Method of determining a recommendation list, information recommendation method and apparatus

Abstract

The present disclosure provides a method for determining a recommendation list, relates to the technical field of computers, and particularly relates to the fields of artificial intelligence and big data. The specific implementation scheme is as follows: obtaining interest resources recalled based on user interests and non-interest resources recalled based on business types, the interest resources and the non-interest resources being provided with recommendation degrees; and distributing the interest resources and the non-interest resources to multiple slots of the recommendation list according to the business types, the recommendation degrees, and a slot control threshold value for controlling the number of non-interest resources. The present disclosure also provides an information recommendation method and device, an electronic device, and a storage medium.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, in particular to the technical field of artificial intelligence and big data. More specifically, the present disclosure provides a method for determining a recommendation list, an information recommendation method, an apparatus, an electronic device and a storage medium. BACKGROUND

[0002] A recommendation system is a tool that helps users quickly find useful information. For example, under the condition that the user's demand is not very clear, the recommendation system can provide high-quality personalized recommendation services for users by using various historical information of the users. SUMMARY

[0003] The present disclosure provides a method for determining a recommendation list, an information recommendation method, an apparatus, an electronic device and a storage medium.

[0004] According to a first aspect, a method for determining a recommendation list is provided, which comprises: obtaining interest resources recalled based on user interest and non-interest resources recalled based on business type, the interest resources and the non-interest resources being provided with a recommendation degree; and allocating the interest resources and the non-interest resources to a plurality of slots of the recommendation list according to the business type, the recommendation degree and a slot control threshold for controlling the number of the non-interest resources in the slots.

[0005] According to a second aspect, an information recommendation method is provided, which comprises: obtaining a recommendation list for a user, wherein the recommendation list comprises a plurality of slots arranged in sequence, and each slot is allocated with a resource containing recommendation information; and recommending the resource in each slot to the user in sequence, wherein the recommendation list is determined according to the method for determining a recommendation list.

[0006] According to a third aspect, an apparatus for determining a recommendation list is provided, which comprises: a first obtaining module configured to obtain interest resources recalled based on user interest and non-interest resources recalled based on business type, the interest resources and the non-interest resources being provided with a recommendation degree; and a determining module configured to allocate the interest resources and the non-interest resources to a plurality of slots of the recommendation list according to the business type, the recommendation degree and a slot control threshold for controlling the number of the non-interest resources in the slots.

[0007] According to a fourth aspect, an information recommendation apparatus is provided, which comprises: a second obtaining module configured to obtain a recommendation list for a user, wherein the recommendation list comprises a plurality of slots arranged in sequence, and each slot is allocated with a resource containing recommendation information; and a recommending module configured to recommend the resource in each slot to the user in sequence, wherein the recommendation list is obtained by training the apparatus for determining a recommendation list.

[0008] According to a fifth aspect, an electronic device is provided, comprising: at least one processor; and a memory connected with the at least one processor in communication; wherein 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 provided by the present disclosure.

[0009] According to a sixth aspect, a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the method provided by the present disclosure is provided.

[0010] According to a seventh aspect, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the method provided by the present disclosure.

[0011] 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 disclosure, nor to limit the scope of the present disclosure. Other features of the present disclosure will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0012] The accompanying drawings are used to better understand the present scheme and do not limit the present disclosure. Among them:

[0013] Figure 1 is an exemplary system architecture schematic diagram to which the method of determining a recommendation list and the information recommendation method can be applied according to an embodiment of the present disclosure;

[0014] Figure 2 is a flowchart of the method of determining a recommendation list according to an embodiment of the present disclosure;

[0015] Figure 3A is a schematic diagram of the method of determining a recommendation list according to an embodiment of the present disclosure;

[0016] Figure 3B is a schematic diagram of a recommendation list according to an embodiment of the present disclosure;

[0017] Figure 4A is a flowchart of the method of determining a recommendation list according to an embodiment of the present disclosure;

[0018] Figure 4B is a schematic diagram of a competition interval according to an embodiment of the present disclosure;

[0019] Figure 4C is a schematic diagram of the method of randomly allocating a slot according to an embodiment of the present disclosure;

[0020] Figure 5 is a flowchart of the information recommendation method according to an embodiment of the present disclosure;

[0021] Figure 6 is a block diagram of a device for determining a recommendation list according to an embodiment of the present disclosure;

[0022] Figure 7 is a block diagram of an information recommendation device according to an embodiment of the present disclosure;

[0023] Figure 8 is a block diagram of an electronic device for a method of determining a recommendation list and / or an information recommendation method according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] Exemplary embodiments of the present disclosure are described herein with reference to the accompanying drawings, which are presented for the purpose of illustration and description. It is to be understood that the embodiments described herein are exemplary only, and various changes and modifications can be made thereto without departing from the scope and spirit of the present disclosure. Also, the descriptions of known functions and constructions are omitted for clarity and conciseness.

