword for which annotation addition is completed, and the example type digital e-commerce user keyword of each group of the not less than one group of example type e-commerce interaction big data belongs to at least one category, and each category corresponds to the example type e-commerce interaction big data of the activity reference information of the not less than one group of example type digital e-commerce user keyword for which annotation addition is completed.

In other words, the big data e-commerce platform system obtains not less than one group of example type e-commerce interaction big data of each kind of example type digital e-commerce user keyword under at least one kind, and each group of example type e-commerce interaction big data is added with the behavior activity reference information of the example type digital e-commerce user keyword in advance; wherein, not less than one group of example type electric business interaction big data of the same kind of example type digital electric business user keywords can be obtained by collecting the interaction data of the same digital electric business user, and the behavior activity degree reference information (behavior activity degree label) of the example type digital electric business user keywords is annotated to be the interaction behavior activity degree of the digital electric business user corresponding to the example type digital electric business user keywords included in the data content when the example type electric business interaction big data is collected. In the embodiment of the invention, the idea of obtaining the example type electric business interaction big data which is annotated and added by the big data electric business platform system can be selected from the electric business interaction big data which is annotated and added, and the electric business interaction big data which is annotated and added can be recorded in the example type cloud sharing server; or, the data can be obtained by adding the annotation to the electric business interaction big data without adding the annotation. In addition, the example electric business interaction big data can be understood as sample electric business interaction big data, and other example data information can be understood as sample data information.

For some exemplary embodiments, the exemplary digital e-commerce user keyword may be a registered member keyword or a guest user keyword; acquiring exemplary page interaction big data of the exemplary digital electric appliance user keyword on the basis that the exemplary digital electric appliance user keyword is a registered member keyword; or obtaining example type voice interaction big data or example type page interaction big data of the example type digital electric telephone user keywords on the basis that the example type digital electric telephone user keywords are the tourist user keywords; as can be seen, the example e-commerce interaction big data may be page interaction big data or voice interaction big data. It is understood that not less than one group of electric business interaction big data can be from the electric business log, and one electric business log can include a plurality of electric business statistics reports, each electric business statistics report corresponding to one group of electric business interaction big data respectively.

Further, the interactive activity level is used for reflecting the service interaction heat or the service interaction frequency of the data electric business user corresponding to the digital electric business user keyword, so that the big data electric business platform system can obtain the example electric business interaction big data of the example digital electric business user keyword in different interactive activity levels in one kind.

And the STEP102 determines an error evaluation index of the behavior preference mining model of the user to be debugged by combining the behavior liveness reference information of the exemplary digital electric power business user keyword and the category corresponding to the exemplary digital electric power business user keyword.

For some exemplary embodiments, for each exemplary e-commerce interaction big data, the big data e-commerce platform system determines the interaction activity of the exemplary digital e-commerce user keyword according to the interaction activity annotated by the behavior activity reference information of the exemplary digital e-commerce user keyword, determines the corresponding interaction activity of the category according to the category to which the exemplary digital e-commerce user keyword belongs, determines a link between the interaction activity of the exemplary digital e-commerce user keyword and the corresponding interaction activity of the category to which the exemplary digital e-commerce user keyword belongs, and obtains an error evaluation index corresponding to the exemplary digital e-commerce user keyword when the set of exemplary e-commerce interaction big data is taken as a debugging exemplary type of the to-be-debugged user behavior preference mining model according to the determined link. The determination idea of the corresponding interactive behavior liveness of the category can be used for retrieving the corresponding interactive behavior liveness set for the category according to the category to which the exemplary digital television user keywords belong; or, according to the behavior activity reference information of at least one example type digital electric power provider user keyword of the kind corresponding to the example type digital electric power provider user keyword, the average value, the maximum/minimum value or the median value of the interaction behavior activity of the example type digital electric power provider user keyword of the kind may be used as the corresponding interaction behavior activity of the kind. The error evaluation index can be an updated variable of the strictness of the multiple regression condition in the model cost index (loss function), the classification processing rule is more rigorous by increasing the strictness of the multiple regression condition, the dispersion degree of the business behavior preference knowledge mapped by the same type of electric business interaction big data in the knowledge characteristic coordinate system is as low as possible, the business behavior preference knowledge mapped by different types of electric business interaction big data in the knowledge characteristic coordinate system is as differentiated as possible, in other words, the strictness of the multiple regression condition is updated to ensure the intra-group centralization and the inter-group differentiation.

For some exemplary embodiments, the error evaluation index corresponding to the exemplary digital electric power consumer keyword may be determined according to an operation relationship (which may be expressed as a comparison result, such as a first operation result (difference calculation) and proportional data) between the interactive behavior activity of the exemplary digital electric power consumer keyword and the corresponding interactive behavior activity of the corresponding category, by an algorithm formula or a form query; or, the error evaluation index can be determined through form query according to the magnitude relation between the interactive behavior activity of the example type digital television business user keyword and the corresponding interactive behavior activity of the affiliated category. For the scheme of determining the error evaluation index by the algorithm formula, the error evaluation index and the comparison result have a first quantization relationship (positive correlation) or a second quantization relationship (negative correlation) based on the algorithm formula; for the idea of determining the error evaluation index for form query, a hierarchical error evaluation index is set in advance. Explaining by taking a scheme with a first quantitative relation between the error evaluation index and a comparison result as an example, in the process of carrying out model debugging by using big interaction data of the example type electric business services with different interaction activity degrees of the same type, the larger the comparison result is, the larger the difference between the interaction activity degree of the example type digital electric business user keyword and the corresponding interaction activity degree of the type to which the keyword belongs is, and the larger the error evaluation index is; equivalently, by improving the processing requirement, when the user behavior preference mining model is debugged by means of multiple groups of example type electric business interaction big data of the same type, the business behavior preference knowledge of example type digital electric business user keywords with large comparison results on the interaction behavior liveness level is intensively mined, and the business behavior preference knowledge which is more beneficial to analysis of different interaction behavior liveness degrees is mined, so that the user behavior preference mining model can mine the business behavior preference knowledge which is more beneficial to identification of different interaction behavior liveness degrees.

Further, for example-type digital electric power business user keywords with different interactive behavior liveness of the same category, by means of the relationship between the interactive behavior liveness of the example-type digital electric power business user keywords and the interactive behavior liveness corresponding to the category, the error evaluation index corresponding to each example-type digital electric power business user keyword can be determined.

And the STEP103, in combination with the error evaluation index, updates a first commonality score between the example type digital electric appliance user keyword and a category corresponding to the example type digital electric appliance user keyword to obtain a model cost of the behavior preference mining model of the user to be debugged.

