Hot object determination method and apparatus, storage medium, and electronic device

Abstract

The present disclosure provides a popular object determination method and device, a storage medium and an electronic device, and relates to the technical field of data processing. The popular object determination method comprises: acquiring feature data of a plurality of candidate objects on different interaction platforms; determining a change trend of the feature data of each candidate object, fitting the change trend, and obtaining a first bucket value of each candidate object according to the fitting result; fusing the feature data of each candidate object on different interaction platforms, and performing cross-platform statistics on the change trend of each candidate object to obtain a second bucket value of each candidate object; and determining a target object of each interaction platform from the candidate objects based on the first bucket value and the second bucket value of each candidate object. The present disclosure can improve the accuracy of the target object.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and in particular to a method, apparatus, storage medium and electronic device for determining hot objects. Background Technology

[0002] Currently, when executing certain activities (exhibitions, promotions, etc.), operators on different interactive platforms need to select a subset of objects (i.e., popular or potentially popular objects) from the total number of objects on each platform to perform corresponding activities. In related technologies, when staff on different interactive platforms select popular or potentially popular objects from the total number of objects, they do so based on the object's recent sales performance and their personal predictions. Summary of the Invention

[0003] However, since the relevant technologies determine popular items based on a single sales data point, and each interactive platform only looks at the platform's data when making decisions about popular items, the determination of popular items is influenced by the predictions of expert experience, making the final determination of popular items inaccurate.

[0004] Therefore, there is a great need for an improved method for identifying popular objects to improve their accuracy.

[0005] In this context, embodiments of the present disclosure are intended to provide a method, apparatus, storage medium, and electronic device for determining popular objects.

[0006] According to a first aspect of this disclosure, a method for determining popular objects is provided, comprising:

[0007] Acquire feature data of multiple candidate objects on different interactive platforms;

[0008] Determine the changing trend of the feature data of each candidate object, fit the changing trend, and obtain the first bucket value of each candidate object based on the fitting result;

[0009] The feature data of each candidate object on different interactive platforms are fused, and the changing trends of each candidate object are statistically analyzed across platforms to obtain the second bucket value of each candidate object.

[0010] Based on the first bucket value and the second bucket value of each candidate object, the target object of each interaction platform is determined from the candidate objects.

[0011] In one implementation, determining the changing trend of the feature data of each candidate object, fitting the changing trend, and obtaining the first bucket value of each candidate object based on the fitting result includes: determining the average value of the feature data of each candidate object within a first preset time period to obtain a recent value; fitting the feature data of each candidate object within a second preset time period and a third preset time period based on the recent value to obtain a first trend value and a second trend value; the second preset time period is less than the third preset time period; converting the recent value, the first trend value, and the second trend value into bucket sorting under different dimensions to obtain the first bucket value of each candidate object; the first bucket value is the bucket score of the recent value, the first trend value, and the second trend value under different dimensions.

[0012] In one implementation, the step of fitting the feature data of each candidate object within a second preset time period and a third preset time period based on the recent value to obtain a first trend value and a second trend value includes: using the least squares method to fit the recent values ​​within the second preset time period and the third preset time period to obtain the first trend value and the second trend value.

[0013] In one embodiment, after fitting the feature data of each candidate object within a second preset time period and a third preset time period based on the recent value to obtain a first trend value and a second trend value, the method further includes: normalizing and truncating the upper and lower limits of the first trend value.

[0014] In one implementation, the step of fusing the feature data of each candidate object on different interactive platforms and performing cross-platform statistics on the changing trends of each candidate object to obtain the second bucket value of each candidate object includes: determining the correlation weight of each candidate object on different interactive platforms based on the feature data of each candidate object on different interactive platforms; and fusing the first bucket values ​​of each candidate object on different interactive platforms based on the correlation weight of each candidate object on different interactive platforms to obtain the second bucket value of each candidate object.

[0015] In one implementation, determining the target object of each interaction platform from the candidate objects based on the first and second bucket values ​​of each candidate object includes: obtaining historical information of each candidate object to obtain a third bucket value; the historical information is used to characterize the predictive feature data corresponding to each candidate object as an initial object; and determining the target object of each interaction platform from the candidate objects based on the first, second, and third bucket values ​​of each candidate object.

