Recommended methods, apparatus, equipment, and storage media

Abstract

This application discloses a recommendation method, apparatus, device, and storage medium; wherein the method includes: obtaining the degree of association between candidate products and at least one target interactive application; obtaining usage information of the at least one target interactive application; determining the preference degree of a target user for the candidate products based at least on the degree of association between the candidate products and the at least one target interactive application, and the usage information of the at least one target interactive application; and selecting recommendable candidate products for recommendation based on the preference degree of each candidate product.

Description

Technical Field

[0001] This application relates to computer technology, and to, but is not limited to, a recommended method, apparatus, device, and storage medium. Background Technology

[0002] With the increasing diversification of value-added services offered by telecom operators, various non-monetary purchase products are bringing users a brand-new experience. "Non-monetary purchase product marketing" represents a development direction for operators in terms of "user retention" and "user value management." Currently, the common approach to non-monetary purchase product marketing by operators is to offer users a package of online internet application membership cards and offline merchandise vouchers. These products are diverse and cover all aspects of users' lives, including food, clothing, housing, transportation, and entertainment. However, they fail to provide differentiated services, causing confusion for users and leading to potential user churn, making it difficult to improve product purchase conversion rates.

[0003] Personalized recommendations, based on massive amounts of data such as basic user information, basic product information, user-product interaction information, and time-space information, mine users' individual interests and purchasing behaviors to recommend information and products that users might be interested in. This approach is widely used in e-commerce, news, and social media. Personalized product recommendations can provide users with products they truly need and are interested in, helping to improve user satisfaction and stickiness. However, the user base for non-monetary product marketing by telecom operators is relatively small, and the amount of user-product interaction data is limited, resulting in significant user and product cold starts, which is insufficient to support the effective implementation of personalized recommendation algorithms in related technologies. Summary of the Invention

[0004] In view of this, the recommendation method, apparatus, device, and storage medium provided in this application can achieve accurate product recommendations for users without requiring a large amount of user-candidate product interaction data, and the recommendation results meet the actual needs of users.

[0005] According to one aspect of the embodiments of this application, a recommendation method is provided, comprising: obtaining the degree of association between candidate products and at least one target interactive application; obtaining usage information of the at least one target interactive application; determining the preference degree of a target user for the candidate products based at least on the degree of association between the candidate products and the at least one target interactive application, and the usage information of the at least one target interactive application; and selecting recommendable candidate products for recommendation based on the preference degree of each candidate product.

[0006] The recommendation apparatus provided in this application includes: a first acquisition module, configured to acquire the degree of association between candidate products and at least one target interactive application; a second acquisition module, configured to acquire usage information of the at least one target interactive application; a determination module, configured to determine the preference degree of a target user for the candidate products based at least on the degree of association between the candidate products and the at least one target interactive application, and the usage information of the at least one target interactive application; and a recommendation module, configured to select recommendable candidate products for recommendation based on the preference degree of each candidate product.

[0007] The electronic device provided in this application includes a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements the method described in this application.

[0008] The computer-readable storage medium provided in this application embodiment stores a computer program thereon, which, when executed by a processor, implements the method described in this application embodiment.

[0009] In this embodiment, the following steps are taken: First, the degree of association between candidate products and at least one target interactive application is obtained. Second, usage information of the target user on at least one target interactive application is obtained. Third, the target user's preference for the candidate products is determined based on at least the degree of association between the candidate products and at least one target interactive application, as well as the usage information of at least one target interactive application. Fourth, based on the preference of each candidate product, recommendable candidate products are selected. Thus, by incorporating the target user's usage information on frequently used target interactive applications, and then jointly determining the target user's preference for the candidate products based on the association between the target interactive applications and candidate products, accurate product recommendations can be achieved without requiring a large amount of user interaction data with candidate products, and the recommendation results meet the user's actual needs.

[0010] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0011] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0012] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0013] Figure 1 A schematic diagram illustrating the implementation flow of a recommended method provided in an embodiment of this application;

[0014] Figure 2 A schematic diagram illustrating the implementation flow of a recommended method provided in an embodiment of this application;

[0015] Figure 3 A schematic diagram illustrating the implementation flow of a recommended method provided in an embodiment of this application;

[0016] Figure 4 A comparison chart of recommendation results before and after the fusion of user proportions provided in this application embodiment;

[0017] Figure 5 A schematic diagram illustrating the implementation flow of a recommended method provided in an embodiment of this application;

[0018] Figure 6 A schematic diagram of the recommended device provided in the embodiments of this application;

[0019] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0022] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0023] This application provides a recommended method applied to an electronic device. This electronic device can be various types of computer devices with information processing capabilities, such as desktop computers, laptops, or handheld computers. It can also be a mobile terminal, such as a mobile phone, in-vehicle computer, tablet computer, or POS machine. Furthermore, it can be a server, such as a physical server or server cluster. The functionality achieved by this method can be implemented by a processor in the electronic device calling program code. The program code can be stored in a computer storage medium. Therefore, the electronic device includes at least a processor and a storage medium.

[0024] Figure 1 The following is a schematic diagram illustrating the implementation flow of the recommended method provided in the embodiments of this application, as shown below. Figure 1 As shown, the method may include the following steps 101 to 104:

[0025] Step 101: Obtain the degree of association between the candidate product and at least one target interactive application.

