Data processing method, device and storage medium based on behavior association model

CN118643220BActive Publication Date: 2024-11-19QINGDAO HAIER INTELLIGENT HOME APPLIANCE TECHNOLOGY CO LTD +1
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202411104342.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-19
Estimated Expiration
2044-08-13

AI Technical Summary

Technical Problem

[0006]本申请实施例提供了一种基于行为关联模型的数据处理方法、装置及存储介质,以至少解决相关技术中,在向用户推荐数据时,存在推荐数据不准确的技术问题

Benefits of technology

[0025] In this embodiment of the present application, an object behavior set is obtained by performing time series processing on the device control data of the first object's control of the home appliance according to the time series processing rules, wherein the object behavior set includes the first behavior record of the first object. ; Perform clustering processing on multiple object behavior sets corresponding to different objects, and obtain the group to which the first object belongs based on the results of the clustering processing; Use a behavior association model to perform clustering on the first behavior record of the first object Calculate and obtain the behavior measurement parameters of the first object, where the behavior measurement parameters include behavior support and behavior confidence; determine the preset control behaviors set for the group to which the first object belongs in different time periods , determine the target control behavior of the first object from the preset control behavior based on the behavior measurement parameter; set the target control behavior set generated based on the target control behavior as the recommended data of the first object, and push the recommendation data of the first object to the control terminal of the first object; that is to say, after binding the user data and the user's operation of the home appliance, this application can collect user data in real time when the user operates the home appliance. The operating behavior records data (i.e., control data), and then the users are divided into groups, and the behavioral correlation calculations are performed on the grouped users respectively to obtain the user's behavior habit data, and finally the user behavior is recommended to the user in real time (i.e., the target control behavior) . The above technical solution solves the technical problem of inaccurate recommended data when recommending data to users, thereby improving the accuracy of recommended data and improving user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118643220B_ABST
    Figure CN118643220B_ABST
Patent Text Reader

Abstract

The present application discloses a data processing method, device and storage medium based on a behavior association model, and relates to the field of electrical digital data processing. The method comprises: performing time sequence processing on device control data of a first object controlling a home appliance according to a time sequence processing rule to obtain an object behavior set; performing clustering processing on multiple object behavior sets corresponding to different objects, and obtaining a group to which the first object belongs based on the result of the clustering processing; calculating a first behavior record of the first object using a behavior association model to obtain a behavior measurement parameter of the first object; determining preset control behaviors set for the group to which the first object belongs in different time periods, and determining a target control behavior of the first object from the preset control behaviors based on the behavior measurement parameter; setting a target control behavior set generated based on the target control behavior as recommended data of the first object, and pushing the recommended data of the first object to a control terminal of the first object.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of electronic digital data processing, and in particular to a data processing method, device and storage medium based on a behavior association model. Background Art

[0002] At present, with the continuous updating and iteration of IoT technology, the home furnishing industry is also making continuous progress and providing users with more efficient home furnishing services through home furnishing products. For example, smart home manufacturers and service providers can provide users with more considerate recommendation services in terms of message recommendations by identifying user groups and user behavior habits.

[0003] However, in the related technologies, user behavior habits are identified directly based on user data and user behavior data. There is a lack of effective analysis and processing of the data, which will recommend inaccurate behavior habit data to users, affecting the effective management of user behavior habits and resulting in a poor user experience.

[0004] Therefore, in the related art, when recommending data to users, there is a technical problem of inaccurate recommended data.

[0005] Regarding the related technologies, when recommending data to users, there is a technical problem of inaccurate recommended data, and no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present application provide a data processing method, device and storage medium based on a behavior association model, so as to at least solve the technical problem in the related art that when recommending data to users, the recommended data is inaccurate.

[0007] According to an embodiment of the present application, a data processing method based on a behavior association model is provided, including: performing time series processing on device control data of a first object controlling a household appliance according to a time series processing rule to obtain an object behavior set, wherein the object behavior set includes a first behavior record of the first object; performing clustering processing on multiple object behavior sets corresponding to different objects, and obtaining a group to which the first object belongs based on a result of the clustering processing; calculating the first behavior record of the first object using a behavior association model to obtain behavior measurement parameters of the first object, wherein the behavior measurement parameters include behavior support and behavior confidence; determining preset control behaviors set for the group to which the first object belongs in different time periods, and determining target control behaviors of the first object from the preset control behaviors based on the behavior measurement parameters; setting a target control behavior set generated based on the target control behavior as recommended data for the first object, and pushing the recommended data of the first object to a control terminal of the first object.

[0008] In an exemplary embodiment, the device control data includes at least a control instruction received by the household appliance and operation data generated by the household appliance executing the control instruction, and the device control data of the first object controlling the household appliance is subjected to timing processing according to a timing processing rule to obtain an object behavior set, including: binding the generation time when the household appliance generates the operation data and the operation data to obtain a second behavior record, wherein the second behavior record includes the bound generation time and the operation data, and the operation data includes at least the object numbers of the different objects, the location data of the different objects, the behavior data of the different objects, and the device type of the household appliance; sorting the generation time according to the timing processing rule to obtain the sorted generation time, wherein the timing processing rule indicates processing in chronological order; sorting multiple second behavior records according to the sorted generation time, and generating the object behavior set based on the sorted multiple second behavior records.

[0009] In an exemplary embodiment, after performing timing processing on device control data of a home appliance controlled by a first object according to a timing processing rule to obtain an object behavior set, the method further includes: for a first behavior record of the first object included in the object behavior set, determining at least a first generation moment and first operation data bound to the first generation moment from the first behavior record, wherein the first operation data at least includes: a first object number of the first object, a first position of the first object, a first object behavior of the first object and a first device type of the home appliance; for multiple first behavior records, when it is determined that the time interval between the multiple first generation moments is less than a preset interval, deleting invalid behavior records from the multiple first behavior records according to the earliest first generation moment to obtain remaining behavior records, wherein the invalid behavior records represent behavior records corresponding to other generation moments in the multiple first behavior records except the earliest first generation moment; and updating the object behavior set based on the remaining behavior records to obtain an updated object behavior set.

[0010] In an exemplary embodiment, timing processing is performed on device control data of a home appliance controlled by a first object according to timing processing rules, including: performing data cleaning processing on the device control data to obtain cleaned device control data; wherein the data cleaning processing includes at least one of the following: null value processing, abnormal value processing; obtaining a verification result of data verification of the cleaned device control data; when it is determined that the verification result passes the verification, performing format standardization processing on the cleaned device control data to obtain standard data, and processing the standard data according to the timing processing rules.

[0011] In an exemplary embodiment, the different objects include at least the first object and the second object, and clustering processing is performed on multiple object behavior sets corresponding to the different objects, and the group to which the first object belongs is obtained based on the result of the clustering processing, including: determining M behavior records that have been assigned to a group from the multiple object behavior sets, and the M behavior records represent the current center point required for clustering grouping; determining a third behavior record that belongs to an unassigned group from the multiple object behavior sets, and calculating the first Euclidean distance between the first eigenvalue of the third behavior record and the second eigenvalue of the M behavior records; determining a fourth behavior record corresponding to the smallest first Euclidean distance from the M behavior records, and assigning a group to the third behavior record according to the group to which the fourth behavior record belongs, and the group to which the third behavior record belongs is the same as the group to which the fourth behavior record belongs. group; re-determine L behavior records that have been assigned to groups, and the L behavior records represent the updated center point required for clustering grouping; determine the fifth behavior record corresponding to the smallest second Euclidean distance from the L behavior records, and assign a group to the sixth behavior record that has not been assigned to a group among the L behavior records according to the group to which the fifth behavior record belongs, and the second Euclidean distance represents the Euclidean distance between the third eigenvalue of the sixth behavior record and the fourth eigenvalue of the L behavior records; if it is determined that the current center point is consistent with the updated center point, then set the belonging group for the first object according to the belonging group of the third behavior record, and set the belonging group for the second object according to the belonging group of the sixth behavior record, wherein the third behavior record represents the behavior record belonging to the first object, and the sixth behavior record represents the behavior record belonging to the second object.

[0012] In an exemplary embodiment, after determining M behavior records that have been assigned to a group from the plurality of object behavior sets, the method further includes: selecting a current value of M from a preset value interval for M, and calculating a silhouette coefficient corresponding to the current value; determining a target value having a maximum silhouette coefficient from a plurality of silhouette coefficients, and determining the target value as the current value of M; wherein the silhouette coefficient S(i) is calculated by the following formula: ; Wherein, a(i) represents the cohesion of the current value, b(i) represents the separation of the current value, and i is a positive integer.

[0013] In an exemplary embodiment, respectively calculating the first Euclidean distance between the first feature value of the third behavior record and the second feature value of the M behavior records includes: normalizing the record value of the third behavior record to obtain the first feature value, wherein the record value of the third behavior record at least includes a second generation time, and second operation data bound to the second generation time, wherein the second operation data at least includes: a second position of the second object, a second object behavior of the second object, and a second device type of the home appliance, and the first feature value at least includes a first time feature value corresponding to the second generation time, a first position feature value of the second position, a first behavior feature value of the second object behavior, and a first type feature value of the second device type; and performing normalization on the M behavior records. The record value of each behavior record in the behavior records is normalized to obtain the second feature value, wherein the record value of each behavior record at least includes a third generation time and third operation data bound to the third generation time, wherein the third operation data represents data of a third object corresponding to each behavior record, and at least includes: a third position of the third object, a third object behavior of the third object and a third device type of the home appliance, and the second feature value at least includes a second time feature value corresponding to the third generation time, a second position feature value of the third position, a second behavior feature value of the third object behavior and a second type feature value of the third device type; the first Euclidean distance between the first feature value and the second feature value is calculated according to the following formula: , where d(m, n) is the first Euclidean distance, w1 is the feature value at the first moment, x1 is the feature value at the first position, y1 is the feature value at the first behavior, z1 is the feature value of the first type, w2 is the feature value at the second moment, x2 is the feature value at the second position, y2 is the feature value of the second behavior, and z2 is the feature value of the second type.

