Method, device, electronic device and medium for generating activity number prediction model

Abstract

The embodiments of the present disclosure disclose a method, device, electronic device and computer-readable medium for generating an activity number prediction model. A specific implementation of the method includes: selecting a training sample from a training sample set; inputting the activity time series features into a time series network to obtain the change trend information of the number of people reached by at least one category within the time interval; inputting the change trend information of the number of people reached, the category information of at least one category, and the resource consumption information of the target activity into a shallow neural network to obtain the predicted number of people reached corresponding to the target activity; based on the difference between the actual number of people reached and the predicted number of people reached, adjusting the parameters of the time series network and the shallow neural network to obtain an activity number prediction model. This implementation achieves accurate prediction of the number of people in an activity.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technology, and in particular to a method, device, electronic device, and computer-readable medium for generating an activity attendance prediction model. Background Art

[0002] Joint activities refer to activities held by multiple parties (such as information delivery) to increase the number of users and other indicators of each party. For joint activities, it is necessary to make predictions in advance about the delivery effect, such as the number of people reached. The relevant estimation method is generally based on the historical data of the same period, using year-on-year and month-on-month methods.

[0003] However, when using the above method to make predictions, the following technical problems often occur:

[0004] First, the effectiveness of the combined effect depends on the activity time, and it is difficult to find historical data with exactly the same activity time for prediction, which leads to low prediction accuracy.

[0005] Second, year-on-year and month-on-month forecasts cannot meet flexible needs. Once the activity time changes, the corresponding historical data needs to be re-searched, which consumes a lot of manpower and material resources. Summary of the invention

[0006] The content section of this disclosure is used to introduce concepts in a brief form, and these concepts will be described in detail in the specific implementation section below. The content section of this disclosure is not intended to identify the key features or essential features of the technical solution for which protection is sought, nor is it intended to limit the scope of the technical solution for which protection is sought. Some embodiments of the present disclosure propose a method, device, electronic device, and medium for generating an activity number prediction model to solve one or more of the technical problems mentioned in the above background technology section.

[0007] In a first aspect, some embodiments of the present disclosure provide a method for generating an activity number prediction model, the method comprising: selecting training samples from a training sample set, the training samples comprising activity timing features and labels corresponding to at least one category corresponding to the target activity in a target time interval, the labels being used to characterize the actual number of people reached by the target activity; inputting the activity timing features into a timing network to obtain information on the changing trend of the number of people reached by at least one category within the time interval; inputting the information on the changing trend of the number of people reached, category information of at least one category, and resource consumption information of the target activity into a shallow neural network to obtain the predicted number of people reached corresponding to the target activity; adjusting parameters of the timing network and the shallow neural network based on the difference between the actual number of people reached and the predicted number of people reached to obtain an activity number prediction model.

[0008] In a second aspect, some embodiments of the present disclosure provide a device for generating an activity number prediction model, the device comprising: a selection unit, configured to select training samples from a training sample set, the training samples comprising activity time series features and labels corresponding to at least one category corresponding to the target activity in a target time interval, the labels being used to characterize the actual number of people reached by the target activity; a change trend information generation unit, configured to input the activity time series features into a timing network, and obtain change trend information of the number of people reached of at least one category within the time interval; a reach number generation unit, configured to input change trend information of the number of people reached, category information of at least one category, and resource consumption information of the target activity into a shallow neural network, and obtain the predicted number of people reached corresponding to the target activity; a model generation unit, configured to adjust parameters of the timing network and the shallow neural network based on the difference between the actual number of people reached and the predicted number of people reached, and obtain an activity number prediction model.

[0009] In a third aspect, some embodiments of the present disclosure provide an electronic device comprising: one or more processors; a storage device on which one or more programs are stored, and when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation manner of the above-mentioned first aspect.

[0010] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium having a computer program stored thereon, wherein when the program is executed by a processor, the method described in any implementation manner of the above-mentioned first aspect is implemented.

[0011] The above-mentioned embodiments of the present disclosure have the following beneficial effects: accurate prediction of the number of people in an activity is achieved by generating an activity number prediction model. In this process, the time series features are processed by a time series network to generate information on the trend of changes in the number of people. On this basis, the three factors of the trend of changes in the number of people, category information, and resource consumption information are comprehensively considered, taking into account the impact of each factor on the number of people reached by the activity. By combining the three factors through a shallow neural network, the accuracy of the model's predicted number of people reached can be improved. In addition, since the structure of the shallow neural network is relatively simple, the calculation speed can be accelerated. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that components and elements are not necessarily drawn to scale.