[0026] In this embodiment, the target interactive application is an application that the user has used. The target interactive application can be any application on the device, or it can be an application selected from multiple candidate interactive applications that meets certain conditions. For example, in some embodiments, the target interactive application can be selected by executing steps 1011 to 1014.

[0027] It should be noted that when the recommendation method is applied to a server, the target interactive application obtained should be an application used by the target user (i.e., the user to be recommended the corresponding candidate product), rather than an application used by other users.

[0028] In some embodiments, determining the target interactive application corresponding to the candidate product and constructing the degree of association between the candidate product and at least one target interactive application can be achieved by performing the following steps 1011 to 1015:

[0029] Step 1011: Determine the name similarity between multiple candidate interactive applications and candidate products.

[0030] When identifying target interactive applications that are related to candidate products, candidate interactive applications belonging to the same category can be first categorized to obtain multiple different sets of candidate interactive applications; then, the name similarity between the candidate product and each candidate interactive application in the set of candidate interactive applications belonging to the same category as the candidate product is determined.

[0031] In some embodiments, determining the name similarity between candidate products and candidate interactive applications can be done by determining the similarity between strings of names. For example, in some embodiments, the similarity between two names is calculated using the Jaccard coefficient. The Jaccard coefficient is calculated as shown in Formula 1, where A and B represent two sets of strings, |A∩B| represents the number of characters in the intersection of the two strings, and |A∪B| represents the number of characters in the union of the two strings.

[0032]

[0033] For example, for the candidate product "AB Video Membership Card", the Jaccard coefficient between the candidate product name "AB Video" and the names of all candidate interactive applications is calculated, resulting in J("AB Video", "AB Video") = 4 / 4 = 1. This determines the name similarity between the candidate product "AB Video Membership Card" and the candidate interactive application "AB Video". Of course, in some embodiments, J("AB Video", "AB Video Membership Card") = 4 / 7 can also be calculated, and this is not a limitation.

[0034] Step 1012: Select the application whose name similarity meets the second condition from multiple candidate interactive applications as the target interactive application.

[0035] After determining the name similarity between each candidate interactive application and the candidate product, the target interactive application is selected based on the name similarity of each candidate interactive application.

[0036] In some embodiments, the second condition is the highest name similarity, for example, selecting the candidate interactive application with the highest similarity as the target interactive application. In other embodiments, the second condition may also be that the name similarity is greater than a certain threshold, for example, selecting the candidate interactive application with a similarity greater than 0.5 as the target interactive application. In the embodiments of this application, the second condition is not limited.

[0037] Step 1013: If the number of target interactive applications is less than the second threshold, determine the functional similarity between the other candidate interactive applications (excluding the target interactive application) and the candidate product.

[0038] Step 1014: Select the application whose functional similarity meets the third condition from other candidate interactive applications as the new target interactive application.

[0039] Understandably, not every candidate product has at least one target interactive application that has been visited by the target user and has a similar name. Therefore, selecting target interactive applications based on the name similarity between candidate products and candidate interactive applications may not be able to identify target interactive applications that meet the quantity requirements.

[0040] In some embodiments, if a target interactive application cannot be selected by calculating name similarity or the number of selected target interactive applications does not meet the requirements, an application whose function is similar to that of the candidate product can be determined as a new target interactive application.

[0041] For example, suppose that for the candidate product "CD Bike 30-day riding card", after calculating the name similarity between the candidate product and each candidate interactive application, no target interactive application that meets the quantity requirement is matched, but the functional similarity between the candidate interactive applications "AC Bike" and "BE Riding" and "CD Bike" meets the third condition, then "AC Bike" and "BE Riding" can be used as new target interactive applications.

[0042] Here, there are no restrictions on the third condition. For example, the third condition can be the highest functional similarity or the functional similarity greater than a certain threshold.

[0043] It should be noted that the execution order of steps 1011 and 1013 is not limited in this embodiment. For example, step 1011 can be executed first, followed by step 1013. That is, the name similarity between candidate products and candidate interactive applications is calculated first, and then the functional similarity between candidate products and candidate interactive applications is calculated. In this way, name similarity is easier to calculate than functional similarity. After the target interactive applications that meet the quantity requirements are determined based on name similarity, the step of calculating functional similarity can be avoided to a certain extent, thereby speeding up the calculation and saving computing power.

[0044] Of course, you can also execute step 1013 first and then step 1011, that is, first calculate the functional similarity between the candidate product and the candidate interactive application, and then calculate the name similarity between the candidate product and the candidate interactive application. In this way, the target interactive application can be determined based on the functional similarity, and there is a high probability that the target interactive application that meets the quantity requirement can be directly determined. This can largely avoid the step of calculating the name similarity, thereby saving computing power.

[0045] Step 1015: Determine the degree of association between the candidate product and the target interactive application based at least on the similarity between the candidate product and the selected target interactive application in meeting the second condition or the third condition.

[0046] In some embodiments, if the name similarity and functional similarity of the finally selected target interactive application both meet the corresponding conditions, the degree of association between the candidate product and the target interactive application can be determined based on the name similarity of the target interactive application; the degree of association between the candidate product and the target interactive application can also be determined based on the functional similarity of the target interactive application; or the degree of association between the candidate product and the target interactive application can be determined by comprehensively considering the name similarity and functional similarity of the target interactive application. This application does not limit this.

