Potential user determination method and device, storage medium and electronic device

Abstract

The present invention discloses a potential user determination method and device, a storage medium and an electronic device, and relates to the field of computer technology. The potential user determination method includes: determining multiple item information according to the item interaction behavior of the target user, converting each item information into an item vector respectively, and constructing a first item sequence composed of multiple item vectors; inputting the first item sequence into a trained intended item prediction model to determine a second item sequence; determining the target item based on the second item sequence, and determining the item acquisition feature of the target user in combination with the information of the target item; inputting the item acquisition feature into a trained probability model to determine the target item acquisition probability of the target user, and determining whether the target user is a potential user to acquire the target item according to the target item acquisition probability. The present disclosure can improve the accuracy of identifying potential users.

Description

Background Art

[0002] In the field of e-commerce, personalized recommendations and precise user behavior guidance can locate the target consumer group, directly display products or services to users, and save users time searching for products or services. However, personalized recommendations often rely on users' active visits, the form of recommendations is single, and the targeted population is relatively broad, making it difficult for e-commerce platforms to actively identify potential users who may obtain products.

[0003] At present, the more common methods for identifying potential users include: the Lookalike model based on seed population expansion, the method based on user tags, and the unsupervised method based on text analysis. However, these methods all have the problem of low accuracy in identifying potential users.

[0004] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0005] The purpose of the present disclosure is to provide a potential user determination method, a potential user determination device, a storage medium and an electronic device, thereby overcoming the problem of low potential user identification accuracy in the related art at least to a certain extent.

[0006] According to one aspect of the present disclosure, a method for determining potential users is provided, including: determining multiple item information based on the item interaction behavior of a target user, converting each item information into an item vector, and constructing a first item sequence composed of multiple item vectors; inputting the first item sequence into a trained intended item prediction model to determine a second item sequence; determining a target item based on the second item sequence, and determining an item acquisition feature of the target user in combination with the information of the target item; inputting the item acquisition feature into a trained probability model to determine the target item acquisition probability of the target user, and determining whether the target user is a potential user for acquiring the target item based on the target item acquisition probability.

[0007] In an exemplary embodiment of the present disclosure, converting each item information into an item vector includes: performing word segmentation processing on the item information to obtain each word segmentation result; inputting each word segmentation result into a trained first conversion model to determine the word segmentation vector corresponding to each word segmentation result; integrating each word segmentation vector to determine the item vector corresponding to the item information.

[0008] In an exemplary embodiment of the present disclosure, converting each item information into an item vector includes: determining the number of views of the item corresponding to each item information to determine the browsing sequence of each item; and inputting the browsing sequence into a trained second conversion model to determine the item vector corresponding to each item information.

[0009] In an exemplary embodiment of the present disclosure, the potential user determination method includes: determining multiple first users; determining a sample input sequence based on multiple item information determined by the item interaction behavior of each of the first users, and determining a sample output sequence based on the intended item information of each of the first users; using the sample input sequence as the input of an intended item prediction model, and using the sample output sequence as the output of the intended item prediction model, so as to train the intended item prediction model.

[0010] In an exemplary embodiment of the present disclosure, determining the target item based on the second item sequence includes: taking the items included in the second item sequence as the target item.

[0011] In an exemplary embodiment of the present disclosure, determining a target item based on the second item sequence includes: determining an item included in the second item sequence as a first candidate item; determining an item similar to the first candidate item as a second candidate item; and determining a set of the first candidate item and the second candidate item as a target item.

[0012] In an exemplary embodiment of the present disclosure, determining an item similar to the first candidate item as a second candidate item includes: calculating the similarity between the first candidate item and other items in an item set; and selecting items in the item set whose similarity is greater than a preset similarity threshold as second candidate items.

[0013] In an exemplary embodiment of the present disclosure, determining the item acquisition characteristics of the target user in combination with the information of the target item includes: determining the first characteristic of the target user based on the item interaction behavior of the target user within a first preset time period; determining the second characteristic of the target user based on the user portrait of the target user; extracting basic information from the information of the target item, and determining the first characteristic of the target item based on the basic information; extracting dynamic information from the information of the target item, and determining the second characteristic of the target item based on the dynamic information; combining the first characteristic and the second characteristic of the target user and the first characteristic and the second characteristic of the target item to determine the item acquisition characteristics of the target user.

[0014] In an exemplary embodiment of the present disclosure, the potential user determination method also includes: determining multiple second users; respectively determining the item acquisition characteristics of each of the second users; determining user information of the existence of intentional behavior within a second preset time period; determining training samples of a probability model based on the item acquisition characteristics of each of the second users and the user information of the existence of intentional behavior within the second preset time period; and training the probability model using the training samples of the probability model.

[0015] In an exemplary embodiment of the present disclosure, respectively determining the item acquisition characteristics of each second user includes: determining the first characteristic of the second user based on the item interaction behavior of the second user within a third preset time period; determining the second characteristic of the second user based on the user portrait of the second user; extracting basic information from the information of the reserved item, and determining the first characteristic of the reserved item based on the basic information; extracting dynamic information from the information of the reserved item, and determining the second characteristic of the reserved item based on the dynamic information; combining the first characteristic and the second characteristic of the second user and the first characteristic and the second characteristic of the reserved item to determine the item acquisition characteristics of the second user.

[0016] In an exemplary embodiment of the present disclosure, determining whether the target user is a potential user for acquiring the target item based on the target item acquisition probability includes: judging whether the target item acquisition probability is greater than a preset probability threshold; if the target item acquisition probability is greater than the preset probability threshold, determining the target user as a potential user for acquiring the target item.

[0017] In an exemplary embodiment of the present disclosure, after determining that the target user is a potential user for obtaining the target item, the user determination method further includes: if the target user has an intention to achieve the target item within a fourth preset time period, pushing first preferential information to the target user; if the target user does not have an intention to achieve the target item within the fourth preset time period, pushing second preferential information to the target user.

[0018] According to one aspect of the present disclosure, a potential user determination device is provided, which includes an item sequence construction module, an intended item prediction module, an acquisition feature determination module and a potential user determination module.

[0019] Specifically, the item sequence construction module is used to determine multiple item information based on the item interaction behavior of the target user, convert each item information into an item vector, and construct a first item sequence composed of multiple item vectors; the intended item prediction module is used to input the first item sequence into a trained intended item prediction model to determine a second item sequence; the acquisition feature determination module is used to determine the target item based on the second item sequence, and determine the item acquisition feature of the target user in combination with the information of the target item; the potential user determination module is used to input the item acquisition feature into a trained probability model to determine the target item acquisition probability of the target user, and determine whether the target user is a potential user to acquire the target item based on the target item acquisition probability.

