Resource recommendation method and device, electronic equipment and storage medium

Abstract

The invention discloses a resource recommendation method and device, electronic equipment and a storage medium, and relates to the technical field of computers, in particular to the field of artificial intelligence such as deep learning and intelligent search. According to the specific implementation scheme, a candidate recommendation resource list corresponding to a target user is obtained; determining diversity preference information of the target user according to the historical resource interaction information of the target user; wherein the diversity preference information refers to preference information of the target user for resource diversity; according to the diversity preference information, reordering the candidate recommendation resource list to obtain a target recommendation resource list; and performing resource recommendation to the target user according to the target recommendation resource list.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computers, in particular to the field of artificial intelligence such as deep learning and intelligent search, and specifically relates to a resource recommendation method and device, an electronic device, and a storage medium. BACKGROUND

[0002] With the rapid development of Internet technology and mobile terminals, the resources available to users are rapidly increasing. For example, in the field of video content, massive video data can be continuously recommended to users through online video platforms, social media, short video applications, and other channels. For another example, in the field of e-commerce, e-commerce platforms can recommend goods to users. SUMMARY

[0003] The present application provides a resource recommendation method and device, an electronic device, and a storage medium. The specific solutions are as follows:

[0004] According to an aspect of the present application, a resource recommendation method is provided, comprising:

[0005] obtaining a candidate recommendation resource list corresponding to a target user;

[0006] determining diversity preference information of the target user according to historical resource interaction information of the target user; wherein the diversity preference information refers to preference information of the target user for resource diversity;

[0007] reordering the candidate recommendation resource list according to the diversity preference information to obtain a target recommendation resource list;

[0008] performing resource recommendation to the target user according to the target recommendation resource list.

[0009] According to another aspect of the present application, a resource recommendation device is provided, comprising:

[0010] an obtaining module configured to obtain a candidate recommendation resource list corresponding to a target user;

[0011] a determining module configured to determine diversity preference information of the target user according to historical resource interaction information of the target user; wherein the diversity preference information refers to preference information of the target user for resource diversity;

[0012] a reordering module configured to reorder the candidate recommendation resource list according to the diversity preference information to obtain a target recommendation resource list;

[0013] a recommendation module configured to perform resource recommendation to the target user according to the target recommendation resource list.

[0014] According to another aspect of the present application, an electronic device is provided, comprising:

[0015] at least one processor; and

[0016] a memory connected with the at least one processor; wherein,

[0017] the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method described in the above embodiments.

[0018] According to another aspect of the present application, a non-transitory computer readable storage medium storing computer instructions is provided, wherein the computer instructions are used to make the computer perform the method described in the above embodiments.

[0019] According to another aspect of the present application, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method described in the above embodiments.

[0020] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0021] The accompanying drawings are used to better understand the present application, and do not limit the present application. Among them:

[0022] Figure 1 a flowchart of a resource recommendation method provided by an embodiment of the present application;

[0023] Figure 2 a flowchart of a resource recommendation method provided by another embodiment of the present application;

[0024] Figure 3 a flowchart of a resource recommendation method provided by another embodiment of the present application;

[0025] Figure 4 a flowchart of a resource recommendation method provided by another embodiment of the present application;

[0026] Figure 5 a resource recommendation process diagram provided by an embodiment of the present application;

[0027] Figure 6 a structural diagram of a resource recommendation device provided by an embodiment of the present application;

[0028] Figure 7This is a block diagram of an electronic device used to implement the resource recommendation method of the embodiments of this application. Detailed Implementation

[0029] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of this application, including various details to aid understanding. These should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this application. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0030] It should be noted that the acquisition, storage, use, and processing of data in this application comply with the relevant provisions of national laws and regulations and do not violate public order and good morals.

[0031] The resource recommendation method, apparatus, electronic device, and storage medium of embodiments of this application are described below with reference to the accompanying drawings.

[0032] Figure 1 This is a flowchart illustrating a resource recommendation method provided in an embodiment of this application.

[0033] The resource recommendation method of this application embodiment can be executed by the resource recommendation device of this application embodiment, which can be configured in an electronic device.

[0034] Among them, electronic devices can be any device with computing capabilities, such as personal computers, mobile terminals, servers, etc. Mobile terminals can be hardware devices with various operating systems, touch screens and / or displays, such as in-vehicle devices, mobile phones, tablets, personal digital assistants, wearable devices, etc.

[0035] like Figure 1 As shown, the resource recommendation method includes:

[0036] Step 101: Obtain the list of candidate recommended resources corresponding to the target user.

[0037] The candidate recommended resource list can be a list of resources to be recommended determined by other methods. For example, the candidate recommended resource list can be a list of resources obtained through fine-tuning.

[0038] For example, the resource to be recommended in this application may be one or more of video resources, text resources, URLs, products, etc., or may be other resources, without limitation.

[0039] The target user can be any user who needs resource recommendations.

[0040] Step 102: Determine the diverse preference information of the target user based on the target user's historical resource interaction information.