[0047] It should be noted that the degree of correlation between candidate products and target interactive applications is directly proportional to their similarity; that is, the higher the similarity, the greater the correlation.

[0048] For example, if the candidate product "AB Video Membership Card" and the target interactive application "AB Video" have different names but belong to the same type of application software, then the association degree between the candidate product "AB Video Membership Card" and the target interactive application "AB Video" is set to 1. Similarly, although "CD Bike" and "AC Bike" and "BE Cycling" are not the same application and have different names, their functions are similar. Therefore, the association degree between the candidate product "CD Bike 30-Day Cycling Card" and the target interactive applications "AC Bike" and "BE Cycling" is set to a value in the range (0, 1), for example, 0.8 in this embodiment.

[0049] In some embodiments, the target interactive application that is associated with the candidate product can also be determined by calculating the semantic similarity between the candidate product and the candidate interactive application.

[0050] Step 102: Obtain usage information of the target user on at least one target interactive application.

[0051] In some embodiments, after identifying the target interactive application corresponding to the candidate product, usage information of the target user on the target interactive application over a period of time can be collected. For example, usage information of the target user on "AB Video" over the past week.

[0052] Step 103: Determine the target user's preference for the candidate product based at least on the degree of association between the candidate product and at least one target interactive application, and the usage information of at least one target interactive application.

[0053] Understandably, in certain specific environments, such as product recommendations by operators, the user base is relatively small, and there is also limited interaction data between users and candidate products. Therefore, if product recommendations are made to users based solely on the interaction data between users and candidate products, the accuracy of the recommendations will be poor.

[0054] Based on this, in this embodiment, the following steps are taken: First, the degree of association between candidate products and at least one target interactive application is obtained. Second, the usage information of the target user on at least one target interactive application is obtained. Third, the target user's preference for the candidate products is determined based on at least the degree of association between the candidate products and at least one target interactive application, as well as the usage information of at least one target interactive application. Fourth, based on the preference of each candidate product, recommended candidate products are selected for recommendation. Thus, by introducing the target user's usage information on frequently used target interactive applications, and then jointly determining the target user's preference for the candidate products based on the association between the target interactive applications and candidate products, accurate product recommendations can be achieved without requiring a large amount of user interaction data with candidate products, and the recommendation results meet the user's actual needs.

[0055] Step 104: Select recommended candidate products based on the preference level of each candidate product.

[0056] In some embodiments, step 104 can be implemented by performing steps 207 to 209 in the following embodiments.

[0057] Figure 2 The following is a schematic diagram illustrating the implementation flow of the recommended method provided in the embodiments of this application, as shown below. Figure 2 As shown, the method may include the following steps 201 to 209:

[0058] Step 201: Obtain the degree of association between the candidate product and at least one target interactive application;

[0059] Step 202: Obtain at least one usage parameter of the target user for at least one target interactive application.

[0060] In some embodiments, the parameters used include at least one of the following: number of uses, duration of use, and data usage.

[0061] Step 203: The correlation between the candidate product and the target interactive application and the usage parameters of the target interactive application are fused to obtain the first fused value of the usage parameters corresponding to the candidate product.

[0062] Here, there is no limitation on the method of fusing the correlation degree and the usage parameters of the target interactive application. For example, in some embodiments, the correlation degree and the usage parameters can be fused by multiplying the correlation degree between the candidate product and the target interactive application and the usage parameters of the target interactive application, thereby obtaining a first fused value of the usage parameter corresponding to the candidate product.

[0063] For example, for the candidate product "CD Bike 30-day riding card", its corresponding target interactive applications are "AC Bike" and "BE Riding", and its correlation with these two target interactive applications is 0.8. Assume that target user A uses the target interactive application "AC Bike" 5 times, for a total duration of 288 minutes, and uses 12M of data; target user A uses the target interactive application "BE Riding" 8 times, for a total duration of 208 minutes, and uses 10.8M of data.

[0064] Therefore, the first fusion value for the number of uses of the target interactive application "AC Bike" corresponding to the candidate product "CD Bike 30-day riding card" is: correlation degree * number of uses of "AC Bike" = (0.8 * 5) = 4 times; correspondingly, the first fusion value for the usage duration is: correlation degree * usage duration of "AC Bike" = (0.8 * 288) = 230.4 minutes; the first fusion value for the data usage is: correlation degree * data usage of "AC Bike" = (0.8 * 12) = 9.6M; the first fusion value for the number of uses of the target interactive application "BE Riding" corresponding to the candidate product "CD Bike 30-day riding card" is: correlation degree * number of uses of "BE Riding" = (0.8 * 8) = 6.4 times; correspondingly, the first fusion value for the usage duration is: correlation degree * usage duration of "BE Riding" = (0.8 * 208) = 166.4 minutes; the first fusion value for the data usage is: correlation degree * data usage of "BE Riding" = (0.8 * 10.8) = 8.64M.

[0065] For example, consider the candidate product "AB Video Membership Card," whose target interactive application is "AB Video," with a correlation degree of 1. Assume target user A uses the target interactive application "AB Video" 21 times, for a total duration of 266 minutes, and uses 8345MB of data. Then, the first fusion value for the number of times the target interactive application "AB Video" is used is: correlation degree * number of times "AB Video" is used = (1 * 21) = 21 times; correspondingly, the first fusion value for the duration of use is: correlation degree * duration of use of "AB Video" = (1 * 266) = 266 minutes; and the first fusion value for the data usage is: correlation degree * data usage of "AB Video" = (1 * 8345) = 8345MB.