[0016] In one implementation, determining the target object of each interaction platform from the candidate objects based on the first bucket value, second bucket value, and third bucket value of each candidate object includes: performing a weighted summation of the first bucket values ​​of each candidate object based on a first preset weight; performing a weighted summation of the weighted summation result, the second bucket value, and the third bucket value based on a second preset weight to obtain the final score of each candidate object; and determining the target object of each interaction platform from the candidate objects based on the final score of each candidate object.

[0017] In one implementation, the feature data includes: user interaction statistics, output data, and input data.

[0018] According to a second aspect of this disclosure, a hotspot object determination apparatus is provided, comprising:

[0019] The feature data acquisition module is configured to acquire feature data of multiple candidate objects on different interactive platforms.

[0020] The first bucket value determination module is configured to determine the changing trend of the feature data of each candidate object, fit the changing trend, and obtain the first bucket value of each candidate object based on the fitting result.

[0021] The feature data of each candidate object on different interactive platforms are fused, and the change trend of the second bucket value determination module of each candidate object is statistically analyzed across platforms to obtain the second bucket value of each candidate object.

[0022] The target object determination module is configured to determine the target objects of each interaction platform from the candidate objects based on the first bucket value and the second bucket value of each candidate object.

[0023] In one implementation, the first bucket value determination module is configured to: determine the average value of the feature data of each candidate object within a first preset time period to obtain a recent value; based on the recent value, fit the feature data of each candidate object within a second preset time period and a third preset time period to obtain a first trend value and a second trend value; the second preset time period is less than the third preset time period; convert the recent value, the first trend value and the second trend value into bucket sorting under different dimensions to obtain the first bucket value of each candidate object; the first bucket value is the bucket score of the recent value, the first trend value and the second trend value under different dimensions.

[0024] In one implementation, the first bucket value determination module is configured to: use the least squares method to fit the recent values ​​within the second preset time period and the third preset time period respectively to obtain the first trend value and the second trend value.

[0025] In one embodiment, the aforementioned popular object determination device further includes a post-processing module configured to: normalize and extract upper and lower limits for the first trend value.

[0026] In one implementation, the second bucket value determination module is configured to: determine the relevance weight of each candidate object across different interactive platforms based on the feature data of each candidate object on different interactive platforms; and, based on the relevance weight of each candidate object across different interactive platforms, fuse the first bucket values ​​of each candidate object across different interactive platforms to obtain the second bucket value of each candidate object.

[0027] In one implementation, the target object determination module is configured to: acquire historical information of each candidate object to obtain a third bucket value; the historical information is used to characterize the predictive feature data corresponding to each candidate object as an initial object; and determine the target object of each interaction platform from the candidate objects based on the first bucket value, the second bucket value and the third bucket value of each candidate object.

[0028] In one implementation, the target object determination module is configured to: perform a weighted summation of the first bucket values ​​of each candidate object based on a first preset weight; perform a weighted summation of the weighted summation result, the second bucket value, and the third bucket value based on a second preset weight to obtain the final score of each candidate object; and determine the target object of each interaction platform from the candidate objects based on the final score of each candidate object.

[0029] In one implementation, the feature data includes: user interaction statistics, output data, and input data.

[0030] According to a third aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the hot object determination method of the first aspect and its possible implementations.

[0031] According to a fourth aspect of this disclosure, an electronic device is provided, comprising: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the hot object determination method of the first aspect and possible implementations thereof by executing the executable instructions.