[0020] In an exemplary embodiment of the present disclosure, the item sequence building module includes a first conversion unit.

[0021] Specifically, the first conversion unit is configured to execute: performing word segmentation processing on the item information to obtain each word segmentation result; inputting each of the word segmentation results into a trained first conversion model to determine the word segmentation vector corresponding to each of the word segmentation results; integrating each of the word segmentation vectors to determine the item vector corresponding to the item information.

[0022] In an exemplary embodiment of the present disclosure, the item sequence building module includes a second conversion unit.

[0023] Specifically, the second conversion unit is configured to perform: determining the number of views of the item corresponding to each item information to determine the browsing sequence of each item; inputting the browsing sequence into a trained second conversion model to determine the item vector corresponding to each item information.

[0024] In an exemplary embodiment of the present disclosure, the potential user determination device further includes a first model training module.

[0025] Specifically, the first model training module is configured to perform: determining multiple first users; determining a sample input sequence based on multiple item information determined by the item interaction behavior of each of the first users, and determining a sample output sequence based on the intended item information of each of the first users; using the sample input sequence as the input of the intended item prediction model, and using the sample output sequence as the output of the intended item prediction model, so as to train the intended item prediction model.

[0026] In an exemplary embodiment of the present disclosure, the feature acquisition determination module includes a first target item determination unit.

[0027] Specifically, the first target item determination unit is used to take items included in the second item sequence as target items.

[0028] In an exemplary embodiment of the present disclosure, the acquisition feature determination module includes a second target item determination unit.

[0029] Specifically, the second target item determination unit is configured to execute: determining items included in the second item sequence as first candidate items; determining items similar to the first candidate items as second candidate items; and determining a set of the first candidate items and the second candidate items as a target item.

[0030] In an exemplary embodiment of the present disclosure, the second target item determination unit includes a similarity calculation subunit and an item determination subunit.

[0031] Specifically, the similarity calculation subunit is used to calculate the similarity between the first candidate item and other items in the item set; the item determination subunit is used to select items in the item set whose similarity is greater than a preset similarity threshold as second candidate items.

[0032] In an exemplary embodiment of the present disclosure, the acquisition feature determination module includes a first acquisition feature determination unit.

[0033] Specifically, the first feature acquisition determination unit is configured to perform: determining the first feature of the target user based on the target user's item interaction behavior within a first preset time period; determining the second feature of the target user based on the user portrait of the target user; extracting basic information from the information of the target item, and determining the first feature of the target item based on the basic information; extracting dynamic information from the information of the target item, and determining the second feature of the target item based on the dynamic information; combining the first and second features of the target user and the first and second features of the target item to determine the item acquisition feature of the target user.

[0034] In an exemplary embodiment of the present disclosure, the potential user determination device further includes a second model training module.

[0035] Specifically, the second model training module is configured to perform: determining multiple second users; respectively determining the item acquisition characteristics of each of the second users; determining user information of the existence of intentional behavior within a second preset time period; determining the training samples of the probability model based on the item acquisition characteristics of each of the second users and the user information of the existence of intentional behavior within the second preset time period; and training the probability model using the training samples of the probability model.

[0036] In an exemplary embodiment of the present disclosure, the second model training module includes a second acquisition feature determination unit.

[0037] Specifically, the second feature acquisition determination unit is configured to perform: determining the first feature of the second user based on the second user's item interaction behavior within a third preset time period; determining the second feature of the second user based on the user portrait of the second user; extracting basic information from the information of the reserved item, and determining the first feature of the reserved item based on the basic information; extracting dynamic information from the information of the reserved item, and determining the second feature of the reserved item based on the dynamic information; combining the first feature and the second feature of the second user and the first feature and the second feature of the reserved item to determine the item acquisition feature of the second user.

[0038] In an exemplary embodiment of the present disclosure, the potential user determination module includes a potential user determination unit.

[0039] Specifically, the potential user determination unit is configured to execute: determining whether the probability of acquiring the target item is greater than a preset probability threshold; if the probability of acquiring the target item is greater than the preset probability threshold, determining the target user as a potential user for acquiring the target item.

[0040] In an exemplary embodiment of the present disclosure, the potential user determination device further includes a preferential information push module.

[0041] Specifically, the preferential information push module is configured to execute: if the target user has an intention to achieve the target item within a fourth preset time period, the first preferential information is pushed to the target user; if the target user does not have an intention to achieve the target item within the fourth preset time period, the second preferential information is pushed to the target user.

[0042] According to one aspect of the present disclosure, a storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the potential user determination method described in any one of the above embodiments is implemented.

[0043] According to one 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 the potential user determination method described in any one of the above embodiments by executing the executable instructions.

[0044] In the technical solutions provided by some embodiments of the present disclosure, the first item sequence determined based on the item interaction behavior of the target user is input into an intended item prediction model to obtain a second item sequence that may characterize the user's interest, the target item is determined based on the second item sequence, and the item acquisition feature of the target user is determined in combination with the information of the target item, and the item acquisition feature is input into a probability model to obtain the target item acquisition probability of the target user, and whether the target user is a potential user to acquire the target item is determined based on the target item acquisition probability. On the one hand, the exemplary embodiment of the present disclosure can accurately determine the potential user by combining the user's item interaction behavior, thereby helping to guide the behavior of the potential user and saving the user time to search for items; on the other hand, by converting the item information into a vector form, it is helpful to analyze and calculate and use the model for processing; on the other hand, the target item is determined by using the second item sequence that the target user may be interested in, and then the item acquisition feature is determined based on the target item, and the item acquisition feature is input into the probability model to determine whether the target user is a potential user, thereby further solving the problem of identifying potential users, using the probability method can provide a more flexible potential user positioning result, and improve the operation effect of the e-commerce platform.