[0066] Step 204: Merge the first fusion value corresponding to each target interactive application to obtain the second fusion value of the usage parameters corresponding to the candidate product.

[0067] Here, the fusion method for the first fusion value is not limited. For example, in some embodiments, the first fusion value can be fused by adding the first fusion values ​​corresponding to each target interactive application to obtain a second fusion value for a certain usage parameter corresponding to the candidate product.

[0068] For example, regarding the parameter "usage duration," the target interactive applications for the candidate product "CD Bike 30-Day Riding Card" are "AC Bike" and "BE Riding." The first combined value for the usage duration of "AC Bike" is 230.4 minutes, and the first combined value for the usage duration of "BE Riding" is 166.4 minutes. Therefore, the second combined value for the usage duration of the candidate product "CD Bike 30-Day Riding Card" is 230.4 + 166.4 = 396.8 minutes; correspondingly, the second combined value for the number of uses is 4 + 6.4 = 10.4 times; and the second combined value for data usage is 9.6 + 8.64 = 18.24 MB.

[0069] Correspondingly, for the candidate product "AB Video Membership Card", the second converged value of its usage time is 266 minutes, the second converged value of its usage frequency is 21 times, and the second converged value of its data usage is 8345M.

[0070] Step 205: Merge the second fusion value corresponding to each candidate product to obtain the total fusion value corresponding to the usage parameters.

[0071] Here, the method of fusing the second fusion value is not limited. For example, in some embodiments, the second fusion value can be fused by adding the second fusion values ​​corresponding to each candidate product to obtain the total fusion value of a certain usage parameter.

[0072] For example, by combining the second fusion value of the number of uses corresponding to the candidate product "CD Bike 30-Day Riding Card" and the second fusion value of the number of uses corresponding to the candidate product "AB Video Membership Card", the total fusion value of the number of uses is 10.4 + 21 = 31.4 times; correspondingly, the second fusion value of the usage time is 396.8 + 266 = 662.8 minutes; and the total fusion value of the data usage is 18.24 + 8345 = 8363.24M.

[0073] Step 206: Determine the target user's preference for the candidate product based at least on the second fusion value of each usage parameter corresponding to the candidate product and the total fusion value corresponding to each usage parameter.

[0074] In some embodiments, step 206 can be achieved by performing steps 3091 to 3092 in the following embodiments, that is, taking the total ratio of candidate products determined in step 3092 as the target user's preference for candidate products.

[0075] Step 207: Sort the candidate products according to the preference of each candidate product to obtain the sorting result.

[0076] In this embodiment, the sorting method for candidate products is not limited. For example, candidate products with higher preference can be ranked first, and candidate products with lower preference can be ranked last (reverse order); of course, candidate products with lower preference can also be ranked first, and candidate products with higher preference can be ranked last.

[0077] Step 208: Select the candidate products whose preference meets the first condition from the ranking results as the first recommended products.

[0078] After sorting the candidate products, recommending them in order of preference, and excluding products with lower preference, can make the recommendation results more in line with the user's actual needs.

[0079] Step 209: Recommend at least the first recommended product as a candidate product.

[0080] Understandably, the number of recommendable candidate products in the ranking results obtained through the recommendation method in steps 207 to 208 may be relatively small. For example, in some cases, the first condition is set to a preference score greater than 0.7, and the required number of recommendations is 9, but only 7 candidate products with a preference score greater than 0.7 are found in the ranking results. In this case, the following steps 2091 to 2094 can be performed to continue selecting recommendable candidate products:

[0081] Step 2091: If the number of recommendable candidate products is less than the threshold, obtain the preference of other users for the candidate products.

[0082] Step 2092: Determine the mass preference of the candidate products based at least on the preferences of other users for the candidate products.

[0083] Here, the preferences of other users for the candidate product are taken into account, thus recommending products to the target users from the perspective of popular preferences.

[0084] In some embodiments, the mass preference of a candidate product can be determined jointly based on the preference of target users for the candidate product and the preference of other users for the candidate product; alternatively, the mass preference of a candidate product can be determined solely based on the preference of other users for the candidate product.

[0085] Step 2093: Based on the popularity of each candidate product, select at least one second recommended product that is different from the first recommended product from among the candidate products.

[0086] In some embodiments, the total number of the second recommended products and the first recommended products is greater than or equal to a threshold.

[0087] For example, if it is necessary to recommend 6 products to each target user, then 6 first recommended products are obtained sequentially from the recommendation methods in steps 207 to 208; if it is necessary to recommend 9 products to each target user, then 7 first recommended products are obtained sequentially from the recommendation methods in steps 207 to 208, and then 2 second recommended products that are different from the first recommended products are selected according to public preference.

[0088] Step 2094: Recommend the first recommended product and at least one second recommended product as recommended candidate products.

[0089] It should be noted that the first recommended product here can be at least one recommended product determined in step 208 above, or it can be all the recommended products determined in step 208 above.

