App-based cross-domain recommendation method and device

Abstract

The present disclosure relates to an APP-based cross-domain recommendation method and device. The method comprises: obtaining all positive combinations of each user, each positive combination comprising a positive e-commerce commodity and a positive community content; obtaining the number of positive combinations shared by each two users as a weight score; obtaining all positive combinations corresponding to all positive e-commerce commodities and all positive community contents; obtaining a similarity score corresponding to each positive combination according to the weight score of each two users; obtaining a target item, the target item comprising a target e-commerce commodity or a target community content; obtaining a target similarity score corresponding to the target item, the target similarity score being greater than a preset threshold; outputting a to-be-recommended item in the positive combination corresponding to the target similarity score; wherein, when the target item is a target e-commerce commodity, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of e-commerce, and particularly relates to a cross-domain recommendation method and device based on APP. BACKGROUND

[0002] The micro-store APP has an e-commerce part and a community part, the e-commerce part includes goods (for example, goods uploaded by a seller in the micro-store APP), and the community part includes content (for example, content (text / video, etc.) created by a user in the APP). Among them, the e-commerce part is similar to Taobao or Jingdong; and the community part is similar to Weibo or Xiaohongshu.

[0003] The existing recommendation method is as follows:

[0004] (1) Goods tagging: manually tagging e-commerce goods with community tags.

[0005] (2) User tagging: according to the positive behavior (such as purchase / click / add to cart, etc.) of a user on e-commerce goods, combining the tags of the goods, and tagging the user through rules. For example, if a user often buys food-related goods, the user is tagged as “food”.

[0006] (3) Content recommendation: according to the tags tagged by the user, find corresponding content for recommendation.

[0007] Disadvantages of the existing scheme:

[0008] (1) High cost: manual goods tagging consumes huge labor costs.

[0009] (2) Low coverage: there are hundreds of millions of goods in the micro-store, and manual goods tagging cannot accurately cover all the goods.

[0010] (3) The information reflected by manual goods tagging is limited, because the tags have limitations and cannot fully reflect the user's intention, and the goods tagging does not consider user behavior.

[0011] (4) The information transmission link is long (goods->tags->content), and there is information loss in each link, which affects the final recommendation result. SUMMARY

[0012] To overcome the problems in the related art, the embodiments of the present disclosure provide a cross-domain recommendation method and device based on APP. The technical solution is as follows:

[0013] According to a first aspect of the embodiments of the present disclosure, a cross-domain recommendation method based on APP is provided, comprising:

[0014] Obtaining all positive combinations of each user, each positive combination including a positive e-commerce good and a positive community content;

[0015] obtaining the number of same positive combinations of each two users as a weight score;

[0016] obtaining all positive combinations corresponding to all positive e-commerce goods and all positive community contents;

[0017] obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each two users;

[0018] obtaining a target item, the target item including: a target e-commerce good or a target community content;

[0019] obtaining a target similarity score corresponding to the target item and greater than a preset threshold;

[0020] outputting a to-be-recommended item in a positive combination corresponding to the target similarity score; wherein, when the target item is a target e-commerce good, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce good.

[0021] The technical scheme provided by the embodiment of the present disclosure can include the following beneficial effects: the embodiment of the present disclosure provides an APP-based cross-domain recommendation method, including: obtaining all positive combinations of each user, each positive combination including: a positive e-commerce good and a positive community content; obtaining the number of same positive combinations of each two users as a weight score; obtaining all positive combinations corresponding to all positive e-commerce goods and all positive community contents; obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each two users; obtaining a target item, the target item including: a target e-commerce good or a target community content; obtaining a target similarity score corresponding to the target item and greater than a preset threshold; outputting a to-be-recommended item in a positive combination corresponding to the target similarity score; wherein, when the target item is a target e-commerce good, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce good. Through the above method, the process of manually tagging goods can be solved, manpower can be saved, and the relationship between more e-commerce goods and community contents can be covered. Furthermore, based on the behavior data of users, the recommendation from e-commerce goods to community contents (or vice versa) is directly and clearly linked.

[0022] In one embodiment, obtaining the similarity score of each of the positive combinations according to the weight score of each two users includes:

[0023] grouping users corresponding to a current positive combination to form a plurality of user pairs; each of the user pairs includes two users;

[0024] obtaining a temporary score of each of the user pairs in the current positive combination according to the weight score of each two users;

[0025] Summing up the temporary scores of all user pairs in the current forward combination, a similarity score corresponding to the current forward combination is obtained.

[0026] In one embodiment, after the similarity score corresponding to each of the forward combinations is obtained according to the weight score of each of the two users, the method further comprises:

[0027] According to the similarity score corresponding to each of the forward combinations, a community content corresponding to each of the e-commerce goods is obtained, and a similarity score between each of the e-commerce goods and the corresponding community content is obtained.

[0028] According to the similarity score corresponding to each of the forward combinations, an e-commerce good corresponding to each of the community contents is obtained, and a similarity score between each of the community contents and the corresponding e-commerce good is obtained.