[0045] 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

[0046] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present disclosure, and together with the specification, are used to explain the principles of the present disclosure. Obviously, the drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0047] Figure 1 A flowchart of a potential user determination method according to an exemplary embodiment of the present disclosure is schematically shown;

[0048] Figure 2 A block diagram schematically shows a potential user determination device according to an exemplary embodiment of the present disclosure;

[0049] Figure 3 Schematically shows a block diagram of an item sequence building module according to an exemplary embodiment of the present disclosure;

[0050] Figure 4 Schematically shows a block diagram of an item sequence building module according to another exemplary embodiment of the present disclosure;

[0051] Figure 5Schematically shows a block diagram of a potential user determination device according to another exemplary embodiment of the present disclosure;

[0052] Figure 6 Schematically shows a block diagram of an acquisition feature determination module according to an exemplary embodiment of the present disclosure;

[0053] Figure 7 Schematically shows a block diagram of an acquisition feature determination module according to another exemplary embodiment of the present disclosure;

[0054] Figure 8 Schematically shows a block diagram of a second target item determination unit according to an exemplary embodiment of the present disclosure;

[0055] Fig. 9 Schematically shows a block diagram of an acquisition feature determination module according to another exemplary embodiment of the present disclosure;

[0056] Fig.10 A block diagram schematically shows a potential user determination device according to another exemplary embodiment of the present disclosure;

[0057] Fig.11 Schematically shows a block diagram of a second model training module according to an exemplary embodiment of the present disclosure;

[0058] Fig.12 Schematically shows a block diagram of a potential user determination module according to an exemplary embodiment of the present disclosure;

[0059] Fig.13 A block diagram schematically shows a potential user determination device according to another exemplary embodiment of the present disclosure;

[0060] Fig.14 A schematic diagram showing a storage medium according to an exemplary embodiment of the present disclosure; and

[0061] Fig.15 A block diagram of an electronic device according to an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0062] 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.

[0063] 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.

[0064] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the steps. For example, some steps can be decomposed, while some steps can be combined or partially combined, so the actual execution order may change according to the actual situation. In addition, the terms "first", "second", "third", "fourth", etc. used in this disclosure are only for the purpose of distinction and should not be used as limitations of the content of this disclosure.

[0065] SMS and email are good ways for e-commerce platforms to promote products. These products can correspond to various items, and these items can be promoted to users in a paid or free manner. SMS and email have the characteristics of low cost and high reach rate. However, in recent years, due to the abuse of SMS and email, the promotion business of e-commerce platforms has been hindered, and the effect of guiding user behavior cannot be achieved well.

[0066] At present, the more common methods of identifying potential users include: the Lookalike model based on seed population expansion, specifically, based on the pre-defined seed population, similar users are mined through user portraits and relationship chains; the method of determining users based on user tags, specifically, using user search behavior information to determine user tags, and determining potential users based on user tags; unsupervised methods based on text analysis, specifically, using user search and access results to generate keyword feature vectors, extracting feature vectors of target items based on text, and using nearest neighbor retrieval methods to mine potential users. In addition, there are some methods based on artificially defined indicators and calculation rules.

[0067] However, the Lookalike model has a wide range of targeted groups and lacks user behavioral characteristics. It is difficult for e-commerce platforms to actively reach potential users who may acquire their products, and the conversion rate is low. In addition to lacking user behavioral characteristics, the user tag-based method is also difficult to handle problems involving sorting. The unsupervised method based on text analysis ignores some user operations. Artificially defining indicators and calculation rules requires a lot of human resources and the method has poor universality. Therefore, it can be seen that the problem of low accuracy in identifying potential users is caused by the defects of the existing processing methods.

[0068] In view of this, the exemplary embodiments of the present disclosure provide a new potential user determination method and device, which can improve the accuracy of potential user identification and thus help guide user behavior.

[0069] It should be noted that the potential user determination method of the exemplary embodiment of the present disclosure can be implemented by a server, that is, the server can implement each step of the potential user determination method of the present disclosure. For example, the server can be a server used by an e-commerce platform. In addition, the server can also be a third-party server serving the e-commerce platform, which is not particularly limited in this exemplary embodiment. In this case, the potential user determination device described below can be configured in the server.

[0070] Figure 1 The flowchart of the potential user determination method according to an exemplary embodiment of the present disclosure is schematically shown. Figure 1 , the potential user determination method may include the following steps:

[0071] S12. Determine multiple item information according to the item interaction behavior of the target user, convert each item information into an item vector, and construct a first item sequence composed of multiple item vectors.

[0072] In the exemplary embodiment of the present disclosure, the target user is a user whose potential user is to be determined. For example, the developer can pre-build a user whitelist, and the target user can be any user in the user whitelist. In addition, the user's item interaction behavior may include but is not limited to the user's browsing behavior of items, the behavior of items being added to the acquisition column, the user's behavior of following items, the user's pre-acquisition behavior of items, the user's collection behavior of items, etc. Among them, the pre-acquisition behavior can also be called the intention behavior, which indicates that the user is interested in acquiring items.

[0073] According to some embodiments of the present disclosure, the server may determine multiple items of information based on one item interaction behavior among the above-listed item interaction behaviors.

[0074] For example, multiple item information can be determined based on browsing behavior. Specifically, first, the server can extract browsing behavior data from a flow table of a data warehouse that stores user data. On the one hand, the server can extract browsing behavior data within a period of time (e.g., 30 days); on the other hand, the browsing behavior data can be represented in the form of a triple, for example, <user ID, browsing time, item ID>. Next, the server can determine multiple item information involved based on these browsing behavior data, where the item information can include item name, item description information, item ID, item supplier information, etc.

[0075] It is easy to understand that, similarly, the server can also determine multiple item information based on the behavior data of items being added to the acquisition column, the behavior data of users paying attention to items, the user's pre-acquisition behavior data of items, or the user's collection behavior data of items.

[0076] In addition, the server can also determine multiple item information by combining multiple behavior data in the above-mentioned behavior data. For example, the server can determine multiple item information by combining the user's attention behavior data on the item and the user's collection behavior data on the item. It should be noted that this is not specifically limited in the present exemplary embodiment.

[0077] After determining multiple items of information, the server can convert each item of information into an item vector to facilitate subsequent analysis and processing.

[0078] According to some embodiments of the present disclosure, first, the server may perform word segmentation processing on each item information to obtain each word segmentation result. Specifically, the existing natural language processing method may be used to perform word segmentation on the item information, and the present disclosure does not impose any special restrictions on the specific implementation method of word segmentation; next, the server may input each word segmentation result obtained into a trained first conversion model to determine the word segmentation vector of each word segmentation result, and the dimension of the word segmentation vector may be, for example, 64. Among them, the first conversion model may be a Word2vev model, however, the first conversion model may also be a GloVe model, an ELMo model, a BERT model, etc., and the present disclosure does not impose any special restrictions on the type of the first conversion model, the selection of training samples, and the training process; then, the server may integrate the analysis vectors output by the first conversion model to determine the item vector corresponding to the item information.

