Recommendation information generation method, device, electronic device and computer readable medium

Abstract

The embodiments of the present disclosure disclose a method, device, electronic device and computer-readable medium for generating recommendation information. A specific implementation of the method includes: in response to detecting an item browsing operation of a target user, obtaining an item recommendation model of the target user, wherein the item recommendation model generates recommended item information through a layered portrait of historical items of the target user; determining at least one initial recommended item information of the target user based on the item browsing operation and the item recommendation model; and generating recommendation information based on the attribute information of the at least one initial recommended item information. This implementation reduces data processing time, reduces the consumption of data processing capabilities, and improves the accuracy of information recommended to users.

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 recommendation information. Background Art

[0002] With the development of information technology, obtaining information through the Internet has become an important way for users to obtain information. In order to improve the efficiency of users obtaining information, information servers can push information to users based on user portrait information or user historical browsing records.

[0003] The existing technology has the following deficiencies when pushing information to users:

[0004] Information servers usually push information to users when they browse related information. In practice, there is a huge amount of information on information websites. In order to search for information that users may be interested in from the huge amount of information, information servers need to use complex information search models to quickly process data. This process requires a lot of data processing capabilities and time, and the accuracy of the recommended information obtained in the end is not high. Summary of the invention

[0005] The content of this disclosure is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The content of this disclosure is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.

[0006] Some embodiments of the present disclosure propose a method, device, electronic device, and computer-readable medium for generating recommendation information to solve 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 recommendation information, the method comprising: in response to detecting an item browsing operation of a target user, obtaining an item recommendation model of the target user, the item recommendation model generating recommended item information through a historical item layered portrait of the target user; determining at least one initial recommended item information of the target user based on the item browsing operation and the item recommendation model; and generating recommendation information based on attribute information of the at least one initial recommended item information.

[0008] In a second aspect, some embodiments of the present disclosure provide a recommendation information generating device, which includes: an item recommendation model acquisition unit, configured to acquire an item recommendation model of the target user in response to detecting an item browsing operation of the target user, the item recommendation model generating recommended item information through a historical item hierarchical portrait of the target user; an initial recommended item information determination unit, configured to determine at least one initial recommended item information of the target user based on the item browsing operation and the item recommendation model; and a recommendation information generation unit, configured to generate recommendation information based on attribute information of the at least one initial recommended item information.

[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 various embodiments of the present disclosure have the following beneficial effects: the accuracy of the recommendation information obtained by the recommendation information generation method of some embodiments of the present disclosure is improved. Specifically, the reason for the low accuracy of the recommendation information is that the item recommendation model consumes a lot of data processing capabilities and time. Based on this, the recommendation information generation method of some embodiments of the present disclosure obtains the item recommendation model of the above-mentioned target user when the item browsing operation of the target user is detected. Among them, the historical item hierarchical portrait describes the item attributes in a hierarchical manner, and divides the items by attributes, which greatly reduces the data dimension and data processing volume. Then, at least one initial recommended item information of the above-mentioned target user is determined through the item browsing operation and the item recommendation model; finally, the recommendation information is generated based on the attribute information of at least one initial recommended item information. Due to the use of the historical item hierarchical portrait, the data processing time is reduced, and the consumption of data processing capabilities is reduced. Therefore, through this application, the accuracy of the information recommended to the user can be improved, and the efficiency of online information push can be guaranteed. 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 recommendation information in some embodiments of the present disclosure;

[0014] Figure 2 is a flow chart of some embodiments of the method for generating recommendation information according to the present disclosure;

[0015] Figure 3 is a flowchart of some embodiments of the item recommendation model training method according to the present disclosure;

[0016] Figure 4 is a schematic diagram of the structure of a layered portrait of a historical object according to the present disclosure;

[0017] Figure 5 are flow charts of other embodiments of the method for generating recommendation information according to the present disclosure;

[0018] Figure 6 It is a schematic diagram of aggregation of attribute information to be processed;

[0019] Figure 7 is a schematic diagram of the structure of some embodiments of the recommendation information generating device according to the present disclosure;

[0020] Figure 8 It is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. DETAILED DESCRIPTION

[0021] 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.

[0022] 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.

[0023] 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.