For some exemplary embodiments, the mining model of the behavior preference of the user to be debugged can be used for behavior preference mining and behavior preference differentiation, and at least one group of matched electric business interaction big data is determined according to the common score (similarity) of the mined business behavior preference knowledge (user behavior characteristics or user operation characteristics); the behavior preference mining function in the user behavior preference mining model to be debugged can be realized by adopting rules such as a moving average unit, and the behavior preference distinguishing function can be realized on the basis of the rules such as a down-sampling unit and/or a feature integration unit. Here, the mining model of the behavior preference of the user to be debugged may be an unmodulated model; alternatively, it may be a pre-debug model. According to the error evaluation index of the example digital electric business user keywords of each group of example electric business interaction big data, updating a first common score between the example digital electric business user keywords and the corresponding types of the example digital electric business user keywords in the model cost of the model mined according to the behavior preference of the user to be debugged to obtain the model cost; the updating of the first commonality score can be the difference between the business behavior preference knowledge of the exemplary digital electric appliance user keyword and the intra-cluster preference knowledge of the category corresponding to the exemplary digital electric appliance user keyword; or updating the Euclidean calculation value of the difference between the business behavior preference knowledge of the example type digital electric appliance user keyword and the intra-cluster preference knowledge of the category corresponding to the example type digital electric appliance user keyword. Here, the in-cluster preference knowledge of the category corresponding to the exemplary digital electric power business user keyword may be obtained from the business behavior preference knowledge of the alternative exemplary digital electric power business user keyword of the category through a K-means clustering rule, and the business behavior preference knowledge of the exemplary digital electric power business user keyword and the business behavior preference knowledge of the alternative exemplary digital electric power business user keyword may be obtained from the corresponding exemplary digital electric power business user keyword through user behavior preference mining model behavior preference mining, respectively.

For example digital electric power business user keywords with different interactive behavior liveness of the same kind, because the error evaluation indexes are determined according to the interactive behavior liveness of the example digital electric power business user keywords, model costs obtained by the example digital electric power business user keywords with each interactive behavior liveness according to the corresponding error evaluation indexes are different. In this way, the first common score is updated through the error evaluation index of each exemplary digital electric power business user keyword to obtain the model cost corresponding to each exemplary digital electric power business user keyword, so that the training expectation of the model to be debugged is guided by updating the classification requirements of some exemplary digital electric power business user keywords, and the user behavior preference mining model can perform enhanced mining on the business behavior preference knowledge of at least one exemplary digital electric power business user keyword with the same type of interactive behavior liveness.

And the STEP104 is used for updating the model variables of the mining model of the user behavior preference to be debugged by combining the model cost so as to enable the model cost generated by the updated mining model of the user behavior preference to be in a stable state.

For some exemplary embodiments, in a cycle phase, updating model variables in the to-be-debugged user behavior preference mining model based on model costs until the model costs of the to-be-debugged user behavior preference mining model are in a stable state, stopping the cycle phase, completing debugging of the to-be-debugged user behavior preference mining model, and obtaining the updated user behavior preference mining model.

In the process of debugging the model cost debugging network by using the example type digital electric appliance user keywords with different interactive behavior liveness degrees, the user behavior preference mining model carries out enhanced mining on the knowledge vectors of the example type digital electric appliance user keywords with different interactive behavior liveness degrees by using the model cost corresponding to the example type digital electric appliance user keywords with different interactive behavior liveness degrees, the updated user behavior preference mining model can mine the business behavior preference knowledge which is more beneficial to identifying different interactive behavior liveness degrees, and therefore under the condition of analyzing the digital electric appliance user keywords with different interactive behavior liveness degrees, the business behavior preference knowledge mined by the updated user behavior preference mining model is used for retrieving and matching the electric appliance business interaction big data, and the matching precision of the electric appliance business interaction big data with different interactive behavior liveness degrees can be improved.

In the embodiment of the invention, the example type electric business interaction big data added with the behavior activity reference information of the example type digital electric business user keywords are obtained, and the example type electric business interaction big data of the example type digital electric business user keywords with different interaction behavior activities in one kind can be obtained; and determining the error evaluation index corresponding to each example type digital electric power business user keyword by means of the relation between the interactive behavior liveness of the example type digital electric power business user keyword and the interactive behavior liveness corresponding to the belonged type. Updating a first commonality score between the exemplary digital electric power business user keyword and a category corresponding to the exemplary digital electric power business user keyword through the error evaluation index of each exemplary digital electric power business user keyword to obtain a model cost of a behavior preference mining model of a user to be debugged; the user behavior preference mining model can conduct reinforced mining on knowledge vectors of example type digital electric appliance user keywords with different interaction behavior liveness degrees by means of model costs corresponding to example type digital electric appliance user keywords with different interaction behavior liveness degrees, and the updated user behavior preference mining model can mine business behavior preference knowledge which is more beneficial to recognition of different interaction behavior liveness degrees. Therefore, model cost is updated for each example type digital electric appliance user keyword by means of the behavior activity reference information, the user behavior preference mining model is debugged based on the model cost, the user behavior preference mining model can conduct reinforced mining on knowledge vectors of the example type digital electric appliance user keywords with different interaction behavior activities, the updated user behavior preference mining model can mine business behavior preference knowledge which is more beneficial to recognition of different interaction behavior activities, and therefore under the condition of analyzing the digital electric appliance user keywords with different interaction behavior activities, the business behavior preference knowledge mined by the updated user behavior preference mining model is searched and matched with the electric appliance business interaction big data, and accuracy of matching of the electric business interaction big data with different interaction behavior activities can be improved.

In some embodiments, the STEP102 may be implemented based on the following description, and may determine the error rating index corresponding to the example digital tv provider user keyword based on the first interactive activity of the example digital tv provider user keyword and the second interactive activity of the category to which the example digital tv provider user keyword belongs.

STEP201, determining a first interactive activity of the example digital television user keyword in combination with the activity reference information of the example digital television user keyword.

For some exemplary embodiments, the behavior activity reference information is annotated with interaction behavior activity of the digital e-commerce user corresponding to the exemplary digital e-commerce user keyword when the exemplary e-commerce interaction big data is collected, and a first interaction behavior activity of the exemplary digital e-commerce user keyword is determined according to the interaction behavior activity annotated by the behavior activity reference information of the exemplary e-commerce interaction big data; for example, the behavior activity reference information of the example type e-commerce interaction big data is 10, that is, the interaction behavior activity of the digital e-commerce user corresponding to the example type digital e-commerce user keyword when the example type e-commerce interaction big data is collected is V _ active _10, and it is determined that the first interaction behavior activity of the example type digital e-commerce user keyword is V _ active _10.

STEP202, determining a second interactive activity liveness of the category corresponding to the example digital television user keyword.

For some exemplary embodiments, the second interactive activity set by the category may be retrieved according to the category corresponding to the exemplary digital television user keyword; or determining the second interactive activity of the type according to the activity reference information of at least one exemplary digital electric power business user keyword of the type corresponding to the exemplary digital electric power business user keyword; for example, a first interactive activity level of at least one exemplary type digital electric utility user keyword is determined according to the activity level reference information of the exemplary type digital electric utility user keyword in not less than one set of exemplary type electric utility interaction big data corresponding to the type of the exemplary type digital electric utility user keyword, and a second interactive activity level of the type is determined according to the first interactive activity level of the at least one exemplary type digital electric utility user keyword. In some embodiments, the average, maximum/minimum, or median value of the first interactive activity liveness of the at least one exemplary digital utility user keyword can be considered as the second interactive activity liveness of the category. The corresponding categories are consistent, and it can be shown that the example type digital e-commerce user keywords correspond to the same digital e-commerce user, and the example type e-commerce service interaction big data of the example type digital e-commerce user keywords are obtained by performing interaction data acquisition on the same digital e-commerce user.