[0014] In an exemplary embodiment, the first Euclidean distance between the first feature value of the third behavior record and the second feature value of the M behavior records is calculated respectively, including: obtaining a first object label of the first object and a second object label of the second object, wherein the first object label is determined based on an object attribute and a first object attribute value of the first object, and the second object label is determined based on an object attribute and a second object attribute value of the second object; normalizing the first object label to obtain a first label feature value, and updating the first feature value by incorporating the first label feature value into the first feature value; and normalizing the second object label to obtain a second label feature value, and updating the second feature value by incorporating the second label feature value into the second feature value; and calculating the first Euclidean distance between the updated first feature value and the updated second feature value.

[0015] In an exemplary embodiment, the first Euclidean distance between the first feature value of the third behavior record and the second feature value of the M behavior records is calculated respectively, including: obtaining a first object label of the first object and a second object label of the second object, wherein the first object label is determined based on an object attribute and a first object attribute value of the first object, and the second object label is determined based on an object attribute and a second object attribute value of the second object; normalizing the first object label to obtain a first label feature value, and updating the first feature value by incorporating the first label feature value into the first feature value with a first weight; normalizing the second object label to obtain a second label feature value, and updating the second feature value by incorporating the second label feature value into the second feature value with a second weight, wherein the first weight is different from the second weight; and calculating the first Euclidean distance between the updated first feature value and the updated second feature value.

[0016] In an exemplary embodiment, the target control behavior of the first object is determined from the preset control behaviors based on the behavior measurement parameters, including: determining a first control behavior from multiple preset control behaviors whose behavior support satisfies a preset support matching rule and whose behavior confidence satisfies a preset confidence matching rule, wherein the first control behavior corresponds to a behavior occurrence time period; when it is determined that the behavior occurrence time period belongs to the preset time period, determining the first control behavior as the target control behavior.

[0017] In an exemplary embodiment, the behavior support is determined to satisfy the preset support matching rule in the following manner: the behavior support is calculated by the following formula: Esupport = sum(count(P(A,B))*K*T); wherein Esupport is the behavior support, K ​​is the behavior coefficient, T is the time attenuation coefficient, and count(P(A,B)) represents the number of times the time factor A and the location factor B appear at the same time; when it is determined that the behavior support is greater than the support threshold, it is determined that the behavior support satisfies the preset support matching rule.

[0018] In an exemplary embodiment, the behavior confidence is determined to satisfy the preset confidence matching rule in the following manner: the behavior confidence is calculated by the following formula: Econfidence = sum(count(P(A,B))*K*T) / sum(count(P(i))*K*T); wherein Econfidence is the behavior confidence, K is the behavior coefficient, T is the time attenuation coefficient, count(P(A,B)) represents the number of times the time factor A and the position factor B appear at the same time, and count(P(i)) represents the number of times the multiple preset control behaviors appear; when it is determined that the behavior confidence is greater than the confidence threshold, it is determined that the behavior confidence satisfies the preset confidence matching rule.

[0019] In an exemplary embodiment, the use of the behavior association model to calculate the first behavior record of the first object to obtain the behavior measurement parameters of the first object includes: obtaining the first number of occurrences of the control behavior of the first object at a specified time and a specified location from the first behavior record; and determining the first number of occurrences as the behavior support.

[0020] In an exemplary embodiment, the use of the behavior association model to calculate the first behavior record of the first object to obtain the behavior measurement parameters of the first object, including: obtaining the first occurrence number of the control behavior of the first object at a specified time and a specified location from the first behavior record, and obtaining the second occurrence number of all control behaviors of the first object from the first behavior record; and determining the ratio of the first occurrence number to the second occurrence number as the behavior confidence.

[0021] According to another aspect of an embodiment of the present application, a data processing device based on a behavior association model is also provided, including: a processing module, used to perform time sequence processing on device control data of a first object controlling a home appliance according to a time sequence processing rule to obtain an object behavior set, wherein the object behavior set includes a first behavior record of the first object; a first obtaining module, used to perform clustering processing on multiple object behavior sets corresponding to different objects, and obtain a group to which the first object belongs based on the result of the clustering processing; a second obtaining module, used to calculate the first behavior record of the first object using the behavior association model to obtain behavior measurement parameters of the first object, wherein the behavior measurement parameters include behavior support and behavior confidence; a determination module, used to determine the preset control behavior set for the group to which the first object belongs in different time periods, and determine the target control behavior of the first object from the preset control behavior based on the behavior measurement parameters; a pushing module, used to set the target control behavior set generated based on the target control behavior as recommended data of the first object, and push the recommended data of the first object to the control terminal of the first object.

[0022] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned data processing method based on the behavior association model when running.

[0023] According to another aspect of an embodiment of the present application, an electronic device is also provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the data processing method based on the behavior association model through the computer program.

[0024] According to another embodiment of the present application, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0025] In an embodiment of the present application, a time sequence processing is performed on the device control data of the first object controlling the home appliance according to the time sequence processing rule to obtain an object behavior set, wherein the object behavior set includes a first behavior record of the first object; clustering processing is performed on multiple object behavior sets corresponding to different objects, and the group to which the first object belongs is obtained based on the result of the clustering processing; the first behavior record of the first object is calculated using a behavior association model to obtain a behavior measurement parameter of the first object, and the behavior measurement parameter includes a behavior support and a behavior confidence; the preset control behavior set for the group to which the first object belongs in different time periods is determined, and the target control behavior of the first object is determined from the preset control behavior based on the behavior measurement parameter; the target control behavior set generated based on the target control behavior is set as the recommended data of the first object, and the recommended data of the first object is pushed to the control terminal of the first object; that is, after binding the user data and the user operating the home appliance, the present application can collect the behavior record data (i.e., control data) of the user operation in real time when the user operates the home appliance, and then group the users, and perform behavior association calculation on the grouped users respectively, so as to obtain the user's behavior habit data, and finally recommend the user behavior (i.e., target control behavior) to the user in real time. The above technical solution solves the technical problem of inaccurate recommended data when recommending data to users, thereby improving the accuracy of recommended data and enhancing user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0027] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0028] Figure 1 It is a hardware environment diagram of a data processing method based on a behavior association model in an embodiment of the present application.

[0029] Figure 2 It is a flowchart of a data processing method based on a behavior association model according to an embodiment of the present application.

[0030] Figure 3 It is a schematic diagram of data flow according to an embodiment of the present application.

[0031] Figure 4It is a schematic diagram of a data processing method based on a behavior association model according to an embodiment of the present application.

[0032] Figure 5 It is a structural block diagram of a data processing device based on a behavior association model according to an embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0034] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0035] According to one aspect of an embodiment of the present application, a data processing method based on a behavior association model is provided. The data processing method based on a behavior association model is widely used in smart home (Smart Home), smart home, smart home device ecology, smart residential (Intelligence House) ecology and other whole-house intelligent digital control application scenarios. Optionally, in this embodiment, the above-mentioned data processing method based on a behavior association model can be applied to Figure 1 In the hardware environment composed of the terminal device 102 and the server 104 shown in FIG. Figure 1 As shown, the server 104 is connected to the terminal device 102 via a network, and can be used to provide services (such as application services, etc.) for the terminal or a client installed on the terminal. A database can be set on the server or independently of the server to provide data storage services for the server 104. Cloud computing and / or edge computing services can be configured on the server or independently of the server to provide data computing services for the server 104.

[0036] The network may include but is not limited to at least one of the following: wired network, wireless network. The wired network may include but is not limited to at least one of the following: wide area network, metropolitan area network, local area network, and the wireless network may include but is not limited to at least one of the following: WIFI (Wireless Fidelity), Bluetooth. The terminal device 102 may be but is not limited to a PC, a mobile phone, a tablet computer, a smart air conditioner, a smart range hood, a smart refrigerator, a smart oven, a smart stove, a smart washing machine, a smart water heater, a smart washing equipment, a smart dishwasher, a smart projection equipment, a smart TV, a smart clothes drying rack, a smart curtain, a smart audio and video, a smart socket, a smart speaker, a smart speaker, a smart fresh air equipment, a smart kitchen and bathroom equipment, a smart bathroom equipment, a smart sweeping robot, a smart window cleaning robot, a smart mopping robot, a smart air purification equipment, a smart steamer, a smart microwave oven, a smart kitchen treasure, a smart purifier, a smart water dispenser, a smart door lock, etc.

[0037] In this embodiment, a data processing method based on a behavior association model is provided, which is applied to the above terminal device. Figure 2 It is a flowchart of a data processing method based on a behavior association model according to an embodiment of the present application, and the process includes the following steps: Step S202, timing processing is performed on the device control data of the first object controlling the home appliance according to the timing processing rule to obtain an object behavior set, wherein the object behavior set includes the first behavior record of the first object; wherein the device control data can be understood as the user's control instructions to the home appliance and the operation data generated by the home appliance executing the control instructions, that is, the device control data at least includes the control instructions received by the home appliance and the operation data generated by the home appliance executing the control instructions.