[0025] In a recommendation system, the to-be-recommended resources for recommendation to a user can be selected from a large amount of resources using a recall model. For example, the recall model can recall candidate resources (which can be referred to as interest resources) based on user interest (for example, the user interest is determined according to user historical browsing records, etc.), and the candidate resources are ranked and presented to the user in the form of a recommendation list.

[0026] In actual application scenarios, some resources are not recalled based on user interest, but are recalled due to some business needs (which can be referred to as non-interest resources). The business needs can include hot push needs, such as major hot events currently occurring, etc., which need to be recommended to the general public.

[0027] Since the non-interest resources have low relevance to the user, if ranking is performed according to the relevance, the non-interest resources are difficult to be ranked into the recommendation list, resulting in that some non-interest resources containing important information or hot information cannot be recommended to the user, and causing poor recommendation effect.

[0028] In the technical solutions of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in the technical solutions comply with relevant laws and regulations, and do not violate public order and good customs.

[0029] In the technical solutions of the present disclosure, the authorization or consent of the user is obtained before the user personal information is acquired or collected.

[0030] Figure 1 is an exemplary system architecture schematic diagram according to an embodiment of the present disclosure, which can apply the method of determining a recommendation list and the information recommendation method. It should be noted that,Figure 1 The system architecture shown is only an example of a system architecture to which the embodiments of the present disclosure can be applied, to help those skilled in the art understand the technical content of the present disclosure, but does not mean that the embodiments of the present disclosure cannot be used in other devices, systems, environments or scenarios.

[0031] As shown in Figure 1 The system architecture 100 according to the embodiment can include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is a medium for providing a communication link between the terminal devices 101, 102, 103 and the server 105. The network 104 can include various connection types, such as wired and / or wireless communication links, etc.

[0032] A user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. The terminal devices 101, 102, 103 can be various electronic devices, including but not limited to smart phones, tablet computers, laptop computers, etc.

[0033] At least one of the method for determining a recommendation list and the information recommendation method provided by the embodiments of the present disclosure can generally be executed by the server 105. Accordingly, the apparatus for determining a recommendation list and the information recommendation apparatus provided by the embodiments of the present disclosure can generally be provided in the server 105. The method for determining a recommendation list and the information recommendation method provided by the embodiments of the present disclosure can also be executed by a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and / or the server 105. Accordingly, the apparatus for determining a recommendation list and the information recommendation apparatus provided by the embodiments of the present disclosure can also be provided in a server or a server cluster different from the server 105 and capable of communicating with the terminal devices 101, 102, 103 and / or the server 105.

[0034] Figure 2 is a flowchart of a method for determining a recommendation list according to an embodiment of the present disclosure.

[0035] As shown in Figure 2 The method 200 for determining a recommendation list can include operations S210-S220.

[0036] At operation S210, interest resources recalled based on user interest and non-interest resources recalled based on business type are obtained.

[0037] For example, the recommendation list can include a plurality of slots (e.g., 14), the interest resources and the non-interest resources can be allocated into the plurality of slots, and the recommendation list to which the resources are allocated can be presented to the user.

[0038] For example, the interest resource can be selected from a resource library based on the user feature using a recall model (e.g., a rank model). The resources in the resource library can include various news, messages, videos, products, etc. in the Internet. Each resource in the resource library has its own feature label, such as the keyword, name, etc. of the resource. The user feature can be generated based on the historical operation behavior (e.g., click, browse, forward, etc.) of the user.

[0039] The similarity between the user feature and the information feature of the resource in the resource library can be evaluated using the recall model. It can be understood that the higher the similarity evaluation value, the higher the relevance of the resource to the user, the higher the interest degree of the user, and the higher the recommendation degree. Therefore, the similarity evaluation value of each resource can be used as the recommendation degree of the resource. The top N (e.g., N = 10) resources with the highest recommendation degree can be determined as the interest resource to be recommended.

[0040] For example, the non-interest resource recalled based on the business type can include a non-interest resource recalled based on the business demand of multiple business types. The business demand can include a hot spot push demand, an application promotion demand, a live streaming flow demand, and a demand to show some major events to all users. According to the business type, the business demand can include a strong insertion type, a hot spot type, and a promotion type, and correspondingly, the business resource can also include a strong insertion resource, a hot spot resource, and a promotion resource. The strong insertion resource is, for example, a resource including major event information, which has a relatively high priority and needs to be shown to all users. The hot spot type is, for example, a resource including hot spot event information, which has a relatively high attention and can be shown to most users. The promotion resource is, for example, a resource including application promotion information, which can be recommended to some specific users.

[0041] For example, for the business demand of the hot spot type, in response to the recall request of the business demand of the hot spot type, the recommendation degree of the hot spot resource can be determined based on the information feature (e.g., click volume, forwarding volume, etc.) of the hot spot resource using the recall model. The top several (e.g., 5) with the highest recommendation degree can be determined as the hot spot resource to be recommended.

[0042] For example, for the business demand of the promotion type, in response to the recall request of the business demand of the promotion type, the similarity between the user feature and the information feature of the promotion resource can be calculated as the recommendation degree of the promotion resource using the recall model. The top several (e.g., 3) with the highest recommendation degree can be determined as the promotion resource to be recommended.