[0041] Historical resource interaction information refers to the interaction behavior information generated between the target user and historical recommended resources. For example, historical resource interaction information may include, but is not limited to, the time, duration, and actions of the interaction between the target user and historical recommended resources.

[0042] For example, the historical recommended resources are videos, the interaction duration can be the viewing time of the video, and the interaction operations can include, but are not limited to, swiping, liking, commenting, and sharing the video.

[0043] For example, historical resource interaction information could be the interaction information between a target user and recommended resources within a target historical period. For instance, historical resource interaction information could be the interaction information between a target user and recommended resources over the past month, or the interaction information between a target user and recommended resources over the past six months, etc.

[0044] It should be noted that the target historical period in this application can be determined according to actual needs, and there are no restrictions on it.

[0045] In this application, diversity preference information may refer to the target user's preference information for resource diversity.

[0046] For example, diversity preference information may include, but is not limited to, diversity preference scores and diversity preference levels. The diversity preference score can be used to characterize the target user's preference for resource diversity; for instance, a higher diversity preference score indicates a greater preference for high diversity among target users.

[0047] In this application, the target user's operation information on historical recommended resources of each resource category can be determined based on the target user's historical resource interaction information, and the target user's diverse preference information can be determined based on the operation information.

[0048] For example, a resource category can refer to a first-level vertical category or a first-level category. For instance, resource categories could include technology, entertainment, sports, finance, etc.

[0049] Step 103: Based on the diversity preference information, reorder the candidate recommended resource list to obtain the target recommended resource list.

[0050] In this application, the number of similar resource intervals corresponding to the target user can be determined based on the diversity preference information, and the positions of similar resources in the candidate recommendation list can be adjusted according to the number of similar resource intervals corresponding to the target user, so that the number of intervals between the adjusted similar resources meets the requirement of the number of similar resource intervals corresponding to the target user, thereby obtaining the target recommendation resource list.

[0051] Similar resources here can refer to resources that belong to the same first-level vertical category.

[0052] Step 104: Recommend resources to the target user based on the target recommended resource list.

[0053] In this application, resources can be recommended to target users sequentially according to the order of resources in the target recommended resource list.

[0054] For example, the resource recommendation method of this application embodiment can be applied to scenarios such as video recommendation and product recommendation.

[0055] In this embodiment, the resource diversity preference of the target user is quantified based on the target user's historical resource interaction information to obtain a diversity preference score. Based on the target user's diversity preference score, the candidate recommended resource list is reordered so that the reordered video recommendation list meets the target user's personalized needs for resource diversity. Recommending based on the reordered video recommendation list can improve the accuracy of video recommendation.

[0056] Figure 2 This is a flowchart illustrating a resource recommendation method provided in another embodiment of this application.

[0057] like Figure 2 As shown, the resource recommendation method includes:

[0058] Step 201: Obtain the list of candidate recommended resources corresponding to the target user.

[0059] Step 202: Determine the diverse preference information of the target user based on the target user's historical resource interaction information.

[0060] In this application, steps 201-202 can be implemented in any of the embodiments of this application, so they will not be described in detail here.

[0061] Step 203: Adjust the baseline similar resource interval parameters based on the diversity preference information to obtain the target similar resource interval parameters.

[0062] Among them, the benchmark similar resource interval parameter refers to the benchmark interval parameter between similar resources in the resource recommendation list.

[0063] For example, the interval parameter may include the number of intervals, the interval distance, etc. Here, the number of intervals may refer to the number of resources in the interval between similar resources, and the interval distance may refer to the distance between similar resources, that is, the number of steps required to get from one resource in the resource recommendation list to its similar resources.

[0064] For example, a resource recommendation list includes resource a1, resource b1, and resource a2. Resource a1 and resource a2 are similar resources. Since resource b1 is between resource a1 and resource a2, the interval between resource a1 and resource a3 is 1. Since resource a1 goes to resource b1 first, and then from resource b1 to resource a2, the interval between resource a1 and resource a2 is 2.

[0065] In addition, the benchmark similar resource interval parameter can be set according to preset rules or determined by other methods, and this application does not limit it.

[0066] In this application, the parameter adjustment amount corresponding to the target user can be determined based on the diverse preference information of the target user, and the benchmark similar resource interval parameter can be adjusted according to the parameter adjustment amount to obtain the target similar resource interval parameter.

[0067] For example, the baseline similar resource interval parameter includes the number of baseline similar resource intervals. If the number of baseline similar resource intervals is 3 and the parameter adjustment is +1, then the number of target similar resource intervals is 3+1=4, which means that the number of baseline similar resource intervals is adjusted from 3 to 4.

[0068] For example, the current parameter adjustment amount for a target user can be determined based on the mapping relationship between diversity preference scores and parameter adjustment amounts.

[0069] For example, the lower the diversity preference score, the smaller the parameter adjustment amount; the lower the diversity preference level, the smaller the parameter adjustment amount.