[0032] The technical solution disclosed herein has the following beneficial effects:

[0033] In the aforementioned method for identifying popular objects, feature data of multiple candidate objects on different interactive platforms are acquired; the changing trend of the feature data of each candidate object is determined, and the changing trend is fitted to obtain the first bucket value of each candidate object based on the fitting result; the feature data of each candidate object on different interactive platforms are fused, and the changing trend of each candidate object is statistically analyzed across platforms to obtain the second bucket value of each candidate object; based on the first bucket value and the second bucket value of each candidate object, the target object for each interactive platform is determined from the candidate objects. Because the evaluation of each candidate object not only uses multi-dimensional feature data but also considers the influence of the interactive platform, a more accurate evaluation of each candidate object is achieved. Attached Figure Description

[0034] The above and other objects, features, and advantages of this disclosure will become readily apparent from the following detailed description of exemplary embodiments, taken in conjunction with the accompanying drawings. Several embodiments of this disclosure are illustrated in the drawings by way of example and not limitation, in which:

[0035] Figure 1 A schematic diagram of a system architecture in this exemplary embodiment is shown;

[0036] Figure 2 This diagram illustrates a method for determining popular objects in this exemplary embodiment;

[0037] Figure 3 This example implementation shows a sub-flowchart for determining the first bucket value.

[0038] Figure 4 This diagram illustrates the determination of the first bucket value in this exemplary embodiment.

[0039] Figure 5 This example implementation shows a sub-flowchart for determining the second bucket value;

[0040] Figure 6 This exemplary embodiment shows a sub-flowchart for determining the target object;

[0041] Figure 7 This exemplary embodiment shows a sub-flowchart for determining the target object;

[0042] Figure 8 This diagram illustrates the structure of a popular object determination device according to this exemplary embodiment.

[0043] Figure 9 A schematic diagram of the structure of an electronic device in this exemplary embodiment is shown. Detailed Implementation

[0044] Exemplary embodiments of this disclosure will be described more fully below with reference to the accompanying drawings.

[0045] The accompanying drawings are schematic illustrations of this disclosure and are not necessarily drawn to scale. Some block diagrams shown in the drawings may be functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in hardware modules or integrated circuits, or in networks, processors, or microcontrollers. Implementations can be carried out in various forms and should not be construed as limited to the examples set forth herein. The features, structures, or characteristics described in this disclosure can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough description of embodiments of this disclosure. However, those skilled in the art will recognize that one or more specific details may be omitted when implementing the technical solutions of this disclosure, or other methods, components, apparatuses, steps, etc., may be used to replace one or more specific details. Invention Overview

[0047] Currently, when executing certain activities (exhibitions, promotions, etc.), operators of different interactive platforms need to select a portion of the objects (i.e. popular objects or potential popular objects) from the full range of objects on different interactive platforms to execute the corresponding activities.

[0048] In related technologies, when staff on different interactive platforms select popular or potentially popular objects from the full dataset, they do so based on the object's recent sales performance and personal predictions. This approach presents the following problems:

[0049] 1) Popular targets are determined based on a single sales data point;

[0050] 2) Objects are displayed on multiple interactive platforms, and each interactive platform only looks at a single set of data when making decisions about popular objects;

[0051] 3) The identification of popular targets is influenced by the predictions based on expert experience;

[0052] In summary, the identified popular targets in related technologies are not accurate enough.

[0053] In view of the above problems, the exemplary embodiments of this disclosure first provide a method for determining popular objects, so as to improve the accuracy of popular objects to a certain extent.

[0054] The following is combined Figure 1 The system architecture and application scenarios of the operating environment of this exemplary implementation are described in an exemplary manner.

[0055] Figure 1A schematic diagram of the system architecture is shown. System architecture 100 may include a server 110 and a popular object identification system 120. Server 110 may be a single server or a cluster of servers, and can be used to store and / or provide backend data on multidimensional feature data of multiple candidate objects on different interaction platforms. Popular object identification system 120 may also be a single server or a cluster of servers, and can process the data obtained from server 110 to determine popular objects based on the processing results. Server 110 and popular object identification system 120 can be connected via wired or wireless communication links for data exchange.

[0056] In one implementation, this exemplary implementation can also be implemented independently based on server 110. For example, after obtaining multidimensional feature data of multiple candidate objects on different interaction platforms, server 110 obtains the target object by executing the above-described popular object determination method.

[0057] In one implementation, this exemplary implementation can also be implemented independently based on the popular object identification system 120. For example, the popular object identification system 120 can obtain multi-dimensional feature data of multiple candidate objects on different interactive platforms from a background database, and obtain the target object by executing the above-described popular object identification method.