[0024] 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".

[0025] 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.

[0026] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0027] Figure 1 It is a schematic diagram of an application scenario of the method for generating recommendation information according to some embodiments of the present disclosure.

[0028] like Figure 1 As shown, the target user can log in to the information website on the terminal device 101 (for example, a mobile phone, tablet computer, etc.) to browse information. When the information server 102 of the information website detects the target user's item browsing operation, it can be considered that the target user has a need for relevant information. At this time, the information server 102 can obtain the item recommendation model of the target user, and determine at least one initial recommended item information of the target user through the item browsing operation and the item recommendation model; wherein the item recommendation model generates recommended item information through the target user's historical item hierarchical portrait. The historical item hierarchical portrait describes the attributes of the item in a hierarchical manner, and divides the items by attributes, which greatly reduces the data dimension and data processing amount. Afterwards, the recommended information is generated based on the attribute information of the initial recommended item information (such as Figure 1 Finally, the information server 102 can send the recommended information to the terminal device 101. Due to the use of historical item hierarchical portraits, the data processing time is reduced, the consumption of data processing capacity is reduced, the timeliness of recommending information to users is improved, and the efficiency of online information push is guaranteed.

[0029] It should be understood that Figure 1 The number of terminal devices 101 and information servers 102 in the embodiment is only for illustration. Any number of terminal devices 101 and information servers 102 may be provided according to implementation requirements.

[0030] Continue to refer Figure 2 , Figure 2 The process 200 of some embodiments of the method for generating recommendation information according to the present disclosure is shown. The method for generating recommendation information includes the following steps:

[0031] Step 201: in response to detecting an item browsing operation of a target user, obtaining an item recommendation model for the target user.

[0032] In some embodiments, the execution subject of the recommendation information generation method (for example Figure 1The information server 102 shown in the figure can detect the target user's item browsing operation through a wired connection or a wireless connection. After detecting the item browsing operation, the execution subject can obtain the item recommendation model of the target user. It should be noted that the above wireless connection method can include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (UltraWide Band) connection, and other wireless connection methods currently known or to be developed in the future.

[0033] In practice, users can log in to information websites through terminal device 101 to browse information (for example, browse mobile phone information, computer information, etc.). When the execution subject detects the target user's item browsing operation (for example, clicking on a certain item information, opening the item information page, etc.), it can be considered that the target user has a need for relevant information. At this time, the execution subject can obtain a pre-trained item recommendation model for the target user. Among them, the item recommendation model can generate recommended item information through the target user's historical item hierarchical portrait. The historical item hierarchical portrait classifies and screens items through hierarchical indexing of item attributes, which greatly reduces the data processing volume and search time for searching items and improves the timeliness of obtaining recommended information.

[0034] Step 202: determining at least one initial recommended item message for the target user based on the item browsing operation and the item recommendation model.

[0035] In some embodiments, the execution entity may convert the target user's item browsing operation into input information of the item recommendation model, so that the item recommendation model processes the input information and outputs at least one initial recommended item message. For example, the execution entity may convert the item browsing operation into a vector or an information sequence, and then input the vector or the information sequence into the item recommendation model to obtain the initial recommended item message.

[0036] Step 203: Generate recommendation information based on the attribute information of the at least one piece of initial recommended item information.

[0037] In some embodiments, after obtaining the initial recommended item information, in order to further improve the accuracy of the recommended information, the execution subject may modify, filter, and perform other operations on the initial recommended item information through the attribute information of the initial recommended item information, thereby obtaining the recommended information, thereby improving the accuracy and effectiveness of the recommended information. For example, the execution subject may filter out multiple main attribute information from the attribute information of the initial recommended item information, and then search for information within a set time period through the main attribute information as the recommended information; or the execution subject may use information that satisfies multiple main attribute information at the same time as the recommended information.

