Sample generation method and device, storage medium and electronic equipment

Abstract

The invention provides a sample generation method and device, electronic equipment and a computer readable storage medium, and relates to the technical field of data processing. The method comprises the steps of obtaining multiple pieces of historical order information, wherein the historical order information comprises user operation behaviors for historical orders and article information of the historical orders; target features and an order timetable are obtained from the historical order information, the order timetable is a timetable of user operation behaviors, and the target features are features affecting interval time of the user operation behaviors; grouping the historical orders according to the target features, and performing community aggregation on the groups of the historical orders according to the order time table to obtain a plurality of communities; and determining the window duration of each community for obtaining the order sample data, generating a window configuration file, and generating the order sample data in real time based on the window configuration file. According to the embodiment of the invention, the window for obtaining the order sample data can be flexibly set, and the timeliness and accuracy of order sample data collection are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a sample generation method, a sample generation device, an electronic device, and a computer-readable storage medium. Background Art

[0002] In e-commerce platforms, platforms, merchants and users are constantly behaving and changing. In order to enable intelligent algorithms to perceive changes in massive data more quickly and provide a better user experience, it is necessary to grasp the dynamic distribution of samples in real time.

[0003] In related technologies, e-commerce platforms usually obtain a unified window based on statistical analysis as the maximum waiting time for positive feedback. User behaviors within the window will be recorded, and user behaviors outside the window will be ignored. Among them, positive feedback refers to the user behaviors that can be collected within the window. However, due to the influence of factors such as product categories, user purchasing power, and promotional activities, there are differences in the time intervals between user behaviors. The use of a unified window cannot adapt to different sample data situations, and thus cannot ensure the timeliness and accuracy of sample generation.

[0004] In order to solve the above problems, the embodiments of the present disclosure provide a sample generation method, a sample generation device, an electronic device, and a computer-readable storage medium. Summary of the invention

[0005] The purpose of the embodiments of the present disclosure is to provide a sample generation method, a sample generation device, an electronic device and a computer-readable storage medium, thereby solving to a certain extent the problems existing in the related technology that the use of a unified window cannot adapt to different sample data situations and cannot ensure the timeliness and accuracy of sample generation.

[0006] According to a first aspect of the present disclosure, a sample generation method is provided, comprising: obtaining a plurality of historical order information, wherein the historical order information comprises user operation behaviors for the historical orders and item information of the historical orders; obtaining target features and an order schedule from the historical order information, wherein the order schedule is a schedule of the user operation behaviors, and the target features are features of the interval time that affects the user operation behaviors; grouping the historical orders according to the target features, and performing community aggregation on each grouping of the historical orders according to the order schedule to obtain a plurality of communities; determining a window duration for each community to obtain order sample data, and generating a window configuration file, so as to generate order sample data in real time based on the window configuration file.

[0007] In an exemplary embodiment of the present disclosure, grouping the historical order information according to the target feature includes: for each of the target features, performing fine-grained grouping of the historical orders according to the distance of the order timetable.

[0008] In an exemplary embodiment of the present disclosure, the grouping of each of the historical order information is community aggregated according to the order timetable to obtain multiple communities, including: for each group of the historical orders, calculating the time consumption distribution within the group of the historical orders according to the order timetable, the time consumption distribution is the distribution of the processing time of each of the historical orders from creation to completion; calculating the inter-group distance of each group based on the time consumption distribution within each group, and performing community aggregation on each group according to the inter-group distance through a community discovery algorithm based on modularity to obtain multiple communities.

[0009] In an exemplary embodiment of the present disclosure, determining the window duration for each community to obtain order sample data includes: for each community, calculating the time distribution of each historical order in the community based on the order timetable, the time distribution being the distribution of processing time of each historical order from creation to completion; in the time distribution, taking the processing time for which the proportion of historical orders exceeds a preset proportion as the window duration for the community to obtain order sample data.

[0010] In an exemplary embodiment of the present disclosure, the real-time generation of order sample data based on the window configuration file includes: receiving an order in real time and obtaining order information; determining the community to which the real-time order belongs in the window configuration file based on the order information, and generating a window for obtaining the real-time order, so as to use the real-time order as order sample data.