[0029] The output of the target similarity score corresponding to the forward combination of the recommended item comprises:

[0030] The output of the target similarity score corresponding to the recommended item.

[0031] In one embodiment, the obtaining of the total forward combination of each user comprises:

[0032] The positive behavior of each user in the e-commerce goods and the community content is counted, and the positive behavior comprises clicking, purchasing, liking or commenting;

[0033] The forward e-commerce goods and the forward community content corresponding to the positive behavior of each user are obtained.

[0034] The total forward combination of each user is obtained according to the forward e-commerce goods and the forward community content corresponding to the positive behavior of each user.

[0035] In one embodiment, the obtaining of the forward e-commerce goods and the forward community content corresponding to the positive behavior of each user comprises:

[0036] The abnormal data in the positive behavior of each user in the e-commerce goods and the community content is deleted, and the abnormal data comprises the positive behavior whose number of positive behaviors satisfies a preset condition.

[0037] The forward e-commerce goods and the forward community content corresponding to the positive behavior of each user are obtained according to the remaining data.

[0038] According to a second aspect of the embodiments of the present disclosure, an APP-based cross-domain recommendation device is provided, comprising:

[0039] The first obtaining module is configured to obtain all positive combinations of each user, each of the positive combinations comprising a positive e-commerce commodity and a positive community content.

[0040] The second obtaining module is configured to obtain the number of positive combinations of each two users as a weight score.

[0041] The third obtaining module is configured to obtain all positive combinations corresponding to all positive e-commerce commodities and all positive community contents.

[0042] The fourth obtaining module is configured to obtain a similarity score corresponding to each of the positive combinations according to the weight score of each two users.

[0043] The fifth obtaining module is configured to obtain a target item, the target item comprising a target e-commerce commodity or a target community content.

[0044] The sixth obtaining module is configured to obtain a target similarity score corresponding to the target item, the target similarity score being greater than a preset threshold.

[0045] The output module is configured to output a to-be-recommended item in a positive combination corresponding to the target similarity score, wherein, when the target item is a target e-commerce commodity, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity.

[0046] In an embodiment, the fourth obtaining module comprises:

[0047] The grouping submodule is configured to group users corresponding to a current positive combination to form a plurality of user pairs, each of the user pairs comprising two users.

[0048] The first obtaining submodule is configured to obtain a temporary score of each of the user pairs in the current positive combination according to the weight score of each two users.

[0049] The second obtaining submodule is configured to total the temporary scores of all user pairs in the current positive combination to obtain a similarity score corresponding to the current positive combination.

[0050] In an embodiment, after the similarity score corresponding to each of the positive combinations is obtained according to the weight score of each two users, the device further comprises a seventh obtaining module and an eighth obtaining module, and the output module comprises an output submodule.

[0051] The seventh obtaining module is configured to obtain a community content corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content according to the similarity score corresponding to each of the positive combinations.

[0052] The eighth obtaining module is configured to obtain, according to the similarity score corresponding to each of the positive combinations, an e-commerce commodity corresponding to each of the community contents, and a similarity score between each of the community contents and the corresponding e-commerce commodity.

[0053] The output sub-module is configured to output the target similarity score corresponding to the to-be-recommended item.

[0054] In an embodiment, the first obtaining module comprises:

[0055] The statistical sub-module is configured to count positive behaviors of each user in e-commerce commodities and community contents, the positive behaviors including clicking, purchasing, liking or commenting;

[0056] The second obtaining sub-module is configured to obtain positive e-commerce commodities and positive community contents corresponding to the positive behaviors of each user.

[0057] The third obtaining sub-module is configured to obtain, according to the positive e-commerce commodities and the positive community contents corresponding to the positive behaviors of each user, all positive combinations of each user.

[0058] In an embodiment, the second obtaining sub-module comprises:

[0059] The deletion sub-module is configured to delete abnormal data in the positive behaviors of each user in e-commerce commodities and community contents, wherein the abnormal data includes positive behaviors whose number of times satisfies a preset condition.

[0060] The fourth obtaining sub-module is configured to obtain, according to the remaining data, the positive e-commerce commodities and the positive community contents corresponding to the positive behaviors of each user.

[0061] According to a third aspect of the embodiments of the present disclosure, an APP-based cross-domain recommendation device is provided, comprising:

[0062] a processor;

[0063] a memory for storing processor-executable instructions;

[0064] The processor is configured to:

[0065] obtain all positive combinations of each user, each of the positive combinations including one positive e-commerce commodity and one positive community content;

[0066] obtain the number of positive combinations of each of two users as a weight score;

[0067] obtain all positive combinations corresponding to all positive e-commerce commodities and all positive community contents;

[0068] According to the weight of each of the two users, a similarity score corresponding to each of the positive combinations is obtained;

[0069] A target item is obtained, the target item including: a target e-commerce commodity or a target community content;

[0070] A target similarity score corresponding to the target item is obtained, the target similarity score being greater than a preset threshold;

[0071] A to-be-recommended item in a positive combination corresponding to the target similarity score is output; when the target item is a target e-commerce commodity, the to-be-recommended item is a community content; and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity.