[0038] The recommendation information obtained by the recommendation information generation method disclosed in some embodiments of the present disclosure has improved timeliness of the recommendation information. Specifically, the reason for the low timeliness of the recommendation information is that the item recommendation model consumes a lot of data processing capabilities and time. Based on this, the recommendation information generation method of some embodiments of the present disclosure obtains the item recommendation model of the target user when the item browsing operation of the target user is detected. Among them, the historical item hierarchical portrait describes the attributes of the item in a hierarchical manner, and divides the items by attributes, which greatly reduces the data dimension and data processing volume. Then, at least one initial recommended item information of the target user is determined through the item browsing operation and the item recommendation model; finally, the recommendation information is generated based on the attribute information of at least one initial recommended item information. Due to the use of the historical item hierarchical portrait, the data processing time is reduced, and the consumption of data processing capabilities is reduced. Therefore, through this application, the accuracy of recommending information to users can be improved, and the efficiency of online information push can be guaranteed.

[0039] Continue to refer Figure 3 , Figure 3 The flowchart 300 of some embodiments of the item recommendation model training method according to the present disclosure is shown. The item recommendation model training method comprises the following steps:

[0040] Step 301: Obtain a set of historical item browsing information of the target user within a set time period.

[0041] Different users may pay attention to different information. In order to obtain the recommended item information of the target user in a targeted manner, the execution subject (e.g. Figure 1 The information server 102 shown in the figure can obtain a historical item browsing information set within a time period set by the target user. The historical item browsing information in the historical item browsing information set can be item information that the user has previously browsed on the information website.

[0042] Step 302, constructing a hierarchical portrait of the historical object through the above historical object browsing information set.

[0043] In order to analyze the historical item browsing information and determine the information that the target user is concerned about, the execution subject can construct a hierarchical portrait of the historical item through the above historical item browsing information set. The hierarchical portrait of the historical item can include multiple layers of attribute information, and each layer of attribute information belongs to a corresponding attribute layer. For example, the hierarchical portrait of the historical item can include attribute information such as XX mobile phone, YY brand, and ZZ store. The attribute layer corresponding to the attribute information of XX mobile phone can be electronic products; the attribute layer corresponding to the YY brand can be brand information; the attribute layer corresponding to the ZZ store can be store information, etc.

[0044] In some optional implementations of some embodiments, the above-mentioned construction of a hierarchical portrait of a historical object through the above-mentioned historical object browsing information set may include: aggregating attributes of the historical object browsing information in the above-mentioned historical object browsing information set through a preset hierarchical index to obtain a hierarchical portrait of the historical object.

[0045] The execution subject can obtain a preset hierarchical index in advance, and then divide the historical item browsing information in the historical item browsing information set by attributes through the hierarchical index, and then cluster the attributes after the attribute division, so as to obtain a hierarchical portrait of the historical item. Figure 4 As shown, the preset hierarchical index may include attribute layer 1, attribute layer 2, attribute layer 3, and attribute layer 4. Attribute layer 1 may represent item type information; attribute layer 2 may represent item category information; attribute layer 3 may represent item brand information; and attribute layer 4 may represent item store name information. Attribute layer 1 may have K attributes; attribute layer 2 may have M attributes; attribute layer 3 may have N attributes; and attribute layer 4 may have T attributes. Each attribute layer may have multiple attribute nodes represented by circles, and each attribute node may represent specific attribute information. For example, when attribute layer 1 represents item type information, the attribute nodes included in attribute layer 1 may be: food, electronic equipment, household appliances, etc. Similarly, attribute layer 2, attribute layer 3, and attribute layer 4 may contain corresponding attribute nodes. The execution entity may divide the historical item browsing information in the historical item browsing information set according to the hierarchical index, and then aggregate the divided attribute nodes to obtain a hierarchical portrait of the historical items of the target user. That is, by Figure 4 The attribute information corresponding to the black attribute nodes in each attribute layer constitutes the hierarchical portrait of historical items, and the white attribute nodes are attribute nodes that are irrelevant to the target user. The hierarchical portrait of historical items can represent the items corresponding to which attribute information the user prefers.

[0046] In some optional implementations of some embodiments, the above-mentioned hierarchical index is obtained by the following steps:

[0047] The first step is to build an initial hierarchical index using preset static attributes.

[0048] The execution subject may first obtain static attributes and construct an initial hierarchical index through the static attributes. The static attributes may include at least one of the following: item type information, item category information, item brand information, item store name information, etc. The initial hierarchical index may include at least one attribute layer, and the attribute layer may include at least one attribute node.