[0058] As can be seen from the above, in this exemplary embodiment, the method for determining popular objects can be executed by the server 110 or the popular object confirmation system 120. This disclosure does not limit this.

[0059] The following is combined Figure 2 The method for determining popular objects in this exemplary embodiment will be described. Figure 2 An exemplary flow of this popular object determination method is shown, which may include the following steps S210 to S240:

[0060] Step S210: Obtain feature data of multiple candidate objects on different interactive platforms;

[0061] Step S220: Determine the changing trend of the feature data of each candidate object, fit the changing trend, and obtain the first bucket value of each candidate object based on the fitting result;

[0062] Step S230: Merge the feature data of each candidate object on different interactive platforms, and perform cross-platform statistics on the changing trends of each candidate object to obtain the second bucket value of each candidate object.

[0063] Step S240: Based on the first bucket value and the second bucket value of each candidate object, determine the target object of each interaction platform from the candidate objects.

[0064] In the aforementioned method for identifying popular targets, feature data of multiple candidate targets on different interactive platforms are acquired; the changing trends of the feature data of each candidate target are determined, and the changing trends are fitted to obtain the first bucket value of each candidate target based on the fitting results; the feature data of each candidate target on different interactive platforms are fused, and the changing trends of each candidate target are statistically analyzed across platforms to obtain the second bucket value of each candidate target; based on the first and second bucket values ​​of each candidate target, the target targets for each interactive platform are determined from the candidate targets. Because the evaluation of each candidate target not only uses multi-dimensional feature data but also considers the influence of the interactive platform, a more accurate evaluation of each candidate target can be achieved.

[0065] The following is about Figure 2 Each step in the process will be explained in detail.

[0066] refer to Figure 2 In step S210, feature data of multiple candidate objects are obtained on different interactive platforms.

[0067] Generally, any object can be displayed on every interactive platform or on some interactive platforms; correspondingly, when any object can be displayed on every interactive platform, the candidate object can be all objects existing on any interactive platform; when any object is displayed on some interactive platforms, the candidate object can be the sum of objects on all interactive platforms.

[0068] The candidate objects vary depending on the interactive platform. When the interactive platform is an e-commerce platform, the candidate objects can be various objects, such as cars and food. When the interactive platform is a cultural dissemination platform, the candidate objects can be various cultural dissemination objects, such as books and open courses. When the interactive platform is an entertainment platform, the candidate objects can be music and videos.

[0069] In one implementation, the feature data includes: user interaction statistics, output data, and input data; wherein, user interaction statistics refer to the data of interactions between users and candidate objects on the interaction platform, such as: exposure UV, product CTR, product CVR, add-to-cart UV, product add-to-cart rate, etc. on the e-commerce platform; output data refers to the data that makes the e-commerce platform profitable due to user transactions on the interaction platform, such as: product GMV, product ARPU, payment UV, etc.; input data refers to the investment in taking certain measures on the interaction platform to improve returns, such as: product advertising ROI, etc.

[0070] Continue to refer to Figure 2 In step S220, the changing trend of the feature data of each candidate object is determined, the changing trend is fitted, and the first bucket value of each candidate object is obtained based on the fitting result.

[0071] The trend of change refers to the increase or decrease of the characteristic data of each candidate object in a short period of time. Since daily characteristic data exhibits significant fluctuations, a short-term average can be used to characterize the increase or decrease of the characteristic data of each candidate object within that short period (on any given day). That is, the average value of the characteristic data of each candidate object within a short period is determined as the increase or decrease of the characteristic data of each candidate object within that short period (on any given day). The time period referred to as "short period" can be determined empirically, such as three days, five days, or one week; no specific limitation is made here.

[0072] Fitting refers to determining the long-term trend of the characteristic data of each candidate object based on the short-term trend of the characteristic data of each candidate object. For example, if the short-term trend is the average value of the characteristic data of each candidate object within a week, then fitting can be based on the weekly trend of the characteristic data of each candidate object to determine the trend of the characteristic data of each candidate object within a month, three months, six months and / or a year.