[0079] According to other embodiments of the present disclosure, first, the server can determine the number of views of the item corresponding to each item information, and determine the view sequence of each item based on the number of views. For example, the view sequence can be a sequence that arranges the items from high to low based on the number of views, or a sequence that arranges the items from low to high based on the number of views. This is not particularly limited in this exemplary embodiment. Next, the server can input the determined view sequence into a trained second conversion model, and the output of the second conversion model is the item vector corresponding to each item information. The second conversion model can be the same type as the first conversion model, which is a Word2vev model. However, the second conversion model can also be different from the first conversion model, which can be a Word2vev model, a GloVe model, an ELMo model, a BERT model, etc.

[0080] In addition, for an item information, the above two embodiments may be combined to determine the item vector. For example, when the dimension of the item vector is determined to be 64 in the above first embodiment and the dimension of the item vector is determined to be 64 in the above second embodiment, the two may be combined into a 128-dimensional vector to represent the item information.

[0081] After converting each item information into an item vector, the server may generate a first item sequence consisting of multiple item vectors. For example, when the item interaction behavior is a browsing behavior, the first item sequence may be a sequence in which the items are arranged from high to low in terms of browsing times, or a sequence in which the items are arranged from low to high in terms of browsing times.

[0082] It should be understood that the above described solution is a solution for constructing the first item sequence after converting each item information into an item vector. However, in other embodiments of the present disclosure, a sequence composed of item information may be constructed first, and then all the item information therein may be vectorized based on the sequence to determine the first item sequence.

[0083] In addition, the first conversion model and / or the second conversion model may be configured in the server. However, it is easy to understand that the first conversion model and / or the second conversion model may also be configured in another server, which is not particularly limited in this exemplary embodiment.

[0084] S14. Input the first item sequence into a trained intended item prediction model to determine a second item sequence.

[0085] In an exemplary embodiment of the present disclosure, the intended item prediction model can be a sequence-to-sequence based neural network model (seq2seq neural network model), and the seq2seq neural network model is an end-to-end model that can predict and process sequences. The exemplary embodiment of the present disclosure adopts a seq2seq neural network model to implement the process of intended item prediction. However, it should be understood that the seq2seq neural network model is not limited to the form. For example, a seq2seq neural network model with an attention mechanism can also be used. In addition, in other embodiments of the present disclosure, other neural network models can also be used to process the first item sequence, and this is not particularly limited in the present exemplary embodiment. Similar to the above-mentioned first conversion model and second conversion model, the intended item prediction model can be configured in a server that executes the exemplary method of the present disclosure, but it can also be configured in other servers other than the server that executes the exemplary method of the present disclosure.

[0086] The exemplary embodiment of the present disclosure may also include a process of training the intended item prediction model.

[0087] First, the server can determine multiple first users. Specifically, it can determine users who have pre-acquisition behavior (intention achievement behavior) within a period of time as first users based on the flow table of the data warehouse storing user data. However, it can also determine the first user based on the user whitelist described in step S12, for example, randomly select a number of users from the user whitelist as the first user. In this case, the number of first users can be determined based on the number of users in the user whitelist and the computer processing power.

[0088] Next, the server can determine multiple item information based on the item interaction behavior of each first user. It should be understood that for each of the multiple first users, multiple item information can be determined based on its item interaction behavior. The method of determining multiple items corresponding to the first user is the same as the method of determining multiple item information corresponding to the target user in step S12, which will not be repeated here. After determining the multiple item information corresponding to the first user, the sample input sequence can be determined based on the multiple item information. Similarly, the multiple item information can be converted into item vectors respectively, and a sequence composed of multiple item vectors can be constructed as the sample input sequence. It should be noted that for each first user, there can be a corresponding sample input sequence.

[0089] In addition, the server may obtain the desired item information of each first user from the data warehouse, and determine the sample output sequence based on the desired item information. The desired items may be sorted based on the number of desired items to generate the sample output sequence. Similarly, there may be a corresponding sample output sequence for each first user. It should be understood that the sample output sequence is also represented in the form of an item vector.

[0090] It should be noted that when the server determines the sample input sequence and the sample output sequence, it uses the item information extracted within a time period, which can be, for example, one month, one quarter, half a year, one year, etc. It is easy to understand that the larger the time period span and the larger the sample data, the better the model prediction effect.

[0091] Subsequently, the server may use the sample input sequence as the input of the intended item prediction model, and use the sample output sequence as the output of the intended item prediction model, so as to train the intended item prediction model.

[0092] In addition, after the intended item prediction model is trained, the method of the exemplary embodiment of the present disclosure may further include obtaining samples for verifying the intended item prediction model, so as to verify the trained intended item prediction model. The specific process of obtaining samples for verifying the model and the process of verifying the model are similar to the above-mentioned training process and will not be repeated here.

[0093] The above describes a scheme for training the intended item prediction model by the server of the exemplary method of the present disclosure. However, according to other embodiments of the present disclosure, the process of training the intended item prediction model can be performed by other servers other than the server executing the exemplary method of the present disclosure, and the training method is similar and will not be repeated here.

[0094] After determining the trained intended item prediction model, the server can input the first item sequence constructed in step S12 into the trained intended item prediction model to determine the second item sequence. It is easy to understand that the second item sequence represents the item sequence that the target user may be interested in based on the target user's item interaction behavior.

[0095] S16. Determine a target item based on the second item sequence, and determine an item acquisition feature of the target user in combination with information of the target item.

[0096] According to some embodiments of the present disclosure, the server may determine the items contained in the sequence from the second item sequence, and use these items as target items.

[0097] According to some other embodiments of the present disclosure, first, the server may determine the items contained in the sequence from the second item sequence, and use these items as first candidate items; next, the server may determine the items similar to the first candidate items as second candidate items, and determine the set of the first candidate items and the second candidate items as the target items.

[0098] Specifically, the process of determining the target item may include:

[0099] First, the server may calculate the similarity between the first candidate item and other items in the item set.

[0100] The item set described in the exemplary embodiment of the present disclosure may be an item set consisting of all items on an e-commerce platform, or may be an item set consisting of all items under a certain category.