[0013] Figure 1 is a schematic diagram of an application scenario of a method for generating an activity number prediction model according to some embodiments of the present disclosure;

[0014] Figure 2 is a flow chart of some embodiments of the method for generating an activity number prediction model according to the present disclosure;

[0015] Figure 3 An exemplary shallow neural network structure is shown;

[0016] Figure 4 is a flow chart of other embodiments of the method for generating an activity number prediction model according to the present disclosure;

[0017] Figure 5 is a schematic diagram of the structure of some embodiments of the device for generating an activity number prediction model according to the present disclosure;

[0018] Figure 6 It is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0019] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0020] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure can be combined with each other.

[0021] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0022] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0023] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0024] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0025] Figure 1It is a schematic diagram of an application scenario of the method for generating an activity attendance prediction model in some embodiments of the present disclosure.

[0026] exist Figure 1 In the application scenario, first, the computing device 101 can select training samples 103 from the training sample set 102. The training sample 103 includes activity time series features 1031 and labels 1032 corresponding to at least one category corresponding to the target activity in the target time interval, and the label 1032 is used to characterize the actual number of people reached by the target activity. Then, the computing device 101 can input the activity time series features 1031 into the timing network 104 to obtain the change trend information 105 of the number of people reached by at least one category in the time interval. On this basis, the computing device 101 can input the change trend information 105 of the number of people reached, the category information 106 of at least one category, and the resource consumption information 107 of the target activity into the shallow neural network 108 to obtain the predicted number of people reached 109 corresponding to the target activity. Based on the difference between the actual number of people reached and the predicted number of people reached, the parameters of the timing network and the shallow neural network are adjusted to obtain an activity number prediction model.

[0027] It should be noted that the computing device 101 can be hardware or software. When the computing device is hardware, it can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device. When the computing device is embodied as software, it can be installed in the hardware devices listed above. It can be implemented as multiple software or software modules for providing distributed services, or as a single software or software module. No specific limitation is made here.

[0028] It should be understood that Figure 1 The number of computing devices in the embodiment is only illustrative. Any number of computing devices may be provided according to implementation requirements.

[0029] Continue to refer Figure 2 , shows a process 200 of some embodiments of the method for generating an activity number prediction model according to the present disclosure. The method for generating an activity number prediction model comprises the following steps:

[0030] Step 201, selecting training samples from a training sample set, the training samples including activity time series features and labels corresponding to at least one category of a target activity in a target time interval, the labels being used to characterize the actual number of people reached by the target activity.

[0031] In some embodiments, the execution subject of the activity number prediction model generation method can (for example) Figure 1The computing device 101 in the example can select training samples from the training sample set. In practice, the sample can be selected randomly or according to certain conditions. Among them, the training sample includes the activity timing features and labels corresponding to at least one category corresponding to the target activity in the target time interval. Among them, the target activity can be various types of activities, such as online promotional activities. The at least one category corresponding to the target activity can be at least one category participating in the target activity. For example, three categories of brand A clothing, brand B cosmetics, and brand C home appliances jointly participate in a festival joint promotional activity. Then, the festival joint promotional activity corresponds to these three categories. In practice, the target activity corresponds to an activity time interval. The target time interval can be a time interval related to the activity time interval. For example, the activity time interval can be directly determined as the target time interval. The activity timing features corresponding to the target time interval can be the activity features corresponding to multiple time points (such as every day) in the target time interval. As an example, the activity features can be the number of collections, the transaction volume, etc.

[0032] In some optional implementations of some embodiments, the target time interval may be a time interval consisting of an activity time interval and a preset time period before the activity. For example, the target time interval may include the activity time interval and the day before the activity. If the activity time interval is from January 2nd to January 5th, then the target time interval may be from January 1st to January 5th. Optionally, the above-mentioned activity characteristics may include but are not limited to at least one of the following data: number of people reached, number of clicks, number of impressions, and the like. Among them, the number of people reached may be the distributed number of people on the 4A model (Aware, Appeal, Act, Advocate). In these implementations, the data of the preset time period before the activity can be considered as the baseline data of the activity. Therefore, by combining the data of the preset time period before the activity, it is easier to capture the trend of changes in the number of people reached, thereby making the prediction of the trend information of changes in the number of people reached more accurate.

[0033] In some optional implementations of some embodiments, the training sample set is obtained by the following steps: for the original training sample in the original training sample set, splitting according to the corresponding activity time interval to generate multiple training samples corresponding to the original training sample to obtain the training sample set. In this way, the problem of insufficient original training sample data can be solved.