[0038] Step S204, clustering the multiple object behavior sets corresponding to different objects, and obtaining the group to which the first object belongs based on the result of the clustering process; optionally, this step is a process of the user determining the group to which he belongs, which can be understood as actively determining the group to which the first object belongs when monitoring the first object controlling the home appliance, or it can be based on the request initiated by the first object to push recommended data, after determining the group to which the first object belongs, and then executing subsequent processes. Alternatively, it can also be to determine the groups to which all objects belong in advance, and then determine the group to which the first object belongs from the groups to which all objects belong. This application does not limit the conditions for determining the group to which the user belongs.

[0039] Step S206, use the behavior association model to calculate the first behavior record of the first object to obtain the behavior measurement parameters of the first object, and the behavior measurement parameters include behavior support and behavior confidence; Step S208, determine the preset control behavior set for the group to which the first object belongs in different time periods, and determine the target control behavior of the first object from the preset control behavior based on the behavior measurement parameters; Step S210, set the target control behavior set generated based on the target control behavior as the recommended data of the first object, and push the recommended data of the first object to the control terminal of the first object.

[0040] Through the above steps, the device control data of the first object controlling the home appliance is processed in time sequence according to the time sequence processing rule to obtain an object behavior set, wherein the object behavior set includes the first behavior record of the first object; clustering processing is performed on multiple object behavior sets corresponding to different objects, and the group to which the first object belongs is obtained based on the result of clustering processing; the first behavior record of the first object is calculated using a behavior association model to obtain a behavior measurement parameter of the first object, and the behavior measurement parameter includes a behavior support and a behavior confidence; the preset control behavior set for the group to which the first object belongs in different time periods is determined, and the target control behavior of the first object is determined from the preset control behavior based on the behavior measurement parameter; the target control behavior set generated based on the target control behavior is set as the recommended data of the first object, and the recommended data of the first object is pushed to the control terminal of the first object; that is, after binding the user data and the user operating the home appliance, the present application can collect the behavior record data (i.e., control data) of the user operation in real time when the user operates the home appliance, and then group the users, and perform behavior association calculation on the grouped users respectively, so as to obtain the user's behavior habit data, and finally recommend the user behavior (i.e., target control behavior) to the user in real time. The above technical solution solves the technical problem of inaccurate recommended data when recommending data to users, thereby improving the accuracy of recommended data and enhancing user experience.

[0041] like Figure 3As shown, firstly, the user is bound to different home appliances respectively, and then the binding relationship after the user binds the device can be sent to the cloud server through wireless technology such as Bluetooth. In this way, when the user controls the home appliance, the control data and binding relationship corresponding to the home appliance can be sent to the cloud server, and the user habit recognition model built into the Hadoop big data platform corresponding to the cloud server is used to group users based on the control data, and the associated behavior of the grouped users is calculated to obtain the user behavior recommended for the user. Among them, the user habit recognition model can be used to implement the above-mentioned method of determining the target control behavior of the first object.

[0042] In an exemplary embodiment, the control data at least includes the control instruction received by the home appliance and the operation data generated by the home appliance executing the control instruction. Then, the implementation process of processing the control data when different objects control the home appliance according to the timing processing rule to obtain the object behavior set in the above step S202 may include: the device control data at least includes the control instruction received by the home appliance and the operation data generated by the home appliance executing the control instruction, performing timing processing on the device control data when the first object controls the home appliance according to the timing processing rule to obtain the object behavior set, including: processing the operation data generated by the home appliance The generation time of each object is bound to the operation data to obtain a second behavior record, wherein the second behavior record includes the generation time and the operation data after binding, and the operation data at least includes the object number of the different objects, the location data of the different objects, the behavior data of the different objects, and the device type of the home appliance; the generation time is sorted according to the timing processing rule to obtain the sorted generation time, wherein the timing processing rule indicates processing in chronological order; the multiple second behavior records are sorted according to the sorted generation time, and the object behavior set is generated based on the sorted multiple second behavior records.

[0043] Exemplarily, the above-mentioned second behavior record is expressed as "User U01 turns on the lights in the living room at 18:00 on July 15, 2022". If multiple second behavior records include "User U01 turns on the lights in the living room at 18:00 on July 15, 2022", "User U01 turns on the lights in the living room at 19:00 on July 15, 2022", and "User U01 turns on the lights in the living room at 15:00 on July 15, 2022", then the object behavior set generated according to the sorted multiple second behavior records is ["User U01 turns on the lights in the living room at 15:00 on July 15, 2022", "User U01 turns on the lights in the living room at 18:00 on July 15, 2022", "User U01 turns on the lights in the living room at 19:00 on July 15, 2022"].

[0044] The first behavior record in the above embodiment can be understood as a behavior record of the object behavior set generated from this embodiment.

[0045] In an exemplary embodiment, after timing processing is performed on device control data of a home appliance controlled by a first object according to a timing processing rule to obtain an object behavior set, further for the first behavior record of the first object included in the object behavior set, at least a first generation moment and first operation data bound to the first generation moment are determined from the first behavior record, wherein the first operation data at least includes: the first object number of the first object, the first position of the first object, the first object behavior of the first object and the first device type of the home appliance; for multiple first behavior records, when it is determined that the time interval between the multiple first generation moments is less than a preset interval, invalid behavior records are deleted from the multiple first behavior records according to the earliest first generation moment to obtain remaining behavior records, wherein the invalid behavior records represent behavior records corresponding to other generation moments in the multiple first behavior records except the earliest first generation moment; the object behavior set is updated based on the remaining behavior records to obtain an updated object behavior set.

[0046] In this embodiment, by deleting invalid behavior records from the second behavior records, invalid associated behavior records can be removed from the record data of the user's home appliance operation behavior, so as to improve the accuracy of the subsequent recommended data.

[0047] In an exemplary embodiment, other technical solutions are further proposed for performing timing processing on device control data of a home appliance controlled by a first object according to timing processing rules, specifically including: performing data cleaning processing on the device control data to obtain cleaned device control data; wherein the data cleaning processing includes at least one of the following: null value processing, abnormal value processing; obtaining a verification result of data verification on the cleaned device control data; when it is determined that the verification result passes the verification, performing format standardization processing on the cleaned device control data to obtain standard data, and processing the standard data according to the timing processing rules.

[0048] Optionally, for format standardization, since different features may have different dimensions and numerical ranges, the data needs to be standardized or normalized so that the clustering algorithm treats each feature fairly.

[0049] Optionally, for example, for the time data in the device control data, it is converted from a timestamp format to a date format such as year, month, and day, and then the standard data is processed according to the time sequence processing rules, that is, the year, month, day, etc. are sorted in chronological order.

[0050] Optionally, when it is determined that the verification result indicates that the control data has not passed the verification, the cleaned device control data is sent to the user, and the user is prompted to verify the device control data.

[0051] In an exemplary embodiment, the different objects include at least the first object and the second object, and clustering processing of multiple object behavior sets corresponding to different objects can be achieved in the following manner, and the group to which the first object belongs is obtained based on the result of clustering processing, including: determining M behavior records that have been assigned to a group from the multiple object behavior sets, and the M behavior records represent the current center point required for clustering grouping; determining a third behavior record that belongs to an unassigned group from the multiple object behavior sets, and calculating the first Euclidean distance between the first eigenvalue of the third behavior record and the second eigenvalue of the M behavior records; determining a fourth behavior record corresponding to the smallest first Euclidean distance from the M behavior records, and assigning a group to the third behavior record according to the group to which the fourth behavior record belongs, and the group to which the third behavior record belongs is associated with the fourth behavior record. The group to which the record belongs is consistent; re-determine L behavior records that have been assigned a group, and the L behavior records represent the updated center point required for clustering grouping; determine the fifth behavior record corresponding to the smallest second Euclidean distance from the L behavior records, and assign a group to the sixth behavior record that has not been assigned a group among the L behavior records according to the group to which the fifth behavior record belongs, and the second Euclidean distance represents the Euclidean distance between the third eigenvalue of the sixth behavior record and the fourth eigenvalue of the L behavior records; if it is determined that the current center point is consistent with the updated center point, then set a belonging group for the first object according to the belonging group of the third behavior record, and set a belonging group for the second object according to the belonging group of the sixth behavior record, wherein the third behavior record represents a behavior record belonging to the first object, and the sixth behavior record represents a behavior record belonging to the second object.

[0052] It is understandable that the above M and L are both positive integers, and M is less than L. For example, 7 behavior records that have been assigned to a group are determined from the plurality of object behavior sets, and after the groups are assigned, 8 behavior records that have been assigned to a group are re-determined. Of course, for a plurality of third behavior records, more L behavior records can also be re-determined, which will not be described in detail in this application.

[0053] It should be noted that, in order to adapt to the calculation of Euclidean distance, the data required for calculating the Euclidean distance must be converted into numerical type in advance, so that the data required by distance-based algorithms such as clustering algorithms can be measurable in Euclidean space.

[0054] Optionally, for the M behavior records that have been assigned to groups, it can be understood that the objects described in the M behavior records have been grouped, and the object corresponding to each behavior record has a group to which it belongs.

[0055] The process of using the third behavior record to update the M behavior records can be understood as adding the third behavior record assigned to the current center point to the behavior record of the corresponding center point to complete the update process of the M behavior records. For example, if the group to which the third behavior record belongs is consistent with the group to which the i-th behavior record belongs among the M behavior records, the third behavior record is added to the i-th behavior record to update the M behavior records.