[0101] The present disclosure can calculate similarity by the cosine similarity principle. Specifically, considering that the second item sequence is composed of item vectors of multiple items, for example, one of the item vectors is recorded as Represent each item in the item set as a vector, select an item and record it as Therefore, the similarity sim(X,Y) between two items can be calculated using the following formula:

[0102]

[0103] Thus, the similarity between the first candidate item and other items in the item set can be calculated respectively.

[0104] In addition to cosine similarity, other embodiments of the present disclosure may also use principles such as Mahalanobis distance, Manhattan distance, and Euclidean distance to calculate the similarity between items, which is not particularly limited in this exemplary embodiment.

[0105] Next, the server may determine the items with similarities greater than a preset similarity threshold as second candidate items. For example, the preset similarity threshold may be set to 0.8. It is easy to understand that each first candidate item may correspond to multiple second candidate items.

[0106] Subsequently, the server may determine a set of the first candidate item and the second candidate item as the target item.

[0107] After determining the target item, the server may determine the item acquisition characteristics of the target user based on the information of the target item.

[0108] On the one hand, the server can determine the first feature of the target user based on the target user's item interaction behavior within a first preset time period (e.g., one day, one week, one month, one quarter, etc.). Taking the example that the item interaction behavior is the target user's item browsing behavior, the server can determine the target user's item browsing sequence, and count the number of times the target user browses each item and the total number of times the target user browses, and use the number of times the target user browses each item and the total number of times the target user browses as the first feature of the target user.

[0109] However, it should be understood that the item interaction behavior can be a combination of one or more of the following: the target user's browsing behavior of the item, the behavior of the item being added to the acquisition bar, the target user's behavior of paying attention to the item, the target user's behavior of pre-acquiring the item, and the target user's behavior of collecting the item. Similarly, the corresponding determined data is the first feature. For example, the number of times the target user adds items to the acquisition bar for each item and the total number of times the target user adds items to the acquisition bar can be used as the first feature of the target user.

[0110] On the other hand, the server can determine the second feature of the target user based on the user portrait of the target user. Specifically, the user portrait of each user can be pre-constructed and stored in the user portrait table of the data warehouse. In this case, first, the server can determine the enumeration features from the user portrait of the target user, such as user account, gender, age group, membership level, region, item acquisition ability level, etc.; then, these enumeration features can be OneHot encoded to generate multiple 0 / 1 features; in addition, the server can extract numerical features from the user portrait of the target user, such as user promotion sensitivity score, user activity score, etc.; next, the numerical features can be spliced ​​with the enumeration features after OneHot encoding to obtain the second feature of the target user.

[0111] On the other hand, the server can extract basic information from the information of the target item and determine the first feature of the target item based on the basic information. Specifically, first, the server can extract the enumerated features of the target item from the pre-stored item feature table, such as item category, brand, color, gender of the intended user, etc.; then, these enumerated features can be one-hot encoded to generate multiple 0 / 1 features; in addition, the server can extract the numerical features of the target item from the item feature table, such as size, weight, etc.; next, the server can concatenate the numerical features of the target item with the enumerated features after one-hot encoding to obtain the first feature of the target item.

[0112] On the other hand, the server can extract dynamic information from the information of the target item, and determine the second feature of the target item based on the dynamic information. Specifically, first, the server can extract the browsed data, the data added to the acquisition column, the data paid attention to, the data collected, the data with intentional behavior, etc. of the target item in the first preset time period from the flow table of the data warehouse; next, the server can determine the corresponding values ​​based on these data, for example, the number of browses, the number of visits by unique visitors (Unique Visitor, UV), the number of additions to the acquisition column, the number of attentions, the number of collections, the number of intentional behaviors, etc.; in addition, the server can obtain the shipment volume of the target item in the first preset time period; in addition, similarly, the server can determine the category or brand corresponding to the target item, and count the total number of browses, the number of visits by unique visitors, the number of additions to the acquisition column, the number of attentions, the number of collections, the number of intentional behaviors, the shipment volume, etc. of the same category or brand. And integrate the above information to determine the second feature of the target item.

[0113] Thus, the first feature and the second feature of the target user and the first feature and the second feature of the target item described above may be combined to determine the item acquisition feature of the target user.

[0114] S18. Input the item acquisition feature into a trained probability model to determine the target item acquisition probability of the target user, and determine whether the target user is a potential user for acquiring the target item based on the target item acquisition probability.

[0115] In an exemplary embodiment of the present disclosure, the probability model can be a gradient boosting decision tree model (GradientBoosting Decision Tree, GBDT), GBDT is an iterative decision tree algorithm, which can be composed of multiple decision trees, and the conclusions of all trees are accumulated to finally obtain the predicted result. However, it should be understood that other probability models can also be used to implement the probability model described in the present disclosure, for example, random forest, XGBoost, etc., and the present disclosure does not impose special restrictions on the probability model. Similar to the models described above, the probability model of the exemplary embodiment of the present disclosure can be configured in the server that executes the exemplary method of the present disclosure, however, it can also be configured in other servers other than the server that executes the exemplary method of the present disclosure.

[0116] The exemplary embodiments of the present disclosure may further include a process of training a probability model.

[0117] First, the server may determine multiple second users. Similar to determining multiple first users in step S14, specifically, the server may determine the second users based on the behavior of the flow table of the data warehouse storing user data within a period of time, however, the second users may also be determined based on the user whitelist described in step S12. It is easy to understand that the multiple second users may include the above-mentioned first user.

[0118] Next, the server may determine the item acquisition characteristics of each second user respectively. For example, the server may determine the item acquisition characteristics of each second user every day.

[0119] Specifically, the server determines the first feature of the second user based on the second user's item interaction behavior within the third preset time period; determines the second feature of the second user based on the second user's user portrait, extracts basic information from the information of the scheduled item, and determines the first feature of the scheduled item based on the basic information; extracts dynamic information from the information of the scheduled item, and determines the second feature of the scheduled item based on the dynamic information; combines the first and second features of the second user and the first and second features of the scheduled item to determine the item acquisition feature of the second user. The third preset time period may be one day, one week, one month, one quarter, etc., and may be the same as the first preset time period mentioned above. In addition, the scheduled item may be a randomly determined item. The specific process is similar to determining the item acquisition feature of the target user in step S16 above, and will not be repeated here.

[0120] In addition, the server can determine user information of intentional behavior within a second preset time period, wherein the second preset time period can be a period of time (for example, 15 days) after determining the second user item acquisition feature, however, the present disclosure does not impose any special restrictions on the second preset time period.