[0090] In this embodiment of the application, if the number of recommended candidate products does not meet the demand, recommendations are made based on the general preference of multiple other users for the candidate products, thereby ensuring that the number of recommended products is sufficient and enriching the content of the product recommendation list.

[0091] Figure 3 The following is a schematic diagram illustrating the implementation flow of the recommended method provided in the embodiments of this application, as shown below. Figure 3 As shown, the method may include the following steps 301 to 310:

[0092] Step 301: Obtain the degree of association between the candidate product and at least one target interactive application;

[0093] Step 302: Obtain at least one usage parameter of at least one target interactive application of the target user;

[0094] Step 303: The correlation between the candidate product and the target interactive application and the usage parameters of the target interactive application are fused to obtain the first fused value of the usage parameters corresponding to the candidate product.

[0095] Step 304: Merge the first fusion value corresponding to each target interactive application to obtain the second fusion value of the usage parameters corresponding to the candidate product;

[0096] Step 305: Merge the second fusion values ​​corresponding to each candidate product to obtain the total fusion value corresponding to the parameters used;

[0097] Step 306: Obtain the first total number of users using at least one target interactive application.

[0098] Understandably, video and music products constitute a large proportion of the candidate products, have a large audience, and consume significant amounts of time and data. Conversely, convenient lifestyle products like "BE Takeout" consume less user time and data. Therefore, when making recommendations based on the preference calculation method in steps 201 to 206 of the above embodiment, the calculated preference for candidate interactive applications like "BE Takeout," which have a smaller actual audience and consume less time and data, is lower. This does not reflect the actual situation.

[0099] Based on this, in the embodiments of this application, not only the user's usage information of the target interactive application is considered, but also the total number of users of the target interactive application is considered, and the user's preference for candidate products that are related to the target interactive application is obtained in a comprehensive manner, so that the recommendation results are more in line with the user's needs.

[0100] Step 307: Merge the first total number of users corresponding to each candidate product to obtain the second total number of users.

[0101] In some embodiments, the total number of users of all candidate products can be merged by adding the first total number of users corresponding to each candidate product to obtain the second total number of users.

[0102] Step 308: Determine the user percentage of the candidate product based on the total number of first users and the total number of second users corresponding to the candidate product.

[0103] In some embodiments, the user share of a candidate product can be determined by calculating the inverse user frequency index (iuf) of the candidate product. The inverse user frequency index is derived from the inverse document frequency index (IRF). The inverse user frequency index (iuf) of candidate product R is calculated as: log(total number of second users / total number of first users). Wherein, the smaller the total number of first users of candidate product R, the larger the inverse user frequency index (iuf) of candidate product R.

[0104] Step 309: Determine the target user's preference for the candidate products based on the user proportion of the candidate products, the second fusion value of each usage parameter corresponding to the candidate products, and the total fusion value corresponding to each usage parameter.

[0105] In some embodiments, step 309 can be achieved by performing the following steps 3091 to 3093:

[0106] Step 3091: Determine the ratio of the second fusion value of the usage parameters corresponding to the candidate product to the total fusion value corresponding to the usage parameters;

[0107] Step 3092: The ratios of each usage parameter corresponding to the candidate product are fused to obtain the total ratio corresponding to the candidate product.

[0108] In some embodiments, the determination of the overall ratio corresponding to candidate product R can be achieved using Formula 2:

[0109] average(dur-R / all-dur+num-R / all-num+flux-R / all-flux)(formula 2);

[0110] Where, dur_R is the second fusion value of the duration of target user A's use of candidate product R, num_R is the second fusion value of the number of times target user A uses candidate product R, flux_R is the second fusion value of the usage traffic of target user A's use of candidate product R, all_dur is the total fusion value of the duration of use, all_num is the total fusion value of the number of uses, and all_flux is the total fusion value of the usage traffic.

[0111] Step 3093: Combine the user percentage of candidate products with the total ratio to obtain the target user's preference for candidate products.

[0112] Here, there is no limitation on the method of combining the user percentage and the total ratio. For example, in some embodiments, the user percentage of candidate product R determined in step 308 can be multiplied by the total ratio of candidate product R determined in step 3092 to obtain the target user's preference for candidate product R.

[0113] Step 310: Select recommended candidate products based on the preference level of each candidate product.

[0114] In this embodiment of the application, by fusing the user proportion of candidate products with the fusion value of the usage parameters of candidate products, products with a smaller audience but actual demand can be recommended to users, which helps to uncover user needs and enrich the content of the product recommendation list. For example Figure 4 The left side of the middle section shows the ranking of a user's preference for a product before the user percentage was integrated. Figure 4 The right side of the middle section shows the user's preference for the product after merging the user percentages of the candidate products. As you can see, by merging the user percentages of the candidate products, the recommendation rankings of the candidate products "G application", "I application", and "L application" have improved to a certain extent, which is obviously more in line with the actual situation.

[0115] In some embodiments, personalized recommendation algorithms require massive amounts of data, especially user-product interaction data, such as user clicks, add-to-cart, and purchase behavior data that reflect user interests. However, in the marketing scenario of operators purchasing products without monetary value (i.e., an example of candidate products), due to the small user base and limited user-product interaction data, the personalized recommendation algorithms in some embodiments become obsolete in certain scenarios.

[0116] Based on this, the following will describe an exemplary application of the embodiments of this application in a practical application scenario.