[0056] In an exemplary embodiment, after determining M behavior records that have been assigned to a group from the plurality of object behavior sets, the method further includes: selecting a current value of M from a preset value interval for M, and calculating a silhouette coefficient corresponding to the current value; determining a target value having a maximum silhouette coefficient from a plurality of silhouette coefficients, and determining the target value as the current value of M; wherein the silhouette coefficient S(i) is calculated by the following formula: ; Where a(i) represents the cohesion of the current value, b(i) represents the separation of the current value, and i is a positive integer. Where i represents the sample point for calculating the silhouette coefficient, which can be understood as the current value mentioned above.

[0057] Silhouette coefficient is often used in computer vision and image analysis to describe and quantify the shape characteristics of objects in images.

[0058] The cohesion used in calculating the silhouette coefficient is a measure of the compactness of an object's shape, and is usually used to describe how tightly the points around the geometric center of an object are distributed. A shape with high cohesion means that its boundary points are closer to the geometric center. The calculation method of cohesion may vary depending on the specific application, but a common formula is to compare the ratio of the area of ​​the object to the length of its boundary with the equivalent circular diameter of the object. The closer this comparison value is to 1, the more compact the shape is.

[0059] Separation is often used to measure the spatial distribution and distance between different objects in an image. High separation means that there are clear boundaries and distances between objects, while low separation may indicate that objects overlap or are close to each other. In some cases, separation can be associated with the distance between the center points of the minimum covering circle or bounding box of the object, or evaluated by calculating the distance distribution between objects.

[0060] In an exemplary embodiment, further, the first Euclidean distance between the first characteristic value of the third behavior record and the second characteristic value of the M behavior records is calculated respectively by the following steps, including: normalizing the record value of the third behavior record to obtain the first characteristic value, wherein the record value of the third behavior record at least includes a second generation moment and second operation data bound to the second generation moment, wherein the second operation data at least includes: the second position of the second object, the second object behavior of the second object and the second device type of the household appliance, and the first characteristic value at least includes a first moment characteristic value corresponding to the second generation moment, a first position characteristic value of the second position, a first behavior characteristic value of the second object behavior and a first type characteristic value of the second device type value; normalize the record value of each behavior record in the M behavior records to obtain the second feature value, wherein the record value of each behavior record at least includes a third generation time and third operation data bound to the third generation time, wherein the third operation data represents data of a third object corresponding to each behavior record, and at least includes: a third position of the third object, a third object behavior of the third object and a third device type of the home appliance, and the second feature value at least includes a second time feature value corresponding to the third generation time, a second position feature value of the third position, a second behavior feature value of the third object behavior and a second type feature value of the third device type; calculate the first Euclidean distance between the first feature value and the second feature value according to the following formula: , where d(m, n) is the first Euclidean distance, w1 is the feature value at the first moment, x1 is the feature value at the first position, y1 is the feature value at the first behavior, z1 is the feature value of the first type, w2 is the feature value at the second moment, x2 is the feature value at the second position, y2 is the feature value of the second behavior, and z2 is the feature value of the second type.

[0061] It should be noted that the calculation process of the second Euclidean distance is similar to the calculation process of the first Euclidean distance, which will not be repeated here.

[0062] Through this embodiment, different users are grouped based on clustering algorithms such as K-Means to identify user groups that meet the recommendations. Behavior association calculations can be performed for users in each group to determine the control behaviors of each group's users on home appliances, thereby improving the accuracy of recommending control behaviors to users.

[0063] It should be noted that the clustering algorithm is an unsupervised learning algorithm commonly used in data analysis and machine learning. Its purpose is to group objects in the data set so that objects in the same group are highly similar to each other, but have low similarity to objects in other groups. In other words, using a clustering algorithm to group users is a process of dividing users into different groups based on different characteristics and similarities.

[0064] There are many types of clustering algorithms, including but not limited to: K-means clustering algorithm: This is one of the most commonly used clustering algorithms and optimizes clustering by iteratively moving cluster centers and reassigning objects to the nearest cluster center.

[0065] Hierarchical clustering algorithm: This method does not require the number of clusters to be specified in advance. It gradually aggregates objects by creating a cluster tree (dendrogram).

[0066] DBSCAN clustering algorithm: A density-based spatial clustering algorithm that can handle clusters of any shape and can identify noisy data.

[0067] Spectral clustering algorithm: Based on the similarity matrix of data, spectral theory in graph theory is used to perform clustering. Clustering algorithms are widely used in market segmentation, social network analysis, biodata science, image segmentation and other fields.

[0068] The principles of clustering algorithms can be summarized as follows: 1. Similarity metric: The clustering algorithm first needs to define a similarity metric, such as Euclidean distance, Manhattan distance, cosine similarity, etc., to evaluate the similarity between objects. The Euclidean distance is selected in this application.

[0069] 2. Select initial centers: Most clustering algorithms require specifying or automatically selecting initial cluster centers. These centers can be points in the data set or randomly generated points, such as the current center point or updated center point mentioned above.

[0070] 3. Assign objects to the nearest center: Based on the similarity measure, assign each object to the nearest cluster center to form an initial cluster.

[0071] 4. Update cluster centers: Recalculate the center of each cluster, which is usually based on the average position of the cluster members or some form of centroid.

[0072] 5. Iterative optimization: Repeat steps 3 and 4 until the cluster center no longer changes significantly or the preset number of iterations is reached.

[0073] 6. Evaluate clustering effect: Use indicators such as silhouette coefficient and Davidson-Botting index to evaluate clustering effect.

[0074] 7. Determine the number of clusters: In some algorithms, the number of clusters needs to be specified in advance, while in other algorithms, such as DBSCAN, the number of clusters is the result of running the algorithm.

[0075] 8. Clustering is completed: When the termination condition is met, the clustering process ends and the final clustering result is obtained.

[0076] Optionally, the clustering process is described through the following steps: Step 1. Data collection: First, it is necessary to collect relevant user data, which may include the user's personal data, behavioral characteristics, purchase history, browsing habits, etc.

[0077] Step 2. Feature Selection: Determine which features are important for grouping users. This may include age, income, purchase frequency, preferred product types, etc.

[0078] Step 3. Data preprocessing: Clean the data, handle missing values, convert data formats, etc. to facilitate clustering algorithm processing.

[0079] Step 4. Select a clustering algorithm: Select an appropriate clustering algorithm based on the characteristics of the data and business requirements. Common clustering algorithms include K-means, hierarchical clustering, DBSCAN, Gaussian mixture model, etc.

[0080] Step 5. Determine the clustering parameters: For some clustering algorithms (such as K-means), the number of groups to be divided (K value) needs to be determined in advance. This can be assisted by methods such as the Elbow Method and the Silhouette Coefficient.

[0081] Step 6. Perform clustering: Use the selected clustering algorithm to perform cluster analysis on the data.

[0082] Step 7. Evaluate clustering results: Use clustering evaluation indicators (such as silhouette coefficient, Davidson-Boting index, etc.) to evaluate the quality of clustering results.

[0083] Step 8. Explain the clustering results: Explain the user groups based on the clustering results and identify the characteristics of each group.

[0084] Step 9. Apply clustering results: Apply clustering results to actual business, such as personalized recommendation, targeted recommendation, etc.

[0085] Step 10. Iterative optimization: Continuously iterate and optimize the clustering process based on business feedback and performance indicators.

[0086] The output of a clustering algorithm is usually a set of user groups, where the members within each group have highly similar characteristics, while the members of different groups have obvious differences in characteristics. Through clustering, companies can more effectively formulate marketing strategies for different user groups.

[0087] In an exemplary embodiment, multiple technical solutions are proposed for respectively calculating the first Euclidean distance between the first feature value of the third behavior record and the second feature value of the M behavior records, specifically including: Solution 1, obtaining the first object label of the first object and the second object label of the second object, wherein the first object label is determined based on the object attribute and the first object attribute value of the first object, and the second object label is determined based on the object attribute and the second object attribute value of the second object; normalizing the first object label to obtain a first label feature value, and updating the first feature value by incorporating the first label feature value into the first feature value; and normalizing the second object label to obtain a second label feature value, and updating the second feature value by incorporating the second label feature value into the second feature value; calculating the first Euclidean distance between the updated first feature value and the updated second feature value.

[0088] Solution 2, obtaining a first object tag of the first object and a second object tag of the second object, wherein the first object tag is determined based on the object attribute of the first object and the first object attribute value, and the second object tag is determined based on the object attribute of the second object and the second object attribute value; normalizing the first object tag to obtain a first tag feature value, and updating the first feature value by incorporating the first tag feature value into the first feature value with a first weight; normalizing the second object tag to obtain a second tag feature value, and updating the second feature value by incorporating the second tag feature value into the second feature value with a second weight, wherein the first weight is different from the second weight; calculating the first Euclidean distance between the updated first feature value and the updated second feature value. In the above solution, the object attribute can be understood as the biological attributes of the user (that is, the first object and other different objects), such as gender, age, height, weight, etc., and the object attribute value is the corresponding specific value. For example, for the attribute of age, the object attribute value can be 10 years old.

[0089] It can be understood that the process of incorporating the first label feature value into the first feature value with the first weight means that when the first label feature value is incorporated into the first feature value in the form of a new category, the product of the first label feature value and the first weight is incorporated into the first feature value in the form of a new category.