[0121] Subsequently, a training sample of the probability model may be determined based on the item acquisition characteristics of each second user and the user information of the intended behavior within the second preset time period, and the probability model may be trained using the training sample.

[0122] Regarding the process of determining the training samples of the probability model, for example, the server can determine the item acquisition features of each second user on January 1, and when the second preset time period is 15 days, determine the user information with intentional behavior from January 1 to January 15. If the item acquisition features and the user information with intentional behavior are stored in the form of a table, the item acquisition feature table and the user information table with intentional behavior can be associated by left join, and the data in the item acquisition feature table that matches the user information table with intentional behavior is recorded as a positive example, and the mismatched data is recorded as a negative example. Next, the negative sampling method can be used to determine the same number of positive examples and negative examples to obtain the training samples of the probability model on January 1. For January 2 to January 30, repeat the above steps of determining the training samples for January 1 to obtain the training samples of the probability model.

[0123] The probability model is trained using the determined training samples. In addition, after the probability model is trained, the method of the exemplary embodiment of the present disclosure may further include obtaining samples for verifying the probability model so as to verify the trained probability model. The specific process of obtaining samples for verifying the probability model and the process of verifying the model are similar to the above-mentioned training process and will not be repeated here.

[0124] The above describes a scheme for training a probability model by a server of the exemplary method of the present disclosure. However, according to other embodiments of the present disclosure, the process of training a probability model can be performed by other servers other than the server executing the exemplary method of the present disclosure, and the training method is similar and will not be repeated here.

[0125] After determining the trained probability model, the server may input the item acquisition feature of the target user determined in step S16 into the trained probability model to determine the target item acquisition probability of the target user.

[0126] In addition, after determining the target item acquisition probability of the target user, the server can determine whether the target user is a potential user for acquiring the target item based on the target item acquisition probability. Specifically, the server can determine whether the target item acquisition probability is greater than a preset probability threshold, for example, the preset probability threshold can be set to 0.5. If the target item acquisition probability is greater than the preset probability threshold, the target user is determined to be a potential user for acquiring the target item.

[0127] For the determined potential users, the exemplary embodiment of the present disclosure also provides a user behavior guidance strategy. Specifically, if the target user has an intention to achieve behavior for the target item within a fourth preset time period (for example, within six months from now), the first preferential information is pushed to the target user; if the target user does not have an intention to achieve behavior for the target item within the fourth preset time period, the second preferential information is pushed to the target user.

[0128] By judging whether the target user has an intention-fulfilling behavior for the target item within the fourth preset time period, the target user can be classified and different user behavior guidance strategies can be implemented. For example, if the target user has an intention-fulfilling behavior for the target item within the fourth preset time period, it can be indicated that the target user is an old user. In this case, the first preferential information can be pushed to the target user based on the strategy for old users, such as doubling points, issuing coupons, etc.; if the target user does not have an intention-fulfilling behavior for the target item within the fourth preset time period, it can be indicated that the target user may be a new user, and the second preferential information can be pushed to the target user based on the strategy for new users, such as issuing coupons. It is easy to understand that the face value of coupons issued to old users and new users may be different.

[0129] 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.

[0130] Furthermore, this example implementation also provides a potential user determination device.

[0131] Figure 2 FIG. 1 schematically shows a block diagram of a potential user determination device according to an exemplary embodiment of the present disclosure. Figure 2 According to an exemplary embodiment of the present disclosure, the potential user determination device 2 may include an item sequence construction module 21, an intended item prediction module 23, an acquisition feature determination module 25 and a potential user determination module 27.

[0132] Specifically, the item sequence construction module 21 can be used to determine multiple item information based on the item interaction behavior of the target user, convert each item information into an item vector, and construct a first item sequence composed of multiple item vectors; the intended item prediction module 23 can be used to input the first item sequence into a trained intended item prediction model to determine the second item sequence; the acquisition feature determination module 25 can be used to determine the target item based on the second item sequence, and determine the item acquisition feature of the target user in combination with the information of the target item; the potential user determination module 27 can be used to input the item acquisition feature into a trained probability model to determine the target user's target item acquisition probability, and determine whether the target user is a potential user to acquire the target item based on the target item acquisition probability.

[0133] Based on the potential user determination device of the exemplary embodiment of the present disclosure, on the one hand, by combining the user's item interaction behavior, it can accurately determine the potential user, saving the user time in searching for items; on the other hand, by converting the item information into a vector form, it is helpful to analyze and process it using a model; on the other hand, the target item is determined by using a second item sequence that the target user may be interested in, and then the item acquisition feature is determined based on the target item, and the item acquisition feature is input into a probability model to determine whether the target user is a potential user. As a result, the problem of identifying potential users can be further better solved, and the use of probability can provide more flexible potential user positioning results, thereby improving the operational effect of the e-commerce platform.

[0134] According to an exemplary embodiment of the present disclosure, referring to Figure 3 , the item sequence construction module 21 may include a first conversion unit 301.

[0135] Specifically, the first conversion unit 301 can be configured to execute: performing word segmentation processing on the item information to obtain each word segmentation result; inputting each of the word segmentation results into a trained first conversion model to determine the word segmentation vector corresponding to each of the word segmentation results; integrating each of the word segmentation vectors to determine the item vector corresponding to the item information.

[0136] According to an exemplary embodiment of the present disclosure, referring to Figure 4 , the item sequence construction module 21 may include a second conversion unit 401.

[0137] Specifically, the second conversion unit 401 can be configured to perform: determining the number of views of the item corresponding to each item information to determine the browsing sequence of each item; inputting the browsing sequence into a trained second conversion model to determine the item vector corresponding to each item information.

[0138] According to an exemplary embodiment of the present disclosure, referring to Figure 5 Compared with the potential user determination device 2, the potential user determination device 5 may include a first model training module 51 in addition to the item sequence construction module 21, the intended item prediction module 23, the feature acquisition determination module 25 and the potential user determination module 27.

[0139] Specifically, the first model training module 51 can be configured to perform: determining multiple first users; determining a sample input sequence based on multiple item information determined by the item interaction behavior of each of the first users, and determining a sample output sequence based on the intended item information of each of the first users; using the sample input sequence as the input of the intended item prediction model, and using the sample output sequence as the output of the intended item prediction model, so as to train the intended item prediction model.

[0140] According to an exemplary embodiment of the present disclosure, referring to Figure 6 , the feature acquisition determination module 25 may include a first target item determination unit 601.

[0141] Specifically, the first target item determining unit 601 may be configured to take items included in the second item sequence as target items.