[0072] According to a fourth aspect of the embodiments of the present disclosure, a computer readable storage medium is provided, and the computer readable storage medium stores computer instructions. The computer instructions are executed by a processor to implement the steps of the method in any one of the first aspect.

[0073] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF DRAWINGS

[0074] The accompanying drawings, which are incorporated into and form part of the specification, illustrate embodiments consistent with the present disclosure and, together with the specification, serve to explain the principles of the present disclosure.

[0075] Figure 1 is a flowchart of an APP-based cross-domain recommendation method according to an exemplary embodiment.

[0076] Figure 2 is a flowchart of an APP-based cross-domain recommendation method according to an exemplary embodiment.

[0077] Figure 3 is a flowchart of an APP-based cross-domain recommendation method according to an exemplary embodiment.

[0078] Figure 4 is a flowchart of an APP-based cross-domain recommendation method according to an exemplary embodiment.

[0079] Figure 5 is a block diagram of an APP-based cross-domain recommendation device according to an exemplary embodiment.

[0080] Figure 6 is a block diagram of an APP-based cross-domain recommendation device 80 according to an exemplary embodiment. DETAILED DESCRIPTION

[0081] The exemplary embodiments will be described in detail herein with reference to the attached drawings. The following description is made with reference to the accompanying drawings in which like reference numerals refer to like elements or features. The following detailed description is not intended to represent all embodiments in accordance with the present disclosure. Rather, they are merely examples of devices and methods in accordance with some aspects of the present disclosure as detailed in the appended claims.

[0082] First, the terms used in the present disclosure are explained:

[0083] (1) Community content (feed) is referred to as content, which is the content (text / video, etc.) created by users in the APP.

[0084] (2) E-commerce item (item) is referred to as commodity, which is the commodity listed by the seller in the micro-store APP.

[0085] (3) Community content tag (tag) is referred to as tag, such as food, star, etc.

[0086] (4) Positive behavior: such as user's click, purchase, like, comment, etc. in the APP.

[0087] Figure 1 is a flowchart of an APP-based cross-domain recommendation method according to an exemplary embodiment, as shown in Figure 1 , the method comprises the following steps S101-S107:

[0088] In step S101, all positive combinations of each user are obtained, each positive combination comprising a positive e-commerce item and a positive community content;

[0089] In one implementation, as shown in Figure 2 , all positive combinations of each user are obtained, comprising the following sub-steps:

[0090] In step S1011, the positive behavior of each user in the e-commerce item and the community content is counted, and the positive behavior includes click, purchase, like or comment;

[0091] In step S1012, the positive e-commerce item and the positive community content corresponding to the positive behavior of each user are obtained;

[0092] In one embodiment, as shown in Figure 3 , the positive e-commerce item and the positive community content corresponding to the positive behavior of each user are obtained, comprising the following sub-steps:

[0093] In step S10121, delete the abnormal data in the positive behavior of each user in the e-commerce goods and community content; wherein the abnormal data includes: the number of positive behaviors satisfying the preset condition;

[0094] In step S10122, the positive e-commerce goods and the positive community content corresponding to the positive behavior of each user are obtained according to the remaining data.

[0095] In step S1013, the total positive combination of each user is obtained according to the positive e-commerce goods and the positive community content corresponding to the positive behavior of each user.

[0096] In this step, first, the positive behavior of the user in the e-commerce and the content is counted, and the abnormal data (such as the user or community content / e-commerce goods with too many or too few clicks, or the user or community content / e-commerce goods with too many or too few purchases, or the user or community content / e-commerce goods with too many or too few likes, or the user or community content / e-commerce goods with too many or too few comments) is removed.

[0097] In step S102, the number of two users having the same positive combination is obtained as the weight score;

[0098] Based on the positive behavior of the user, the weight score (user_score) of the behavior of two users is calculated:

[0099] Use (positive goods, positive content) as a positive combination (positive pair), for example: use (positive goods A, positive content B) as a positive pair. The weight score of two users A and B is the number of the same positive pairs, as shown in Table 1, which is used to calculate the weight score of A and B, wherein A and B have common positive combinations (e-commerce goods A, community content B), (e-commerce goods C, community content D) and (e-commerce goods E, community content F), so the weight score of A and B is 3 points.

[0100] Table 1

[0101]

[0102]

[0103] Specifically, the user with a large number of positive behaviors and the user with a small number of positive behaviors have different information amounts for the same positive behavior pair. That is: assuming that user A clicks all community content and e-commerce goods, then he simultaneously clicks AB cannot explain anything. Assuming that user B only clicks a few times, then he simultaneously clicks AB, which indicates that the similarity of AB is likely to be high.

[0104] In step S103, all positive combinations corresponding to all positive e-commerce goods and all positive community content are obtained;

[0105] All positive pairs (positive e-commerce goods, positive community content) are calculated.