In some embodiments that can be independent, the second interactive behavior activity can also be used to determine the in-cluster preference knowledge of the category in a subsequent digital power grid user analysis model debugging process, and the business behavior preference knowledge extracted by the exemplary digital power grid user keywords of the second interactive behavior activity of the category through the user behavior preference mining model is used as the in-cluster preference knowledge of the category.

And determining an error evaluation index for the exemplary digital electric power business user keyword as a model cost for debugging the exemplary digital electric power business user keyword based on the first interactive behavior liveness of the exemplary digital electric power business user keyword and the second interactive behavior liveness of the category to which the exemplary digital electric power business user keyword belongs, so that the error evaluation index corresponding to each exemplary digital electric power business user keyword is determined by means of the relation between the interactive behavior liveness of the exemplary digital electric power business user keyword and the interactive behavior liveness corresponding to the category to which the exemplary digital electric power business user keyword belongs.

STEP203, determining a first comparison result of the example type digital electric appliance user keyword under the corresponding category by combining the first interactive activity of the example type digital electric appliance user keyword and the second interactive activity of the category corresponding to the example type digital electric appliance user keyword.

For some example embodiments, for each example type of e-commerce interaction big data, determining a first comparison result of an example type of digital e-commerce user keyword under a category of the example type of e-commerce interaction big data according to an operational relationship between a first interaction behavior activity of the example type of digital e-commerce user keyword and a second interaction behavior activity of the category to which the example type of digital e-commerce user keyword belongs; the operational relationship between the first interactive activity and the second interactive activity may be an operational relationship for calculating a comparison result therebetween, such as the first operational result and/or the ratio data.

In some embodiments, the first comparison result of the exemplary digital television user keyword in the category is determined by the interactive activity liveness first operation result, and the STEP203 may be implemented based on the following STEPs 2031 and 2032.

And the STEP2031 is used for determining a first operation result of the first interactive activity liveness of the example type digital electric appliance user keyword and a second interactive activity liveness of the type corresponding to the example type digital electric appliance user keyword.

For some exemplary embodiments, for each exemplary electric business interaction big data, the first interaction activity degree of the exemplary digital electric business user keyword of the exemplary electric business interaction big data is V1, and the second interaction activity degree of the category to which the exemplary digital electric business user keyword belongs is V2, so as to obtain the first operation result of V1-V2.

STEP2032, using the mapping value of the first operation result as a first comparison result of the exemplary digital television user keyword under the corresponding category.

For some exemplary embodiments, the determined mapping value of the first operation result is used as a first comparison result of the keyword of the exemplary digital electric company user in the category; the first comparison result can reflect the difference of the interactive behavior activity between the first interactive behavior activity of the corresponding example-type digital electric appliance user keyword and the second interactive behavior activity of the belonging category, which is also called the differentiation of the interactive behavior activity; the larger the first comparison result is, the larger the distinction degree of the interactive activity between the first interactive activity degree of the exemplary digital television user keyword and the second interactive activity degree of the category to which the keyword belongs is.

In the embodiment of the invention, according to the mapping value of the first operation result between the first interactive behavior activity degree of the example type digital electric power business user keyword of each example type electric power business interaction big data and the second interactive behavior activity degree of the affiliated category, the first comparison result of the example type digital electric power business user keyword under the affiliated category can be determined.

In some embodiments, the STEP203 may be implemented based on the following STEPs 2033 and 2034 by determining the first comparison result of the exemplary digital television user keyword in the category by using the interactive activity ratio data.

And the STEP2033 is used for determining proportion data between the first interactive activity degree of the example digital television user keyword and the second interactive activity degree of the category corresponding to the example digital television user keyword.

For some example embodiments, for each example electricity business interaction big data, the ratio data between V3/V4 is calculated with V3 being the larger one and V4 being the smaller one of the first interaction activity of the example digital electricity business user keyword and the second interaction activity of the category to which the example digital electricity business user keyword corresponds.

STEP2034, using a mapping value of the first operation result of the ratio data and the selected value as a first comparison result of the exemplary digital electric telephone user keyword under the corresponding category.

For some exemplary embodiments, a mapping value of a first operation result of the ratio data and a selected value is calculated, and the mapping value is used as a first comparison result of the exemplary digital television user keyword in the category; the first comparison result can reflect the discrimination of the interactive behavior activity between the first interactive behavior activity of the corresponding exemplary digital electric appliance user keyword and the second interactive behavior activity of the affiliated type; the larger the first comparison result is, the larger the distinction degree of the interactive activity between the first interactive activity degree of the exemplary digital television user keyword and the second interactive activity degree of the category to which the keyword belongs is.

In the embodiment of the invention, according to the proportion data between the first interactive activity degree of the example type digital electric power business user keyword of each example type electric power business interaction big data and the second interactive activity degree of the affiliated category, the first comparison result of the example type digital electric power business user keyword under the affiliated category can be determined.

STEP204, determining the error evaluation index corresponding to the exemplary digital electric telephone user keyword according to the first comparison result of the exemplary digital electric telephone user keyword.

For some example embodiments, for each example type e-commerce interaction big data, the first comparison result of the example type digital e-commerce user keyword may be directly used as an error evaluation index corresponding to the example type digital e-commerce user keyword; or, the first comparison result may be updated based on at least one set variable, and the calculation result may be used as an error evaluation index corresponding to the exemplary digital power grid user keyword; wherein, the at least one set variable includes but is not limited to: a first constraint value (upper limit) of the severity of a preset multiple regression condition; a second constraint (lower limit) of the severity of the predetermined multiple regression condition. The first constraint value and/or the second constraint value of the preset multiple regression condition strictness degree can be updated according to the debugging example type set and the model cost index.

For some exemplary embodiments, the error evaluation index of the exemplary digital power provider user keyword and the first comparison result may be in a relationship of a first quantization relationship, in other words, the larger the first comparison result is, the larger the error evaluation index is, and the smaller the first comparison result is, the smaller the error evaluation index is; in view of more rigorous requirements on multivariate regression analysis, the service behavior preference knowledge dispersity of the same type of electricity and business interaction big data mapped in the knowledge characteristic coordinate system is as low as possible; therefore, in the process of carrying out model debugging by means of the same type of example type electric business interaction big data with different interaction behavior activeness, the user behavior preference mining model can strengthen the learning of example type digital electric business user keywords with a large first comparison result, and mining the business behavior preference knowledge which is more beneficial to analyzing different interaction behavior activeness, so that the user behavior preference mining model can mine the business behavior preference knowledge which is more beneficial to identifying different interaction behavior activeness.

Therefore, the error evaluation index corresponding to the example type digital power grid user keyword is determined through the first comparison result of the example type digital power grid user keyword, so that the model cost is updated through the error evaluation index, and further, the business behavior preference knowledge which is more beneficial to analysis of different interaction behavior activeness is mined out through the model cost by the user behavior preference mining model.