[0049] The second step is to adjust the attribute nodes of the attribute layer in the above initial hierarchical index through knowledge attributes to obtain the above hierarchical index.

[0050] In practice, new items often appear on information websites. For example, a new food, a new electronic device, etc. These new items may contain attribute information that was not previously available, or may take into account multiple attribute information from the past. To this end, the execution subject may set these new attribute information as knowledge attributes. That is, the above-mentioned knowledge attributes are attributes generated based on the historical item browsing information of the above-mentioned target user. The execution subject may adjust the attribute nodes of the attribute layer to obtain the above-mentioned hierarchical index. Among them, the above-mentioned adjustment may include addition, aggregation or splitting, etc. The amount of data processing can be greatly reduced through hierarchical indexing. For example, there are 10 million items. Then each of these 10 million items can be divided into attribute nodes in each attribute layer in the hierarchical index. Accordingly, each attribute layer of the hierarchical index can correspond to 10 million items. When determining the recommended item information through the item recommendation model, the execution subject can determine the information that the user may need according to the attribute layer, and the number of attribute nodes contained in each attribute layer is much smaller than the number of items. In this way, it can be processed layer by layer or multiple layers can be processed simultaneously to determine the items that the target user may need. Then, the attribute nodes of the items that the user may need in each attribute layer are aggregated, and then the items that meet multiple attribute nodes of the user can be screened out. Then, the items can be considered as the recommended item information needed by the target user. In this way, the hierarchical index of items (i.e. attribute information) reduces the amount of data processing, improves the data processing speed, realizes the rapid search of information, facilitates the online push of information, and ensures the timeliness of information push.

[0051] In some optional implementations of some embodiments, the above-mentioned aggregating attributes of the historical item browsing information in the above-mentioned historical item browsing information set by using a preset hierarchical index to obtain a hierarchical portrait of the historical item may include the following steps:

[0052] In the first step, for the historical item browsing information in the historical item browsing information set, the initial item hierarchical portrait of the item corresponding to the historical item browsing information is determined through the hierarchical index.

[0053] The execution subject can determine the attribute layer where the attribute information of the item corresponding to the historical item browsing information is located through the hierarchical index to obtain the initial item hierarchical portrait. At this time, the black attribute nodes in each attribute layer corresponding to the initial item hierarchical portrait are usually scattered.

[0054] In the second step, a hierarchical portrait of a historical item is generated by aggregating the attributes of the initial hierarchical portrait set of the historical item corresponding to the above historical item browsing information set.

[0055] Each initial item hierarchical portrait in the initial item hierarchical portrait set has a corresponding initial item hierarchical portrait. The execution subject can aggregate the attributes of the corresponding black attribute nodes in the initial item hierarchical portrait set to generate a historical item hierarchical portrait. Then the historical item hierarchical portrait can represent the attribute distribution of the information required by the target user. Figure 4 In the example, there are two adjacent black attribute nodes in attribute layer 2. There are four adjacent black attribute nodes in attribute layer 4. Adjacent attribute nodes can form an attribute cluster, which means that the target user is interested in items belonging to the attribute cluster.

[0056] Step 303: training an item recommendation model based on the above-mentioned hierarchical portraits of historical items and the historical item browsing information set.

[0057] The execution entity can perform model training based on the existing multiple intelligent algorithms through the layered portraits of historical items and the historical item browsing information collection to obtain an item recommendation model.

[0058] In some optional implementations of some embodiments, the training of the item recommendation model based on the historical item layered portraits and the historical item browsing information set may include the following steps:

[0059] The first step is to obtain the historical item identifier of the historical item browsing information in the above historical item browsing information set.

[0060] The execution subject can search for the historical item identifier corresponding to the historical item through the historical item browsing information. The historical item identifier can be the serial number, item number, etc. of the corresponding item.

[0061] The second step is to convert the above-mentioned layered portraits of historical objects into historical portrait feature vectors.

[0062] The execution subject can convert the hierarchical portrait of the historical item into a historical portrait feature vector. For example, the historical portrait feature vector can be a determinant, where the rows represent the attribute layer and the columns represent the order of the attribute nodes in the attribute layer. The numbers in the determinant represent the attribute nodes in the attribute layer corresponding to the number of rows of the number and the number of columns of the data.