[0106] In step S104, the similarity score of each positive combination is obtained according to the weight score of each pair of users;

[0107] In one embodiment, as shown in FIG. 1, the similarity score of each positive combination is obtained according to the weight score of each pair of users, including the following sub-steps: Figure 4

[0108] In step S1041, the users corresponding to the current positive combination are grouped to form a plurality of user pairs; each user pair includes two users.

[0109] All positive pairs of different users (at least two) with positive behavior are combined.

[0110] In step S1042, the temporary score of each user pair in the current positive combination is obtained according to the weight score of each pair of users.

[0111] For example, each user pair calculates the temporary score according to the formula 1 / (1+user_score).

[0112] In step S1043, the temporary scores of all user pairs in the current positive combination are summed up to obtain the similarity score corresponding to the current positive combination.

[0113] The more people have positive behavior on the combination of (good A, content B), the more likely it is that the similarity of the two is high.

[0114] For example, Table 2, the positive combination is (e-commerce good X, community content Y), and the users corresponding to (e-commerce good X, community content Y) are A, B, C, D, E and F. A, B, C, D, E and F are combined in pairs, wherein A and B form a group, C and D form a group, and E and F form a group. According to Table 1, the user_score of A and B is 3, and the temporary score of A and B is calculated to be 0.25 by 1 / (1+user_score). Similarly, the user_score of C and D is 5, and the temporary score of C and D is calculated to be 0.2 by 1 / (1+user_score); the user_score of E and F is 9, and the temporary score of E and F is calculated to be 0.1 by 1 / (1+user_score). Finally, the similarity score of (e-commerce good X, community content Y) is 0.55 by summing up the temporary scores of A and B, C and D, and E and F.​

[0115] Table 2

[0116]

[0117] In step S105, a target item is obtained, the target item including: a target e-commerce commodity or a target community content;

[0118] In step S106, a target similarity score greater than a preset threshold is obtained, the target similarity score corresponding to the target item;

[0119] In step S107, a to-be-recommended item in a positive combination corresponding to the target similarity score is outputted; wherein, when the target item is the target e-commerce commodity, the to-be-recommended item is the community content, and when the target item is the target community content, the to-be-recommended item is the e-commerce commodity.

[0120] The target item can be a commodity purchased by a user or a community content browsed by the user. Based on the similarity score of the positive combination obtained above, the user behavior and the similarity score can be used for recommendation. For example, if a user A purchases an e-commerce commodity A, the most similar community contents B, C and D of the commodity A are found and recommended.

[0121] The embodiment of the present disclosure provides an APP-based cross-domain recommendation method, including: obtaining all positive combinations of each user, each positive combination including: a positive e-commerce commodity and a positive community content; obtaining the number of same positive combinations of each two users as a weight score; obtaining all positive combinations corresponding to all positive e-commerce commodities and all positive community contents; obtaining a similarity score corresponding to each positive combination according to the weight score of each two users; obtaining a target item, the target item including: a target e-commerce commodity or a target community content; obtaining a target similarity score greater than a preset threshold, the target similarity score corresponding to the target item; outputting a to-be-recommended item in a positive combination corresponding to the target similarity score; wherein, when the target item is the target e-commerce commodity, the to-be-recommended item is the community content, and when the target item is the target community content, the to-be-recommended item is the e-commerce commodity. Through the above method, the process of manually labeling commodities can be solved, manpower can be saved, and more relationships between e-commerce commodities and community contents can be covered. Further, based on the behavior data of the user, the recommendation from the e-commerce commodity to the community content (or vice versa) is directly clear.

[0122] Further, the weight score of the user in the cross-domain calculation in the disclosure is calculated, and the traditional user weight calculation method only considers all click behaviors of the user as the score, without cross-domain and without constructing (commodity / content) combinations. The similarity scores of the community content and the e-commerce commodities are calculated, and in the process of calculating the similarity of the e-commerce commodities and the community content, the two are distinguished, and only the similarity score of one domain corresponding to the other domain is calculated. The traditional calculation method only considers the similarity within the domain.

[0123] In one embodiment,

[0124] After obtaining the similarity score corresponding to each of the positive combinations according to the weight score of each of the two users, the community content corresponding to each of the e-commerce commodities and the similarity score between each of the e-commerce commodities and the corresponding community content can be obtained according to the similarity score corresponding to each of the positive combinations; and the e-commerce commodity corresponding to each of the community contents and the similarity score between each of the community contents and the corresponding e-commerce commodities can be obtained according to the similarity score corresponding to each of the positive combinations; at this time, the target similarity score corresponding to the recommended item is directly output.

[0125] After obtaining the similarity score corresponding to each of the positive combinations, the similar community content corresponding to a certain e-commerce commodity and the score can be obtained. Or the similar e-commerce commodity corresponding to the community content and the score.

[0126] The following is an apparatus embodiment of the disclosure, which can be used to execute the method embodiment of the disclosure.

[0127] Figure 5 is a block diagram of an APP-based cross-domain recommendation device according to an exemplary embodiment. The device can be realized as part or all of an electronic device by software, hardware, or a combination of the two. As Figure 5 shown, the APP-based cross-domain recommendation device includes:

[0128] The first obtaining module 11 is configured to obtain all positive combinations of each user, and each of the positive combinations includes a positive e-commerce commodity and a positive community content.