[0090] Optionally, incorporating the first tag feature value into the first feature value to update the first feature value can be understood as incorporating the first tag feature value into the first feature value in the form of a new category. For example, the first feature value includes the first moment feature value of the corresponding category at the second generation moment, the first position feature value under the corresponding category of the second position, the first behavior feature value under the corresponding category of the second object behavior, and the first type feature value under the corresponding category of the second device type. Incorporating the first tag feature value into the first feature value in the form of a new category means incorporating the first tag feature value under the tag category into the first feature value.

[0091] It can be understood that the updated first feature value includes at least the first moment feature value corresponding to the second generation moment, the first position feature value of the second position, the first behavior feature value of the second object behavior, the first type feature value of the second device type and the first label feature value. The updated second feature value includes at least the second moment feature value corresponding to the third generation moment, the second position feature value of the third position, the second behavior feature value of the third object behavior, the second type feature value of the third device type and the second label feature value.

[0092] Through the above scheme, based on the algorithm principle of the clustering algorithm, different objects are grouped and the group to which each object belongs is determined. The method of behavior recommendation can be implemented for the objects in each group respectively, thereby improving the personalization of the recommended data.

[0093] In an exemplary embodiment, the implementation process of determining the target control behavior of the first object from the preset control behaviors based on the behavior measurement parameters specifically includes: determining a first control behavior from multiple preset control behaviors whose behavior support satisfies a preset support matching rule and whose behavior confidence satisfies a preset confidence matching rule, wherein the first control behavior corresponds to a behavior occurrence time period; when it is determined that the behavior occurrence time period belongs to the preset time period, determining the first control behavior as the target control behavior.

[0094] The determination that the time period of the behavior occurs belongs to the preset time period may be, for example, that the preset time period is consistent with the time period of the behavior. For example, if the time period of the behavior occurs at 2 pm, the preset time period is also 2 pm.

[0095] In an exemplary embodiment, it can be determined in the following manner that the behavior support satisfies the preset support matching rule: the behavior support is calculated by the following formula: Esupport = sum(count(P(A,B))*K*T); wherein Esupport is the behavior support, K ​​is the behavior coefficient, T is the time attenuation coefficient, and count(P(A,B)) represents the number of times the time factor A and the location factor B appear at the same time; when it is determined that the behavior support is greater than the support threshold, it is determined that the behavior support satisfies the preset support matching rule.

[0096] It is understandable that the values ​​of K, T, and Esupport are [0, 1]. Esupport is, for example, [0, 99]. This application does not limit the maximum value of the behavior support. The behavior coefficient can be preset, and the artificial experience value can be used.

[0097] In an exemplary embodiment, it can be determined in the following manner that the behavior confidence satisfies the preset confidence matching rules: the behavior confidence is calculated by the following formula: Econfidence = sum(count(P(A,B))*K*T) / sum(count(P(i))*K*T); wherein Econfidence is the behavior confidence, K is the behavior coefficient, T is the time attenuation coefficient, count(P(A,B)) represents the number of times the time factor A and the position factor B appear at the same time, and count(P(i)) represents the number of times the multiple preset control behaviors appear; when it is determined that the behavior confidence is greater than the confidence threshold, it is determined that the behavior confidence satisfies the preset confidence matching rules.

[0098] It can be understood that the values ​​of K, T, and Econfidence above are [0, 1].

[0099] In an exemplary embodiment, the use of the behavior association model to calculate the first behavior record of the first object to obtain the behavior measurement parameters of the first object can also be achieved through the following process: obtaining the first number of occurrences of the control behavior of the first object at a specified time and a specified location from the first behavior record; and determining the first number of occurrences as the behavior support.

[0100] Optionally, for the specified time and the specified location, it can be understood that it is set based on the historical time and the historical location. In this way, based on the user's control behavior at the historical time and the historical location, the control behavior recommended for the user at the same time and the same location within a certain period of time in the future can be reasonably set. For example, referring to the user's historical control behavior of opening the window in the living room at 10 o'clock on October 10, 2021, the predicted control behavior of opening the window in the living room at 10 o'clock every day can be set in the prediction process. Among them, the specified time and the specified location can be changed.

[0101] In an exemplary embodiment, the process of calculating the first behavior record of the first object using the behavior association model to obtain the behavior measurement parameters of the first object can also be implemented by the following steps: obtaining the first occurrence number of the control behavior of the first object at a specified time and a specified location from the first behavior record, and obtaining the second occurrence number of all control behaviors of the first object from the first behavior record; and determining the ratio of the first occurrence number to the second occurrence number as the behavior confidence.

[0102] Optionally, in one embodiment, the high-frequency control behaviors of users in each time period may be sorted according to the groups to which the users belong, and then push rules may be configured according to application scenarios to provide users with a butler service that reminds them of recommended behaviors.

[0103] In order to better understand the process of the above-mentioned data processing method based on the behavior association model, the process of the above-mentioned data processing method based on the behavior association model is described below in combination with an optional embodiment, but it is not used to limit the technical solution of the embodiment of the present application.

[0104] In this embodiment, combined Figure 4 A data processing method based on a behavior association model is provided. Figure 4 is a schematic diagram of a data processing method based on a behavior association model according to an embodiment of the present application. It should be noted that in this embodiment, the following contents need to be defined: A: a set of operation record data when a user operates a home appliance (i.e., the above-mentioned object behavior set). The behavior record of the operation record data set is represented as a(x), i.e., the behavior record of a certain home appliance operation by the user, a(a1, a2, ..., a k ) represents the 1st to kth home appliance operation records of the user, such as the home appliance operation records of the user turning on the air conditioner, the user turning off the air conditioner, and the user turning on the TV.

[0105] B: User effective behavior result data set (i.e., corresponding to the above updated object behavior set). The associated behavior record of the user effective behavior result data set is represented as b(x), i.e., the associated behavior record of the user for a certain effective home appliance operation, b(b1, b2, ..., bk ) represents the 1st to kth valid home appliance operation behavior records of the user. For example, valid behavior records such as user ID, behavior time period, user turning on the air conditioner; user ID, behavior time period, user air conditioner selecting smart mode, etc.

[0106] G: User grouping table. Use g(x) to represent a user and the group data to which the user belongs, g(g1, g2, …, g k ) represents the grouping data of users 1 to k. For example, Zhang San is the first user group, Li Si is the second user group, Wang Wu is the third user group, etc.

[0107] C: User cluster center point. Use c(x) to represent the data of a user cluster center point, c(c1, c2, ..., c k ) represents the data of the center points of the 1st to kth user clusters. The center point of the cluster is the center point required for clustering grouping.

[0108] H: User behavior habit data set. Use h(x) to represent a user's behavior habit data, h(h1, h2, ..., h k ) represents the user's 1st to kth behavior habit records. For example, user ID, behavior time period, living room, user turning on the air conditioner, user air conditioner selecting smart mode, support, confidence, door opening and other behavior habits.

[0109] like Figure 4 As shown, the specific steps are as follows: Step S401, user binding. Figure 3 As shown, the user is bound to the home appliance.

[0110] Step S402: Collect data on the user's operation of the home appliance and process the data according to certain rules.

[0111] First, the existing bound user data can be standardized to identify unique users. The standardization process may include data cleaning, data formatting, etc.

[0112] Furthermore, the user's home appliance operation record data can be associated and processed.

[0113] like Figure 3 As shown in the figure, after the user operates the home appliance on the device side, the device backplane collects the operation record data and sends it to the cloud server in real time through the network. The data is transmitted to the data platform through Kafka for data calculation.

[0114] Specifically, it can be processed according to the time sequence processing rules: for the behavior record a(k), the user's home appliance operation data is processed for behavior association according to the time sequence processing rules to obtain the user's effective behavior result data set B1 (which can be understood as the above-mentioned object behavior set) shown in the following Table 1. Among them, the time in Table 1 corresponds to the above-mentioned generation time, the position in Table 1 corresponds to the above-mentioned position data, the device in Table 1 corresponds to the device type of the above-mentioned home appliance device, and the behavior in Table 1 represents the control behavior.

[0115] Table 1

[0116]

[0117] In the home appliance operation records, if the same device performs the same action twice in a row within a certain time interval, the second action record is considered invalid relative to the first action record.

[0118] The third record 'turn on the air conditioner' in Table 1 has the same action as the second record 'turn on the air conditioner'. Therefore, the third record is an invalid associated action record relative to the second record. By deleting the invalid associated action record in the valid action result data set B1, the user valid action result data set B2 shown in Table 2 below can be obtained.

[0119] Table 2

[0120]

[0121] Step S403: Determine the user tag. The user tag set B3 set for the user is shown in Table 3 below.

[0122] Table 3

[0123]

[0124] Among them, the labels in Table 3 correspond to weight coefficients, and the value range of the weight coefficients is [0, 1].

[0125] Step S404: Perform user grouping. The user grouping calculation can be performed using the K-Means clustering algorithm model based on the user's effective home appliance operation record data set B2. To avoid confusion with the parameters of the formula below, the K value of the K-Means clustering algorithm model is replaced by M here.

[0126] For example, if you plan to divide users into M groups, continue to perform the following steps: Step 1, select the initial center point.

[0127] Based on experience, M samples (i.e., M grouped users) are selected from the users as the initial center points of the M cluster groups. These center points can be called Means.

[0128] Step 2: Calculate the distance.

[0129] According to the user's effective home appliance operation record data set B2, the Euclidean distance is used to calculate the distance between each ungrouped user and the M sample center point users. The calculation formula is: , where the corresponding multidimensional feature vector of point m is [x1, y1, z1…], and the corresponding multidimensional feature vector of point n is [x2, y2, z2…].