[0034] As an example, suppose the time span of joint activity JA is N days, the start date is B, and the end date is B+N, where B+N represents the date N days after B. Then we can split activity JA into the following sub-activities:

[0035] Activity Name Event start time Event end time JA1 B B+1 JA2 B B+2 JA3 B B+3 …… …… …… J A B B+N

[0036] In this way, the actual activity JA becomes N samples JA1, JA2, ..., JA. The resource consumption, display volume, and click volume of each sample are gradually increasing. The category can be the same as JA. The number of people reached is based on B-1 days, and the number of people reached every day is added up (this requires that the historical data records the number of people reached every day without duplication, and only newly reached users are recorded as the number of people reached on that day). The start date of all the sub-activities above is B. It can be understood that in practice, the start date of a sub-activity can be any day in the activity. In this way, the number of training samples is greatly increased. This in turn improves the prediction accuracy of the model.

[0037] Step 202: Input the activity time series features into the time series network to obtain the change trend information of the number of people reached by at least one category within the time interval.

[0038] In some examples, the above-mentioned execution entity can input the activity time series features into the time series network to obtain the change trend information of the number of people reached in at least one category within the time interval. Among them, as an example, the time series network can be a recurrent neural network, a long short-term memory network LSTM (Long Short-Term Memory), etc. Among them, the LSTM forget gate, input gate and output gate are used to capture contextual features. In practice, the activity time interval of the target activity is a continuous time interval, generally ranging from a few days to dozens of days. The recurrent neural network can capture the trend of changes in time series features and can support any time length.

[0039] Step 203, input the change trend information of the number of people reached, the category information of at least one category, and the resource consumption information of the target activity into the shallow neural network to obtain the predicted number of people reached corresponding to the target activity.

[0040] In some embodiments, the execution entity may input the trend information of the number of people reached, the category information of at least one category, and the resource consumption information of the target activity into a shallow neural network to obtain the predicted number of people reached corresponding to the target activity. The category information may be the brand or name corresponding to the category, etc. The resource consumption information of the target activity may be the time cost value, monetary cost value, etc. of the target activity.

[0041] In some embodiments, a shallow neural network is relative to a deep neural network, and its depth is generally shallower. As an example, Figure 3An exemplary shallow neural network is shown. The bottom layer is the input layer, where N represents the output of the sequential network above, which mainly captures contextual information. BC represents category information, which should reflect the impact of the brand or product list on the number of people reached in the category in the target activity. Cost is the resource consumption information of the target activity. The larger the Cost, the more people are reached. In practice, Cost can also be subdivided into on-site and off-site resource consumption information as needed.

[0042] The middle layer consists of three neurons (h 1 、h 2 、h 3 ). The output layer consists of four neurons (A shown in the figure) 1 , A 2 , A 3 , A 4 ).

[0043] Step 204, based on the difference between the actual number of people reached and the predicted number of people reached, adjust the parameters of the time series network and the shallow neural network to obtain an active number prediction model.

[0044] In some embodiments, the difference can be transferred back from the output layer through back propagation, stochastic gradient descent, etc., and the parameters of the time series network and the shallow neural network can be adjusted to obtain an activity number prediction model.

[0045] In some embodiments, an accurate prediction of the number of people in an activity is achieved by generating an activity number prediction model. In this process, the time series features are processed by a time series network to generate information on the trend of changes in the number of people. On this basis, the three factors of the trend of changes in the number of people, category information, and resource consumption information are comprehensively considered, because each factor has an impact on the number of people reached by the activity. By combining the three factors through a shallow neural network, the accuracy of the number of people reached predicted by the model can be improved. In addition, since the structure of the shallow neural network is relatively simple, the calculation speed can be accelerated.

[0046] Further references Figure 4 , which shows a process 400 of another embodiment of a method for generating a model for predicting the number of people in an activity. The process 400 of the method for generating a model for predicting the number of people in an activity includes the following steps:

[0047] Step 401, selecting training samples from a training sample set, the training samples including activity time series features and labels corresponding to at least one category of a target activity in a target time interval, the labels being used to characterize the actual number of people reached by the target activity.

[0048] Step 402, input the activity time series features into the time series network to obtain the change trend information of the number of people reached by at least one category within the time interval.

[0049] Step 403, input the trend information of the number of people reached, the category information of at least one category, and the resource consumption information of the target activity into a shallow neural network to obtain the predicted number of people reached corresponding to the target activity.

[0050] Step 404: Based on the difference between the actual number of people reached and the predicted number of people reached, adjust the parameters of the time series network and the shallow neural network to obtain an adjusted model.