[0129] The second obtaining module 12 is configured to obtain the number of positive combinations of each of the two users as the weight score.

[0130] The third obtaining module 13 is configured to obtain all positive combinations corresponding to all positive e-commerce commodities and all positive community contents.

[0131] The fourth obtaining module 14 is configured to obtain the similarity score corresponding to each of the positive combinations according to the weight score of each of the two users.

[0132] The fifth obtaining module 15 is configured to obtain a target item, where the target item comprises a target e-commerce commodity or a target community content.

[0133] The sixth obtaining module 16 is configured to obtain a target similarity score corresponding to the target item, where the target similarity score is greater than a preset threshold.

[0134] The output module 17 is configured to output a to-be-recommended item in a positive combination corresponding to the target similarity score, where when the target item is the target e-commerce commodity, the to-be-recommended item is the community content, and when the target item is the target community content, the to-be-recommended item is the e-commerce commodity.

[0135] In an embodiment, the fourth obtaining module comprises:

[0136] The grouping submodule is configured to group users corresponding to a current positive combination to form a plurality of user pairs, where each user pair comprises two users.

[0137] The first obtaining submodule is configured to obtain a temporary score of each user pair in the current positive combination according to a weight score of each of the two users.

[0138] The second obtaining submodule is configured to add up the temporary scores of all user pairs in the current positive combination to obtain a similarity score corresponding to the current positive combination.

[0139] In an embodiment, after the similarity score corresponding to each positive combination is obtained according to the weight score of each of the two users, the device further comprises a seventh obtaining module and an eighth obtaining module, and the output module comprises an output submodule.

[0140] The seventh obtaining module is configured to obtain, according to the similarity score corresponding to each positive combination, a community content corresponding to each e-commerce commodity and a similarity score between each e-commerce commodity and the corresponding community content.

[0141] The eighth obtaining module is configured to obtain, according to the similarity score corresponding to each positive combination, an e-commerce commodity corresponding to each community content and a similarity score between each community content and the corresponding e-commerce commodity.

[0142] The output submodule is configured to output a to-be-recommended item corresponding to the target similarity score.

[0143] In an embodiment, the first obtaining module comprises:

[0144] The counting submodule is configured to count positive behaviors of each user in e-commerce commodities and community contents, where the positive behaviors comprise clicking, purchasing, liking, or commenting.

[0145] a second obtaining sub-module, configured to obtain a positive e-commerce commodity and a positive community content corresponding to a positive behavior of each user;

[0146] a third obtaining sub-module, configured to obtain all positive combinations of each user according to the positive e-commerce commodity and the positive community content corresponding to the positive behavior of each user.

[0147] In an embodiment, the second obtaining sub-module comprises:

[0148] a deleting sub-module, configured to delete abnormal data in the positive behavior of each user in the e-commerce commodity and the community content; wherein the abnormal data comprises a positive behavior whose number of times satisfies a preset condition;

[0149] a fourth obtaining sub-module, configured to obtain a positive e-commerce commodity and a positive community content corresponding to a positive behavior of each user according to the remaining data.

[0150] According to a third aspect of the embodiments of the present disclosure, an APP-based cross-domain recommendation device is provided, comprising:

[0151] a processor;

[0152] a memory for storing processor-executable instructions;

[0153] wherein the processor is configured to:

[0154] obtain all positive combinations of each user, each of the positive combinations comprising a positive e-commerce commodity and a positive community content;

[0155] obtain a number of positive combinations of each two users as a weight score;

[0156] obtain all positive combinations corresponding to all positive e-commerce commodities and all positive community contents;

[0157] obtain a similarity score corresponding to each of the positive combinations according to the weight score of each of the two users;

[0158] obtain a target item, the target item comprising a target e-commerce commodity or a target community content;

[0159] obtain a target similarity score corresponding to the target item, the target similarity score being greater than a preset threshold;

[0160] output a to-be-recommended item in a positive combination corresponding to the target similarity score; wherein when the target item is a target e-commerce commodity, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity.

[0161] The processor can be further configured to:

[0162] According to the weight score of each two users, a similarity score of each of the forward combinations is obtained, including:

[0163] Grouping the users corresponding to the current forward combination to form a plurality of user pairs, each of the user pairs including two users;

[0164] According to the weight score of each two users, a temporary score of each of the user pairs in the current forward combination is obtained;

[0165] Summing up the temporary scores of all the user pairs in the current forward combination to obtain a similarity score corresponding to the current forward combination.

[0166] In one embodiment, after the processor obtains the similarity score corresponding to each of the forward combinations according to the weight score of each two users, the processor can be further configured to:

[0167] According to the similarity score corresponding to each of the forward combinations, obtaining a community content corresponding to each of the e-commerce goods, and a similarity score between each of the e-commerce goods and the corresponding community content;

[0168] According to the similarity score corresponding to each of the forward combinations, obtaining an e-commerce good corresponding to each of the community contents, and a similarity score between each of the community contents and the corresponding e-commerce good;

[0169] The output of the to-be-recommended items in the forward combination corresponding to the target similarity score includes:

[0170] Outputting the to-be-recommended items corresponding to the target similarity score.