[0063] The third step is to take the above historical item identification and the historical portrait feature vector corresponding to the above historical item identification as input, and take the historical item information corresponding to the above historical item identification as output, and train to obtain an item recommendation model.

[0064] The execution subject can use a variety of existing intelligent algorithms to take the historical item identification and the historical portrait feature vector corresponding to the above historical item identification as input, and the historical item information corresponding to the historical item identification as output, to train the item recommendation model. Among them, the historical item information can be the name, function and other information of the historical item, and the historical item information can be represented by vectors, images and the like. In addition, the execution subject can also take the historical item identification and the historical portrait feature vector corresponding to the above historical item identification as input, and take the user's click information on the target item browsing page as output to train the item recommendation model. According to actual conditions, there can be other ways to train the item recommendation model, which will not be described one by one here. From the above process of training the item recommendation model, it can be seen that the item recommendation model can construct a corresponding relationship between the historical item identification, the historical portrait feature vector and the historical item information.

[0065] Further references Figure 5 , which shows a process 500 of another embodiment of a method for generating recommendation information. The process 500 of the method for generating recommendation information includes the following steps:

[0066] Step 501: in response to detecting an item browsing operation of a target user, obtaining an item recommendation model of the target user.

[0067] The content of step 501 is the same as that of step 201 and will not be repeated here.

[0068] Step 502: Obtain the target item identifier of the target item browsing information corresponding to the above item browsing operation.

[0069] The execution subject may first obtain the target item identifier of the target item browsing information corresponding to the item browsing operation. The target item identifier may be the serial number, item number, etc. of the corresponding item.

[0070] Step 503: Input the target item identifier and the target historical portrait feature vector corresponding to the target item identifier into the item recommendation model to obtain at least one candidate recommended item information.

[0071] From the above description, it can be seen that the item recommendation model can construct the correspondence between the historical item identification, the historical portrait feature vector and the historical item information. Therefore, the execution subject can input the target item identification and the target historical portrait feature vector corresponding to the above target item identification into the above item recommendation model to obtain at least one candidate recommended item information.

[0072] Step 504: Match the at least one piece of candidate recommended item information based on the target item identifier to obtain at least one piece of initial recommended item information.

[0073] After obtaining at least one candidate recommended item information, the execution subject can match at least one candidate recommended item information again through the target item identifier to obtain the initial recommended item information, thereby further improving the effectiveness of the initial recommended item information. For example, the target item identifier corresponds to a mobile phone, and at least one candidate recommended item information includes a mobile phone, a tablet computer, a basketball, etc. After matching through the target item identifier, the initial recommended item information obtained can be a mobile phone and a tablet computer.

[0074] Step 505: Obtain the to-be-processed attribute information of the at least one piece of initial recommended item information.

[0075] The execution subject may obtain the attribute information of the item corresponding to each piece of initial recommended item information, and set the attribute information as the attribute information to be processed.

[0076] Step 506: Aggregate the attribute information set to be processed corresponding to the at least one piece of initial recommended item information to determine at least one target attribute set.

[0077] The execution subject may perform an aggregation operation on the attribute information set to be processed to determine at least one target attribute set. Figure 6 This is a schematic diagram of the aggregation of attribute information to be processed. Figure 6 As shown, the execution subject can aggregate the attribute information set to be processed to obtain Figure 6 The closed interval A, closed interval B and closed interval C in . Then the closed interval A, closed interval B and closed interval C correspond to the combination of three target attributes.

[0078] Step 507: Set the item information corresponding to the at least one target attribute set as recommendation information.

[0079] The execution subject may set the item information satisfying the closed interval A, the closed interval B and / or the closed interval C as the recommended information. Figure 6 It can be seen that closed interval A, closed interval B and closed interval C have an intersection, which contains two common attribute nodes ( Figure 6 The execution subject can set the item information that satisfies the two common attribute nodes as the recommended information, further improving the accuracy and effectiveness of the recommended information. It should be noted that the recommended information can be the same as the historical item browsing information of the target user, or it can be different from the historical item browsing information.

[0080] Step 508: construct an information push window based on the recommendation information, and send the information push window to the terminal device where the target user is located.