[0011] According to a second aspect of the present disclosure, a sample generation device is provided, including: a historical information acquisition module, used to acquire multiple historical order information, the historical order information includes user operation behavior for historical orders and item information of historical orders; a feature and schedule acquisition module, used to acquire target features and order schedule from the historical order information, the order schedule is the schedule of the user operation behavior, and the target feature is the feature of the interval time that affects the user operation behavior; a grouping aggregation module, used to group the historical orders according to the target features, and perform community aggregation on each group of the historical orders according to the order schedule to obtain multiple communities; a sample generation module, used to determine the window duration for each community to obtain order sample data, and generate a window configuration file to generate order sample data in real time based on the window configuration file.

[0012] In an exemplary embodiment of the present disclosure, the grouping aggregation module is specifically used to: for each of the target features, perform fine-grained grouping of the historical orders according to the distance of the order schedule; for each group of the historical orders, calculate the intra-group time distribution of the historical orders according to the order schedule, and the time distribution is the distribution of the processing time of each historical order from creation to completion; calculate the inter-group distance of each group based on the intra-group time distribution, and perform community aggregation on each group according to the inter-group distance through a community discovery algorithm based on modularity to obtain multiple communities.

[0013] In an exemplary embodiment of the present disclosure, the sample generation module is specifically used to: receive orders in real time and obtain order information; determine the community to which the real-time order belongs in the window configuration file based on the order information, and generate a window for obtaining the real-time order, so as to use the real-time order as order sample data.

[0014] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute any one of the above-mentioned methods by executing the executable instructions.

[0015] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, any one of the methods described above is implemented.

[0016] The exemplary embodiments of the present disclosure may have some or all of the following beneficial effects:

[0017] In the sample generation method provided by the exemplary implementation of the present disclosure, a plurality of historical order information is obtained, and the historical order information includes user operation behaviors for historical orders and item information of historical orders; a target feature and an order schedule are obtained from the historical order information, the order schedule is a schedule of user operation behaviors, and the target feature is a feature of the interval time that affects the user operation behaviors; the historical orders are grouped according to the target feature, and the groups of each historical order are community aggregated according to the order schedule to obtain multiple communities; the window duration for each community to obtain order sample data is determined, and a window configuration file is generated to generate order sample data in real time based on the window configuration file. The present disclosure can group historical orders with similar time consumption distances into one community by grouping and community aggregation of historical orders according to target features and time consumption distribution, and generate a window configuration file by determining the window corresponding to each community, so that the window can be flexibly configured to adapt the window to different real-time orders, so as to solve the problem that for real-time samples with shorter time consumption, the backend state linearly increases due to the excessively long cache window, thus becoming a bottleneck of task stability, and for real-time orders with longer time consumption, the user operation behavior cannot be obtained due to the excessively short cache window, thereby improving the timeliness and accuracy of real-time sample generation.

[0018] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0020] Figure 1 A schematic diagram schematically shows a flow chart of a sample generation method according to an embodiment of the present disclosure;

[0021] Figure 2 A schematic diagram schematically shows a method framework of a sample generation method according to a specific embodiment of the present disclosure;

[0022] Figure 3 A block diagram of a sample generating device according to an embodiment of the present disclosure is schematically shown;

[0023] Figure 4 A schematic structural diagram of an electronic device in an embodiment of the present disclosure is shown schematically. DETAILED DESCRIPTION

[0024] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0025] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0026] Intelligent algorithms require large-scale, accurate, and real-time machine learning training. On e-commerce platforms, the platform, merchants, and users are constantly behaving and changing. Therefore, in order to allow the algorithm to more quickly perceive the subtle changes in massive data and provide a better user experience, it is necessary to grasp the dynamic distribution of samples in real time and ensure that the model is continuously updated.

[0027] Related technologies mainly provide real-time training samples by building a real-time sample platform to support online learning of the model. However, the current real-time sample platform obtains a unified window based on statistical analysis as the maximum waiting time for positive feedback. User behaviors within the window will be recorded, and user behaviors outside the window will be ignored. Among them, positive feedback refers to the user behavior that can be collected within the window.