[0142] According to an exemplary embodiment of the present disclosure, referring to Figure 7 , the feature acquisition determination module 25 may include a second target item determination unit 701.

[0143] Specifically, the second target item determination unit 701 may be configured to execute: determining items included in the second item sequence as first candidate items; determining items similar to the first candidate items as second candidate items; and determining a set of the first candidate items and the second candidate items as a target item.

[0144] According to an exemplary embodiment of the present disclosure, referring to Figure 8 The second target item determination unit 701 may include a similarity calculation subunit 8001 and an item determination subunit 8003 .

[0145] Specifically, the similarity calculation subunit 8001 can be used to calculate the similarity between the first candidate item and other items in the item set; the item determination subunit 8003 can be used to select items in the item set whose similarity is greater than a preset similarity threshold as second candidate items.

[0146] According to an exemplary embodiment of the present disclosure, referring to Fig. 9 , the acquisition feature determination module 25 may include a first acquisition feature determination unit 901.

[0147] Specifically, the first acquisition feature determination unit 901 can be configured to perform: determining the first feature of the target user based on the target user's item interaction behavior within a first preset time period; determining the second feature of the target user based on the user portrait of the target user; extracting basic information from the information of the target item, and determining the first feature of the target item based on the basic information; extracting dynamic information from the information of the target item, and determining the second feature of the target item based on the dynamic information; combining the first feature and the second feature of the target user and the first feature and the second feature of the target item to determine the item acquisition feature of the target user.

[0148] According to an exemplary embodiment of the present disclosure, referring to Fig.10 Compared with the potential user determination device 2, the potential user determination device 10 may include a second model training module 101 in addition to the item sequence construction module 21, the intended item prediction module 23, the feature acquisition determination module 25 and the potential user determination module 27.

[0149] Specifically, the second model training module 101 can be configured to perform: determining multiple second users; respectively determining the item acquisition characteristics of each of the second users; determining user information of intentional behavior within a second preset time period; determining training samples of the probability model based on the item acquisition characteristics of each of the second users and the user information of intentional behavior within the second preset time period; and training the probability model using the training samples of the probability model.

[0150] It should be understood that the second model training module 101 may also be included in the potential user determination device 5 , and correspondingly, the first model training module 51 may also be included in the potential user determination device 10 .

[0151] According to an exemplary embodiment of the present disclosure, referring to Fig.11 The second model training module 101 may include a second feature acquisition determination unit 111.

[0152] Specifically, the second acquisition feature determination unit 111 can be configured to perform: determining the first feature of the second user based on the second user's item interaction behavior within a third preset time period; determining the second feature of the second user based on the user portrait of the second user; extracting basic information from the information of the reserved item, and determining the first feature of the reserved item based on the basic information; extracting dynamic information from the information of the reserved item, and determining the second feature of the reserved item based on the dynamic information; combining the first feature and the second feature of the second user and the first feature and the second feature of the reserved item to determine the item acquisition feature of the second user.

[0153] According to an exemplary embodiment of the present disclosure, referring to Fig.12 , the potential user determination module 27 may include a potential user determination unit 121 .

[0154] Specifically, the potential user determination unit 121 may be configured to execute: determining whether the probability of acquiring the target item is greater than a preset probability threshold; if the probability of acquiring the target item is greater than the preset probability threshold, determining the target user as a potential user for acquiring the target item.

[0155] According to an exemplary embodiment of the present disclosure, referring to Fig.13 Compared with the potential user determination device 2, the potential user determination device 13 may include a preferential information push module 131 in addition to the item sequence construction module 21, the intended item prediction module 23, the feature acquisition determination module 25 and the potential user determination module 27.

[0156] Specifically, the discount information push module 131 is configured to execute: if the target user has an intention to achieve the target item within a fourth preset time period, the first discount information is pushed to the target user; if the target user does not have an intention to achieve the target item within the fourth preset time period, the second discount information is pushed to the target user.

[0157] It should be noted that the preferential information push module 131 may also be included in the potential user determination device 5 or the potential user determination device 10 .

[0158] Since the functional modules of the program running performance analysis device according to the embodiment of the present invention are the same as those in the above-mentioned method invention embodiment, they will not be described in detail here.

[0159] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the above "Exemplary Method" section of the present specification.

[0160] refer to Fig.14As shown, a program product 1400 for implementing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0161] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The 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 (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable 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 disk, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0162] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a 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.

[0163] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0164] Program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0165] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0166] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0167] Refer to the following Fig.15 15 to 16. An electronic device 1500 according to this embodiment of the present invention is described. Fig.15 The electronic device 1500 shown is merely an example and should not limit the functions and scope of use of the embodiments of the present invention.

[0168] like Fig.15 As shown, the electronic device 1500 is in the form of a general computing device. The components of the electronic device 1500 may include, but are not limited to: the at least one processing unit 1510, the at least one storage unit 1520, a bus 1530 connecting different system components (including the storage unit 1520 and the processing unit 1510), and a display unit 1540.

[0169] The storage unit stores program codes, which can be executed by the processing unit 1510, so that the processing unit 1510 performs the steps according to various exemplary embodiments of the present invention described in the above “Exemplary Method” section of this specification. For example, the processing unit 1510 can perform the following steps: Figure 1Step S12 shown in: determining multiple item information based on the item interaction behavior of the target user, converting each item information into an item vector, and constructing a first item sequence composed of multiple item vectors; step S14: inputting the first item sequence into a trained intended item prediction model to determine a second item sequence; step S16: determining the target item based on the second item sequence, and determining the item acquisition feature of the target user in combination with the information of the target item; step S18: inputting the item acquisition feature into a trained probability model to determine the target item acquisition probability of the target user, and determining whether the target user is a potential user to acquire the target item based on the target item acquisition probability.

[0170] The storage unit 1520 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 15201 and / or a cache storage unit 15202 , and may further include a read-only storage unit (ROM) 15203 .

[0171] The storage unit 1520 may also include a program / utility 15204 having a set (at least one) of program modules 15205, such program modules 15205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0172] Bus 1530 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0173] The electronic device 1500 may also communicate with one or more external devices 1600 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1500, and / or communicate with any device that enables the electronic device 1500 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1550. In addition, the electronic device 1500 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 1560. As shown, the network adapter 1560 communicates with other modules of the electronic device 1500 via a bus 1530. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1500, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0174] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0175] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to an exemplary embodiment of the present invention, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0176] 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.