[0130] Step 3: Clustering and grouping.

[0131] According to the distance between each ungrouped user and the M sample center point users, the users are assigned to the center point users closest to them and added to the group corresponding to the center point.

[0132] Step 4: Calculate the center point.

[0133] After all users are assigned to the corresponding groups, recalculate the center point of each group.

[0134] Calculation formula: C1=(g(1)+ g(2)+…+ g(n)) / count(G(x)).

[0135] Step 5: Iterate until convergence.

[0136] Continue to iterate steps 2 to 4 until the clustering algorithm converges. In other words, if the center point of each group does not change after multiple iterations, and the center point of the group calculated this time is consistent with the center point of the group calculated last time, then the clustering algorithm has converged.

[0137] According to the characteristic data after clustering of each group, the corresponding user group data is divided as shown in Table 4.

[0138] Table 4

[0139]

[0140] Step S405, performing association algorithm calculation. User behavior association calculation can be performed based on the user effective behavior result data set B2, including the following steps.

[0141] Step 1: Calculate the user behavior support. The value range of the behavior support is not limited, for example, it is set to [0, 99].

[0142] Behavior support is the threshold of the number of times two different factors appear simultaneously in the statistical grouping, and the two factors are 'time + location' and 'behavior'.

[0143] P(A, B) is used to indicate that factor A and factor B appear at the same time. The calculation formula is count(P(A, B)). Group by user, time + location, and behavior, and calculate the support statistics for each behavior. As shown in Table 5 below, at 17:00, user U01 turned on the air conditioner in the living room 2 times, so the support is 2.

[0144] Table 5

[0145]

[0146] Step 2: Calculate the user behavior confidence. The value range of behavior confidence is [0, 1].

[0147] The behavior confidence indicates the threshold of the number of times two different factors appear at the same time in the statistical grouping. The two factors are 'time + location' and 'behavior'. P(ALL) is used to represent all valid user behavior records, and the calculation formula is count(P(A, B)) / count(P(ALL)). Group by user, time + location, and behavior, and calculate the confidence data corresponding to each behavior action. As shown in Table 6 below, at 17:00, the number of times user U01 turned on the air conditioner in the living room was 1, that is, count(P(A, B)) was 1, and at 17:00, the total number of times user U01 performed all device operations in the living room was 5, that is, count(P(ALL)) was 5, so the behavior confidence is 20%.

[0148] Table 6

[0149]

[0150] Step S406: Marking user habits. Data with a behavior support greater than or equal to 4 and a behavior confidence greater than or equal to 40% are identified as user habit record data.

[0151] Step S407, recommending the user's habitual behavior, which can be reminded and executed by the smart home butler.

[0152] According to the group to which the user belongs, the recommended behavior of the user in the corresponding time period is determined as shown in Table 7 below.

[0153] Table 7

[0154]

[0155] Furthermore, data can be recorded based on user behavior habits to provide different recommended behaviors to the user. As shown in Table 8 below, the recommended behavior for user U01 in the living room at 19:00 is to turn on the air conditioner.

[0156] Table 8

[0157]

[0158] Step S408: The user performs the recommended habitual behavior.

[0159] Through the above embodiment, based on user data and user home appliance operation behavior data, user label processing is first performed, and user groups are grouped according to the clustering algorithm to standardize the user habit range pushed to different groups of people and improve data accuracy. Then, the association algorithm is used to perform association analysis on users, and effective user habit data is calculated according to the support parameter and confidence parameter to perform user habit recommendation. In this way, the clustering algorithm and the association algorithm can be combined to obtain recommended data, thereby improving the accuracy of the recommended data.

[0160] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0161] Figure 5 is a structural block diagram of a data processing device based on a behavior association model according to an embodiment of the present application; Figure 5 As shown, it includes: a processing module 502, which is used to process the control data when different objects control household appliances according to the timing processing rules to obtain an object behavior set; a first obtaining module 504, which is used to cluster the multiple object behavior sets corresponding to different objects, and obtain the group to which the first object belongs based on the result of the clustering processing; a second obtaining module 506, which is used to calculate the first behavior record of the first object using a behavior association model to obtain the behavior measurement parameters of the first object, and the behavior measurement parameters include behavior support and behavior confidence; a determination module 508, which is used to determine the preset control behavior set for the group to which the first object belongs in different time periods, and determine the target control behavior of the first object from the preset control behavior based on the behavior measurement parameters; a push module 510, which is used to set the target control behavior set generated based on the target control behavior as the recommended data of the first object, and push the recommended data of the first object to the control terminal of the first object.

[0162] Through the above-mentioned device and the above-mentioned steps, by performing time sequence processing on the device control data of the first object controlling the home appliance according to the time sequence processing rule, an object behavior set is obtained, wherein the object behavior set includes the first behavior record of the first object; clustering processing is performed on multiple object behavior sets corresponding to different objects, and the group to which the first object belongs is obtained based on the result of clustering processing; the first behavior record of the first object is calculated using a behavior association model to obtain a behavior measurement parameter of the first object, and the behavior measurement parameter includes a behavior support and a behavior confidence; the preset control behavior set for the group to which the first object belongs in different time periods is determined, and the target control behavior of the first object is determined from the preset control behavior based on the behavior measurement parameter; the target control behavior set generated based on the target control behavior is set as the recommended data of the first object, and the recommended data of the first object is pushed to the control terminal of the first object; that is, after binding the user data and the user operating the home appliance, the present application can collect the behavior record data (i.e., control data) of the user operation in real time when the user operates the home appliance, and then group the users, and perform behavior association calculation on the grouped users respectively, so as to obtain the user's behavior habit data, and finally recommend the user behavior (i.e., target control behavior) to the user in real time. The above technical solution solves the technical problem of inaccurate recommended data when recommending data to users, thereby improving the accuracy of recommended data and enhancing user experience.

[0163] In an exemplary embodiment, the device control data includes at least the control instructions received by the household appliance and the operation data generated by the household appliance executing the control instructions, and the processing module is also used to: bind the generation time when the household appliance generates the operation data and the operation data to obtain a second behavior record, wherein the second behavior record includes the bound generation time and the operation data, and the operation data includes at least the object numbers of the different objects, the location data of the different objects, the behavior data of the different objects, and the device type of the household appliance; sort the generation time according to the timing processing rule to obtain the sorted generation time, wherein the timing processing rule indicates processing in chronological order; sort multiple second behavior records according to the sorted generation time, and generate the object behavior set based on the sorted multiple second behavior records.

[0164] In an exemplary embodiment, the processing module is also used to: for the first behavior record of the first object included in the object behavior set, at least determine from the first behavior record a first generation moment and first operation data bound to the first generation moment, wherein the first operation data at least includes: the first object number of the first object, the first position of the first object, the first object behavior of the first object and the first device type of the household appliance; for multiple first behavior records, when it is determined that the time interval between the multiple first generation moments is less than a preset interval, delete invalid behavior records from the multiple first behavior records according to the earliest first generation moment to obtain remaining behavior records, wherein the invalid behavior records represent behavior records corresponding to other generation moments in the multiple first behavior records except the earliest first generation moment; and update the object behavior set based on the remaining behavior records to obtain an updated object behavior set.

[0165] In an exemplary embodiment, the processing module is also used to: perform data cleaning processing on the device control data to obtain cleaned device control data; wherein the data cleaning processing includes at least one of the following: null value processing, abnormal value processing; obtain the verification result of data verification on the cleaned device control data; when it is determined that the verification result passes the verification, perform format standardization processing on the cleaned device control data to obtain standard data, and process the standard data according to the timing processing rules.

[0166] In an exemplary embodiment, the different objects include at least the first object and the second object, and the first obtaining module is further used to: determine M behavior records that have been assigned to a group from the multiple object behavior sets, and the M behavior records represent the current center point required for clustering grouping; determine a third behavior record that belongs to an unassigned group from the multiple object behavior sets, and respectively calculate the first Euclidean distance between the first eigenvalue of the third behavior record and the second eigenvalue of the M behavior records; determine a fourth behavior record corresponding to the smallest first Euclidean distance from the M behavior records, and assign a group to the third behavior record according to the group to which the fourth behavior record belongs, and the group to which the third behavior record belongs is consistent with the group to which the fourth behavior record belongs; and re-determine L behaviors that have been assigned to a group. records, the L behavior records represent the updated center points required for clustering grouping; determine the fifth behavior record corresponding to the smallest second Euclidean distance from the L behavior records, and assign a group to the sixth behavior record that has not been assigned a group in the L behavior records according to the group to which the fifth behavior record belongs, the second Euclidean distance represents the Euclidean distance between the third eigenvalue of the sixth behavior record and the fourth eigenvalue of the L behavior records; if it is determined that the current center point is consistent with the updated center point, then set the belonging group for the first object according to the belonging group of the third behavior record, and set the belonging group for the second object according to the belonging group of the sixth behavior record, wherein the third behavior record represents the behavior record belonging to the first object, and the sixth behavior record represents the behavior record belonging to the second object.

[0167] In an exemplary embodiment, the first obtaining module is further used to: select the current value of K from the preset value interval of K, and calculate the silhouette coefficient corresponding to the current value; determine the target value with the maximum silhouette coefficient from multiple silhouette coefficients, and determine the target value as the current value of K; wherein the silhouette coefficient S(i) is calculated by the following formula: ; Wherein, a(i) represents the cohesion of the current value, b(i) represents the separation of the current value, and i is a positive integer.