[0070] It is understandable that in this application, different users may have different diversity preference information, and therefore the target similarity resource interval parameters of different users may also be different.

[0071] Step 204: Based on the target similarity resource interval parameter, reorder the candidate recommended resource list to obtain the target recommended resource list.

[0072] In this application, the positions of resources ranked lower in the similar resource list can be adjusted according to the target similar resource interval parameter so that the positions of the adjusted similar resources meet the requirements of the target similar resource interval parameter, thereby obtaining the target recommended resource list.

[0073] For example, if the interval between similar resources is 4, and the resources in the candidate recommended resource list are ordered from front to back as A1, A2, B1, C1, D1, E1, and resources A1 and A2 are similar resources, then the position of resource A2, which is ranked later in the list of resources A1 and A2, can be adjusted to be after resource E1, so that the interval between resources A1 and A2 is 4. The adjusted resource list is A1, B1, C1, D1, E1, A2.

[0074] For example, since there may be more than two similar resources in the candidate recommendation list, adjustments can be made for two similar resources each time to facilitate adjustments.

[0075] Step 205: Recommend resources to the target user based on the target recommended resource list.

[0076] In this application, step 205 can be implemented in any of the embodiments of this application, so it will not be described in detail here.

[0077] In this embodiment, the baseline similar resource interval parameter can be personalized based on diverse preference information to obtain a personalized target similar resource interval parameter. Then, based on the target similar resource interval parameter, the candidate recommended resource list can be reordered to achieve personalized rearrangement of the candidate recommended resource list, meeting the personalized needs of different users for resource diversity. Therefore, resource recommendation based on the rearranged recommended resource list can improve the accuracy of resource recommendation.

[0078] Since users' preferences for resource diversity may differ across historical periods, in order to improve the accuracy of diversity preference information, one embodiment of this application can combine resource interaction information from historical periods of different durations to determine diversity preference information, thereby reordering the candidate recommended resource list based on the diversity preference information. The following will combine... Figure 3 To explain, Figure 3 This is a flowchart illustrating a resource recommendation method provided in another embodiment of this application.

[0079] like Figure 3 As shown, the resource recommendation method includes:

[0080] Step 301: Obtain the list of candidate recommended resources corresponding to the target user.

[0081] In this application, step 301 can be implemented in any of the embodiments of this application, so it will not be described in detail here.

[0082] Step 302: Determine the first information gain corresponding to the first historical period based on the resource interaction information within the first historical period.

[0083] In this application, the aforementioned historical resource interaction information may include resource interaction information within a first historical period and resource interaction information within a second historical period, wherein the duration of the first historical period may be shorter than the duration of the second historical period.

[0084] For example, the duration of the first historical period can be less than the first duration threshold, the duration of the second historical period can be greater than the second duration threshold, and the first duration threshold is less than the second duration threshold.

[0085] For example, the first historical period is the past two days, and the second historical period is the past three months.

[0086] It should be noted that the first duration threshold and the second duration threshold can be determined according to actual needs, and this application does not impose any restrictions on them.

[0087] In this application, the first information gain corresponding to the first historical time period can be used to characterize the target user's preference for resource diversity within the first historical time period. For example, the larger the first information gain, the higher the target user's preference for diversity. If the resource distribution of each resource category accessed by the target user in the first historical time period is relatively uniform, the first information gain is larger; conversely, the first information gain is smaller.

[0088] In this application, based on Shannon entropy theory, diversity information entropy can be introduced to assess users' diverse preferences over a historical period.

[0089] In some embodiments, the first interest probability of a target user for each resource category can be determined based on resource interaction information within a first historical period, and the first information gain can be determined based on the first interest probability.

[0090] The probability of a target user's first interest in a resource category can refer to the probability that the target user is interested in that resource category in the first historical period.

[0091] For example, the interaction information of the target user with resources of each resource category can be determined based on the resource interaction information within the first historical period, and the first interest probability of the target user with resources of the same resource category can be determined based on the interaction information with resources of the same resource category.

[0092] For example, based on resource interaction information within the first historical period, a deep learning model can be used to determine the probability of a target user's first interest in a resource category.

[0093] As an example, the first information gain can be calculated using the following formula (1).

[0094]

[0095] Where short_entropy represents the first information gain, or short-term information gain; P(g|u) represents the probability of target user u’s first interest in resource category g, that is, the probability that target user u is interested in resource category g; G represents the set of resource categories.

[0096] Therefore, by determining the probability that a target user is interested in each resource category based on the resource interaction information within the first historical period, and by determining the information gain corresponding to the first historical period based on the probability of interest in each resource category, it is possible to quantify the diversity tendency of the target user in resource interaction information in the short term, thereby improving the accuracy of diversity preference information.

[0097] Step 303: Determine the second information gain corresponding to the second historical period based on the resource interaction information within the second historical period.