[0117] In this embodiment, based on deep packet inspection (DPI) data, user behavior when using mobile applications (APPs) is deeply analyzed to uncover user interest in non-monetary purchase products offered by mobile operators, such as online internet application membership cards and offline merchandise vouchers. This allows for personalized recommendations of these non-monetary purchase products. Through text analysis, a mapping relationship and mapping factors between non-monetary purchase products and APPs are constructed. These factors are mapped to user preference for non-monetary purchase products based on three dimensions: APP usage time, APP usage frequency, and APP data usage (i.e., some examples of usage parameters). Combined with the number of APP users, a weighted preference for non-monetary purchase products is obtained, leading to recommendations of potentially interesting non-monetary purchase products for the user.

[0118] like Figure 5 As shown, the personalized recommendation process for operators' non-monetary purchase products (hereinafter referred to as "products") is completed in four steps:

[0119] Step 1:

[0120] (1) Analyze the product name and APP name, and use the Jaccard coefficient to find the APP that maps to the product;

[0121] (2) Due to the complexity and diversity of data, if some products are not mapped to the APP using the method in (1), then the APP with the same function should be used to map them.

[0122] (3) Set a mapping factor (i.e., degree of association) for each pair (product, APP) to represent the degree of mapping relationship between the product and the APP. The value of the mapping factor is in the range (0,1]. For two completely identical products and APPs, the mapping factor is 1; for (product, APP) pairs with different names but belonging to the same application software, the mapping factor is 1; for (product, APP) pairs belonging to different application software but having the same function, the mapping factor can be set to a value in the range (0,1).

[0123] Step Two:

[0124] (1) Take the apps that are mapped to the product obtained in step one as the target app set. From the DPI data, calculate the usage information (i.e., usage data) of each user for each app in the target app set (APP1, APP2, ..., APPn) in the past week, including usage duration, number of uses, and data usage. For example, calculate the usage duration, number of uses, and data usage of user A for AB video in the past week;

[0125] (2) Multiply the mapping factor of each APP in step (3) by the usage duration of that APP (i.e., an example of the first fusion value), and sum the usage durations of APPs mapped to the same product to obtain the virtual usage duration of the user for that product (i.e., an example of the second fusion value); similarly, multiply the number of times an APP mapped to the same product is used by the APP mapping factor and sum them to obtain the virtual number of times the user uses that product; multiply the usage traffic of APPs mapped to the same product by the APP mapping factor and sum them to obtain the virtual usage traffic of the user for that product.

[0126] Step 3:

[0127] (1) For each user, calculate the user's preference for each product based on the user's virtual usage time, virtual usage frequency, and virtual usage data usage. The user's preference for a particular product is the average of the sum of the user's percentage of virtual usage time for that product, the user's percentage of virtual usage frequency for that product, and the user's percentage of virtual usage data usage for that product. The sum of the user's preferences for all products is 1.

[0128] (2) Calculate the total number of virtual users for all products and the number of virtual users for each product. Here, virtual users refer to users who use an APP that has a mapping relationship with the product.

[0129] (3) Calculate the inverse user frequency index iuf for each product. Multiply the index by the user’s preference for the product to obtain the user’s weighted preference for the product.

[0130] Step Four:

[0131] (1) For each user, sort the user’s weighted preference for all products in reverse order (only select products with non-zero weighted preference);

[0132] (2) Calculate the average weighted preference score (i.e., the mass preference score) of each product, and arrange all products in reverse order according to this score;

[0133] (3) Suppose that N products need to be recommended to each user. Take the top N products from (1) in order. If there are fewer than N products, take the required number of products from (2) in order.

[0134] In this embodiment, we first propose analyzing user behavior on an app using DPI data to calculate users' weighted preference for non-monetary purchase products from operators, and then providing a personalized overall solution for recommending non-monetary purchase products. Secondly, we construct a mapping relationship between non-monetary purchase products and the app, and establish a method for setting mapping factors. Finally, we propose a calculation logic for users' weighted preference for non-monetary purchase products. Thus, even with limited user-product interaction data, we provide operators with an efficient product recommendation method.

[0135] It should be noted that although the steps of the method in this application are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps; or steps from different embodiments may be combined into a new technical solution.

[0136] Based on the foregoing embodiments, this application provides a recommended device, which includes the included modules and the units included in each module, which can be implemented by a processor; of course, it can also be implemented by specific logic circuits; in the implementation process, the processor can be a central processing unit (CPU), microprocessor (MPU), digital signal processor (DSP) or field programmable gate array (FPGA), etc.

[0137] Figure 6 This is a schematic diagram of the structure of the recommended device in the embodiments of this application, such as... Figure 6 As shown, the device 600 includes a first acquisition module 601, a second acquisition module 602, a determination module 603, and a recommendation module 604, wherein:

[0138] The first acquisition module 601 is used to acquire the degree of association between candidate products and at least one target interactive application; the second acquisition module 602 is used to acquire the usage information of the at least one target interactive application; the determination module 603 is used to determine the target user's preference for the candidate products based at least on the degree of association between the candidate products and the at least one target interactive application, and the usage information of the at least one target interactive application; the recommendation module 604 is used to select recommendable candidate products for recommendation based on the preference of each candidate product.