It can be understood that, in the embodiment of the present invention, a first interactive behavior liveness determined by using the behavior liveness reference information of the example type e-commerce service interaction big data and a second interactive behavior liveness of the kind to which the example type digital e-commerce user keyword belongs are creatively used, and an error evaluation index is determined and determined according to the first comparison result, so that the first common score is updated by using the error evaluation index to obtain a model cost, and further, a user behavior preference mining model is mined out through the model cost to develop business behavior preference knowledge more conducive to analysis of different interactive behavior liveness.

In some embodiments, the STEP202 may be implemented based on the following STEPs, wherein the STEP determines the second interactive activity of the category corresponding to the example digital electric business user keyword according to the reference information of the activity of the example digital electric business user keyword in the at least one group of the example electric business interaction big data corresponding to the category corresponding to the example digital electric business user keyword.

STEP301, in the exemplary cloud sharing server, determines that the data content contains at least one group of target setting e-commerce interaction big data samples of alternative exemplary digital e-commerce user keywords corresponding to the category corresponding to the exemplary digital e-commerce user keyword.

For some example embodiments, the example cloud sharing server is used for recording a set electric business interaction big data sample of a plurality of sets of digital electric business user keywords of which the data contents comprise behavior activity reference information with completed annotation adding, and the example cloud sharing server stores the set electric business interaction big data sample of a plurality of alternative example digital electric business user keywords, wherein each alternative example digital electric business user keyword corresponds to no less than one set of set electric business interaction big data sample of behavior activity reference information with completed annotation adding. And determining at least one group of target set electric business interaction big data samples in the example type cloud sharing server according to whether the electric business interaction big data content contains the example type digital electric business user keywords belonging to the same type as the example type digital electric business user keywords. The determined at least one group of target setting electric business interaction big data samples, and the example digital electric business user keywords included in the data content can be used for debugging the debugging example of the user behavior preference mining model.

STEP302, determining a second interaction activity of the type corresponding to the example type digital electric business user keyword by combining the reference information of the activity of the candidate example type digital electric business user keyword in the at least one group of target setting electric business interaction big data sample.

For some exemplary embodiments, for each category of the exemplary digital e-commerce user keywords, according to the reference information of the activity level of the exemplary digital e-commerce user keywords of each set of the at least one set of the target-setting e-commerce interaction big data samples of the category, the activity levels of the interaction behaviors respectively corresponding to the digital e-commerce users of the category when the at least one set of the exemplary e-commerce interaction big data samples are collected are determined, and the second activity level of the interaction behavior is determined for the category by means of the activity levels of the interaction behaviors of the at least one set of the exemplary digital e-commerce user keywords of the exemplary e-commerce interaction big data. Determining a second interactive behavior liveness of the category according to the interactive behavior liveness relation of the keywords of the exemplary digital electric appliance user, for example, taking a median value of the interactive behavior liveness as the second interactive behavior liveness; alternatively, the second interactive activity level of the category may be determined according to the interactive activity level calculation result of the exemplary digital tv provider user keyword. For example, a registered member keyword corresponds to three exemplary electric business interaction big data, the activity reference information is 10, 20, and 30, and a second interaction activity degree using the registered member keyword as a category is determined by the interaction activity degrees of the three exemplary electric business interaction big data, for example, a median value 20 of the interaction activity degrees is used as the second interaction activity degree of the category.

In the embodiment of the invention, the second interactive activity of the category can be determined according to the activity reference information matched with the exemplary digital electric power business user keyword to the exemplary digital electric power business user keyword of the same category, so that the corresponding error evaluation index can be determined by means of the error between the first interactive activity of each exemplary digital electric power business user keyword and the second interactive activity of the category.

In some embodiments that may be independent, the STEP302 may be implemented based on the following STEPs 3021 and 3022, and the second interactive activity level of the category corresponding to the exemplary digital television user keyword is determined by an average value of the interactive activity level of the corresponding at least one interactive activity level.

STEP3021, determining an interaction behavior activity mean value of a category corresponding to the example type digital electric utility user keyword by combining with the behavior activity reference information of the alternative example type digital electric utility user keyword in the at least one group of target setting electric utility interaction big data sample.

For some exemplary embodiments, for each category of the exemplary digital e-commerce user keywords, the reference information of the activity of the digital e-commerce user keywords of the exemplary digital e-commerce user keywords of the category of the at least one group of the target setting electric business interaction big data samples is used to determine the activity of the interaction behavior corresponding to the digital e-commerce user corresponding to the category when the at least one group of the exemplary electric business interaction big data is collected, and the average value of the activity of the interaction behavior of the category of the digital e-commerce user corresponding to the at least one group of the exemplary electric business interaction big data is calculated by means of the determined activity of the interaction behavior of the digital e-commerce user corresponding to the at least one group of the exemplary electric business interaction big data.

STEP3022, taking the average value of the interactive activity liveness of the category corresponding to the exemplary digital e-commerce user keyword as the second interactive activity liveness.

For some exemplary embodiments, the determined average value of the interactive activity liveness is used as the second interactive activity liveness of the category.

In the embodiment of the invention, the average value of the interactive activity liveness of the exemplary digital electric appliance user keyword of the same type is determined according to the behavior liveness reference information of the exemplary digital electric appliance user keyword matched with the same type of the exemplary digital electric appliance user keyword, the average value of the interactive activity liveness is taken as the second interactive activity liveness of the type, so that the corresponding error evaluation index is determined by the error of the interactive activity liveness of each exemplary digital electric appliance user keyword and the average value of the interactive activity liveness of the affiliated type, and the knowledge extraction of the exemplary digital electric appliance user keyword with large error of the average value of the interactive activity liveness is strengthened by the user behavior preference mining model.

In some embodiments that may be independent, a second level of interactive activity of the category to which the exemplary digital television user keyword corresponds is determined at a target level of interactive activity corresponding to the at least one level of interactive activity, and the STEP302 may be implemented based on STEPs 3023 and 3024 below.

And STEP3023, determining the target interaction behavior activity of the type corresponding to the example type digital electric appliance user keyword by combining the behavior activity reference information of the alternative example type digital electric appliance user keyword in the at least one group of target setting electric appliance business interaction big data sample.

For some exemplary embodiments, for each category of the exemplary digital e-commerce user keywords, the behavior activity reference information of the exemplary digital e-commerce user keywords of the e-commerce interaction big data sample is set according to at least one group of targets of the category, the interaction behavior activity corresponding to the digital e-commerce user corresponding to the category when at least one group of the exemplary e-commerce interaction big data is collected is determined, and the target interaction behavior activity of the category is determined by the determined interaction behavior activity of the digital e-commerce user corresponding to at least one group of the exemplary e-commerce interaction big data of the category; wherein the target interactive activity level may be a maximum interactive activity level or a minimum interactive activity level of the exemplary digital electric power company user keyword of the category.

STEP3024, regarding the target interactive behavior liveness of the category corresponding to the exemplary digital e-commerce user keyword as the second interactive behavior liveness.

For some example embodiments, the determined maximum interactive activity level may be used as the second interactive activity level of the category; alternatively, the determined minimum interactive activity level may be used as the second interactive activity level of the category.