[0098] In this application, the second information gain corresponding to the second historical period can be used to characterize the target user's preference for resource diversity within the second historical period. For example, the larger the second information gain, the higher the target user's preference for diversity. If the resource distribution of each resource category accessed by the target user in the second historical period is relatively uniform, the second information gain is larger; conversely, the second information gain is smaller. For example, if the duration of the second historical period is greater than the second duration threshold, the second information gain can be referred to as the long-term information gain.

[0099] In some embodiments, the second interest probability of a target user for each resource category can be determined based on resource interaction information within a second historical period, and the second information gain can be determined based on the second interest probability.

[0100] The probability of a target user having a second interest in a resource category can refer to the probability that the target user is interested in that resource category in a second historical period.

[0101] For example, the interaction information of the target user with resources of each resource category can be determined based on the resource interaction information in the second historical period, and the second interest probability of the target user with resources of the same resource category can be determined based on the interaction information with resources of the same resource category.

[0102] For example, based on resource interaction information within a second historical period, a deep learning model can be used to determine the probability of a target user's second interest in a particular resource category.

[0103] For example, the second information gain can be calculated using formula (1) based on the target user's second interest probability for each resource category, which will not be elaborated here.

[0104] If a user's interaction with the top-ranked resource categories over a long period is relatively high, it indicates a lower preference for diversity. Therefore, to improve the accuracy of the second information gain, optionally, the initial information gain for the second historical period can be determined based on the target user's second interest probability for each resource category. The number of resources in each resource category within the resource interaction information during the second historical period can also be determined. The initial information gain can then be adjusted based on the number of resources in each category to obtain the second information gain. Thus, adjusting the initial information gain in conjunction with the number of resources in each resource category can improve the accuracy of the second information gain.

[0105] For example, the initial information gain corresponding to the second historical period can be determined by using the above formula (1) based on the second interest probability of the target user for each resource category.

[0106] For example, if the difference between the number of resources in each resource category is less than a preset quantity threshold, it indicates that the target user has a high preference for resource diversity. In this case, the initial information gain can be increased to obtain the second information gain.

[0107] For example, the resource categories can be sorted in descending order of the number of resources in each category to obtain a resource category ranking. The ratio between the sum of the number of resources in the first preset number of resource categories in the resource category ranking and the sum of the number of resources in each resource category can be determined. The initial information gain can then be adjusted based on this ratio to obtain a second information gain.

[0108] Therefore, by adjusting the initial information gain based on the user's preference for the resource categories ranked first, the accuracy of the second information gain can be further improved.

[0109] As an example, if the ratio is greater than a preset threshold, it indicates that the target user is more interested in resource categories ranked higher in the resource category ranking and has a lower inclination towards resource diversity. In this case, the initial information gain can be reduced to obtain a second information gain. The adjustment amount can be determined based on the difference between the ratio and the preset threshold; for example, the larger the difference, the larger the adjustment amount.

[0110] For example, if the total number of resources in the top 4 resource categories that interact with the target user in the resource category ranking accounts for more than 50%, the initial information gain can be reduced to obtain the second information gain.

[0111] As an example, if the ratio is less than or equal to a preset threshold, it indicates that the target user has a higher preference for resource diversity. In this case, the initial information gain can be increased to obtain a second information gain. The adjustment amount can be determined based on the difference between the preset threshold and the ratio; for example, the larger the difference, the larger the adjustment amount.

[0112] Therefore, by determining the probability that a target user is interested in each resource category based on the resource interaction information in the second historical period, and by determining the information gain corresponding to the second historical period based on the probability of interest in each resource category, it is possible to quantify the diversity tendency of the target user in resource interaction information in the long term, thereby improving the accuracy of diversity preference information.

[0113] Step 304: Determine the diversity preference score based on the first information gain and the second information gain.

[0114] In this application, the first information gain and the second information gain can be fused to obtain a diversity preference score.

[0115] For example, the diversity preference score can be obtained by multiplying the first information gain by the second information gain.

[0116] Step 305: Obtain diversity preference information based on diversity preference scores.

[0117] In some embodiments, diversity preference scores can be used as diversity preference information.

[0118] In some embodiments, diversity preference levels can be determined based on diversity preference scores, and then diversity preference scores, diversity preference levels, etc., can be used as diversity preference information.

[0119] Step 306: Based on the diversity preference information, reorder the candidate recommended resource list to obtain the target recommended resource list.

[0120] Step 307: Recommend resources to the target user based on the target recommended resource list.

[0121] In this application, steps 306-307 can be implemented in any of the embodiments of this application, so they will not be described in detail here.

[0122] In this embodiment, a first information gain is determined based on resource interaction information within a first historical period, and a second information gain is determined based on resource interaction information within a second historical period. Then, a diversity preference score is determined based on both the first and second information gains. Thus, a short-term information gain is determined based on short-term resource interaction information, and a long-term information gain is determined based on long-term resource interaction information. The diversity preference score is then determined based on both the short-term and long-term information gains. By considering users' preferences for resource diversity in both the short and long term, the accuracy of diversity preference information can be improved.