[0168] In an exemplary embodiment, the first obtaining module is also used to: normalize the record value of the third behavior record to obtain the first feature value, wherein the record value of the third behavior record at least includes a second generation time and second operation data bound to the second generation time, wherein the second operation data at least includes: the second position of the second object, the second object behavior of the second object and the second device type of the household appliance, and the first feature value at least includes a first time feature value corresponding to the second generation time, a first position feature value of the second position, a first behavior feature value of the second object behavior and a first type feature value of the second device type; record each of the M behavior records The value is normalized to obtain the second feature value, wherein the record value of each behavior record at least includes the third generation time and the third operation data bound to the third generation time, wherein the third operation data represents the data of the third object corresponding to each behavior record, and at least includes: the third position of the third object, the third object behavior of the third object and the third device type of the home appliance, and the second feature value at least includes the second time feature value corresponding to the third generation time, the second position feature value of the third position, the second behavior feature value of the third object behavior and the second type feature value of the third device type; the first Euclidean distance between the first feature value and the second feature value is calculated according to the following formula: , where d(m, n) is the first Euclidean distance, w1 is the feature value at the first moment, x1 is the feature value at the first position, y1 is the feature value at the first behavior, z1 is the feature value of the first type, w2 is the feature value at the second moment, x2 is the feature value at the second position, y2 is the feature value of the second behavior, and z2 is the feature value of the second type.

[0169] In an exemplary embodiment, the first obtaining module is also used to: obtain a first object label of the first object and a second object label of the second object, wherein the first object label is determined based on the object attribute and the first object attribute value of the first object, and the second object label is determined based on the object attribute and the second object attribute value of the second object; normalize the first object label to obtain a first label feature value, and update the first feature value by incorporating the first label feature value into the first feature value; and normalize the second object label to obtain a second label feature value, and update the second feature value by incorporating the second label feature value into the second feature value; calculate the first Euclidean distance between the updated first feature value and the updated second feature value.

[0170] In an exemplary embodiment, the first obtaining module is also used to: obtain a first object label of the first object and a second object label of the second object, wherein the first object label is determined based on the object attribute and the first object attribute value of the first object, and the second object label is determined based on the object attribute and the second object attribute value of the second object; normalize the first object label to obtain a first label feature value, and update the first feature value by incorporating the first label feature value into the first feature value with a first weight; normalize the second object label to obtain a second label feature value, and update the second feature value by incorporating the second label feature value into the second feature value with a second weight, wherein the first weight is different from the second weight; calculate the first Euclidean distance between the updated first feature value and the updated second feature value.

[0171] In an exemplary embodiment, the determination module is also used to: determine a first control behavior from multiple preset control behaviors, the behavior support satisfying a preset support matching rule, and the behavior confidence satisfying a preset confidence matching rule, wherein the first control behavior corresponds to a behavior occurrence time period; when it is determined that the behavior occurrence time period belongs to a preset time period, determine the first control behavior as the target control behavior.

[0172] In an exemplary embodiment, the determination module is also used to: determine whether the behavior support satisfies the preset support matching rule by: calculating the behavior support by the following formula: Esupport = sum(count(P(A,B))*K*T); wherein Esupport is the behavior support, K ​​is the behavior coefficient, T is the time attenuation coefficient, and count(P(A,B)) represents the number of times the time factor A and the location factor B appear at the same time; when it is determined that the behavior support is greater than the support threshold, it is determined that the behavior support satisfies the preset support matching rule.

[0173] In an exemplary embodiment, the determination module is also used to: determine whether the behavior confidence satisfies the preset confidence matching rules in the following manner: calculate the behavior confidence by the following formula: Econfidence = sum(count(P(A,B))*K*T) / sum(count(P(i))*K*T); wherein Econfidence is the behavior confidence, K is the behavior coefficient, T is the time attenuation coefficient, count(P(A,B)) represents the number of times the time factor A and the position factor B appear at the same time, and count(P(i)) represents the number of times the multiple preset control behaviors appear; when it is determined that the behavior confidence is greater than the confidence threshold, it is determined that the behavior confidence satisfies the preset confidence matching rules.

[0174] In an exemplary embodiment, the second obtaining module is further used to: obtain a first occurrence number of the control behavior of the first object at a specified time and a specified location from the first behavior record; and determine the first occurrence number as the behavior support.

[0175] In an exemplary embodiment, the second obtaining module is also used to: obtain the first number of occurrences of the control behavior of the first object at a specified time and a specified location from the first behavior record, and obtain the second number of occurrences of all control behaviors of the first object from the first behavior record; and determine the ratio of the first number of occurrences to the second number of occurrences as the behavior confidence.

[0176] An embodiment of the present application further provides a storage medium, which includes a stored program, wherein the program executes any of the above methods when it is run.

[0177] Optionally, in this embodiment, the storage medium can be configured to store program codes for executing the following steps: S1, processing the control data when different objects control household appliances according to timing processing rules to obtain an object behavior set; S2, for a first object included in the different objects, determining from the grouping results the group to which the first object belongs, wherein the grouping results are obtained by grouping the different objects using a clustering algorithm and behavior features corresponding to the object behavior set; S3, determining a target control behavior of the first object within a preset time period from a plurality of control behaviors corresponding to the group to which the first object belongs based on behavior measurement parameters, wherein the behavior measurement parameters include at least a behavior support degree of the target control behavior and a behavior confidence degree of the target control behavior; and determining the target control behavior as the recommended data for the first object.

[0178] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0179] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0180] Optionally, in this embodiment, the processor can be configured to perform the following steps through a computer program: S1, processing the control data when different objects control household appliances according to a timing processing rule to obtain an object behavior set; S2, for a first object included in the different objects, determining from the grouping results the group to which the first object belongs, wherein the grouping results are obtained by grouping the different objects using a clustering algorithm and behavior features corresponding to the object behavior set; S3, determining a target control behavior of the first object within a preset time period from a plurality of control behaviors corresponding to the group to which the first object belongs based on behavior measurement parameters, wherein the behavior measurement parameters include at least a behavior support degree of the target control behavior and a behavior confidence degree of the target control behavior; and determining the target control behavior as the recommended data for the first object.

[0181] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0182] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps in any one of the above method embodiments are implemented.

[0183] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in any of the above method embodiments are implemented.

[0184] The embodiments of the present application also provide a computer program, which includes computer instructions, which are stored in a computer-readable storage medium; a processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the steps in any one of the above method embodiments.

[0185] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0186] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0187] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A data processing method based on a behavior association model, characterized in that: include: Performing time sequence processing on the device control data of the first object controlling the home appliance according to the time sequence processing rule to obtain an object behavior set, wherein the object behavior set includes a first behavior record of the first object; performing clustering processing on a plurality of object behavior sets corresponding to different objects, and obtaining a group to which the first object belongs based on a result of the clustering processing; Using a behavior association model to calculate the first behavior record of the first object, to obtain behavior measurement parameters of the first object, wherein the behavior measurement parameters include behavior support and behavior confidence; Determining preset control behaviors set for the group to which the first object belongs in different time periods, and determining a target control behavior of the first object from the preset control behaviors based on the behavior measurement parameters; setting a target control behavior set generated based on the target control behavior as recommended data for the first object, and pushing the recommended data for the first object to a control terminal of the first object; The determining the target control behavior of the first object from the preset control behaviors based on the behavior measurement parameter includes: determining a first control behavior whose behavior support satisfies a preset support matching rule and whose behavior confidence satisfies a preset confidence matching rule from multiple preset control behaviors, wherein the first control behavior corresponds to a behavior occurrence time period; and in the case of determining that the behavior occurrence time period belongs to a preset time period, determining the first control behavior as the target control behavior; The using behavior association model to calculate the first behavior record of the first object to obtain the behavior measurement parameter of the first object includes: obtaining a first occurrence number of the control behavior of the first object at a specified time and a specified location from the first behavior record; determining the first occurrence number as the behavior support; The method of using the behavior association model to calculate the first behavior record of the first object to obtain the behavior measurement parameters of the first object also includes: obtaining the first occurrence number of the control behavior of the first object at a specified time and a specified location from the first behavior record, and obtaining the second occurrence number of all control behaviors of the first object from the first behavior record; and determining the ratio of the first occurrence number to the second occurrence number as the behavior confidence.

2. The data processing method based on the behavior association model according to claim 1 is characterized in that: The device control data at least includes a control instruction received by the home appliance and operation data generated by the home appliance executing the control instruction. The device control data of the home appliance controlled by the first object is subjected to time sequence processing according to a time sequence processing rule to obtain an object behavior set, including: Binding the generation time of the operation data generated by the household appliance to the operation data to obtain a second behavior record, wherein the second behavior record includes the generation time and the operation data after binding, and the operation data includes at least the object number of the different object, the location data of the different object, the behavior data of the different object, and the device type of the household appliance; Sorting the generation time according to the time sequence processing rule to obtain the sorted generation time, wherein the time sequence processing rule indicates processing in chronological order; The plurality of second behavior records are sorted according to the sorted generation time, and the object behavior set is generated based on the sorted plurality of second behavior records.