[0081] After determining the recommended information, the execution subject can construct an information push window based on the recommended information and send the information push window to the terminal device where the target user is located, such as Figure 1 shown.

[0082] Further references Figure 7 As an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of a device for generating recommendation information. These device embodiments are similar to Figure 2 Corresponding to the method embodiments shown, the device can be specifically applied to various electronic devices.

[0083] like Figure 7 As shown, the recommendation information generating device 700 of some embodiments includes: an item recommendation model acquiring unit 701, an initial recommended item information determining unit 702, and a recommendation information generating unit 703. Among them, the item recommendation model acquiring unit 701 is configured to, in response to detecting an item browsing operation of a target user, acquire an item recommendation model of the target user, wherein the item recommendation model generates recommended item information through a layered portrait of historical items of the target user; the initial recommended item information determining unit 702 is configured to determine at least one initial recommended item information of the target user based on the item browsing operation and the item recommendation model; and the recommendation information generating unit 703 is configured to generate recommendation information based on attribute information of the at least one initial recommended item information.

[0084] In an optional implementation of some embodiments, the recommendation information generating device 700 may further include an item recommendation model building unit (not shown in the figure), which is configured to build an item recommendation model. The item recommendation model building unit may include: a historical item browsing information collection subunit (not shown in the figure), a historical item hierarchical portrait building subunit (not shown in the figure), and an item recommendation model training subunit (not shown in the figure). The historical item browsing information collection subunit is configured to obtain the historical item browsing information collection within the time period set by the target user; the historical item hierarchical portrait building subunit is configured to build a historical item hierarchical portrait through the historical item browsing information collection; the item recommendation model training subunit is configured to train the item recommendation model based on the historical item hierarchical portrait and the historical item browsing information collection.

[0085] In an optional implementation of some embodiments, the above-mentioned historical object hierarchical portrait construction subunit may include: a historical object hierarchical portrait construction module (not shown in the figure), which is configured to perform attribute aggregation on the historical object browsing information in the above-mentioned historical object browsing information set through a preset hierarchical index to obtain a historical object hierarchical portrait.

[0086] In an optional implementation of some embodiments, the above-mentioned historical item hierarchical portrait construction module may include: an initial item hierarchical portrait determination submodule (not shown in the figure) and a historical item hierarchical portrait generation submodule (not shown in the figure). Among them, the initial item hierarchical portrait determination submodule is configured to determine the initial item hierarchical portrait of the item corresponding to the historical item browsing information in the above-mentioned historical item browsing information set through the above-mentioned hierarchical index; the historical item hierarchical portrait generation submodule is configured to generate the historical item hierarchical portrait by performing attribute aggregation on the initial item hierarchical portrait set corresponding to the above-mentioned historical item browsing information set.

[0087] In an optional implementation of some embodiments, the above-mentioned item recommendation model training subunit may include: a historical item identification acquisition module (not shown in the figure), a historical portrait feature vector acquisition module (not shown in the figure) and an item recommendation model training module (not shown in the figure). Among them, the historical item identification acquisition module is configured to obtain the historical item identification of the historical item browsing information in the above-mentioned historical item browsing information set; the historical portrait feature vector acquisition module is configured to convert the above-mentioned historical item layered portrait into a historical portrait feature vector; the item recommendation model training module is configured to take the above-mentioned historical item identification and the historical portrait feature vector corresponding to the above-mentioned historical item identification as input, and take the historical item information corresponding to the above-mentioned historical item identification as output, and train to obtain the item recommendation model.

[0088] In an optional implementation of some embodiments, the recommendation information generating device 700 may include a hierarchical index building unit (not shown in the figure), which is configured to build a hierarchical index. The hierarchical index building unit may include: an initial hierarchical index building subunit (not shown in the figure) and a hierarchical index building subunit (not shown in the figure). The initial hierarchical index building subunit is configured to build an initial hierarchical index through preset static attributes. The static attributes include at least one of the following: item type information, item category information, item brand information. The initial hierarchical index includes at least one attribute layer, and the attribute layer includes at least one attribute node. The hierarchical index building subunit is configured to adjust the attribute nodes of the attribute layer in the initial hierarchical index through knowledge attributes to obtain the hierarchical index. The knowledge attributes are attributes generated according to the historical item browsing information of the target user. The adjustment includes aggregation or splitting.