[0123] Figure 4 This is a flowchart illustrating a resource recommendation method provided in another embodiment of this application.

[0124] like Figure 4 As shown, the resource recommendation method includes:

[0125] Step 401: Obtain the list of candidate recommended resources corresponding to the target user.

[0126] Step 402: Determine the first information gain corresponding to the first historical period based on the resource interaction information within the first historical period.

[0127] Step 403: Determine the second information gain corresponding to the second historical period based on the resource interaction information within the second historical period.

[0128] In this application, steps 401-403 can be implemented in any of the embodiments of this application, so they will not be described in detail here.

[0129] Step 404: Determine the first resource interaction behavior parameters corresponding to each recommendation density based on historical resource interaction information.

[0130] Among them, the first resource interaction behavior parameter corresponding to the recommendation density is used to represent the interaction between the target user and the resource under the recommendation density.

[0131] The first resource interaction behavior parameter can include positive interaction behavior parameters, negative interaction behavior parameters, etc., so as to quantify the diversity preferences through positive and negative interaction behavior parameters, thereby improving the accuracy of diversity preference information.

[0132] In addition, the interaction parameters of the first resource may differ depending on the type of resource.

[0133] For example, if the resource is a video, then the first resource interaction behavior parameters may include, but are not limited to, the satisfactory playback rate, the fast scrolling rate, etc.

[0134] Among them, the satisfactory playback rate can refer to the proportion of users who watch the video completely (or reach a set percentage of the playback progress), reflecting the continued attraction of the video content to users. It can be seen that the satisfactory playback rate is a positive interactive behavior parameter. The fast swipe rate can refer to the proportion of users who quickly swipe to skip the video, reflecting that the content is not attractive enough in the first few seconds. The fast swipe rate is a negative interactive behavior parameter.

[0135] Recommendation density can be used to represent the density of similar resources among historical resource recommendations presented to a target user. Recommendation density can also be used to measure resource diversity; higher recommendation density corresponds to lower resource diversity, and vice versa.

[0136] For example, the distribution of similar resource intervals can be used to represent the recommendation density. For instance, the recommendation density can be divided as follows:

[0137]

[0138] Where x represents the number of similar video intervals, and the value of x is an integer. According to the above division, the recommendation density has three levels: high, medium and low.

[0139] It should be noted that the above-mentioned recommendation density division is only an example and should not be regarded as a limitation of this application. Furthermore, the maximum number of similar video intervals can be determined based on actual conditions, and this application does not impose any limitations on this.

[0140] Taking video recommendation as an example, for users who are not sensitive to resource diversity, high recommendation density has little impact on the satisfactory playback rate and the quick swipe rate. However, for users who are sensitive to resource diversity, high recommendation density will cause them to swipe quickly, increasing the quick swipe rate and decreasing the satisfactory playback rate. Therefore, the satisfactory playback rate and quick swipe rate of high recommendation density can be used to fit diversity preferences in order to reorder the candidate recommended resource list.

[0141] In this application, the historical resource interaction information may include resource interaction information within a third historical period. Based on the resource interaction information within the third historical period, the first resource interaction behavior parameters corresponding to each recommendation density within the third historical period can be determined.

[0142] For example, the third historical period can be shorter than the first historical period, so that the interaction between the target user and the resource under different recommendation intensities in the short term can be determined based on the resource interaction information in the short term.

[0143] To obtain more granular user behavior, for example, the first resource interaction behavior parameter of the target user can be determined based on the resource interaction information in the third historical time period, under the number of similar video intervals in the third historical time period.

[0144] Step 405: Determine the second resource interaction behavior parameters based on the weight parameters corresponding to each recommendation density and the first resource interaction behavior parameters.

[0145] In this application, the first resource interaction behavior parameter can be weighted according to the weight parameter corresponding to each recommendation density, and the second resource interaction behavior parameter can be determined according to the weighting result.

[0146] For example, the first resource interaction behavior parameter is weighted according to the weight parameter corresponding to each recommendation density to obtain a first sum value, and a second sum value is determined among the weight parameters corresponding to each recommendation density. Then, the second resource interaction behavior parameter is determined according to the ratio between the first sum value and the second sum value.

[0147] It can be seen that the second resource interaction behavior parameter can be the weighted average of the first resource interaction behavior parameters corresponding to each recommendation density.

[0148] The weight parameters corresponding to each recommendation density can be determined through grid search or other methods, and this application does not limit this.

[0149] As an example, the second resource interaction behavior parameter can be calculated using the following formula (2):

[0150]

[0151] Among them, P avg P(x) represents the second resource interaction behavior parameter; P(x) represents the first resource interaction behavior parameter when the number of similar video intervals is x; L represents the maximum number of similar video intervals; w x This represents the weight parameter corresponding to the number of similar video intervals, x.

[0152] Taking video recommendation as an example, the satisfaction rate, which has a relatively low recommendation density, has a greater influence. Therefore, high relevance can be ensured while maximizing the satisfaction rate.