In the embodiment of the invention, the target interactive behavior liveness of the exemplary digital electric power business user keyword of the same type is determined according to the behavior liveness reference information matched with the exemplary digital electric power business user keyword of the same type and is used as the second interactive behavior liveness of the type, so that the corresponding error evaluation index is determined by means of the error between the interactive behavior liveness of each exemplary digital electric power business user keyword and the target interactive behavior liveness of the type, and the knowledge extraction of the exemplary digital electric power business user keyword with large error with the target interactive behavior liveness is strengthened by the user behavior preference mining model. For example, when the maximum interactive activity degree is taken as the second interactive activity degree of the category, the user behavior preference mining model strengthens the knowledge refinement of the example type digital electric appliance user keywords with large error from the maximum interactive activity degree, in other words, the user behavior preference mining model strengthens the knowledge refinement of the example type digital electric appliance user keywords with small interactive activity degree.

In some embodiments, the STEP204 may be implemented based on the following relevant contents by determining a quantization difference between the first comparison result of the example digital tv user keyword and the maximum first comparison result of the digital tv user keyword of the same category, and updating the quantization difference by setting an update indication to obtain an error evaluation index.

STEP401, determining the largest first comparison result from the first comparison results of the plurality of digital television subscriber user keywords of the same category as the second comparison result.

For some exemplary embodiments, the types of the plurality of digital electric utility user keywords of the same type are identical to the types of the exemplary digital electric utility user keywords, an exemplary digital electric utility user keyword having a largest first comparison result is determined among the plurality of exemplary digital electric utility user keywords matching the exemplary digital electric utility user keywords of one type, and a value of the first comparison result of the exemplary digital electric utility user keyword is assigned to the second comparison result; in other words, the second comparison result represents the largest first comparison result under the category to which this exemplary digital electric company user keyword belongs.

STEP402, determining a quantization difference between the first comparison result and the second comparison result of the exemplary digital television user keyword.

For some example embodiments, for each example digital utility user keyword, determining a quantified difference between a first comparison result for the example digital utility user keyword and a second comparison result under the category; wherein the quantized difference value is representable based on a first operation result of the first comparison result and the second comparison result; alternatively, the ratio data representation may be based on the first comparison result and the second comparison result; alternatively, the representation may be based on a combination of the first operation result of the first comparison result and the second comparison result and the ratio data, such as the first operation result of calculating the first comparison result and the second comparison result, represented by the ratio data of the calculation result and the second comparison result.

And the STEP403, updating the quantization difference value by adopting a set updating instruction, and obtaining an error evaluation index corresponding to the keyword of the exemplary digital television user.

For some exemplary embodiments, updating the quantization difference between the first comparison result of the exemplary digital electric power business user keyword and the second comparison result of the category to obtain an error evaluation index corresponding to the exemplary digital electric power business user keyword by using a set updating instruction; wherein the setting update indication may include at least one preset model learning variable, and the at least one preset model learning variable includes but is not limited to: a first constraint value of the severity of a preset multiple regression condition; a second constraint value of the severity of the predetermined multiple regression condition. The first constraint value and/or the second constraint value of the preset multiple regression condition strictness degree can be updated according to the debugging example type set and the model cost index. Here, the setting update indication and the update mode of the quantized difference value using the setting update indication may be exemplarily set according to the requirement in combination with the user behavior preference mining model debugging set.

In the embodiment of the invention, a quantitative difference between a first comparison result of an exemplary digital electric appliance user keyword and a maximum first comparison result (namely a second comparison result) of a category corresponding to the exemplary digital electric appliance user keyword is determined, the quantitative difference is updated by using a set variable as an updating variable, and an error evaluation index of the exemplary electric appliance service interaction big data is determined, so that a model cost is updated by using the error evaluation index, and a service behavior preference mining model is mined by using the model cost to more contribute to activity analysis of different interaction behaviors.

In some embodiments, before updating the first commonality score based on the error score index, that is, before STEP103, STEP105, STEP106, STEP107 are further included.

STEP105, combining the to-be-debugged user behavior preference mining model, performing behavior preference mining on the example type digital electric appliance user keywords, and obtaining the service behavior preference knowledge of the example type digital electric appliance user keywords.

For some exemplary embodiments, behavior preference mining is performed on big data of each exemplary type of electric business interaction as a debugging exemplary type through a to-be-debugged user behavior preference mining model, so that business behavior preference knowledge of each exemplary type of digital electric business user keyword is obtained.

And the STEP106 is used for mining the behavior preference of the alternative example type digital electric appliance user keywords which are consistent with the types corresponding to the example type digital electric appliance user keywords by combining the to-be-debugged user behavior preference mining model to obtain the business behavior preference knowledge of the types corresponding to the example type digital electric appliance user keywords.

For some exemplary embodiments, through the to-be-debugged user behavior preference mining model, the behavior preference mining is performed on at least one group of target setting electric business interaction big data samples of at least one alternative exemplary digital electric business user keyword of which the category is consistent with that corresponding to the exemplary digital electric business user keyword, so as to obtain business behavior preference knowledge of the category corresponding to the exemplary digital electric business user keyword, for example, intra-cluster preference knowledge of the category corresponding to the exemplary digital electric business user keyword.

STEP107, determining the first commonality score between the business behavior preference knowledge of the example digital e-telephone user keyword and the business behavior preference knowledge of the category corresponding to the example digital e-telephone user keyword.

For some exemplary embodiments, the service behavior preference knowledge of the standardized exemplary digital electric appliance user keyword and the service behavior preference knowledge of the category corresponding to the standardized exemplary digital electric appliance user keyword are subjected to multiplication (such as cross multiplication) and inverse cosine calculation, and a common score between the service behavior preference knowledge of the exemplary digital electric appliance user keyword and the service behavior preference knowledge of the category corresponding to the exemplary digital electric appliance user keyword is determined according to a calculation result, so as to obtain a first common score.

According to the error evaluation index of the example digital electric business user keywords of each group of example electric business interaction big data, updating a first common score between the example digital electric business user keywords and the corresponding types of the example digital electric business user keywords in the model cost of the model mined according to the behavior preference of the user to be debugged to obtain the model cost; the updating of the first common score can be the difference between the service behavior preference knowledge for updating the example type digital electric appliance user keyword and the intra-cluster preference knowledge of the type corresponding to the example type digital electric appliance user keyword; or updating the Euclidean calculation value of the difference between the business behavior preference knowledge of the example type digital electric appliance user keyword and the intra-cluster preference knowledge of the category corresponding to the example type digital electric appliance user keyword.

In the embodiment of the invention, based on a to-be-debugged user behavior preference mining model, behavior preference mining is respectively carried out on the example type digital electric appliance user keyword and the alternative example type digital electric appliance user keyword of the type corresponding to the example type digital electric appliance user keyword to obtain the service behavior preference knowledge of the example type digital electric appliance user keyword and the service behavior preference knowledge of the type corresponding to the example type digital electric appliance user keyword, and a first common score between the service behavior preference knowledge of the example type digital electric appliance user keyword and the service behavior preference knowledge of the type corresponding to the example type digital electric appliance user keyword can be determined, so that the first common score can be updated by means of an error evaluation index to obtain the model cost; in view of more rigorous requirements on multivariate regression analysis, the service behavior preference knowledge dispersity of the electric business interaction big data mapped in the knowledge characteristic coordinate system (AI characteristic space) is as low as possible, so that the user behavior preference mining model can strengthen the knowledge refinement of the exemplary digital electric business user keywords for setting the interactive behavior liveness, and the updated user behavior preference mining model can mine the service behavior preference knowledge more beneficial to identification of different interactive behavior liveness.