[0043] It can be understood that the specific manner of determining the recommendation degree of the hotspot resource is different from that of determining the recommendation degree of the interest resource, and therefore the hotspot resource and the interest resource cannot be directly sorted in a unified manner according to the recommendation degree. The recommendation degree of the promotion resource is usually much lower than that of the interest resource, and if the promotion resource is distributed according to the arrangement order from high to low of the recommendation degree, the promotion resource is difficult to be distributed to the recommendation list.

[0044] In operation S220, the interest resource and the non-interest resource are distributed to the plurality of slots of the recommendation list according to the service type, the recommendation degree, and the slot control threshold.

[0045] For example, the slot control threshold can be used to determine the number of slots occupied by each type of non-interest resource in the recommendation list. For example, for the strong insertion resource, the slot control threshold can be set to 10%, and it can be determined that the number of slots occupied by the strong insertion resource is 1 (14*10% = 1.4, and the number of slots should be an integer 1). For the hotspot resource, the slot control threshold can be set to 40%, and it can be determined that the number of slots occupied by the hotspot resource is 4. For the promotion resource, the slot control threshold can be set to 10%, and it can be determined that the number of slots occupied by the promotion resource is 1.

[0046] It can be understood that the non-interest resource can be finely divided into a plurality of types of resources according to the service type, and different strategies can be adopted for recommending different types of resources. By setting the slot control threshold, the number of slots occupied by each type of non-interest resource can be limited, and the recommendation effect of the interest resource can be avoided from being affected by too many non-interest resources.

[0047] For example, the number of slots occupied by the strong insertion resource is 1, and the strong insertion resource with the highest recommendation degree can be distributed to a specified slot (for example, a top slot) of the recommendation list.

[0048] For example, the number of slots occupied by the hotspot resource is 4. The 4 hotspot resources can be distributed to 4 random slots of the recommendation list except the specified slot by using a random distribution manner.

[0049] For example, the number of slots occupied by the promotion resource is 1, and the recommendation degree of the target promotion resource can be increased by a weight, for example, multiplied by a weight greater than 1 to obtain a new recommendation degree. According to the comparison of the new recommendation degree and the recommendation degree of the interest resource, the interest resource and the target promotion resource are sorted according to the comparison result, and the interest resource and the target promotion resource are distributed to the remaining slots of the recommendation list except the specified slot and the random slot according to the sorting result.

[0050] According to the embodiments of the present disclosure, corresponding recommendation strategies are adopted for different types of resources, and the number of slots occupied by various types of resources is limited, so that the recommendation rules can be refined and the recommendation effect can be improved.

[0051] The following will be described in combination with Figures 3A-3B, 4A-4C illustrate a method for determining a recommended list according to an embodiment of the present disclosure.

[0052] Figure 3A is a schematic diagram of a method for determining a recommended list according to an embodiment of the present disclosure.

[0053] As shown in Figure 3A , different allocation strategies can be adopted for allocating slots for strong insertion resources 310, hot spot resources 320 and promoted resources 330. For example, a specified position allocation strategy 311 can be adopted for strong insertion resources 310, a random allocation strategy 321 can be adopted for hot spot resources 320, and a weighted ranking allocation strategy 331 can be adopted for promoted resources 330. The allocation results of various resources can be fused and ranked to obtain a ranked recommended list 340.

[0054] For example, the recommended list 340 has 10 slots, and the number of slots occupied by strong insertion resources 310, hot spot resources 320 and promoted resources 330 can be 1, 4 and 2 respectively according to the slot control threshold. The specified position allocation strategy 311 can be adopted to allocate 1 strong insertion resource to the top slot (the first slot) of the recommended list 340. The random allocation strategy 321 can be adopted to allocate 4 hot spot resources 320 to 4 random slots (e.g. the 2nd, 4th, 8th and 10th slots) except the top slot. The weighted ranking allocation strategy 331 can be adopted to multiply the recommendation degrees of the 2 promoted resources by a weight greater than 1 respectively to obtain new recommendation degrees. According to the new recommendation degrees, the 2 promoted resources are ranked together with the interest resources to obtain a ranking result of the promoted resources and the interest resources, and the ranking result is fused into the remaining slots of the recommended list 340 except the specified slots and the random slots, so as to obtain the ranked recommended list 340.

[0055] Figure 3B is a schematic diagram of a recommended list according to an embodiment of the present disclosure.

[0056] As shown in Figure 3B , the recommended list 340 can be determined according to the method for determining a recommended list as shown in Figure 3A . The strong insertion resource is located at the top slot. The first hot spot resource to the fourth hot spot resource are randomly allocated to the 2nd, 4th, 8th and 10th slots respectively. The first interest resource to the fourth interest resource and the first promoted resource are allocated to the 3rd, 5th, 6th, 7th and 9th slots in the order. It can be understood that the new recommendation degree of the first promoted resource is between the recommendation degrees of the third interest resource and the fourth interest resource. The promoted resource with the increased weight also includes the second promoted resource, but since the new recommendation degree of the second promoted resource is less than the recommendation degree of the fourth interest resource, the recommended list 340 has limited slots, and therefore the second promoted resource is not allocated to a slot.