[0177] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary technical means in the art that are not disclosed in the present disclosure. The specification and examples are to be considered as exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

[0178] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for determining potential users, characterized in that: include: Determine multiple item information according to the item interaction behavior of the target user, convert each item information into an item vector, and construct a first item sequence composed of multiple item vectors sorted by browsing times; the interaction behavior includes at least one of the user's browsing behavior of the item, the behavior of the item being added to the acquisition column, the user's attention to the item, the user's pre-acquisition behavior of the item, and the user's collection behavior of the item; the item information includes the item name, item description information, item ID, and item supplier information; Inputting the first item sequence into a trained intended item prediction model to determine a second item sequence; wherein the intended item prediction model is a seq2seq neural network model; the second item sequence represents an item sequence that the target user may be interested in based on the target user's item interaction behavior; wherein determining the target user's interest in interaction behavior includes the user's pre-acquisition behavior of the item; Determine the target item based on the second item sequence, and determine the item acquisition feature of the target user in combination with the information of the target item; wherein determining the target item includes: determining the items included in the second item sequence as first candidate items; calculating the similarity between the first candidate item and other items in the item set; taking the items in the item set whose similarity is greater than a preset similarity threshold as second candidate items; and determining the set of the first candidate item and the second candidate item as the target item; The item acquisition feature is input into a trained probability model to determine the target item acquisition probability of the target user, and whether the target user is a potential user for acquiring the target item is determined based on the target item acquisition probability; wherein the probability model is a GBDT gradient boosting decision tree model.

2. The potential user identification method according to claim 1, characterized in that: Converting each of the item information into an item vector comprises: Performing word segmentation processing on the item information to obtain word segmentation results; Inputting each of the word segmentation results into a trained first conversion model to determine a word segmentation vector corresponding to each of the word segmentation results; The word segmentation vectors are integrated to determine the item vector corresponding to the item information.

3. The potential user identification method according to claim 1, characterized in that: Converting each of the item information into an item vector comprises: Determine the number of views of the item corresponding to each of the item information to determine the viewing sequence of each item; The browsing sequence is input into a trained second conversion model to determine the item vector corresponding to each item information.

4. The potential user identification method according to claim 1, characterized in that: The potential user determination method comprises: determining a plurality of first users; Determine a sample input sequence according to a plurality of item information determined by the item interaction behaviors of each of the first users, and determine a sample output sequence according to the intended item information of each of the first users; The sample input sequence is used as the input of the intended item prediction model, and the sample output sequence is used as the output of the intended item prediction model to train the intended item prediction model.

5. The potential user identification method according to claim 1, characterized in that: Determining the target item based on the second item sequence includes: The items included in the second item sequence are used as target items.

6. The potential user identification method according to claim 1, characterized in that: Determining the item acquisition characteristics of the target user in combination with the information of the target item includes: Determining a first feature of the target user according to the object interaction behavior of the target user within a first preset time period; determining a second feature of the target user based on the user portrait of the target user; Extracting basic information from the information of the target object, and determining a first feature of the target object according to the basic information; Extracting dynamic information from the information of the target object, and determining a second feature of the target object according to the dynamic information; The first feature and the second feature of the target user and the first feature and the second feature of the target item are combined to determine the item acquisition feature of the target user.

7. The potential user identification method according to claim 1, characterized in that: The potential user determination method further includes: determining a plurality of second users; respectively determining item acquisition characteristics of each of the second users; Determining user information of the intended behavior within a second preset time period; Determine a training sample of the probability model based on the item acquisition characteristics of each of the second users and the user information of the user having the intention behavior within the second preset time period; The probability model is trained using the training samples of the probability model.

8. The potential user identification method according to claim 7, characterized in that: Respectively determining the item acquisition characteristics of each of the second users includes: determining a first feature of the second user according to the item interaction behavior of the second user within a third preset time period; determining a second feature of the second user based on the user portrait of the second user; Extracting basic information from the information of the scheduled item, and determining a first feature of the scheduled item based on the basic information; Extracting dynamic information from the information of the predetermined item, and determining a second feature of the predetermined item according to the dynamic information; The first feature and the second feature of the second user and the first feature and the second feature of the predetermined item are combined to determine the item acquisition feature of the second user.

9. The potential user identification method according to claim 1, characterized in that: Determining whether the target user is a potential user for acquiring the target item according to the target item acquisition probability includes: Determining whether the probability of obtaining the target item is greater than a preset probability threshold; If the target item acquisition probability is greater than the preset probability threshold, the target user is determined as a potential user for acquiring the target item.

10. The potential user identification method according to claim 9, characterized in that: After determining that the target user is a potential user for acquiring the target item, the potential user determination method further includes: If the target user has an intention to achieve the target item within a fourth preset time period, push the first discount information to the target user; If the target user does not have any intended behavior for the target item within the fourth preset time period, the second discount information is pushed to the target user.

11. A potential user identification device, characterized in that: include: An item sequence construction module is used to determine multiple item information according to the item interaction behavior of the target user, convert each item information into an item vector, and construct a first item sequence composed of multiple item vectors sorted by browsing times; the interaction behavior includes at least one of the user's browsing behavior of the item, the behavior of the item being added to the acquisition column, the user's attention behavior of the item, the user's pre-acquisition behavior of the item, and the user's collection behavior of the item; the item information includes the item name, item description information, item ID, and item supplier information; The intended item prediction module is used to input the first item sequence into a trained intended item prediction model to determine a second item sequence; wherein the intended item prediction model is a seq2seq neural network model; the second item sequence represents a sequence of items that the target user may be interested in based on the target user's item interaction behavior; wherein the interaction behavior that the target user is interested in includes the user's pre-acquisition behavior of the item; an acquisition feature determination module, configured to determine a target item based on the second item sequence, and determine the item acquisition feature of the target user in combination with the information of the target item; wherein the determination of the target item comprises: determining an item included in the second item sequence as a first candidate item; calculating the similarity between the first candidate item and each other item in the item set; determining an item in the item set whose similarity is greater than a preset similarity threshold as a second candidate item; and determining the set of the first candidate item and the second candidate item as a target item; A potential user determination module is used to input the item acquisition features into a trained probability model to determine the target item acquisition probability of the target user, and determine whether the target user is a potential user for acquiring the target item based on the target item acquisition probability; wherein the probability model is a GBDT gradient boosting decision tree model.

12. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the potential user determination method according to any one of claims 1 to 10 is implemented.

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

Images