[0028] However, due to the influence of factors such as product categories, user purchasing power, and promotional activities, the time intervals between user behaviors vary. The above method has the following problems: (1) The real-time window needs to maintain sufficient cache for data management. An excessively long cache window causes the FLINK (distributed processing framework) backend state to grow linearly, becoming a bottleneck for task stability. (2) The accuracy of user behavior and the delay of user behavior are inversely correlated and mutually restrained. The use of a unified and fixed window setting cannot be applied to and take into account all situations, and therefore does not have good business performance.

[0029] In order to solve the problems existing in the above method, this example embodiment proposes a sample generation method, a sample generation device, an electronic device and a computer-readable storage medium. The technical solution of the embodiment of the present disclosure is described in detail below:

[0030] This exemplary embodiment first provides a sample generation method. Figure 1 As shown, the sample generation method specifically includes the following steps:

[0031] Step S110: Acquire multiple historical order information, where the historical order information includes user operation behaviors for historical orders and item information of historical orders;

[0032] Step S120: obtaining target features and order schedule from historical order information, wherein the order schedule is a schedule of user operation behaviors, and the target features are features of interval time that affect user operation behaviors;

[0033] Step S130: grouping the historical orders according to the target features, and performing community aggregation on the groups of the historical orders according to the order timetable to obtain multiple communities;

[0034] Step S140: Determine the window duration for each community to obtain order sample data, generate a window configuration file, and generate order sample data in real time based on the window configuration file.

[0035] The present invention groups historical orders according to target features and time-consuming distribution and aggregates them into communities, so that historical orders with similar time consumption distances can be divided into a community, and a window configuration file is generated by determining the window corresponding to each community, so that the window can be flexibly configured to adapt to different real-time orders. This can solve the problem that for real-time samples with shorter time consumption, the backend state linearly grows due to an excessively long cache window, thus becoming a bottleneck for task stability, and the problem that for real-time orders with longer time consumption, the user operation behavior cannot be obtained due to an excessively short cache window, thereby improving the timeliness and accuracy of real-time sample generation.

[0036] Next, in another embodiment, the above steps are described in more detail.

[0037] In step S110, a plurality of historical order information is obtained, where the historical order information includes user operation behaviors for historical orders and item information of historical orders.

[0038] The disclosed embodiment is used to provide real-time samples to the e-commerce platform, and improves the timeliness and accuracy of sample generation by flexibly setting the window for collecting real-time samples. Specifically, the window for collecting real-time samples is determined based on historical data. The historical data is the historical order information, that is, the historical orders and order information of the e-commerce platform over the past period of time.

[0039] Exemplarily, the above historical order information includes user operation behaviors and item information for historical orders. Among them, the above user operation information may include product exposure, clicks, adding to shopping carts, following, ordering, payment, and product shipment, order completion, etc. The above item information may include product categories, discount information, product prices, etc. It should be noted that the above scenarios are only exemplary, and the protection scope of the embodiments of the present disclosure is not limited to this.

[0040] In step S120, target features and order schedule are obtained from historical order information, where the order schedule is a schedule of user operation behaviors, and the target features are features of interval time that affect user operation behaviors.

[0041] The disclosed embodiment determines the window for sample generation through the above historical orders and their order information, and the window is mainly affected by the time taken by each order processing flow, so it is necessary to extract the above target feature from the features of the historical orders, and the target feature is the feature of the interval time that affects the user's operation behavior, for example, the target feature may include price power coefficient, historical transaction price, preferential price, user purchasing power, commodity category and other features. The above order timetable is the timetable of user operation behavior, and the time interval between user operation behaviors can be obtained from the order timetable, for example, the time taken by the user from adding to cart to placing an order can be obtained, so that the window can be determined more accurately.

[0042] After obtaining the target features and order schedule from the historical order information, in order to facilitate subsequent processing, the embodiment of the present disclosure may also perform normalization on the target features. For example, the target features may be normalized by the minmax method, and each column feature in the sample data may be scaled to a uniform scale, such as normalizing the target features to the interval of (0,1). In order to make the algorithm training process more convergent while also reducing noise and the risk of overfitting, preferably, the embodiment of the present disclosure may perform the above normalization on the window in units of 5 minutes.