[0171] In one embodiment, the processor obtains all the forward combinations of each user, including:

[0172] Counting the forward behaviors of each user in e-commerce goods and community contents, the forward behaviors including clicking, purchasing, liking, or commenting;

[0173] Obtaining forward e-commerce goods and forward community contents corresponding to the forward behaviors of each user;

[0174] Obtaining all the forward combinations of each user according to the forward e-commerce goods and the forward community contents corresponding to the forward behaviors of each user.

[0175] In one embodiment, the processor obtains the forward e-commerce goods and the forward community contents corresponding to the forward behaviors of each user, including:

[0176] delete abnormal data in the positive behaviors of each user in the e-commerce goods and the community content; wherein the abnormal data includes positive behaviors whose number of times meets a preset condition;

[0177] obtain the positive e-commerce goods and the positive community content corresponding to the positive behaviors of each user according to the residual data.

[0178] As to the apparatus in the above embodiments, the specific manners in which various modules perform operations have been described in detail in the embodiments of the method, and thus will not be described in detail here.

[0179] Figure 6 is a block diagram of an apparatus 80 for APP-based cross-domain recommendation according to an example embodiment, which is suitable for a terminal device. For example, the apparatus 80 can be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.

[0180] The apparatus 80 can include one or more of the following components: a processing component 802, a memory 804, a power supply component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, a sensor component 814, and a communication component 816.

[0181] The processing component 802 usually controls overall operations of the apparatus 80, such as operations associated with displaying, making phone calls, data communications, camera operations, and recording operations. The processing component 802 can include one or more processors 820 to execute instructions to complete all or part of steps of the methods described above. Further, the processing component 802 can include one or more modules to facilitate interaction between the processing component 802 and other components. For example, the processing component 802 can include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0182] The memory 804 is configured to store various types of data to support operations of the apparatus 80. Examples of these data include instructions for any application or methods operating on the apparatus 80, contact data, phonebook data, messages, pictures, videos, etc. The memory 804 can be realized by any type of volatile or nonvolatile storage devices or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0183] Power component 806 provides power to the various components of the device 80. The power component 806 can include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 80.

[0184] The multimedia component 808 includes a display for the device 80 to display graphics, text, and other information for a user. The display can include both a touch-sensitive component (e.g., a touch screen, a touchpad, and so forth) and a non-touch-sensitive component. The display can also include, for example, a liquid crystal display (LCD) and a touch panel (TP). If the display includes a touch panel, the display can be implemented as a touch screen to receive input from a user. The touch panel includes one or more touch sensors to sense touch, swiping, and gestures on the touch panel. The touch sensors can not only sense the boundary of a touch or swipe, but also detect duration and pressure associated with the touch or swipe. In some embodiments, the multimedia component 808 includes a front-facing camera and / or a rear-facing camera. When the device 80 is in operation, the front-facing camera and / or the rear-facing camera can receive external multimedia data. Each camera can be a fixed optical lens system or have a focal length and optical zoom capability.

[0185] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC) that is configured to receive an external audio signal when the device 80 is in an operation mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signal can be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, the audio component 810 also includes a speaker for outputting audio signals.

[0186] The I / O interface 812 provides an interface between the processing component 802 and peripheral interface modules, which can be a keyboard, a click wheel, buttons, and so forth. The buttons can include, but are not limited to, a home button, a volume button, a start button, and a lock button.

[0187] The sensor component 814 includes one or more sensors to provide the device 80 with state assessments of various aspects. For example, the sensor component 814 can detect an open / closed state of the device 80, relative positioning of components, such as a display and keypad of the device 80, a change in position of the device 80 or a component of the device 80, presence or absence of user contact with the device 80, orientation or acceleration / deceleration of the device 80, and temperature changes of the device 80. The sensor component 814 can include proximity sensor(s) configured to detect presence of nearby objects without any physical contact. The sensor component 814 can also include a light sensor, such as a CMOS or CCD image sensor, used for imaging applications. In some embodiments, the sensor component 814 can further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0188] The communication component 816 is configured to facilitate wired or wireless communication between the device 80 and another device. The device 80 can access a wireless network based on a communication standard, such as WiFi, 2G, or 3G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra-WideBand (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0189] In an exemplary embodiment, the device 80 can be implemented using one or more Application-Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, or other electronic components, for performing the above-described methods.

[0190] In an exemplary embodiment, a non-transitory computer-readable storage medium, such as the memory 804 including instructions stored therein, is also provided. The instructions may, for example, be executable by the processor 820 of the device 80 to perform the above-described methods. The non-transitory computer-readable storage medium may, for example, be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disc, and an optical data storage device, and the like.