The embodiment of the invention provides a debugging method of a user behavior preference mining model, on the basis that an example type digital electric appliance user keyword is a registered member keyword, the alternative example type digital electric appliance user keyword is the registered member keyword with different interaction behavior liveness, and the related design thought is as follows.

STEP101, obtaining example e-commerce business interaction big data of activity reference information of the example digital e-commerce user keyword with completed annotation adding.

In some optional designs, the example digital e-commerce user keyword is a registered member keyword, and the example e-commerce business interaction big data is page interaction big data of the registered member keyword; the behavior activity reference information of the example digital electric business user keyword in the example electric business interaction big data is annotated with interaction behavior activity of the registered member keyword when the registered member keyword is collected to the example electric business interaction big data.

STEP201, determining a first interactive activity of the example digital television user keyword in combination with the activity reference information of the example digital television user keyword.

In some optional designs, the first interactive activity level of the registered member keyword is determined based on the activity reference information.

STEP501, in the exemplary e-commerce interaction big database, determines that the data content comprises not less than one group of target setting e-commerce interaction big data examples of the registered member keywords with different interaction behavior liveness.

In some optional designs, in an exemplary electronic commerce interaction big database storing a plurality of sets of set electronic commerce interaction big data examples of registered member keywords of which data contents include annotated and added behavior activity reference information, at least one set of target set electronic commerce interaction big data examples collected at different interaction behavior activities corresponding to the same digital electronic commerce user (the same registered member keyword) is determined; the example digital electric business user keywords are registered member keywords, and the alternative example digital electric business user keywords in the example electric business interaction big database are registered member keywords with different interaction activity liveness.

STEP502, determining a second interactive activity of the registered member keyword in combination with the different interactive activity of the registered member keyword annotated in the at least one group of target set telecommunication service interaction big data sample.

For some exemplary embodiments, according to the interactive activity labels in at least one group of target setting electric business interaction big data samples of the registered member keyword, a first interactive activity of the registered member keyword in each group of target setting electric business interaction big data samples is determined, and a second interactive activity using the registered member keyword as the category of the exemplary digital electric business user keyword is determined according to the first interactive activity of the target setting electric business interaction big data samples.

It can be understood that, in STEP501 to STEP502, under the condition of performing demand mining on e-commerce business interaction big data, according to at least one behavior activity reference information of at least one set of target e-commerce business interaction big data sample of the same registered member keyword, a second interaction behavior activity with the registered member keyword as a digital e-commerce user keyword kind can be determined; setting a first interactive behavior activity of the e-commerce business interaction big data sample in each group of targets by means of the registered member keywords, and determining a corresponding error evaluation index according to the error of the corresponding second interactive behavior activity; the user behavior preference mining model can further conduct reinforced mining on knowledge vectors of the interactive big data with different interactive behavior activeness, and the updated user behavior preference mining model can mine the interactive big data service behavior preference knowledge which is more beneficial to recognition of different interactive behavior activeness; in other words, when the E-commerce business interaction big data matching with different interaction activity degrees is carried out by means of the user behavior preference mining model, the precision of the E-commerce business interaction big data matching with different interaction activity degrees can be improved.

STEP203, determining a first comparison result of the example type digital electric appliance user keyword under the corresponding category by combining the first interactive behavior liveness of the example type digital electric appliance user keyword and the second interactive behavior liveness of the category corresponding to the example type digital electric appliance user keyword.

For some example embodiments, for each example type of e-commerce interaction big data, a first comparison result of the registered member keyword included in the example type of e-commerce interaction big data content under the category of the registered member keyword as the example type of digital e-commerce user keyword is determined according to an operational relationship between a first interaction activity of the registered member keyword and a second interaction activity of the registered member keyword included in the example type of e-commerce interaction big data content.

STEP204, determining the error evaluation index corresponding to the example digital television user keyword in combination with the first comparison result of the example digital television user keyword.

For some exemplary embodiments, the error evaluation index corresponding to the exemplary digital electric utility user keyword is determined directly or by updating according to a first comparison result of the registered member keyword included in the exemplary electric utility interaction big data content under the category that the registered member keyword is the exemplary digital electric utility user keyword.

And the STEP103, in combination with the error evaluation index, updates a first commonality score between the example type digital electric appliance user keyword and a category corresponding to the example type digital electric appliance user keyword to obtain a model cost of the behavior preference mining model of the user to be debugged.

For some exemplary embodiments, according to the error evaluation index of the registered member keyword included in each group of exemplary e-commerce interaction big data content, the first commonality score is updated in the model cost of the model mining the behavior preference of the user to be debugged to obtain the model cost.

And the STEP104 is used for updating the model variables of the mining model of the user behavior preference to be debugged by combining the model cost so as to enable the model cost generated by the updated mining model of the user behavior preference to be in a stable state.

For some exemplary embodiments, in a cycle phase, updating model variables in the to-be-debugged user behavior preference mining model based on model costs until the model costs of the to-be-debugged user behavior preference mining model are in a stable state, stopping the cycle phase, completing debugging of the to-be-debugged user behavior preference mining model, and obtaining the updated user behavior preference mining model.

In the embodiment of the invention, on the basis that the example digital electric power business user keywords are registered member keywords, the example electric power business interaction big data are human page interaction big data, model cost is updated for each page interaction big data by means of behavior liveness reference information, a user behavior preference mining model is debugged based on the model cost, so that the user behavior preference mining model can carry out reinforced mining on knowledge vectors of page interaction behaviors of the registered member keywords with different interaction behavior liveness, and the updated user behavior preference mining model can mine interaction big data business behavior preference knowledge more beneficial to recognition of different interaction behavior liveness; therefore, under the condition of user keyword analysis of the digital e-commerce with different interactive behavior liveness, the service behavior preference knowledge mined by the updated user behavior preference mining model is searched and matched with the e-commerce service interaction big data, and the accuracy of service interaction big data matching of e-commerce meeters with different interactive behavior liveness can be improved.

The embodiment of the invention provides an E-commerce transaction method based on big data, which is characterized in that a user behavior preference mining model is adopted to mine behavior preference, and keyword recognition of a digital E-commerce user is realized based on mined business behavior preference knowledge.

STEP601, obtaining the first e-commerce business interaction big data containing the keywords of the digital e-commerce user to be analyzed.

For some exemplary embodiments, in the whole flow of the digital electric utility user keyword recognition, the first electric utility interaction big data may be electric utility interaction big data whose data content includes the digital electric utility user keyword to be analyzed, and the first electric utility interaction big data includes a restriction condition of the digital electric utility user keyword to be analyzed, which is set in advance.

STEP602, performing behavior preference mining on the digital e-commerce user keyword to be analyzed by combining with the user behavior preference mining model to obtain the service behavior preference knowledge of the digital e-commerce user keyword to be analyzed.