[0043] The acquisition process of the above order timetable can be as follows: through feature screening, extraction, purchase and re-brush, and the production of purchase and order time, a complete time-consuming sample table from purchase to order is obtained, that is, the order timetable. Among them, the features are usually obtained from the feature logs downloaded from the online feature system, and the purchase and order logs are issued by the order system.

[0044] In step S130, the historical orders are grouped according to the target features, and community aggregation is performed on the groups of historical orders according to the order timetable to obtain multiple communities.

[0045] In this step, the embodiment of the present disclosure groups historical orders and aggregates them into communities, so as to aggregate historical orders with similar time consumption into the same community, so as to subsequently generate a window configuration file by determining the window of each community to determine the window for sample generation.

[0046] Exemplarily, the process of grouping historical orders according to target features can be implemented as follows: for each target feature, the historical orders are grouped in fine-grained manner according to the distance of the order schedule. Specifically, the target features can be grouped in fine-grained manner according to the order schedule, with each group being an independent group, and the time consumption distribution within the group is calculated as the attribute of each group. For example, based on the time consumption distribution of different commodity categories, commodity categories with a close time consumption distribution distance can be divided into the same group, thereby obtaining multiple fine-grained groupings for commodity categories.

[0047] After grouping the historical orders according to the target features as described above, the disclosed embodiment may further perform divergence calculation, feature combination, and graph calculation on the grouped historical orders to convert the relevant data of the historical orders into a format that conforms to the input of the subsequent community aggregation algorithm.

[0048] The following is a further explanation of the above divergence calculation, feature combination and graph calculation:

[0049] For each of the above groupings, it is necessary to measure the differences in the time consumption distribution between the groups. For example, the embodiment of the present disclosure can use the JS divergence that satisfies the symmetry to obtain the time consumption distribution characteristics as the distance between the groups for measurement. The specific formula of JS divergence is as follows:

[0050]

[0051] Among them, P and Q are the time consumption distribution of different groups, and KL is the relative entropy.

[0052] In order to make the window expression conform to the input of the community discovery algorithm, after the divergence calculation, each group needs to be processed in other ways, such as the above-mentioned feature combination and graph calculation processing. Specifically, taking the community discovery algorithm as the Louvain algorithm (community discovery algorithm based on modularity) as an example, in order to make the window expression usable as the input of the Louvain algorithm, the disclosed embodiment can map the multi-dimensional feature label to a square matrix, and at the same time can also implement a distributed computing method to improve computing efficiency.

[0053] The above graph computing process can be implemented based on Graph X, which is Spark's graph computing API. It can implement graph construction and parallel graph computing from RDD, simplify graph analysis tasks, and facilitate parallelization of graph computing and the introduction of graph algorithms by expressing each of the above groups in graphs. Taking the community discovery algorithm as Louvain algorithm as an example, the above process can be specifically implemented as follows: Nebula Graph provides the Louvain graph computing algorithm based on Graph X, and users can implement the parallelization of the Louvain algorithm by submitting Spark tasks.

[0054] After obtaining the input that conforms to the community discovery algorithm, illustratively, the above-mentioned process of performing community aggregation on the groupings of historical orders according to the order schedule to obtain multiple communities can be implemented as follows: for each group of historical orders, the time consumption distribution within the group of historical orders is calculated according to the order schedule, and the above-mentioned time consumption distribution is the distribution of the processing time of each historical order from creation to completion; based on the time consumption distribution within each group, the inter-group distance of each group is calculated, and based on the inter-group distance, each group is community aggregated through a community discovery algorithm based on modularity to obtain multiple communities.

[0055] Taking the above-mentioned Louvain algorithm as an example, the Louvain algorithm is a community discovery algorithm based on modularity. Its basic idea is that each node in the network tries to traverse the community labels of all neighbors and selects the community label that maximizes the modularity increment. After maximizing the modularity, each community is regarded as a new node, and the above process is repeated until the modularity no longer increases. The disclosed embodiment can perform community aggregation on the window obtained by grouping based on the Louvain algorithm to obtain a community that is easier to characterize and has more cohesive features. In addition, since the original Louvain algorithm has no hyperparameters, it is not convenient to optimize the training process. Preferably, the disclosed embodiment can allow the community algorithm to converge as quickly as possible while meeting business needs by adding a modular Q gain lower bound for community fusion.