[0191] A non-transitory computer-readable storage medium, when instructions stored therein are executed by a processor of a device, enables the device to perform an APP-based cross-domain recommendation method as described above, the method including:

[0192] obtaining all positive combinations of each user, each of the positive combinations comprising a positive e-commerce commodity and a positive community content;

[0193] obtaining a weight score of each two users having the same positive combination as a weight score;

[0194] obtaining all positive combinations corresponding to all positive e-commerce commodities and all positive community contents;

[0195] obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each two users;

[0196] obtaining a target item, the target item comprising a target e-commerce commodity or a target community content;

[0197] obtaining a target similarity score corresponding to the target item, the target similarity score being greater than a preset threshold;

[0198] outputting a to-be-recommended item in a positive combination corresponding to the target similarity score; wherein, when the target item is a target e-commerce commodity, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity.

[0199] In one embodiment, obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each two users comprises:

[0200] grouping users corresponding to a current positive combination to form a plurality of user pairs, each of the user pairs comprising two users;

[0201] obtaining a temporary score of each of the user pairs in the current positive combination according to the weight score of each two users;

[0202] summing up the temporary scores of all user pairs in the current positive combination to obtain a similarity score corresponding to the current positive combination.

[0203] In one embodiment, after obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each two users, the method further comprises:

[0204] obtaining a community content corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content according to the similarity score corresponding to each of the positive combinations;

[0205] obtaining an e-commerce commodity corresponding to each of the community contents and a similarity score between each of the community contents and the corresponding e-commerce commodity according to the similarity score corresponding to each of the positive combinations;

[0206] The outputted to-be-recommended item in the positive combination corresponding to the target similarity score comprises:

[0207] The outputted to-be-recommended item in the positive combination corresponding to the target similarity score comprises:

[0208] In an embodiment, the obtaining of the total positive combination of each user comprises:

[0209] The positive behavior of each user in the e-commerce goods and the community content comprises clicking, purchasing, liking, or commenting;

[0210] The positive behavior of each user in the e-commerce goods and the community content comprises clicking, purchasing, liking, or commenting;

[0211] The total positive combination of each user is obtained according to the positive e-commerce goods and the positive community content corresponding to the positive behavior of each user.

[0212] In an embodiment, the obtaining of the positive e-commerce goods and the positive community content corresponding to the positive behavior of each user comprises:

[0213] The abnormal data in the positive behavior of each user in the e-commerce goods and the community content is deleted, wherein the abnormal data comprises a positive behavior whose number of times of positive behavior satisfies a preset condition;

[0214] The positive e-commerce goods and the positive community content corresponding to the positive behavior of each user are obtained according to the remaining data.

[0215] Other embodiments of the present disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any variations, uses or adaptations of the disclosure other than those expressly disclosed herein. It is to be understood that the specification and examples are illustrative only and not restrictive of the present disclosure as claimed. The true scope and spirit of the present disclosure is indicated by the following claims.

[0216] It should be understood that the present disclosure is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made to the embodiments without departing from the scope thereof. The scope of the present disclosure is limited only by the claims that follow.

Claims

1. A method for cross-domain recommendation based on APP, characterized in that, The method comprises the following steps: obtaining all positive combinations of each user, each of the positive combinations comprising a positive e-commerce commodity and a positive community content; obtaining the number of two users having the same positive combination as a weight score; obtaining all positive combinations corresponding to all positive e-commerce commodities and all positive community contents; obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each of the two users; obtaining a target item, the target item comprising a target e-commerce commodity or a target community content; obtaining a target similarity score corresponding to the target item and greater than a preset threshold; outputting a to-be-recommended item in a positive combination corresponding to the target similarity score; wherein, when the target item is a target e-commerce commodity, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity; wherein, obtaining a similarity score of each of the positive combinations according to the weight score of each of the two users comprises: grouping users corresponding to a current positive combination to form a plurality of user pairs, each of the user pairs comprising two users; obtaining a temporary score of each of the user pairs in the current positive combination according to the weight score of each of the two users; and adding up the temporary scores of all user pairs in the current positive combination to obtain a similarity score corresponding to the current positive combination. The method further comprises the following steps after obtaining the similarity score corresponding to each of the positive combinations according to the weight score of each of the two users: obtaining community content corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content according to the similarity score corresponding to each of the positive combinations; obtaining e-commerce commodities corresponding to each of the community contents and a similarity score between each of the community contents and the corresponding e-commerce commodities according to the similarity score corresponding to each of the positive combinations; and outputting the to-be-recommended item corresponding to the target similarity score. 2.The APP-based cross-domain recommendation method according to claim 1, characterized in that, The method further comprises the following steps after obtaining the similarity score corresponding to each of the positive combinations according to the weight score of each of the two users: obtaining community content corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content according to the similarity score corresponding to each of the positive combinations; obtaining e-commerce commodities corresponding to each of the community contents and a similarity score between each of the community contents and the corresponding e-commerce commodities according to the similarity score corresponding to each of the positive combinations; and outputting the to-be-recommended item corresponding to the target similarity score. 3.The APP-based cross-domain recommendation method according to claim 1, characterized in that, The method further comprises the following steps after obtaining the similarity score corresponding to each of the positive combinations according to the weight score of each of the two users: obtaining community content corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content according to the similarity score corresponding to each of the positive combinations; obtaining e-commerce commodities corresponding to each of the community contents and a similarity score between each of the community contents and the corresponding e-commerce commodities according to the similarity score corresponding to each of the positive combinations; and outputting the to-be-recommended item corresponding to the target similarity score.