For some exemplary embodiments, the user behavior preference mining model can be obtained by debugging based on the debugging method of the user behavior preference mining model provided in the embodiments, the user behavior preference mining model can be used for behavior preference mining and behavior preference distinguishing, and at least one group of matched electric business interaction big data is determined according to the common score of the mined business behavior preference knowledge; the behavior preference mining function in the user behavior preference mining model to be debugged can be realized by adopting rules such as a moving average unit, and the keyword recognition function of the digital electric telephone user can be realized on the basis of the rules such as a down-sampling unit and/or a feature integration unit. And inputting the first e-commerce business interaction big data containing the digital e-commerce user keywords to be analyzed into a user behavior preference mining model for behavior preference mining to obtain business behavior preference knowledge of the digital e-commerce user keywords to be analyzed.

STEP603, combined with the knowledge of the business behavior preference, retrieving at least one group of second e-business interaction big data paired with the first e-business interaction big data in a setting cloud sharing server.

For some exemplary embodiments, the cloud sharing server is configured to include at least one group of second e-commerce interaction big data of personalized e-commerce service requirements added with the prior user keyword, and the data content of the at least one group of second e-commerce interaction big data may include the prior user keyword in different activity degrees of interaction behavior from the digital e-commerce user keyword to be analyzed. The interactive behavior activity of the prior user keyword can be the same as that of the digital electric appliance user keyword to be analyzed, and can also be different from that of the digital electric appliance user keyword to be analyzed. The cloud sharing server is set to be a static big data set comprising a large amount of electric business interaction big data, and each group of electric business interaction big data is correspondingly added with personalized electric business service requirements of digital electric business user keywords; wherein the personalized e-commerce service requirement may be information for identifying a keyword of the digital e-commerce user. At least one group of matched second electric business interaction big data is retrieved in a set cloud sharing server by means of business behavior preference knowledge of the first electric business interaction big data, the at least one group of matched second electric business interaction big data is electric business interaction big data of prior user keywords at different interaction behavior liveness degrees, and the business behavior preference knowledge of the second electric business interaction big data, which is mined by a user behavior preference mining model, meets specified requirements. For example, the business behavior preference knowledge specifying requirements for interaction with the first e-commerce business big data differs by less than a preset threshold.

For example, the digital electric business user keyword to be analyzed and the prior user keyword which is in different interaction behavior liveness with the digital electric business user keyword to be analyzed can be respectively loaded into the user behavior preference mining model, the business behavior preference knowledge is extracted, electric business interaction big data matching is carried out according to the business behavior preference knowledge of the digital electric business user keyword to be analyzed and the common score of the prior user keyword (set user keyword), and at least one group of second electric business interaction big data which is matched with the first electric business interaction big data of the electric business interaction big data to be analyzed is determined in the set cloud sharing server.

For some exemplary embodiments, the setting cloud sharing servers may be divided according to the interaction activity of the e-commerce interaction big data, in other words, each setting cloud sharing server stores the e-commerce interaction big data of a corresponding setting interaction activity interval, so as to obtain the e-commerce interaction big data of the prior user keyword in the setting interaction activity interval, which is matched with the digital e-commerce user keyword to be analyzed; the size of each interactive behavior activity interval can be flexibly set.

And the STEP604 determines the personalized e-commerce service requirements of the digital e-commerce user keywords to be analyzed in combination with the personalized e-commerce service requirements of the prior user keywords with different interactive behavior liveness degrees in the set cloud sharing server.

For some exemplary embodiments, according to at least one group of retrieved prior user keywords with different interaction behavior liveness included in the second e-commerce service interaction big data, the personalized e-commerce service requirement of the prior user keyword is determined in the set cloud sharing server, and the personalized e-commerce service requirement of the digital e-commerce user keyword to be analyzed is determined according to the determined personalized e-commerce service requirement.

Updating the model cost of the mining model of the behavior preference of the user to be debugged through the error evaluation index of the registered member keyword, so that the mining model of the behavior preference of the user can carry out enhanced mining on the knowledge vector of the exemplary digital electric appliance user keyword with different interactive behavior liveness during debugging, and the mining model of the behavior preference of the user which is updated can mine the business behavior preference knowledge which is more beneficial to identifying different interactive behavior liveness; in this way, under the condition of keyword analysis of digital electric utility users with different interactive activity degrees, whether the interactive activity degree of the digital electric utility user keyword to be analyzed for retrieval is the same as the interactive activity degree of the prior user keyword in the set cloud sharing server or not, the electric utility interaction big data corresponding to the prior user keyword of the same digital electric utility user can be accurately matched with the digital electric utility user keyword to be analyzed, in other words, the embodiment of the invention supports the retrieval of the electric utility interaction big data in the set cloud sharing server by using the electric utility interaction big data of the digital electric utility user keyword to be analyzed with different interactive activity degrees.

In the embodiment of the invention, the digital e-commerce user keyword can be understood as the ID of the digital e-commerce user or the label information of the digital e-commerce user and is used for distinguishing different digital e-commerce users, the personalized e-commerce service requirement can reflect the differentiated service requirements of different to-be-analyzed digital e-commerce user keywords, and the personalized e-commerce service requirement can guide service optimization upgrading or big data recommendation. Further, under the condition of analyzing the digital e-commerce user keywords with different interactive behavior liveness degrees, the matching e-commerce service interaction big data is retrieved through the service behavior preference knowledge mined by the updated user behavior preference mining model, so that the matching precision of the e-commerce service interaction big data with different interactive behavior liveness degrees can be improved, and the precision of determining the personalized e-commerce service requirements of the digital e-commerce user keywords to be analyzed based on the personalized e-commerce service requirements of the matching e-commerce service interaction big data is improved.

Based on the same or similar inventive concepts, please refer to fig. 2 in combination, and an architectural schematic diagram of an application environment 30 of the big data-based e-commerce transaction method is further provided, which includes a big data e-commerce platform system 10 and a cloud sharing server 20 that communicate with each other, and the big data e-commerce platform system 10 and the cloud sharing server 20 implement or partially implement the technical solutions described in the above method embodiments when running.

Further, a computer-readable storage medium is provided, on which a program is stored, which when executed by a processor implements the above-described method.

In the embodiments provided in the embodiments of 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 illustrative only, as the flowcharts 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, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent 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 such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a network device, or the like) to execute 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), a magnetic disk or an optical disk, and 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 a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A big data-based E-commerce transaction method is applied to a big data E-commerce platform system and comprises the following steps:

obtaining first e-commerce business interaction big data containing a digital e-commerce user keyword to be analyzed; performing behavior preference mining on the digital electric appliance user keywords to be analyzed by combining a user behavior preference mining model to obtain business behavior preference knowledge of the digital electric appliance user keywords to be analyzed;

retrieving at least one group of second e-commerce interaction big data matched with the first e-commerce interaction big data in a set cloud sharing server in combination with the business behavior preference knowledge; the cloud sharing server is set to comprise at least one group of E-commerce service interaction big data of personalized E-commerce service requirements of the added annotation-finished prior user keywords; the at least one group of second e-commerce interaction big data comprises prior user keywords with different interaction behavior activeness;

and determining the personalized e-commerce service requirements of the digital e-commerce user keywords to be analyzed by combining the personalized e-commerce service requirements of the prior user keywords with different interactive behavior liveness degrees in the set cloud sharing server.