[0057] Figure 4A is a flowchart of a method of determining a recommendation list according to one embodiment of the present disclosure.

[0058] The embodiment can be a specific implementation of a random allocation strategy for hotspot resources. The random allocation strategy for hotspot resources can be implemented using a roulette algorithm, and the specific implementation is described with reference to operations S421 to operation S426 as shown in FIG. 4. Figure 4A

[0059] In operation S421, for each hotspot resource queue, a competition winning probability of the hotspot resource queue is determined according to the number of hotspot resources in the hotspot resource queue.

[0060] For example, the recall model can recall multiple hotspot resource queues (which can be referred to as queues hereinafter) according to a multi-path recall strategy (for example, according to a strategy of the latest in time, the highest in heat, etc.). The multiple hotspot resource queues can include queue 1, queue 2, queue 3, and the like. Each hotspot resource queue can include multiple hotspot resources, or can be empty due to no recalled resources.

[0061] For example, for queue 1 (queue1), the competition winning probability of queue 1 can be calculated according to the following formula (1).

[0062] p_queue1 = queue1_num / (queue1_num + queue2_num +... + queueN_num) (1)

[0063] Wherein, p_queue1 represents the competition winning probability of queue 1 (queue1), queue1_num represents the number of resources in queue 1, queue2_num represents the number of resources in queue 2, and queueN_num represents the number of resources recalled by the Nth queue.

[0064] Similarly, for queue 2, queue 3,..., the competition winning probability of each can be calculated. It can be understood that among the multiple hotspot resource queues, the greater the competition winning probability of the hotspot resource queue, the greater the probability of being selected as a target queue for resource allocation.

[0065] In operation S422, a competition interval of each hotspot resource queue is determined according to the competition winning probability.

[0066] For example, the multiple hotspot resource queues have an arrangement order. For example, queue 1, queue 2,..., queue N. According to the competition winning probability of the multiple queues, the competition interval of each queue can be determined.

[0067] ​For example, there are four queues in total, the competition winning probability of queue 1 is 0.3, the competition winning probability of queue 2 is 0.2, the competition winning probability of queue 3 is 0.3, and the competition winning probability of queue 4 is 0.2. The competition interval of each queue can be determined according to the cumulative probability. For example, the competition interval of queue 1 is [0, 0.3], the competition interval of queue 2 is [0.3, 0.5] (0.5 is the cumulative probability of queue 1 and queue 2), the competition interval of queue 3 is [0.5, 0.8] (0.8 is the cumulative probability of queue 1, queue 2 and queue 3), and the competition interval of queue 4 is [0.8, 1] (1 is the cumulative probability of the four queues).

[0068] In operation S423, a random slot is determined from the first remaining slot according to the generated random number.

[0069] For example, a random number in the range of [0, 1] is generated, and the position of the random slot can be determined according to the product of the random number and the number of slots of the recommendation list. For example, the recommendation list has 10 slots, and the generated random number is 0.2, and the position of the random slot can be determined as the second position (10*0.2=2). If the second slot is already occupied, a new random number can be generated, for example, the new random number is 0.5, and the position of the random slot can be determined as the fifth position. If the fifth position is also occupied, the above steps can be repeated until the position of the random slot is determined, or the number of repetitions reaches an upper limit (for example, 10 times), and it can be determined that there is no position allocated to the hot resource.

[0070] In operation S424, a target queue is determined from the plurality of hot resource queues according to the interval range in which the random number falls.

[0071] For example, for the generated random number, if the position of the random slot can be allocated according to the random number (for example, the random number is 0.2, and the random slot is determined to be the second position, and the second position is not occupied), the target queue can be determined according to the competition interval to which the random number belongs. For example, the random number 0.2 belongs to the competition interval [0, 0.3] of queue 1, and queue 1 can be determined as the target queue, and the target resource can be selected from queue 1 and allocated to the random slot (the second position).

[0072] It can be understood that operation S424 can be a step of competition of each queue, and the empty queue (without recalled resources) can not participate in the competition, because even if a position is allocated to the empty queue, there is no resource to fill.

[0073] In operation S425, a target hot resource meeting a preset condition is determined from the target queue and allocated to the random slot.

[0074] For example, if queue 1 is the target queue, the resource at the first position in the queue can be selected as the target resource. However, if the resource at the first position does not meet the diversity constraints (e.g., the assigned hot resources cannot be two consecutive short videos), the resource at the second position can be evaluated to determine whether it meets the constraints, until a target resource that meets the constraints is determined and then the target resource is assigned to the random slot determined by operation S423.

[0075] In operation S426, determine whether the number of digits occupied by the hot resource has reached the upper limit. If so, the process ends; otherwise, return to operation S423 for a new random number, until the number of digits occupied by the hot resource reaches the upper limit.