4. An APP-based cross-domain recommendation apparatus, characterized in that, The method further comprises the following steps after obtaining the similarity score corresponding to each of the positive combinations according to the weight score of each of the two users: obtaining community content corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content according to the similarity score corresponding to each of the positive combinations; obtaining e-commerce commodities corresponding to each of the community contents and a similarity score between each of the community contents and the corresponding e-commerce commodities according to the similarity score corresponding to each of the positive combinations; and outputting the to-be-recommended item corresponding to the target similarity score. The first obtaining module is configured to obtain all positive combinations of each user, each of the positive combinations comprising a positive e-commerce commodity and a positive community content; the obtaining of all positive combinations of each user comprises: counting positive behaviors of each user in e-commerce commodities and community contents, the positive behaviors comprising clicking, purchasing, liking, or commenting; obtaining positive e-commerce commodities and positive community contents corresponding to the positive behaviors of each user; and obtaining all positive combinations of each user according to the positive e-commerce commodities and the positive community contents corresponding to the positive behaviors of each user; The second obtaining module is configured to obtain a number of positive combinations of each two users as a weight score of each two users; The third obtaining module is configured to obtain all positive combinations corresponding to all positive e-commerce commodities and all positive community contents; The fourth obtaining module is configured to obtain a similarity score corresponding to each of the positive combinations according to the weight score of each two users; The fifth obtaining module is configured to obtain a target item, the target item comprising a target e-commerce commodity or a target community content; The sixth obtaining module is configured to obtain a target similarity score corresponding to the target item, the target similarity score being greater than a preset threshold; The output module is configured to output a to-be-recommended item in a positive combination corresponding to the target similarity score; when the target item is a target e-commerce commodity, the to-be-recommended item is a community content; and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity. The fourth obtaining module comprises: a grouping submodule configured to group users corresponding to a current positive combination to form a plurality of user pairs; each of the user pairs comprises two users; a first obtaining submodule configured to obtain a temporary score of each of the user pairs in the current positive combination according to the weight score of each two users; and a second obtaining submodule configured to total the temporary scores of all the user pairs in the current positive combination to obtain a similarity score corresponding to the current positive combination. 5.The APP-based cross-domain recommendation apparatus of claim 4, wherein, After the similarity score corresponding to each of the positive combinations is obtained according to the weight score of each two users, the device further comprises a seventh obtaining module and an eighth obtaining module, and the output module comprises an output submodule; the seventh obtaining module is configured to obtain, according to the similarity score corresponding to each of the positive combinations, community contents corresponding to each of the e-commerce commodities and a similarity score between each of the e-commerce commodities and the corresponding community content; the eighth obtaining module is configured to obtain, according to the similarity score corresponding to each of the positive combinations, e-commerce commodities corresponding to each of the community contents and a similarity score between each of the community contents and the corresponding e-commerce commodity; and the output submodule is configured to output a to-be-recommended item corresponding to the target similarity score.

6. An APP-based cross-domain recommendation apparatus, characterized in that, The device comprises: a processor; a memory configured to store processor-executable instructions; and the processor is configured to: Obtaining all positive combinations of each user, each of the positive combinations comprising a positive e-commerce commodity and a positive community content; obtaining the number of two users having the same positive combination as a weight score; obtaining all positive combinations corresponding to all positive e-commerce commodities and all positive community contents; obtaining a similarity score corresponding to each of the positive combinations according to the weight score of each of the two users; obtaining a target item, the target item comprising a target e-commerce commodity or a target community content; obtaining a target similarity score corresponding to the target item, the target similarity score being greater than a preset threshold; outputting a to-be-recommended item in the positive combination corresponding to the target similarity score; wherein, when the target item is a target e-commerce commodity, the to-be-recommended item is a community content, and when the target item is a target community content, the to-be-recommended item is an e-commerce commodity; wherein, Obtaining the similarity score of each of the positive combinations according to the weight score of each of the two users comprises: grouping users corresponding to a current positive combination to form a plurality of user pairs, each of the user pairs comprising two users; obtaining a temporary score of each of the user pairs in the current positive combination according to the weight score of each of the two users; and summing up the temporary scores of all the user pairs in the current positive combination to obtain the similarity score corresponding to the current positive combination. The obtaining of all positive combinations of each user comprises: counting positive behaviors of each user in e-commerce commodities and community contents, the positive behaviors comprising clicking, purchasing, liking, or commenting; obtaining positive e-commerce commodities and positive community contents corresponding to the positive behaviors of each user; and obtaining all positive combinations of each user according to the positive e-commerce commodities and the positive community contents corresponding to the positive behaviors of each user.

7. A computer readable storage medium having stored thereon computer instructions, wherein, The instruction is executed by the processor to implement the steps of the method in any one of claims 1 to 3.

Images