[0051] In some embodiments, the specific implementation of steps 401-404 and the technical effects thereof can be referred to in Figure 2 The corresponding embodiments will not be described in detail here.

[0052] Step 405: determine whether the preset index of the adjusted model meets the preset conditions.

[0053] In some embodiments, as an example, the effect of the model can be evaluated by wmape (Weighted Mean Absolute Percentage Error). Wherein, wmape can be calculated by the following formula:

[0054]

[0055] Where m is the number of fitting points in the test set, y i is the actual value, p i is the fitted value or predicted value. On this basis, as an example, the preset condition may be: the wmape value is greater than 0.46.

[0056] Step 406: In response to the preset condition being met, the adjusted model is determined as an activity number prediction model.

[0057] In some embodiments, when the value of wmape meets a preset threshold range, the adjusted model is determined as an activity number prediction model.

[0058] from Figure 4 It can be seen that Figure 2 Compared with the description of some corresponding embodiments, Figure 4 The process 400 of the method for generating an activity number prediction model in some corresponding embodiments embodies the step of evaluating the model, so as to ensure the accuracy of the predicted value obtained by the model.

[0059] Further references Figure 5 As an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of an activity number prediction model generation device, and these device embodiments are Figure 2 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0060] like Figure 5 As shown, some embodiments of the active number prediction model generation device 500 include a selection unit 501, which is configured to select training samples from a training sample set, and the training samples include activity time series features and labels corresponding to at least one category corresponding to the target activity in the target time interval, and the labels are used to characterize the actual number of people reached by the target activity. The change trend information generation unit 502 is configured to input the activity time series features into the time series network to obtain the change trend information of the number of people reached by at least one category within the time interval. The number of people reached generation unit 503 is configured to input the change trend information of the number of people reached, the category information of at least one category, and the resource consumption information of the target activity into a shallow neural network to obtain the predicted number of people reached corresponding to the target activity. The model generation unit 504 is configured to adjust the parameters of the time series network and the shallow neural network based on the difference between the actual number of people reached and the predicted number of people reached, so as to obtain an active number prediction model.

[0061] In an optional implementation of some embodiments, the activity timing characteristics corresponding to at least one category in a target time interval are determined by the following steps: determining at least one of the following data for each category in the at least one category at each time point in the target activity interval: number of people reached, number of clicks, and number of impressions, the target time interval including the activity time interval corresponding to the target activity and a preset time period before the activity, to obtain the activity timing characteristics.

[0062] In an optional implementation of some embodiments, the training sample set is obtained by the following steps: for the original training samples in the original training sample set, splitting according to the corresponding activity time interval to generate multiple training samples corresponding to the original training samples, to obtain the training sample set.

[0063] In an optional implementation of some embodiments, the model generation unit 504 is further configured to adjust parameters of the time series network and the shallow neural network based on the difference between the actual number of people reached and the predicted number of people reached to obtain an adjusted model; determine whether the preset indicators of the adjusted model meet preset conditions; in response to meeting the preset conditions, determine the adjusted model as an active number prediction model.

[0064] It is understood that the units described in the device 500 are similar to those described in the reference Figure 2 Therefore, the operations, features and beneficial effects described above for the method are also applicable to the device 500 and the units included therein, and will not be described in detail here.

[0065] Reference below Figure 6 , which shows an electronic device (eg, Figure 1A schematic diagram of the structure of the computing device 600 in FIG. Figure 6 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.

[0066] like Figure 6 As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0067] Typically, the following devices may be connected to the I / O interface 605: an input device 606 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 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 6 The electronic device 600 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. Figure 6 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0068] In particular, according to some embodiments of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, some embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In some such embodiments, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of some embodiments of the present disclosure are executed.

[0069] It should be noted that the computer-readable medium recorded in some embodiments of the present disclosure 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 some embodiments of the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In some embodiments of the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, 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.

[0070] 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.

[0071] The computer-readable medium may be included in the electronic device; or it may exist independently without being installed in the electronic device. The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device: selects training samples from the training sample set, the training samples include activity timing features and labels corresponding to at least one category corresponding to the target activity in the target time interval, and the labels are used to characterize the actual number of people reached by the target activity; inputs the activity timing features into the timing network to obtain the trend information of the number of people reached by at least one category in the time interval; inputs the trend information of the number of people reached, the category information of at least one category, and the resource consumption information of the target activity into the shallow neural network to obtain the predicted number of people reached corresponding to the target activity; based on the difference between the actual number of people reached and the predicted number of people reached, adjusts the parameters of the timing network and the shallow neural network to obtain an activity number prediction model.