[0153] Therefore, based on the weight parameters corresponding to each recommendation density, the second resource interaction behavior parameter can be determined by weighted averaging of the first resource interaction behavior parameters under each recommendation density, which can improve the accuracy of the second resource interaction behavior parameter.

[0154] Step 406: Determine the diversity preference score based on the first information gain, the second information gain, and the second resource interaction behavior parameter.

[0155] In some embodiments, the first information gain, the second information gain, and the second resource interaction behavior parameters can be fused to obtain a diversity preference score.

[0156] For example, the first information gain, the second information gain, and the second resource interaction behavior parameter can be multiplied together to obtain the diversity preference score.

[0157] For example, in video recommendation, the diversity preference score can be calculated using the following formula (3):

[0158] diversity_q=s avg *k avg *short_entropy*long_entropy (3)

[0159] Where diversity_q represents the diversity preference score; s avgk represents the weighted average of the satisfactory playback rate. avg represents the weighted average of the fast slip rate; short_entropy represents the first information gain; long_entropy represents the second information gain. Where s avg and k avg It can be calculated using the formula (2) above.

[0160] For example, the diversity preference score can be obtained by multiplying the Nth power of the first information gain, the Nth power of the second information gain, and the power of the second resource interaction behavior parameter. Here, N can be a negative integer less than 0, such as N = -2.

[0161] Therefore, by fusing the first information gain, the second information gain, and the second resource interaction behavior parameters, a diversity preference score is obtained. Since the diversity information gain in historical periods of different durations and the resource interaction behavior parameters of users under different recommendation densities are taken into account, the accuracy of diversity preference quantification can be improved.

[0162] Step 407: Obtain diversity preference information based on diversity preference scores.

[0163] Step 408: Based on the diversity preference information, the candidate recommended resource list is reordered to obtain the target recommended resource list.

[0164] Step 409: Recommend resources to the target user based on the target recommended resource list.

[0165] In this application, steps 407-409 can be implemented in any of the embodiments of this application, so they will not be described in detail here.

[0166] In this embodiment, by determining the first resource interaction behavior parameter corresponding to each recommendation density based on historical resource interaction information, and by determining the second resource interaction behavior parameter based on the weight parameter corresponding to each recommendation density and the first resource interaction behavior parameter, the diversity preference score can be determined based on the information gain corresponding to multiple different historical periods and the second resource interaction behavior parameter, thereby improving the accuracy of the diversity preference score.

[0167] To facilitate understanding of the resource recommendation method in this application, the following will combine... Figure 5 To explain, Figure 5 This is a schematic diagram of a resource recommendation process provided in an embodiment of this application.

[0168] like Figure 5As shown, when recommending resources, resources can be recalled first, and then the recalled resources can be coarsely ranked and finely ranked to obtain the fine ranking result. Based on the user's historical resource interaction information, a diversity preference score is obtained. Based on the diversity preference score, the baseline similar resource interval parameter is adjusted to obtain a personalized similar resource interval parameter. Using the personalized similar resource interval parameter, the fine ranking result is reordered to obtain a distribution list. Then, resources are distributed to users and displayed according to the distribution list.

[0169] Here, the personalized similar resource interval parameter can be understood as the target similar resource interval parameter mentioned above; the distribution list can be understood as the target recommended resource list in the above embodiment. Furthermore, the specific process of obtaining the diversity preference score based on the user's historical resource interaction information can be found in the above embodiment, and therefore will not be repeated here.

[0170] To implement the above embodiments, this application also proposes a resource recommendation device. Figure 6 This is a schematic diagram of the structure of a resource recommendation device provided in an embodiment of this application.

[0171] like Figure 6 As shown, the resource recommendation device 600 includes:

[0172] The acquisition module 610 is used to acquire the list of candidate recommended resources corresponding to the target user;

[0173] The determining module 620 is used to determine the diversity preference information of the target user based on the target user's historical resource interaction information; wherein, the diversity preference information refers to the target user's preference information for resource diversity;

[0174] The reordering module 630 is used to reorder the candidate recommended resource list according to the diversity preference information to obtain the target recommended resource list;

[0175] The recommendation module 640 is used to recommend resources to the target user based on the target recommended resource list.

[0176] Optionally, the reordering module 630 is used for:

[0177] Based on the aforementioned diversity preference information, the baseline similar resource interval parameters are adjusted to obtain the target similar resource interval parameters;

[0178] Based on the target similarity resource interval parameter, the candidate recommended resource list is reordered to obtain the target recommended resource list.