3. The data processing method based on the behavior association model according to claim 2 is characterized in that: After performing time sequence processing on the device control data of the first object controlling the home appliance according to the time sequence processing rule to obtain the object behavior set, the method further includes: For a first behavior record of the first object included in the object behavior set, at least determine a first generation time and first operation data bound to the first generation time from the first behavior record, wherein the first operation data at least includes: a first object number of the first object, a first position of the first object, a first object behavior of the first object, and a first device type of the home appliance; For a plurality of first behavior records, when it is determined that the time interval between the plurality of first generation moments is less than a preset interval, invalid behavior records are deleted from the plurality of first behavior records according to the earliest first generation moment to obtain remaining behavior records, wherein the invalid behavior records represent behavior records corresponding to other generation moments in the plurality of first behavior records except the earliest first generation moment; The object behavior set is updated based on the remaining behavior records to obtain an updated object behavior set.

4. The data processing method based on the behavior association model according to claim 1 is characterized in that: Performing time sequence processing on the device control data of the home appliance controlled by the first object according to the time sequence processing rule includes: Performing data cleaning on the device control data to obtain cleaned device control data; Wherein, the data cleaning process includes at least one of the following: null value processing, abnormal value processing; Obtaining a verification result of data verification of the cleaned device control data; When it is determined that the verification result passes the verification, the format of the cleaned device control data is standardized to obtain standard data, and the standard data is processed according to the timing processing rule.

5. The data processing method based on the behavior association model according to claim 1 is characterized in that: The different objects include at least the first object and the second object, clustering the multiple object behavior sets corresponding to the different objects, and obtaining the group to which the first object belongs based on the result of the clustering, including: Determine M behavior records that have been assigned to groups from the plurality of object behavior sets, wherein the M behavior records represent current center points required for clustering grouping; Determine a third behavior record belonging to an unassigned group from the plurality of object behavior sets, and calculate a first Euclidean distance between a first eigenvalue of the third behavior record and a second eigenvalue of the M behavior records; Determine a fourth behavior record corresponding to the smallest first Euclidean distance from the M behavior records, and assign a group to the third behavior record according to the group to which the fourth behavior record belongs, wherein the group to which the third behavior record belongs is consistent with the group to which the fourth behavior record belongs; Re-determine L behavior records that have been assigned to the group, wherein the L behavior records represent updated center points required for clustering grouping; Determine a fifth behavior record corresponding to the smallest second Euclidean distance from the L behavior records, and assign a group to a sixth behavior record that is not assigned a group among the L behavior records according to the group to which the fifth behavior record belongs, wherein the second Euclidean distance represents the Euclidean distance between the third eigenvalue of the sixth behavior record and the fourth eigenvalue of the L behavior records; If it is determined that the current center point and the updated center point are consistent, the belonging group is set for the first object according to the belonging group of the third behavior record, and the belonging group is set for the second object according to the belonging group of the sixth behavior record, wherein the third behavior record represents the behavior record belonging to the first object, and the sixth behavior record represents the behavior record belonging to the second object.

6. The data processing method based on the behavior association model according to claim 5 is characterized in that: After determining M behavior records that have been assigned to groups from the plurality of object behavior sets, the method further includes: Selecting a current value of M from a preset value interval for M, and calculating a silhouette coefficient corresponding to the current value; Determine a target value having a maximum silhouette coefficient from a plurality of silhouette coefficients, and determine the target value as the current value of M; Among them, the silhouette coefficient S(i) is calculated by the following formula: , Wherein, a(i) represents the cohesion of the current value, b(i) represents the separation of the current value, and i is a positive integer.

7. The data processing method based on the behavior association model according to claim 5 is characterized in that: Respectively calculating the first Euclidean distance between the first characteristic value of the third behavior record and the second characteristic values ​​of the M behavior records, including: Normalizing the record value recorded by the third behavior to obtain the first feature value, wherein the record value recorded by the third behavior at least includes a second generation time and second operation data bound to the second generation time, wherein the second operation data at least includes: a second position of the second object, a second object behavior of the second object, and a second device type of the household appliance, and the first feature value at least includes a first time feature value corresponding to the second generation time, a first position feature value of the second position, a first behavior feature value of the second object behavior, and a first type feature value of the second device type; Normalizing the record value of each behavior record in the M behavior records to obtain the second feature value, wherein the record value of each behavior record at least includes a third generation time and third operation data bound to the third generation time, wherein the third operation data represents data of a third object corresponding to each behavior record, and at least includes: a third position of the third object, a third object behavior of the third object, and a third device type of the household appliance, and the second feature value at least includes a second time feature value corresponding to the third generation time, a second position feature value of the third position, a second behavior feature value of the third object behavior, and a second type feature value of the third device type; The first Euclidean distance between the first eigenvalue and the second eigenvalue is calculated according to the following formula: , Among them, d(m, n) is the first Euclidean distance, w1 is the feature value at the first moment, x1 is the feature value at the first position, y1 is the feature value at the first behavior, z1 is the feature value of the first type, w2 is the feature value at the second moment, x2 is the feature value at the second position, y2 is the feature value of the second behavior, and z2 is the feature value of the second type.

8. The data processing method based on the behavior association model according to claim 5 is characterized in that: Respectively calculating the first Euclidean distance between the first characteristic value of the third behavior record and the second characteristic values ​​of the M behavior records, including: Acquire a first object tag of the first object and a second object tag of the second object, wherein the first object tag is determined based on an object attribute and a first object attribute value of the first object, and the second object tag is determined based on an object attribute and a second object attribute value of the second object; Normalizing the first object label to obtain a first label feature value, and updating the first feature value by incorporating the first label feature value into the first feature value; and performing normalization processing on the second object label to obtain a second label feature value, and updating the second feature value by incorporating the second label feature value into the second feature value; A first Euclidean distance between the updated first eigenvalue and the updated second eigenvalue is calculated.

9. The data processing method based on the behavior association model according to claim 5 is characterized in that: Respectively calculating the first Euclidean distance between the first characteristic value of the third behavior record and the second characteristic values ​​of the M behavior records, including: Acquire a first object tag of the first object and a second object tag of the second object, wherein the first object tag is determined based on an object attribute and a first object attribute value of the first object, and the second object tag is determined based on an object attribute and a second object attribute value of the second object; Normalizing the first object label to obtain a first label feature value, and updating the first feature value by incorporating the first label feature value into the first feature value with a first weight; Normalizing the second object label to obtain a second label feature value, and updating the second feature value by incorporating the second label feature value into the second feature value with a second weight, wherein the first weight is different from the second weight; A first Euclidean distance between the updated first eigenvalue and the updated second eigenvalue is calculated.

10. The data processing method based on the behavior association model according to claim 1, characterized in that: Determine that the behavior support meets the preset support matching rule in the following manner: The behavior support is calculated by the following formula: Esupport=sum(count(P(A,B))*K*T); Wherein, Esupport is the behavior support, K ​​is the behavior coefficient, T is the time decay coefficient, and count(P(A,B)) represents the number of times the time factor A and the location factor B appear at the same time; In the case where it is determined that the behavior support is greater than the support threshold, it is determined that the behavior support satisfies a preset support matching rule.

11. The data processing method based on the behavior association model according to claim 1, characterized in that: Determine whether the behavior confidence satisfies the preset confidence matching rule by: The behavior confidence is calculated by the following formula: Econfidence= sum(count(P(A,B))*K*T) / sum(count(P(i))*K*T); Wherein, Econfidence is the behavior confidence, K is the behavior coefficient, T is the time attenuation coefficient, count(P(A,B)) represents the number of times the time factor A and the location factor B appear at the same time, and count(P(i)) represents the number of times the multiple preset control behaviors appear; In the case where it is determined that the behavior confidence is greater than the confidence threshold, it is determined that the behavior confidence satisfies a preset confidence matching rule.

12. A data processing device based on a behavior association model, characterized in that: include: a processing module, configured to perform time sequence processing on device control data of a first object controlling a home appliance according to a time sequence processing rule to obtain an object behavior set, wherein the object behavior set includes a first behavior record of the first object; A first obtaining module is used to perform clustering processing on a plurality of object behavior sets corresponding to different objects, and obtain the group to which the first object belongs based on the result of the clustering processing; a second obtaining module, configured to calculate the first behavior record of the first object using a behavior association model to obtain behavior measurement parameters of the first object, wherein the behavior measurement parameters include behavior support and behavior confidence; a determination module, configured to determine preset control behaviors set for the group to which the first object belongs in different time periods, and determine a target control behavior of the first object from the preset control behaviors based on the behavior measurement parameters; A push module, configured to set a target control behavior set generated based on the target control behavior as recommended data of the first object, and push the recommended data of the first object to a control terminal of the first object; The determination module is further configured to determine, from a plurality of preset control behaviors, a first control behavior whose behavior support satisfies a preset support matching rule and whose behavior confidence satisfies a preset confidence matching rule, wherein the first control behavior corresponds to a behavior occurrence time period; and when it is determined that the behavior occurrence time period belongs to a preset time period, the first control behavior is determined as the target control behavior; The second obtaining module is further used to: obtain a first occurrence number of the control behavior of the first object at a specified time and a specified location from the first behavior record; and determine the first occurrence number as the behavior support; The second obtaining module is also used to: obtain the first occurrence number of the control behavior of the first object at the specified time and the specified location from the first behavior record, and obtain the second occurrence number of all control behaviors of the first object from the first behavior record; and determine the ratio of the first occurrence number to the second occurrence number as the behavior confidence.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method described in any one of claims 1 to 11 when executed.

Citation Information

Patent Citations

  • Smart home scene recommendation method and smart home equipment

    CN114371629A

  • Operation information determination method and device, storage medium and electronic device

    CN114417988A

  • Air conditioner parameter recommendation method and air conditioner

    CN114462467A

  • Best-Bet Recommendations

    US20110029509A1