[0089] In an optional implementation of some embodiments, the initial recommended item information determination unit 702 may include: a target item identification acquisition subunit (not shown in the figure), a candidate recommended item information acquisition subunit (not shown in the figure), and an initial recommended item information determination subunit (not shown in the figure). Among them, the target item identification acquisition subunit is configured to obtain the target item identification of the target item browsing information corresponding to the above-mentioned item browsing operation; the candidate recommended item information acquisition subunit is configured to input the above-mentioned target item identification and the target historical portrait feature vector corresponding to the above-mentioned target item identification into the above-mentioned item recommendation model to obtain at least one candidate recommended item information; the initial recommended item information determination subunit is configured to match the above-mentioned at least one candidate recommended item information based on the above-mentioned target item identification to obtain at least one initial recommended item information.

[0090] In an optional implementation of some embodiments, the recommendation information generation unit 703 may include: a to-be-processed attribute information acquisition subunit (not shown in the figure), a target attribute set determination subunit (not shown in the figure), and a recommendation information generation subunit (not shown in the figure). The to-be-processed attribute information acquisition subunit is configured to acquire the to-be-processed attribute information of the initial recommended item information in the at least one initial recommended item information; the target attribute set determination subunit is configured to aggregate the to-be-processed attribute information set corresponding to the at least one initial recommended item information and determine at least one target attribute set; the recommendation information generation subunit is configured to set the item information corresponding to the at least one target attribute set as the recommendation information.

[0091] In an optional implementation of some embodiments, the recommendation information generating device 700 may further include: an information push unit (not shown in the figure), configured to construct an information push window based on the recommendation information, and send the information push window to the terminal device where the target user is located.

[0092] It is understood that the units described in the device 700 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 700 and the units included therein, and will not be described in detail here.

[0093] like Figure 8As shown, the electronic device 800 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage device 808 into a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the electronic device 800 are also stored. The processing device 801, the ROM 802, and the RAM 803 are connected to each other via a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0094] Typically, the following devices may be connected to the I / O interface 805: input devices 806 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; output devices 807 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 808 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 809. The communication device 809 may allow the electronic device 800 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 8 The electronic device 800 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 8 Each block shown in the figure may represent one device, or may represent multiple devices as required.

[0095] 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 809, or installed from the storage device 808, or installed from the ROM 802. When the computer program is executed by the processing device 801, the above-mentioned functions defined in the method of some embodiments of the present disclosure are executed.

[0096] It should be noted that the computer-readable medium 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 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, the 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, the computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to 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.

[0097] 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.

[0098] 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: in response to detecting an item browsing operation of a target user, obtains an item recommendation model of the target user, wherein the item recommendation model generates recommended item information through a layered image of the target user's historical items; determines at least one initial recommended item information of the target user based on the item browsing operation and the item recommendation model; and generates recommendation information based on the attribute information of the at least one initial recommended item information.

[0099] 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 the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer 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).

[0100] 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.

[0101] The units described in some embodiments of the present disclosure may be implemented by software or hardware. The units described may also be provided in a processor, for example, may be described as: a processor including an item recommendation model acquisition unit, an initial recommended item information determination unit, and a recommendation information generation unit. The names of these units do not, in some cases, constitute limitations on the units themselves, for example, the recommendation information generation unit may also be described as a "unit for users to generate recommendation information".

[0102] 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.

[0103] 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 recommendation information, comprising: In response to detecting an item browsing operation of a target user, obtaining an item recommendation model of the target user, wherein the item recommendation model generates recommended item information through a layered portrait of historical items of the target user; Determine at least one piece of initial recommended item information for the target user based on the item browsing operation and the item recommendation model; generating recommendation information based on the attribute information of the at least one piece of initial recommended item information; The item recommendation model is trained by the following steps: Acquire a set of historical item browsing information within a time period set by the target user; Performing attribute aggregation on the historical item browsing information in the historical item browsing information set through a preset hierarchical index to obtain a hierarchical portrait of the historical item; Training an item recommendation model based on the hierarchical portraits of historical items and the historical item browsing information set; The method of aggregating the attributes of the historical item browsing information in the historical item browsing information set by using a preset hierarchical index to obtain a hierarchical portrait of the historical item includes: For the historical item browsing information in the historical item browsing information set, determining an initial item hierarchical portrait of the item corresponding to the historical item browsing information through the hierarchical index; By performing attribute aggregation on the initial item hierarchical portrait set corresponding to the historical item browsing information set, a historical item hierarchical portrait is generated, wherein the hierarchical index includes at least one attribute layer, the attribute layer includes at least one attribute node, and each attribute node represents specific attribute information.