[0179] Optionally, the historical resource interaction information includes resource interaction information within a first historical period and resource interaction information within a second historical period, wherein the duration of the first historical period is shorter than the duration of the second historical period; the determining module 620 is used for:

[0180] Based on the resource interaction information within the first historical period, a first information gain corresponding to the first historical period is determined; wherein, the first information gain is used to characterize the target user's tendency toward resource diversity within the first historical period;

[0181] Based on the resource interaction information within the second historical period, a second information gain corresponding to the second historical period is determined; wherein, the second information gain is used to characterize the target user's tendency toward resource diversity within the second historical period;

[0182] The diversity preference score is determined based on the first information gain and the second information gain; wherein the diversity preference score is used to characterize the degree of preference of the target user for resource diversity;

[0183] Based on the diversity preference score, obtain the diversity preference information.

[0184] Optionally, module 620 is defined for:

[0185] Based on the historical resource interaction information, a first resource interaction behavior parameter corresponding to each recommendation density is determined; wherein, the recommendation density is used to represent the density between similar resources in the historical resource recommendation list recommended to the target user;

[0186] The second resource interaction behavior parameters are determined based on the weight parameters corresponding to each recommendation density and the first resource interaction behavior parameters.

[0187] The diversity preference score is determined based on the first information gain, the second information gain, and the second resource interaction behavior parameter.

[0188] Optionally, module 620 is defined for:

[0189] The first resource interaction behavior parameter is weighted according to the weight parameter to obtain the first sum value;

[0190] Determine the second sum value among the weight parameters corresponding to each recommendation density;

[0191] The second resource interaction behavior parameter is determined based on the ratio between the first sum and the second sum.

[0192] Optionally, module 620 is defined for:

[0193] The diversity preference score is obtained by fusing the first information gain, the second information gain, and the second resource interaction behavior parameter.

[0194] Optionally, module 620 is defined for:

[0195] Based on the resource interaction information within the first historical period, determine the first interest probability of the target user for each resource category;

[0196] The first information gain is determined based on the first interest probability.

[0197] Optionally, module 620 is defined for:

[0198] Based on the resource interaction information within the second historical period, determine the second interest probability of the target user for each resource category;

[0199] The second information gain is determined based on the second interest probability.

[0200] Optionally, module 620 is defined for:

[0201] Based on the second interest probability, determine the initial information gain corresponding to the second historical period;

[0202] Determine the number of resources for each resource category in the resource interaction information within the second historical period;

[0203] The initial information gain is adjusted based on the number of resources in each resource category to obtain the second information gain.

[0204] Optionally, module 620 is defined for:

[0205] The resource categories are sorted in descending order of resource quantity to obtain a resource category ranking.

[0206] Determine the ratio between the sum of the resources in the first preset number of resource categories in the resource category sorting and the sum of the resources in each resource category;

[0207] The initial information gain is adjusted based on the ratio to obtain the second information gain.

[0208] It should be noted that the explanation of the above-mentioned resource recommendation method embodiment also applies to the resource recommendation device of this embodiment, so it will not be repeated here.

[0209] In this embodiment, the resource diversity preference of the target user is quantified based on the target user's historical resource interaction information to obtain a diversity preference score. Based on the target user's diversity preference score, the candidate recommended resource list is reordered so that the reordered video recommendation list meets the target user's personalized needs for resource diversity. Recommending based on the reordered video recommendation list can improve the accuracy of video recommendation.

[0210] According to embodiments of this application, this application also provides an electronic device, a readable storage medium, and a computer program product.

[0211] Figure 7 A schematic block diagram of an example electronic device 700 that can be used to implement embodiments of this application is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the application described and / or claimed herein.

[0212] like Figure 7 As shown, device 700 includes a computing unit 701, which can perform various appropriate actions and processes based on a computer program stored in ROM (Read-Only Memory) 702 or a computer program loaded from storage unit 708 into RAM (Random Access Memory) 703. RAM 703 can also store various programs and data required for the operation of device 700. The computing unit 701, ROM 702, and RAM 703 are interconnected via bus 704. I / O (Input / Output) interface 705 is also connected to bus 704.

[0213] Multiple components in device 700 are connected to I / O interface 705, including: input unit 706, such as keyboard, mouse, etc.; output unit 707, such as various types of monitors, speakers, etc.; storage unit 708, such as disk, optical disk, etc.; and communication unit 709, such as network card, modem, wireless transceiver, etc. Communication unit 709 allows device 700 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0214] The computing unit 701 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, CPUs (Central Processing Units), GPUs (Graphics Processing Units), various special-purpose AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, DSPs (Digital Signal Processors), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as resource recommendation methods. For example, in some embodiments, the resource recommendation method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on device 700 via ROM 702 and / or communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the resource recommendation method described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform a resource recommendation method by any other suitable means (e.g., by means of firmware).

[0215] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application-Specific Standard Products), SOCs (System-on-Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0216] The program code used to implement the methods of this application may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0217] In the context of this application, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, RAM, ROM, EPROM (Electrically Programmable Read-Only Memory) or flash memory, optical fiber, CD-ROM (Compact Disc Read-Only Memory), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0218] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0219] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include LANs (Local Area Networks), WANs (Wide Area Networks), the Internet, and blockchain networks.