[0056] In step S140, the window duration for each community to obtain the order sample data is determined, and a window configuration file is generated to generate the order sample data in real time based on the window configuration file.

[0057] In this step, the embodiment of the present disclosure merges the community groups obtained above within the group and generates a window configuration file, so as to generate order sample data in real time based on the window configuration file. Exemplarily, the process of determining the window duration for each community to obtain order sample data can be implemented as follows: for each community, the time distribution of each historical order in the community is calculated according to the order timetable, and the above time distribution is the distribution of the processing time of each historical order from creation to completion; in the time distribution, the processing time of the historical orders whose proportion exceeds the preset proportion is used as the window duration for the community to obtain order sample data. For example, assuming that the above preset proportion is 80%, if the order processing time of the group of more than 80% of the historical orders in a community is 10 minutes, 10 minutes is set as the window of the community.

[0058] The above-mentioned window configuration file is generated based on the windows of each community determined above. When the window configuration file is obtained, the above-mentioned real-time generation of order sample data based on the window configuration file can be realized through the following process: receiving orders in real time and obtaining order information; determining the community to which the real-time order belongs in the window configuration file based on the order information, and generating a window for obtaining the real-time order, so as to use the real-time order as order sample data. For example, when a real-time order added by a user is received, the community to which the real-time order belongs can be determined based on the commodity category, preferential activities and purchasing power of the real-time order, and the window of the community can be obtained in the window configuration file, and the window can be used as the window of the real-time order to generate order sample data in real time.

[0059] Preferably, a method framework diagram of a specific embodiment of the sample generation method of the embodiment of the present disclosure can be as follows: Figure 2 As shown, this specific embodiment can achieve online dynamic adjustment through the following methods: (1) realize window configuration interface, specifically, the above window configuration file can be placed in the distributed configuration center to facilitate real-time update, loading and visualization of the configuration; (2) use Pod granularity cache window to reduce Flink pod pressure, and use the same window group for all sessions under the same pod. By overloading the open method, apply for off-heap memory through Byte Buffer, and use Unsafe to operate off-heap memory; (3) during feature extraction, extract the window features cached in the memory, and extract the personalized window (i.e., the window obtained based on the window configuration file) by indexing the off-heap window configuration file; (4) by customizing the window trigger based on count and time, according to the window time obtained through the window configuration file, set a personalized window trigger for the user operation behavior of adding a purchase; (5) set AB grouping for the window. For example, output window settings for different categories according to different hyperparameters, which is convenient for real-time comparison of the effects of different windows and obtain more accurate window configuration. Figure 2 The other parts in have been described in detail in the corresponding positions of the above sample generation method, so they will not be repeated again.

[0060] In addition, through the above Figure 2 The sample effects generated by the specific embodiment shown in the figure are analyzed, and it is found that compared with the related technologies, it has certain advantages in terms of coverage, negative example delay and accuracy. The above sample effect analysis process is completed based on the following indicators:

[0061] a) Coverage analysis:

[0062] Evaluate data integrity: correlation rate stability, output stability, and output distribution stability.

[0063] Real-time Observer monitoring and early warning at category granularity.

[0064] b) Negative delay analysis:

[0065] Evaluating data timeliness: Output delay distribution

[0066] Real-time Observer monitoring and early warning at category granularity.

[0067] c) Accuracy analysis:

[0068] Evaluation data accuracy: True Positive (positive samples predicted by the model as positive), False Negative (positive samples predicted by the model as negative), False Positive (negative samples predicted by the model as positive), True Negative (negative samples predicted by the model as negative).