2. The method according to claim 1, wherein: The training of the item recommendation model based on the historical item layered portrait and the historical item browsing information set includes: Obtaining the historical item identifier of the historical item browsing information in the historical item browsing information set; Converting the layered portrait of the historical object into a historical portrait feature vector; The historical item identification and the historical portrait feature vector corresponding to the historical item identification are taken as input, and the historical item information corresponding to the historical item identification is taken as output, so as to train and obtain an item recommendation model.

3. The method according to claim 1, wherein: The hierarchical index is obtained by the following steps: Constructing an initial hierarchical index through preset static attributes, wherein the static attributes include at least one of the following: item type information, item category information, and item brand information, and the initial hierarchical index includes at least one attribute layer; The hierarchical index is obtained by adjusting the attribute nodes of the attribute layer in the initial hierarchical index through knowledge attributes, wherein the knowledge attributes are attributes generated according to the historical item browsing information of the target user, and the adjustment includes aggregation or splitting.

4. The method according to claim 1, wherein: The determining, based on the item browsing operation and the item recommendation model, at least one piece of initial recommended item information for the target user includes: Obtaining a target item identifier of target item browsing information corresponding to the item browsing operation; Inputting the target item identifier and the target historical portrait feature vector corresponding to the target item identifier into the item recommendation model to obtain at least one candidate recommended item information; The at least one piece of candidate recommended item information is matched based on the target item identifier to obtain at least one piece of initial recommended item information.

5. The method according to claim 1, wherein: The generating the recommendation information based on the attribute information of the at least one piece of initial recommended item information includes: Acquire the to-be-processed attribute information of the initial recommended item information in the at least one piece of the initial recommended item information, wherein the attribute information of the item corresponding to each piece of the initial recommended item information is acquired, and the obtained attribute information is set as the to-be-processed attribute information; Aggregating a set of to-be-processed attribute information corresponding to the at least one piece of initial recommended item information to determine at least one target attribute set; The item information corresponding to the at least one target attribute set is set as recommendation information.

6. The method according to claim 1, wherein: The method further comprises: An information push window is constructed based on the recommendation information, and the information push window is sent to the terminal device where the target user is located.

7. A device for generating recommendation information, comprising: An item recommendation model acquisition unit is configured to acquire an item recommendation model of the target user in response to detecting an item browsing operation of the target user, wherein the item recommendation model generates recommended item information through a layered portrait of historical items of the target user; an initial recommended item information determining unit, configured to determine at least one piece of initial recommended item information for the target user based on the item browsing operation and an item recommendation model; a recommendation information generating unit, configured to generate recommendation information based on the attribute information of the at least one piece of initial recommended item information; The item recommendation model is trained by the following steps: Acquire a set of historical item browsing information within a time period set by the target user; Performing attribute aggregation on the historical item browsing information in the historical item browsing information set through a preset hierarchical index to obtain a hierarchical portrait of the historical item; Training an item recommendation model based on the hierarchical portraits of historical items and the historical item browsing information set; The method of aggregating the attributes of the historical item browsing information in the historical item browsing information set by using a preset hierarchical index to obtain a hierarchical portrait of the historical item includes: For the historical item browsing information in the historical item browsing information set, determining an initial item hierarchical portrait of the item corresponding to the historical item browsing information through the hierarchical index; By performing attribute aggregation on the initial item hierarchical portrait set corresponding to the historical item browsing information set, a historical item hierarchical portrait is generated, wherein the hierarchical index includes at least one attribute layer, the attribute layer includes at least one attribute node, and each attribute node represents specific attribute information.

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 6.

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 6 is implemented.

Images