2. The method of claim 1, wherein the method for debugging the user behavior preference mining model comprises:

obtaining example type e-commerce business interaction big data of behavior activity reference information of example type digital e-commerce user keywords of which annotation addition is completed;

determining an error evaluation index of a to-be-debugged user behavior preference mining model by combining the behavior liveness reference information of the exemplary digital electric power business user keywords and the corresponding categories of the exemplary digital electric power business user keywords;

updating a first commonality score between the example type digital electric power business user keyword and a category corresponding to the example type digital electric power business user keyword by combining the error evaluation index to obtain a model cost of the behavior preference mining model of the user to be debugged;

and updating the model variables of the mining model of the user behavior preference to be debugged by combining the model cost so as to enable the model cost generated by the updated mining model of the user behavior preference to be in a stable state.

3. The method as claimed in claim 2, wherein the determining an error evaluation index of the mining model of the behavioral preference of the user to be debugged by combining the reference information of the behavioral activity of the exemplary digital electric power provider user keyword and the category corresponding to the exemplary digital electric power provider user keyword comprises:

determining a first interactive behavior liveness of the exemplary digital electric appliance user keyword in combination with the behavior liveness reference information of the exemplary digital electric appliance user keyword;

determining a second interactive behavior activity degree of a category corresponding to the example type digital television user keyword;

determining a first comparison result of the example type digital electric power business user keyword under the corresponding category by combining the first interactive behavior activity of the example type digital electric power business user keyword and the second interactive behavior activity of the category corresponding to the example type digital electric power business user keyword;

and determining the error evaluation index corresponding to the exemplary digital electric power provider user keyword by combining the first comparison result of the exemplary digital electric power provider user keyword.

4. The method of claim 3, wherein determining the second interactive activity of the category corresponding to the example digital television carrier user keyword comprises:

in an exemplary cloud sharing server, determining that the data content contains at least one group of target setting electric business interaction big data samples of alternative exemplary digital electric business user keywords which are consistent with the types corresponding to the exemplary digital electric business user keywords; wherein the exemplary cloud sharing server is used for recording a plurality of sets of set e-commerce interaction big data examples of digital e-commerce user keywords of which the data contents comprise behavior activity reference information of which annotation addition is completed;

and determining second interactive activity of the type corresponding to the example type digital electric appliance user keyword by combining with the reference information of the activity of the alternative example type digital electric appliance user keyword in the at least one group of target setting electric appliance business interaction big data sample.

5. The method as claimed in claim 4, wherein the determining the second interactive activity level of the category corresponding to the exemplary digital electric utility user keyword in combination with the activity reference information of the alternative exemplary digital electric utility user keyword in the at least one set of target setting electric business interaction big data sample comprises:

determining an interactive behavior activity mean value of a category corresponding to the example type digital electric power business user keyword by combining with the behavior activity reference information of the alternative example type digital electric power business user keyword in the at least one group of target setting electric power business interaction big data sample;

and taking the average value of the interactive behavior liveness of the type corresponding to the example type digital electric appliance user keyword as the second interactive behavior liveness.

6. The method as claimed in claims 4-5, wherein the determining the second interactive activity level of the category corresponding to the exemplary digital electric utility user keyword in combination with the activity reference information of the alternative exemplary digital electric utility user keyword in the at least one set of target setting electric business interaction big data sample comprises: determining target interaction behavior activity of a type corresponding to the example type digital electric appliance user keyword by combining with the behavior activity reference information of the alternative example type digital electric appliance user keyword in the at least one group of target setting electric appliance business interaction big data sample; taking the target interactive behavior activity of the type corresponding to the example type digital television business user keyword as the second interactive behavior activity;

wherein, the determining a first comparison result of the example-type digital electric power business user keyword under the corresponding category by combining the first interactive activity of the example-type digital electric power business user keyword and the second interactive activity of the category corresponding to the example-type digital electric power business user keyword comprises: determining a first operation result of a first interactive behavior liveness of the example type digital electric appliance user keyword and a second interactive behavior liveness of a category corresponding to the example type digital electric appliance user keyword; and taking the mapping value of the first operation result as a first comparison result of the exemplary digital electric power business user keyword under the corresponding category.

7. The method of claim 3, wherein determining the error rating index corresponding to the example digital tv provider user keyword in combination with the first comparison result of the example digital tv provider user keyword comprises:

determining a maximum first comparison result from the first comparison results of the plurality of digital electric company user keywords of the same category as a second comparison result; wherein a category of the plurality of digital electric power consumer keywords of the same category is identical to a category of the exemplary digital electric power consumer keyword;

determining a quantized difference between the first comparison result and the second comparison result of the exemplary digital television user keyword;

and updating the quantization difference value by adopting a set updating instruction to obtain an error evaluation index corresponding to the keyword of the exemplary digital electric power business user.

8. The method according to claim 2, wherein before the updating, with reference to the error evaluation index, the first commonality score between the example-type digital electric power business user keyword and the category corresponding to the example-type digital electric power business user keyword to obtain the model cost of the to-be-debugged user behavior preference mining model, the method further comprises:

performing behavior preference mining on the exemplary digital electric power business user keywords by combining the behavior preference mining model of the user to be debugged to obtain business behavior preference knowledge of the exemplary digital electric power business user keywords;

performing behavior preference mining on alternative example type digital electric appliance user keywords which are consistent with the types corresponding to the example type digital electric appliance user keywords by combining the to-be-debugged user behavior preference mining model to obtain business behavior preference knowledge of the types corresponding to the example type digital electric appliance user keywords;

determining the first common score between the business behavior preference knowledge of the example digital electric appliance user keyword and the business behavior preference knowledge of the category corresponding to the example digital electric appliance user keyword.

9. The method as claimed in claim 4, wherein the determining, in the exemplary cloud sharing server, at least one target-setting e-commerce interaction big data sample whose data content includes an alternative exemplary digital e-commerce user keyword corresponding to a category corresponding to the exemplary digital e-commerce user keyword is determined to be not less than one group of the exemplary digital e-commerce user keyword based on the exemplary digital e-commerce user keyword being a registered member keyword, the alternative exemplary digital e-commerce user keyword being the registered member keyword with different interaction behavior liveness includes: in the exemplary electric business interaction big database, determining that the data content comprises not less than one group of target setting electric business interaction big data examples of the registered member keywords with different interaction behavior liveness;

the determining, by combining the reference information of the activity levels of the alternative exemplary digital electric appliance user keywords in the at least one set of target setting electric business interaction big data samples, a second interaction activity level of a category corresponding to the exemplary digital electric appliance user keyword includes: and determining a second interactive activity degree of the registered member keywords by combining different interactive activity degrees of the registered member keywords annotated in the at least one group of target setting e-commerce interaction big data samples.

10. A big data E-commerce platform system is characterized by comprising a processor and a memory; the processor is communicatively connected to the memory, and the processor is configured to read the computer program from the memory and execute the computer program to implement the method of any one of claims 1 to 9.

Images