[0076] For example, based on the slot control threshold, the number of slots occupied by hot resources can be determined (e.g., 4). If the number of currently allocated hot resources is less than 4, a new random number can be generated. For the new random number, operation S423 is returned until all 4 random slots are allocated hot resources.

[0077] For example, during the return operation S423, and the repeated execution of the above operations S423 to S426, each queue can be set with a maximum number of slots. For example, the maximum number of slots for queue 1 is 2. It can participate in the above competition to occupy a random slot first. During the repeated execution of the above operations, it can participate in the above competition to occupy another random slot. After the slots are occupied to the maximum limit, queue 1 can no longer participate in the above competition.

[0078] Figure 4B This is a schematic diagram of a competition zone according to an embodiment of the present disclosure.

[0079] For example, there are four queues. The probability of winning the competition for queue 1 is 0.3, for queue 2 it is 0.2, for queue 3 it is 0.3, and for queue 4 it is 0.2. The competition interval for each queue can be determined by summing the probabilities.

[0080] like Figure 4B As shown, interval 401 can be the competition interval for queue 1, with a range of [0, 0.3]. Interval 402 can be the competition interval for queue 2, with a range of [0.3, 0.5]. Interval 403 can be the competition interval for queue 3, with a range of [0.5, 0.8]. Interval 404 can be the competition interval for queue 4, with a range of [0.8, 1].

[0081] Figure 4C This is a schematic diagram of a method for randomly allocating slots according to an embodiment of the present disclosure.

[0082] like Figure 4CAs shown, queue 1, queue 2, queue 3, …, can be hotspot resource queues. Each queue has a respective contention winning interval, and the contention intervals of each queue range from [0, 1], and the sum of all contention intervals is 1.

[0083] For multiple queues, operation S410 can be performed to determine a target queue according to the contention interval in which the random number falls.

[0084] For example, a random number ranging from [0, 1] can be generated, the target contention interval is determined according to the interval in which the random number falls, and the target queue is determined according to the target contention interval. For example, the random number is 0.2, and falls into the contention interval [0, 0.3] of queue 1, so queue 1 can be determined as the target queue.

[0085] For the target queue, operation S420 can be performed to determine that the target resource is allocated to a random slot from the target queue.

[0086] For example, in queue 1, it is determined whether the resource of the first slot meets the diversity constraint condition. If it meets, the resource can be allocated to a random slot. The random slot can be determined according to the random number. For example, the random number is 0.2, and the random slot can be determined as the second slot. Otherwise, it is determined whether the resource of the second slot meets the diversity constraint condition, until the target resource meeting the diversity constraint condition is selected. If the resources in queue 1 all do not meet the diversity constraint condition, the interest resource can be used as a bottom line. For example, the interest resource is selected to be allocated to a random slot.

[0087] According to an embodiment of the present disclosure, the present disclosure also provides an information recommendation method.

[0088] Figure 5 is a flowchart of an information recommendation method according to an embodiment of the present disclosure.

[0089] As Figure 5 shown, the information recommendation method includes operation S510 to operation S520.

[0090] In operation S510, a recommendation list for a user is obtained, wherein the recommendation list includes a plurality of slots arranged in sequence, and each slot is allocated with a resource containing recommendation information.

[0091] In operation S520, the resources in the plurality of slots are recommended to the user in the order of the plurality of slots in the recommendation list.

[0092] For example, the recommendation list is determined according to the method for determining a recommendation list described above. The recommendation list can include interest resources recalled based on user interest and non-interest resources recalled based on business demand. The non-interest resources are, for example, classified according to business type, and the non-interest resources of different business types are recommended according to different strategies. For example, resources of a strong insertion type can be located in a top slot of the recommendation list. Resources of a hot spot type are allocated to slots in a random allocation manner. Resources of a promotion type are reordered with the interest resources after weighted processing and then allocated to slots of the recommendation list.

[0093] According to an embodiment of the present disclosure, the resources in the recommendation list are recommended in order, which can ensure the diversity and recommendation effect of the recommendation information and improve user experience.

[0094] According to an embodiment of the present disclosure, the present disclosure further provides a device for determining a recommendation list and an information recommendation device.

[0095] Figure 6 FIG. 6 is a block diagram of a device for determining a recommendation list according to an embodiment of the present disclosure.

[0096] As shown in FIG. 6, the device 600 for determining a recommendation list includes a first acquisition module 601 and a determination module 602. Figure 6

[0097] The first acquisition module 601 is configured to acquire interest resources recalled based on user interest and non-interest resources recalled based on business type, and the interest resources and the non-interest resources are each provided with a recommendation degree.

[0098] The determination module 602 is configured to allocate the interest resources and the non-interest resources to a plurality of slots of a recommendation list according to a business type, a recommendation degree, and a slot control threshold value for controlling a number of positions of the non-interest resources.

[0099] The business type includes a strong insertion type, a hot spot type, and a promotion type, and the non-interest resources include strong insertion resources, hot spot resources, and promotion resources; and the determination module includes a first determination unit, a first allocation unit, a second allocation unit, and a third allocation unit.