[0073] After obtaining the fitting results (the long-term trend of the feature data of each candidate object), in order to further reduce the volatility of the feature data of each candidate object, the above trend and fitting results can be transformed into bucket sorting under different dimensions to obtain the bucket sorting of the trend and fitting results under different dimensions. Here, bucket sorting can be understood as bucket sorting. Bucket sorting is a sorting algorithm that works by dividing the array into a finite number of buckets, and then sorting each bucket individually (possibly using other sorting algorithms or continuing to use bucket sorting recursively).

[0074] When the above fitting results (long-term trend) are divided into two time periods (e.g., one month and three months) and the dimension is 2, the trend and fitting results are sorted into 6 bins under different dimensions; that is, the trend is sorted into bins under two dimensions, the fitting results within one month are sorted into bins under two dimensions, and the fitting results within three months are sorted into bins under two dimensions.

[0075] Continue to refer to Figure 2 In step S230, the feature data of each candidate object on different interactive platforms are fused, and the change trend of each candidate object is statistically analyzed across platforms to obtain the second bucket value of each candidate object.

[0076] When candidate objects are displayed / exhibited on multiple interactive platforms, the characteristic data of the candidate objects may differ depending on the interactive platform. At the same time, there are also commonalities. Therefore, in order to more accurately determine the target object, this application embodiment considers the interactive platform factor.

[0077] Feature data fusion is achieved by calculating the relevance weights of the secondary categories to which candidate objects belong across different interactive platforms.

[0078] Cross-platform statistics on trends can be obtained by dividing the recent values ​​of the relevance weights of other interactive platforms into buckets.

[0079] Continue to refer to Figure 2 In step S240, the target objects of each interaction platform are determined from the candidate objects based on the first bucket value and the second bucket value of each candidate object.

[0080] The target objects are those that each interactive platform uses to display and exhibit during the event. These can be understood as popular or potentially popular objects, and no specific definition is made here.

[0081] In one implementation, the target object of each interaction platform can be determined from the candidate objects by using the first bucket value and the second bucket value; alternatively, in order to be compatible with new objects, expert prediction (third bucket value) can be incorporated into the above to determine the target object of each interaction platform from the candidate objects.

[0082] In one implementation, since the daily feature data fluctuates significantly, the aforementioned first bucket value can be determined based on the mean of the feature data; (Refer to...) Figure 3 As shown, step S220 above may include the following steps:

[0083] Step S310: Determine the average value of the feature data of each candidate object within the first preset time period to obtain the recent value.

[0084] The first preset duration is determined based on experience; for example, three days, five days, or one week, without being limited here. Here, the first preset duration can be understood as the aforementioned short period.

[0085] Since the feature data of each candidate object includes three categories, and each candidate object is displayed on multiple interactive platforms, the average value is determined according to the interactive platform and the category of feature data. For example, if the first preset duration is one week, the interactive platform is platform i, the candidate object is object A, and the feature data is the output data Date from the three categories, the average value of the feature data of object A within the first preset duration is as shown in the following formula (1):

[0086]

[0087] Among them, Data cidj For object A, the output data Date is the specific value of the date dj in the interactive platform ci; AvgData cidj For data cidj The average value within the first preset duration; t is the number of days within the first preset duration.

[0088] Step S320: Based on recent values, fit the feature data of each candidate object within the second preset time period and the third preset time period respectively to obtain the first trend value and the second trend value.

[0089] The second preset duration is shorter than the third preset duration.

[0090] The second and third preset durations are determined based on experience; for example, if the second preset duration is one month, the third preset duration is three months; if the second preset duration is three months...

[0091] The third preset duration is six months.

[0092] In one implementation, the first trend value and the second trend value can be obtained by fitting the change slopes of a second preset duration and a third preset duration based on recent values; specifically, step S320 may include the following steps:

[0093] The least squares method is used to fit the recent values ​​within the second and third preset time periods to obtain the first trend value and the second trend value.

[0094] The least squares regression method can be used to calculate the first and second trend values ​​using the following formulas (2) to (4), with the second preset duration being T as an example:

[0095]

[0096]

[0097]

[0098] Where vgX represents the average number of days on the X-axis; AvgY cidj Represents Y-axis AvgData cidj The average value over T days; TrendTData cidj This indicates the linear slope of the curve in the XY coordinate system, which is fitted using the least squares method based on the changes in data over T days.