[0139] In some embodiments, the usage information of the target interactive application includes at least one usage parameter; the device further includes a fusion module, which is configured to fuse the correlation between the candidate product and the target interactive application and the usage parameter of the target interactive application to obtain a first fusion value of the usage parameter corresponding to the candidate product; and to fuse the first fusion value corresponding to each target interactive application to obtain a second fusion value of the usage parameter corresponding to the candidate product; and to fuse the second fusion value corresponding to each candidate product to obtain a total fusion value corresponding to the usage parameter; a determination module 603 is configured to determine the target user's preference for the candidate product based at least on the second fusion value of each usage parameter corresponding to the candidate product and the total fusion value corresponding to each usage parameter.

[0140] In some embodiments, the second acquisition module 602 is further configured to acquire the first total number of users using the at least one target interactive application; the determination module 603 is further configured to determine the target user's preference for the candidate product based on the first total number of users, the second fusion value of each of the usage parameters corresponding to the candidate product, and the total fusion value corresponding to each of the usage parameters.

[0141] In some embodiments, the fusion module is used to fuse the first total number of users corresponding to each candidate product to obtain a second total number of users; the determining module 603 is used to determine the user proportion of the candidate product based on the first total number of users corresponding to the candidate product and the second total number of users; the determining module 603 is also used to determine the target user's preference for the candidate product based on the user proportion of the candidate product, the second fusion value of each usage parameter corresponding to the candidate product, and the total fusion value corresponding to each usage parameter.

[0142] In some embodiments, the determining module 603 is used to determine the ratio of the second fusion value of the usage parameter corresponding to the candidate product to the total fusion value corresponding to the usage parameter; the fusion module is used to fuse the ratio of each usage parameter corresponding to the candidate product to obtain the total ratio corresponding to the candidate product; and to fuse the user proportion of the candidate product with the total ratio to obtain the target user's preference for the candidate product.

[0143] In some embodiments, the apparatus further includes a sorting module and a selection module. The sorting module is used to sort the candidate products according to the preference degree of each candidate product to obtain a sorting result. The selection module is used to select candidate products that meet the preference degree conditions from the sorting result as first recommended products. The recommendation module 604 is used to recommend at least the first recommended product as a recommendable candidate product.

[0144] In some embodiments, a first acquisition module is configured to acquire the preference of other users for the candidate products when the number of recommendable candidate products is less than a threshold; a determination module 603 is configured to determine the popular preference of the candidate products based at least on the preference of other users for the candidate products; a selection module is configured to select at least one second recommended product that is different from the first recommended product from the candidate products based on the popular preference of each candidate product; and a recommendation module 604 is configured to recommend the first recommended product and the at least one second recommended product as recommendable candidate products.

[0145] In some embodiments, the determining module 603 is configured to determine the name similarity between a plurality of candidate interactive applications and the candidate product; the selecting module is configured to select an application whose name similarity satisfies a second condition from the plurality of candidate interactive applications as the target interactive application; the determining module 603 is configured to determine the functional similarity between the other candidate interactive applications (excluding the target interactive application) and the candidate product when the number of target interactive applications is less than a second threshold; the selecting module is configured to select an application whose functional similarity satisfies a third condition from the other candidate interactive applications as the new target interactive application; and the determining module 603 is configured to determine the degree of association between the candidate product and the target interactive application based at least on the similarity between the candidate product and the selected target interactive application satisfying either the second condition or the third condition.

[0146] The descriptions of the above device embodiments are similar to those of the above method embodiments, and have similar beneficial effects. For technical details not disclosed in the device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0147] It should be noted that, in the embodiments of this application... Figure 6The module division shown in the recommended device is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, exist as separate physical units, or be integrated into one unit with two or more units. The integrated units can be implemented in hardware, as software functional units, or a combination of both.

[0148] It should be noted that, in the embodiments of this application, if the above-described methods are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0149] This application provides an electronic device. Figure 7 This is a schematic diagram of the hardware entity of the electronic device according to an embodiment of this application, such as... Figure 7 As shown, the electronic device 700 includes a memory 701 and a processor 702. The memory 701 stores a computer program that can run on the processor 702. When the processor 702 executes the program, it implements the steps in the method provided in the above embodiments.

[0150] It should be noted that the memory 701 is configured to store instructions and applications executable by the processor 702, and can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data and video communication data) in the processor 702 and various modules in the electronic device 700. It can be implemented by flash memory or random access memory (RAM).

[0151] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the method provided in the above embodiments.

[0152] This application provides a computer program product containing instructions that, when run on a computer, cause the computer to perform the steps in the method provided in the above-described method embodiments.

[0153] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium, storage medium, and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0154] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" mentioned throughout the specification mean that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, they will not be repeated here.

[0155] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three kinds of relationships. For example, object A and / or object B can represent three situations: object A exists alone, object A and object B exist simultaneously, and object B exists alone.

[0156] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0157] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple modules or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or modules can be electrical, mechanical, or other forms.

[0158] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.

[0159] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.