[0100] The first determination unit is configured to determine a first number of positions of the strong insertion resources, a second number of positions of the hot spot resources, and a third number of positions of the promotion resources according to the slot control threshold value.

[0101] The first allocation unit is configured to allocate the strong insertion resources of the first number of positions to a specified slot of the plurality of slots.

[0102] The second allocation unit is configured to randomly allocate the hot spot resources of the second number of positions to random slots of first remaining slots, where the first remaining slots are remaining slots of the plurality of slots except the specified slot. ​

[0103] The third distribution unit is configured to distribute the third number of promotion resources to the second remaining slot after weighting processing of the recommendation degree of the third number of promotion resources.

[0104] The hotspot resource includes a plurality of hotspot resource queues; the second distribution unit includes a first determination subunit, a second determination subunit, and a third determination subunit

[0105] The first determination subunit is configured to determine, for a hotspot resource queue, a probability of winning competition of the hotspot resource queue according to a number of hotspot resources in the hotspot resource queue.

[0106] The second determination subunit is configured to determine a random slot from the first remaining slot according to the generated random number.

[0107] The third determination subunit is configured to determine a target queue from the plurality of hotspot resource queues according to the random number and the probability of winning competition.

[0108] The first distribution subunit is configured to determine a target hotspot resource that meets a preset condition from the target queue and distribute the target hotspot resource to the random slot, and return the second determination subunit for a new generated random number until the number of distributed target hotspot resources is equal to the second number of slots.

[0109] The third determination subunit is configured to determine a competition interval of the hotspot resource queue according to the probability of winning competition, and determine the target queue according to the competition interval to which the random number belongs.

[0110] The third distribution unit includes a second distribution subunit, a calculation subunit, a sorting subunit, and a third distribution subunit.

[0111] The second distribution subunit is configured to distribute a weight to the third number of promotion resources.

[0112] The calculation subunit is configured to calculate a new recommendation degree of the third number of promotion resources according to the recommendation degree and the weight of the third number of promotion resources.

[0113] The sorting subunit is configured to sort the third number of promotion resources and the interest resource according to the new recommendation degree and the recommendation degree of the interest resource.

[0114] The third distribution subunit is configured to distribute the third number of promotion resources and the interest resource to the second remaining slot according to a sorting result.

[0115] The first acquisition module includes a first recall unit, a second determination unit, a second recall unit, a third determination unit, a third recall unit, and a fourth determination unit.

[0116] The first recall unit is configured to calculate, using the recall model, a first similarity between the user feature and information features of the plurality of candidate interest resources, as a recommendation degree of the candidate interest resources.

[0117] The second determination unit is configured to determine, according to the first similarity, at least one candidate interest resource from the plurality of candidate interest resources as the interest resource.

[0118] The second recall unit is configured to, for a hot topic type service, calculate, using the recall model, a respective recommendation degree of a plurality of candidate hot topic resources based on respective information features of the plurality of candidate hot topic resources, in response to a recall requirement of the service.

[0119] The third determination unit is configured to determine, according to the respective recommendation degrees of the plurality of candidate hot topic resources, at least one candidate hot topic resource from the plurality of candidate hot topic resources as the non-interest resource.

[0120] The third recall unit is configured to calculate, using the recall model, a second similarity between the user feature and information features of the plurality of candidate promotion resources, as a recommendation degree of the candidate promotion resources, in response to a recall requirement of the service.

[0121] The fourth determination unit is configured to determine, according to the second similarity, at least one candidate promotion resource from the plurality of candidate promotion resources as the non-interest resource.

[0122] Figure 7 FIG. 7 is a block diagram of an information recommendation device according to an embodiment of the present disclosure.

[0123] As shown in FIG. 7, the information recommendation device 700 can include a second acquisition module 701 and a recommendation module 702. Figure 7

[0124] The second acquisition module 701 is configured to acquire a recommendation list for a user, wherein the recommendation list includes a plurality of slots arranged in sequence, and the slots are assigned with resources containing recommendation information.

[0125] The recommendation module 702 is configured to recommend, in sequence, respective resources in the plurality of slots to the user.

[0126] The recommendation list is determined according to the device for determining the recommendation list.

[0127] According to an embodiment of the present disclosure, the present disclosure further provides an electronic device, a readable storage medium and a computer program product.

[0128] Figure 8 ​A schematic block diagram of an example electronic device 800 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptops, desktops, tablets, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown here, their connections and relationships, and their functions, are meant to be examples only, and are not intended to limit the implementations of the present disclosure described and / or claimed in this document.