[0072] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including 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 a user's computer, partially on a user's computer, as a separate software package, partially on a 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 via 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., via the Internet using an Internet service provider).

[0073] 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.

[0074] The units described in some embodiments of the present disclosure may be implemented by software or by hardware. The described units may also be provided in a processor, for example, may be described as: a processor including a selection unit, a change trend information generation unit, a reach number generation unit, and a model generation unit. Among them, the names of these units do not constitute a limitation on the unit itself in some cases, for example, the selection unit may also be described as a "unit for selecting training samples from a training sample set".

[0075] 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.

[0076] The above descriptions are only some preferred embodiments of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, the above-mentioned features are replaced with the technical features with similar functions disclosed in the embodiments of the present disclosure (but not limited to) and the technical solutions formed.

Claims

1. A method for generating a joint activity number prediction model, comprising: A training sample is selected from a training sample set, wherein the training sample includes an activity time series feature and a label corresponding to at least one category corresponding to a target activity in a target time interval, wherein the label is used to characterize the actual number of people reached by the target activity, and the activity time series feature corresponding to the at least one category in the target time interval is determined by the following steps: determining at least one of the following data of each category in the at least one category at each time point in the target activity interval: the number of people reached, the number of clicks, and the number of impressions; Inputting the activity time series features into a time series network to obtain information on the change trend of the number of people reached by the at least one category within the time interval; Inputting the change trend information of the number of reachable people, the category information of the at least one category, and the resource consumption information of the target activity into a shallow neural network to obtain the predicted number of reachable people corresponding to the target activity; Based on the difference between the actual number of people reached and the predicted number of people reached, the parameters of the time series network and the shallow neural network are adjusted to obtain a joint activity number prediction model.

2. The method according to claim 1, wherein: The target time interval includes the activity time interval corresponding to the target activity and a preset time period before the activity.

3. The method according to claim 1, wherein: The training sample set is obtained by the following steps: The original training samples in the original training sample set are split according to the corresponding activity time intervals to generate multiple training samples corresponding to the original training samples, thereby obtaining the training sample set.

4. The method according to claim 1, wherein: The method of adjusting parameters of the time series network and the shallow neural network based on the difference between the actual number of people reached and the predicted number of people reached to obtain an activity number prediction model includes: Based on the difference between the actual number of people reached and the predicted number of people reached, adjusting parameters of the time series network and the shallow neural network to obtain an adjusted model; Determining whether the preset indicators of the adjusted model meet the preset conditions; In response to a preset condition being met, the adjusted model is determined as the activity number prediction model.

5. The method according to claim 3, wherein: The splitting according to the corresponding activity time interval to generate a plurality of training samples corresponding to the original training sample includes: Determine whether there is abnormal data in the data corresponding to each time point in the activity time interval corresponding to the original training sample; In response to the absence of abnormal data, taking the endpoints of the activity time interval as the starting time points and taking the respective time points as the ending time points, to obtain a plurality of sub-intervals; Obtain at least one item of data and the number of people reached corresponding to each sub-interval to generate the corresponding multiple training samples.

6. A device for generating a joint activity number prediction model, comprising: The selection unit is configured to select a training sample from a training sample set, wherein the training sample includes an activity time series feature and a label corresponding to at least one category corresponding to a target activity in a target time interval, wherein the label is used to characterize the actual number of people reached by the target activity, and the activity time series feature corresponding to the at least one category in the target time interval is determined by the following steps: determining at least one of the following data of each category in the at least one category at each time point in the target activity interval: the number of people reached, the number of clicks, and the number of impressions; A change trend information generating unit is configured to input the activity time series feature into a time series network to obtain change trend information of the number of people reached by the at least one category within the time interval; A reach generation unit is configured to input the reach change trend information, the category information of the at least one category, and the resource consumption information of the target activity into a shallow neural network to obtain a predicted reach corresponding to the target activity; The model generation unit is configured to adjust the parameters of the time series network and the shallow neural network based on the difference between the actual number of people reached and the predicted number of people reached, so as to obtain a joint activity number prediction model.

7. The device according to claim 6, wherein: The activity timing characteristics corresponding to the at least one category in the target time interval are determined by the following steps: Determine at least one of the following data for each of the at least one category at each time point in the target activity interval: number of people reached, number of clicks, and number of impressions. The target time interval includes the activity time interval corresponding to the target activity and the preset time period before the activity, to obtain the activity timing characteristics.

8. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

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

Images