[0220] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service system that addresses the shortcomings of traditional physical hosts and VPS (Virtual Private Server) services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers integrated with blockchain technology.

[0221] According to embodiments of this application, this application also provides a computer program product that, when an instruction processor in the computer program product is executed, performs the resource recommendation method proposed in the above embodiments of this application.

[0222] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this application can be achieved, and this is not limited herein.

[0223] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A resource recommendation method, comprising: Obtain the list of candidate recommended resources for the target user; Based on the target user's historical resource interaction information, the target user's diversity preference information is determined; wherein, the diversity preference information refers to the target user's preference information for resource diversity. Based on the diverse preference information, the candidate recommended resource list is reordered to obtain the target recommended resource list; Based on the target recommended resource list, resources are recommended to the target user.

2. The method as described in claim 1, wherein, The step of reordering the candidate recommended resource list based on the diverse preference information to obtain the target recommended resource list includes: Based on the aforementioned diversity preference information, the baseline similar resource interval parameters are adjusted to obtain the target similar resource interval parameters; Based on the target similarity resource interval parameter, the candidate recommended resource list is reordered to obtain the target recommended resource list.

3. The method as described in claim 1, wherein, The historical resource interaction information includes resource interaction information within a first historical period and resource interaction information within a second historical period, wherein the duration of the first historical period is shorter than the duration of the second historical period. The step of determining the diverse preference information of the target user based on the target user's historical resource interaction information includes: Based on the resource interaction information within the first historical period, a first information gain corresponding to the first historical period is determined; wherein, the first information gain is used to characterize the target user's tendency toward resource diversity within the first historical period; Based on the resource interaction information within the second historical period, a second information gain corresponding to the second historical period is determined; wherein, the second information gain is used to characterize the target user's tendency toward resource diversity within the second historical period; The diversity preference score is determined based on the first information gain and the second information gain; wherein the diversity preference score is used to characterize the degree of preference of the target user for resource diversity; Based on the diversity preference score, obtain the diversity preference information.

4. The method of claim 3, wherein, Determining the diversity preference score based on the first information gain and the second information gain includes: Based on the historical resource interaction information, a first resource interaction behavior parameter corresponding to each recommendation density is determined; wherein, the recommendation density is used to represent the density between similar resources in the historical resource recommendation list recommended to the target user; The second resource interaction behavior parameters are determined based on the weight parameters corresponding to each recommendation density and the first resource interaction behavior parameters. The diversity preference score is determined based on the first information gain, the second information gain, and the second resource interaction behavior parameter.

5. The method of claim 4, wherein, The step of determining the second resource interaction behavior parameter based on the weight parameters corresponding to each recommendation density and the first resource interaction behavior parameter includes: The first resource interaction behavior parameter is weighted according to the weight parameter to obtain the first sum value; Determine the second sum value among the weight parameters corresponding to each recommendation density; The second resource interaction behavior parameter is determined based on the ratio between the first sum and the second sum.

6. The method of claim 4, wherein, The step of determining the diversity preference score based on the first information gain, the second information gain, and the second resource interaction behavior parameter includes: The diversity preference score is obtained by fusing the first information gain, the second information gain, and the second resource interaction behavior parameter.

7. The method of claim 3, wherein, The step of determining the first information gain corresponding to the first historical period based on the resource interaction information within the first historical period includes: Based on the resource interaction information within the first historical period, determine the first interest probability of the target user for each resource category; The first information gain is determined based on the first interest probability.

8. The method of claim 3, wherein, The step of determining the second information gain corresponding to the second historical period based on the resource interaction information within the second historical period includes: Based on the resource interaction information within the second historical period, determine the second interest probability of the target user for each resource category; The second information gain is determined based on the second interest probability.

9. The method of claim 8, wherein, Determining the second information gain based on the second interest probability includes: Based on the second interest probability, determine the initial information gain corresponding to the second historical period; Determine the number of resources for each resource category in the resource interaction information within the second historical period; The initial information gain is adjusted based on the number of resources in each resource category to obtain the second information gain.

10. The method of claim 9, wherein, The step of adjusting the initial information gain based on the resource quantity of each resource category to obtain the second information gain includes: The resource categories are sorted in descending order of resource quantity to obtain a resource category ranking. Determine the ratio between the sum of the resources in the first preset number of resource categories in the resource category sorting and the sum of the resources in each resource category; The initial information gain is adjusted based on the ratio to obtain the second information gain.

11. A resource recommendation device, comprising: The acquisition module is used to obtain a list of candidate recommended resources corresponding to the target user; The determining module is used to determine the diversity preference information of the target user based on the target user's historical resource interaction information; wherein, the diversity preference information refers to the target user's preference information for resource diversity; The reordering module is used to reorder the candidate recommended resource list according to the diversity preference information to obtain the target recommended resource list; The recommendation module is used to recommend resources to the target user based on the target recommended resource list.

12. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-10.

13. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-10.

14. A computer program product comprising a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1-10.

Images