[0129] As shown in Figure 8 The device 800 includes a computing unit 801 that can perform various appropriate actions and processes in accordance with a computer program stored in a read-only memory (ROM) 802 or a computer program loaded into a random access memory (RAM) 803 from a storage unit 808. Various programs and data required for the operation of the device 800 can also be stored in the RAM 803. The computing unit 801, the ROM 802, and the RAM 803 are connected to each other through a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0130] Various components in the device 800 are connected to the I / O interface 805, including an input unit 806, such as a keyboard, a mouse, etc.; an output unit 807, such as various types of displays, speakers, etc.; the storage unit 808, such as a magnetic disk, an optical disk, etc.; and a communication unit 809, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 809 allows the device 800 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0131] The computing unit 801 can be various general and / or special purpose processing components with processing and computing capabilities. Some examples of the computing unit 801 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 801 performs various methods and processes described above, such as the method of determining a recommendation list and / or the information recommendation method. For example, in some embodiments, the method of determining a recommendation list and / or the information recommendation method can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 808. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 800 via the ROM 802 and / or the communication unit 809. When the computer program is loaded onto the RAM 803 and executed by the computing unit 801, one or more steps of the method of determining a recommendation list and / or the information recommendation method described above can be performed. Alternatively, in other embodiments, the computing unit 801 can be configured to perform the method of determining a recommendation list and / or the information recommendation method by any other suitable means, such as by means of firmware.

[0132] Various implementations of the systems and techniques described above can be realized in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a complex programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0133] Program code for carrying out methods of the present disclosure can be written in any combination of one or more programming languages. The program code can be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the program code, when executed by the processor or controller, produces a means for implementing the functions / acts specified in the flowcharts and / or block diagrams. The program code can be executed entirely on a machine, partially on a machine, partially on a machine and partially on a remote machine or entirely on a remote machine or server.

[0134] In the context of the present disclosure, a machine-readable medium can be a tangible medium that contains or stores a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include but is not limited to an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0135] To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; 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 acoustic, speech, or tactile input.

[0136] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0137] The computer system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[0138] It should be understood that the various forms of flow shown above can be used to reorder, add, or remove steps. For example, the steps recited in the present disclosure can be performed in parallel, in series, or in a different order, as long as the desired results of the technology disclosed in the present disclosure are achieved, which is not limited herein.

[0139] The specific implementation described above does not constitute a limitation on the protection scope of the present disclosure. 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 replacements, and improvements made within the spirit and principles of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for determining a recommendation list, comprising: The system retrieves interest resources based on user interests and non-interest resources based on business type. Both interest resources and non-interest resources are assigned a recommendation level. Non-interest resources include forced insertion resources, hot topic resources, and promotional resources. The first occupancy number of the forced insertion resource, the second occupancy number of the hot spot resource, and the third occupancy number of the promotion resource are determined according to the slot control threshold of each business type. The business types include forced insertion type, hot spot type, and promotion type. The forced insertion resources of the first placeholder are allocated to a designated slot among the plurality of slots; The hotspot resources of the second placeholder are randomly allocated to random slots in the first remaining slots, wherein the first remaining slots are the remaining slots among the plurality of slots excluding the designated slot; The promotional resources in the third place are weighted by recommendation degree and then reordered with the interest resources. Based on the sorting result, the promotional resources and the interest resources are allocated to the second remaining slots, wherein the second remaining slots are the remaining slots among the multiple slots excluding the designated slots and the random slots. The step of randomly allocating the hotspot resources of the second occupancy slot to random slots in the first remaining slot includes: For each of the multiple hot resource queues, the probability of winning the competition for that hot resource queue is determined based on the number of hot resources in that hot resource queue. Based on the probability of winning the competition, the competition range of the hot resource queue is determined; Based on the competition interval to which the generated random number belongs, the target queue is determined from the plurality of hot resource queues; and Target hotspot resources that meet preset conditions are determined from the target queue and allocated to the random slots.

2. The method of claim 1, wherein, The step of randomly allocating the hotspot resources of the second occupancy bit to random slots in the first remaining slots further includes: Based on the generated random number, a random slot is determined from the first remaining slot; and The determined target hotspot resources are allocated to the random slots, and for a newly generated random number, the step of determining a random slot from the first remaining slots based on the random number is returned until the number of allocated target hotspot resources equals the second placeholder number.

3. The method of claim 1, wherein, The step of weighting the promotional resources in the third slot with their recommendation scores, re-sorting them with the interest resources, and allocating the promotional resources and interest resources to the second remaining slots according to the sorting results includes: The promotional resource allocation weight for the third placeholder; Calculate the new recommendation score of the promotional resources in the third place based on their recommendation score and weight. Based on the new recommendation score and the recommendation score of the interest resources, the promoted resources in the third place and the interest resources are ranked; and Based on the sorting results, the promotional resources of the third occupant and the interest resources are allocated to the second remaining slot.

4. The method of claim 1, wherein, The acquisition of interest resources recalled based on user interests and non-interest resources recalled based on business type includes: A recall model is used to calculate the first similarity between user features and information features of multiple candidate interest resources, which is used as the recommendation score of the candidate interest resources; and Based on the first similarity, at least one candidate interest resource is determined from the plurality of candidate interest resources as the interest resource.

5. The method of claim 1, wherein, The acquisition of interest resources recalled based on user interests and non-interest resources recalled based on business type includes: for popular business types... In response to the recall requirement of this service, a recall model is used to calculate the recommendation degree of each of the multiple candidate hot resources based on their respective information features; and Based on the recommendation degree of each of the multiple candidate hot resources, at least one candidate hot resource is determined from the multiple candidate hot resources as the non-interest resource.