[0069] In addition, in the model effect analysis, the performance of this specific embodiment is also more advantageous. Specifically, through offline model indicator analysis, the offline experimental training platform is searched, and the same hourly test set is used to train the AB model to simulate online. Through experimental observation, it is found that the window model samples obtained by this specific embodiment are smoother and more in line with user usage scenarios, and the CVR (order line / click) and CTR (click / request number) indicators are significantly improved. Through online model indicator analysis, the online AB experimental platform compares the window models of related technologies and this specific embodiment, and finds that after the dynamic adjustment of the latter, CVR, CTR, liquidity, and richness are significantly improved.

[0070] It should be noted that although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0071] Furthermore, in this exemplary embodiment, a sample generation device is also provided, referring to Figure 3 As shown, the sample generation device 300 may include a historical information acquisition module 310, a feature and schedule acquisition module 320, a grouping aggregation module 330 and a sample generation module 340. Among them:

[0072] A historical information acquisition module 310 is used to acquire a plurality of historical order information, wherein the historical order information includes user operation behaviors for historical orders and item information of historical orders;

[0073] The feature and schedule acquisition module 320 is used to acquire target features and order schedules from historical order information, where the order schedule is a schedule of user operation behaviors, and the target features are features of the interval time that affects user operation behaviors;

[0074] A grouping aggregation module 330 is used to group historical orders according to target features, and to perform community aggregation on each group of historical orders according to an order schedule to obtain multiple communities;

[0075] The sample generation module 340 is used to determine the window duration for each community to obtain order sample data, generate a window configuration file, and generate order sample data in real time based on the window configuration file.

[0076] In an exemplary embodiment of the present disclosure, the above-mentioned grouping aggregation module is specifically used to: for each target feature, fine-grained grouping of historical orders according to the distance of the order schedule; and, for each group of historical orders, calculating the time consumption distribution within the group of historical orders according to the order schedule, the time consumption distribution is the distribution of the processing time of each historical order from creation to completion; based on the time consumption distribution within each group, the distance between each group is calculated, and according to the distance between groups, community aggregation is performed on each group through a community discovery algorithm based on modularity to obtain multiple communities.

[0077] In an exemplary embodiment of the present disclosure, the above-mentioned sample generation module is specifically used to: for each community, calculate the time distribution of each historical order in the community according to the order timetable, and the time distribution is the distribution of the processing time of each historical order from creation to completion; in the time distribution, the processing time of the historical orders whose proportion exceeds the preset proportion is used as the window time for the community to obtain order sample data.

[0078] In addition, the sample generation module is also used to: receive orders in real time and obtain order information; determine the community to which the real-time order belongs in the window configuration file based on the order information, and generate a window for obtaining the real-time order to use the real-time order as order sample data.

[0079] The specific details of each module or unit in the above sample generation device have been described in detail in the corresponding table sample generation method, so they will not be repeated here.

[0080] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0081] Figure 4 Schematic diagram of the structure of an electronic device in the embodiment of the present disclosure. Figure 4 , which shows a structural schematic diagram of an electronic device 400 suitable for implementing the embodiments of the present disclosure. Figure 4 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0082] like Figure 4 As shown, the electronic device 400 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 401, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 402 or a program loaded from a storage device 408 to a random access memory (RAM) 403 to implement the voice control method of the embodiment described in the present disclosure. In the RAM 403, various programs and data required for the operation of the electronic device 400 are also stored. The processing device 401, the ROM 402, and the RAM 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0083] Typically, the following devices may be connected to the I / O interface 405: an input device 406 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 407 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 408 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 409. The communication device 409 may allow the electronic device 400 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4The electronic device 400 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0084] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart, thereby implementing the voice control method as described above. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 409, or installed from the storage device 408, or installed from the ROM 402. When the computer program is executed by the processing device 401, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0085] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0086] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0087] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0088] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device:

[0089] Acquire multiple historical order information, where the historical order information includes user operation behaviors for historical orders and item information of historical orders;

[0090] Obtaining target features and order timetable from historical order information, wherein the order timetable is a timetable of user operation behaviors, and the target features are features of interval time that affect user operation behaviors;

[0091] The historical orders are grouped according to the target features, and the groups of historical orders are aggregated into communities according to the order timetable to obtain multiple communities.