[0160] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0161] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0162] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0163] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0164] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0165] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A recommendation method, characterized in that, The method includes: Obtain the degree of correlation between the candidate product and at least one target interactive application; Obtain usage information of the at least one target interactive application; The target user's preference for the candidate product is determined based at least on the degree of association between the candidate product and the at least one target interactive application, as well as the usage information of the at least one target interactive application. Based on the preference level of each candidate product, recommendable candidate products are selected and recommended. The usage information of the target interactive application includes at least one usage parameter; Determining the target user's preference for the candidate products based at least on the degree of association between the candidate products and the at least one target interactive application, and the usage information of the at least one target interactive application, includes: The correlation between the candidate product and the target interactive application and the usage parameters of the target interactive application are fused to obtain a first fused value of the usage parameters corresponding to the candidate product; The first fusion value corresponding to the same usage parameter of multiple target interactive applications corresponding to each candidate product is fused to obtain the second fusion value of the usage parameter corresponding to each candidate product. The second fusion values ​​corresponding to the same usage parameter of all the candidate products are fused to obtain the total fusion value corresponding to the usage parameter. The target user's preference for the candidate products is determined based on the second fusion value of each usage parameter corresponding to each candidate product and the total fusion value corresponding to each usage parameter.

2. The method according to claim 1, characterized in that, The step of determining the target user's preference for the candidate products based on the second fusion value of each usage parameter corresponding to each candidate product and the total fusion value corresponding to each usage parameter includes: Obtain the first total number of users using the at least one target interactive application; The target user's preference for the candidate product is determined based on the first total number of users, the second fusion value of each of the usage parameters corresponding to the candidate product, and the total fusion value corresponding to each of the usage parameters.

3. The method according to claim 2, characterized in that, The step of determining the target user's preference for the candidate products based on the first total number of users, the second fusion value of each usage parameter corresponding to the candidate products, and the total fusion value corresponding to each usage parameter includes: The first total number of users corresponding to each candidate product is merged to obtain the second total number of users; The user percentage of the candidate products is determined based on the total number of the first user and the total number of the second user corresponding to the candidate products. The target user's preference for the candidate products is determined based on the user percentage of the candidate products, the second fusion value of each of the usage parameters corresponding to the candidate products, and the total fusion value corresponding to each of the usage parameters.

4. The method according to claim 3, characterized in that, The step of determining the target user's preference for the candidate products based on the user proportion of the candidate products, the second fusion value of each of the usage parameters corresponding to the candidate products, and the total fusion value corresponding to each of the usage parameters includes: Determine the ratio of the second fusion value of the usage parameter corresponding to the candidate product to the total fusion value corresponding to the usage parameter; The ratios of each of the usage parameters corresponding to the candidate products are fused to obtain the total ratio corresponding to the candidate products; The user percentage of the candidate products and the total ratio are combined to obtain the target user's preference for the candidate products.

5. The method according to any one of claims 1 to 4, characterized in that, The usage parameters include at least one of the following: number of uses, usage duration, and usage data volume.

6. The method according to claim 1, characterized in that, The step of selecting recommendable candidate products based on the preference level of each candidate product includes: The candidate products are sorted according to the preference level of each candidate product to obtain the sorting result; From the ranking results, select the candidate products that meet the preference criteria as the first recommended products; At least the first recommended product should be recommended as a candidate product.

7. The method according to claim 6, characterized in that, The step of recommending at least the first recommended product as a recommendable candidate product includes: If the number of recommended candidate products is less than a threshold, obtain the preference of other users for the candidate products; The popular preference of the candidate products is determined at least based on the preference of the other users for the candidate products; Based on the popularity of each candidate product, at least one second recommended product that is different from the first recommended product is selected from each candidate product; The first recommended product and the at least one second recommended product are recommended as candidate products.

8. The method according to claim 1, characterized in that, The process of constructing the association between the candidate product and at least one target interactive application includes: Determine the name similarity between multiple candidate interactive applications and the candidate products; From the plurality of candidate interactive applications, the application whose name similarity meets the second condition is selected as the target interactive application; If the number of target interactive applications is less than a second threshold, determine the functional similarity between the other candidate interactive applications (excluding the target interactive application) and the candidate product. From the other candidate interactive applications, select the application whose functional similarity meets the third condition as the new target interactive application; The degree of association between the candidate product and the target interactive application is determined based at least on the similarity between the candidate product and the selected target interactive application in satisfying the second condition or the third condition.

9. A recommended device, characterized in that, include: The first acquisition module is used to acquire the degree of association between the candidate product and at least one target interactive application; The second acquisition module is used to acquire usage information of the at least one target interactive application; The determining module is used to determine the target user's preference for the candidate product based at least on the degree of association between the candidate product and the at least one target interactive application, and the usage information of the at least one target interactive application; The recommendation module is used to select recommendable candidate products based on the preference level of each candidate product. The usage information of the target interactive application includes at least one usage parameter. The determining module is specifically configured to: fuse the correlation between the candidate product and the target interactive application with the usage parameter of the target interactive application to obtain a first fusion value for the usage parameter corresponding to the candidate product; fuse the first fusion values ​​corresponding to the same usage parameter of multiple target interactive applications corresponding to each candidate product to obtain a second fusion value for the usage parameter corresponding to each candidate product; fuse the second fusion values ​​corresponding to the same usage parameter of all candidate products to obtain a total fusion value for the usage parameter; and determine the target user's preference for the candidate product based on the second fusion value for each usage parameter corresponding to each candidate product and the total fusion value for each usage parameter.

10. An electronic device comprising a memory and a processor, the memory storing a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the method according to any one of claims 1 to 8.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 8.

Images