6. The method of claim 1, wherein, The acquisition of interest resources recalled based on user interests and non-interest resources recalled based on business type includes: for promotional businesses, In response to the recall requirement of this service, a recall model is used to calculate a second similarity between user characteristics and information features of multiple candidate promotional resources, which serves as the recommendation score of the candidate promotional resources; and Based on the second similarity, at least one candidate promotion resource is determined from the plurality of candidate promotion resources as the non-interest resource.

7. An information recommendation method, comprising: Obtain a recommendation list for a user, wherein the recommendation list includes multiple slots arranged in sequence, and each slot is allocated resources containing recommendation information; and In the order described above, the resources in each of the multiple slots are recommended to the user. The recommended list is determined by the method according to any one of claims 1 to 6.

8. An apparatus for determining a recommendation list, comprising: The first acquisition module is used to acquire interest resources recalled based on user interests and non-interest resources recalled based on business type. Both the interest resources and the non-interest resources are set with a recommendation degree. The non-interest resources include forced insertion resources, hot resources and promotional resources. The slot determination module is used to determine the first slot size of the forced insertion resource, the second slot size of the hot spot resource, and the third slot size of the promotion resource according to the slot control threshold of each business type. The business types include forced insertion type, hot spot type, and promotion type. The first allocation module is used to allocate the forced insertion resources of the first placeholder to a specified slot among the plurality of slots; The second allocation module is used to randomly allocate the hotspot resources of the second placeholder to random slots in the first remaining slots, wherein the first remaining slots are the remaining slots among the plurality of slots excluding the designated slots; The third allocation module is used to reorder the promotional resources in the third placeholder with the interest resources after weighting them by recommendation degree, and allocate the promotional resources and the interest resources to the second remaining slots according to the sorting result, wherein the second remaining slots are the remaining slots among the plurality of slots excluding the designated slots and the random slots; The second allocation module is further configured to, for each of the multiple hot resource queues, determine the probability of winning the competition for that hot resource queue based on the number of hot resources in that hot resource queue; determine the competition interval of the hot resource queue based on the probability of winning the competition; determine a target queue from the multiple hot resource queues based on the competition interval to which the generated random number belongs; and determine target hot resources that meet preset conditions from the target queue and allocate them to the random slot.

9. The apparatus of claim 8, wherein, The second allocation module includes: A random slot determination unit is configured to determine a random slot from the first remaining slots based on the generated random number; and The first allocation unit is used to allocate the determined target hotspot resources to the random slots, and return to the second determination subunit for the generated new random number, until the number of allocated target hotspot resources is equal to the second placeholder number.

10. The apparatus of claim 8, wherein, The third allocation module includes: The second allocation unit is used to allocate weights to the promotion resources of the third occupancy. The calculation unit is used to calculate the new recommendation degree of the promotional resources in the third place based on the recommendation degree and weight of the promotional resources in the third place. A sorting unit is configured to sort the promoted resources and the interest resources in the third placeholder based on the new recommendation score and the recommendation score of the interest resources; and The third allocation unit is used to allocate the promotional resources of the third occupant and the interest resources to the second remaining slot according to the sorting result.

11. The apparatus of claim 8, wherein, The first acquisition module includes: The first recall unit is used to calculate a first similarity between user features and information features of multiple candidate interest resources using a recall model, as the recommendation degree of the candidate interest resources; and An interest resource determination unit is configured to determine at least one candidate interest resource as the interest resource from the plurality of candidate interest resources based on the first similarity.

12. The apparatus of claim 8, wherein, The first acquisition module includes: The second recall unit is used to, in response to the recall needs of services of a certain type, calculate the recommendation degree of each of the multiple candidate hot resources based on their respective information features using a recall model; and The hotspot resource determination unit is used to determine at least one candidate hotspot resource from the plurality of candidate hotspot resources as the non-interest resource based on the recommendation degree of each of the plurality of candidate hotspot resources.

13. The apparatus according to claim 8, wherein, The first acquisition module includes: The third recall unit, in response to the recall requirement of this service, uses a recall model to calculate a second similarity between user characteristics and information features of multiple candidate promotional resources, as the recommendation degree of the candidate promotional resources; and The promotion resource determination unit is used to determine at least one candidate promotion resource from the plurality of candidate promotion resources as the non-interest resource based on the second similarity.

14. An information recommendation device, comprising: The second acquisition module is used to acquire a recommendation list for a user, wherein the recommendation list includes multiple slots arranged in sequence, and each slot is allocated resources containing recommendation information; and The recommendation module is used to recommend the resources in the plurality of slots to the user in the order stated above. The recommended list is determined by the apparatus according to any one of claims 8 to 13.

15. 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 to 7.

16. 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 to 7.

17. A computer program product comprising a computer program stored on at least one of a readable storage medium and an electronic device, the computer program implementing the method according to any one of claims 1 to 7 when executed by a processor.

Images