[0092] Determine the window duration for each community to obtain order sample data, generate a window configuration file, and generate order sample data in real time based on the window configuration file.

[0093] Optionally, when the above one or more programs are executed by the electronic device, the electronic device may also execute other steps described in the above embodiments.

[0094] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0095] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0096] The units involved in the embodiments described in the present disclosure may be implemented by software or hardware, wherein the name of a unit does not, in some cases, limit the unit itself.

[0097] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0098] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0099] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0100] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0101] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. A sample generation method, characterized in that: include: Acquire multiple historical order information, wherein the historical order information includes user operation behaviors for historical orders and item information of historical orders; Acquire a target feature and an order schedule from the historical order information, wherein the order schedule is a schedule of the user's operation behavior, and the target feature is a feature of the interval time that affects the user's operation behavior; The historical orders are grouped according to the target feature, and community aggregation is performed on each group of the historical orders according to the order timetable to obtain a plurality of communities; Determine the window duration for each community to obtain the order sample data, generate a window configuration file, and generate the order sample data in real time based on the window configuration file.

2. The sample generation method according to claim 1, characterized in that: The grouping of the historical orders according to the target feature includes: For each of the target features, the historical orders are grouped in a fine-grained manner according to the distance of the order timetable.

3. The sample generation method according to claim 1, characterized in that: The grouping of each of the historical orders is subjected to community aggregation according to the order timetable to obtain multiple communities, including: For each group of the historical orders, calculate the time consumption distribution within the group of the historical orders according to the order timetable, wherein the time consumption distribution is the distribution of the processing time from creation to completion of each of the historical orders; The inter-group distances of each group are calculated based on the time consumption distribution within each group, and community aggregation is performed on each group through a community discovery algorithm based on modularity according to the inter-group distances to obtain multiple communities.

4. The sample generation method according to claim 1, characterized in that: Determining the window duration for each community to obtain order sample data includes: For each community, calculate the time consumption distribution of each of the historical orders in the community according to the order timetable, where the time consumption distribution is the distribution of the processing time of each of the historical orders from creation to completion; In the time consumption distribution, the processing time in which the proportion of the historical orders exceeds the preset proportion is used as the window time for the community to obtain order sample data.

5. The sample generation method according to claim 1, characterized in that: Generating order sample data in real time based on the window configuration file includes: Receive orders and obtain order information in real time; The community to which the real-time order belongs is determined in the window configuration file according to the order information, and a window for obtaining the real-time order is generated to use the real-time order as order sample data.

6. A sample generating device, characterized in that: include: A historical information acquisition module, used to acquire a plurality of historical order information, wherein the historical order information includes user operation behaviors for historical orders and item information of historical orders; A feature and schedule acquisition module, used to acquire target features and an order schedule from the historical order information, wherein the order schedule is a schedule of the user's operation behavior, and the target feature is a feature of the interval time that affects the user's operation behavior; A grouping and aggregation module, used to group the historical orders according to the target feature, and perform community aggregation on each group of the historical orders according to the order timetable to obtain multiple communities; The sample generation module is used to determine the window duration for each community to obtain order sample data, generate a window configuration file, and generate order sample data in real time based on the window configuration file.

7. The sample generating device according to claim 6, characterized in that: The packet aggregation module is specifically used for: For each of the target features, the historical orders are finely grouped according to the distance of the order timetable; For each group of the historical orders, calculate the time consumption distribution within the group of the historical orders according to the order timetable, wherein the time consumption distribution is the distribution of the processing time from creation to completion of each of the historical orders; The inter-group distances of each group are calculated based on the time consumption distribution within each group, and community aggregation is performed on each group through a community discovery algorithm based on modularity according to the inter-group distances to obtain multiple communities.

8. The sample generating device according to claim 6, characterized in that: The sample generation module is specifically used for: Receive orders and obtain order information in real time; The community to which the real-time order belongs is determined in the window configuration file according to the order information, and a window for obtaining the real-time order is generated to use the real-time order as order sample data.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

10. An electronic device, characterized in that: include: processor; A memory, configured to store executable instructions of the processor; The processor is configured to perform the method of any one of claims 1 to 5 by executing the executable instructions.

Images