[0099] Step S330: Convert the recent value, the first trend value, and the second trend value into bucket sorting under different dimensions to obtain the first bucket value of each candidate object.

[0100] Among them, the first bucket value is the bucket score of the recent value, the first trend value and the second trend value under different dimensions.

[0101] Dimensions include global, first-level category, and second-level category; here, global refers to all objects in a certain interactive platform; first-level category refers to the category to which a certain object belongs; second-level category refers to a more detailed classification under the first-level category; for example, if the first-level category is women's clothing, then T-shirts, tops, bottoms, etc. are second-level categories.

[0102] In one implementation, the first bucket value can be determined by the following formula (5):

[0103]

[0104] Among them, AvgDataOrd cidj Indicates AvgData cidj Sort the numerical values ​​of all objects in descending order under the interactive platform CI DateDJ; Count cidj This indicates the total number of objects under the ci date dj interactive platform; AvgDataRS cidj Indicates AvgData cidj Global bucket sorting.

[0105] Similarly, AvgDataRSCate1 cidj AvgDataRSCate2 cidj They represent AvgData respectively cidj The sorting is performed in buckets for the first-level category and the second-level category. Order and Count represent the descending order of the first-level category and the second-level category, and the total number of objects.

[0106] Similarly, TrendT1NormDataRS cidj TrendT1NormDataRSCate1 cidj TrendT1NormDataRSCate2 cidj These represent the bucket sorting of TrendT1NormData in the global, first-level category, and second-level category categories, respectively.

[0107] TrendT2NormDataRS cidj TrendT2NormDataRSCate1 cidj TrendT2NormDataRSCate2 cidj These represent the bucket sorting of TrendT2NormData and TrendT1NormData respectively, in the global, first-level category, and second-level category categories.

[0108] In summary, as Figure 4 As shown, each feature data will eventually be converted into 9 bucket values, and these 9 bucket values ​​will form the first bucket value.

[0109] In one implementation, since large differences in the feature data of different objects can lead to distortion of the calculated first and second trend values, the feature data of different objects can be processed accordingly to reduce distortion. After step S320 above, the following steps may be included:

[0110] Normalize and truncate the upper and lower limits of the first trend value.

[0111] Because significant differences in data between different objects can distort the calculated trend value, for example, the daily average GMV might be in the tens of thousands for object A and the tens of thousands for object B. If object B increases by 100% month-on-month while object A increases by 50% month-on-month, the difference in value range will cause the trend value of object A to be much greater than that of object B. In one implementation, the normalization and upper and lower limit truncation of the first trend value can be achieved using the following formula (6):

[0112]

[0113] Among them, TrendT1NormData cidj This represents the normalized slope, suitable for comparing the slope of the same value for objects of different sizes.

[0114] Normalization, or data normalization, is a means of scaling the dimensions of data to make them comparable. When scaling the dimensions, Min-max normalization or Z-Score normalization can be used, and no specific choice is made here.

[0115] Upper and lower limit truncation, also known as data truncation, is a method of truncating the lower and upper limits of a data range into two columns of data: the lower limit and the upper limit.

[0116] In one implementation, to make the target object more accurate, the interaction platform factor can be incorporated into the target object determination process; see reference. Figure 5 As shown, step S230 above may include the following steps:

[0117] Step S510: Determine the correlation weight of each candidate object across different interactive platforms based on the feature data of each candidate object on different interactive platforms.

[0118] The relevance weight represents the similarity of the feature data of candidate objects across different interactive platforms; in one implementation, it can be determined using the following formula (7):

[0119]

[0120] in, This indicates the sales percentage of the second-level category cate2k under the interactive platform ci within the third preset time period; This indicates the sales percentage of the second-level category cate2k under the interactive platform cm within the third preset time period; This indicates the relevance weight of the interaction platform cm to the interaction platform ci in the secondary category cate2k.

[0121] Step S520: Based on the relevance weights of each candidate object across different interactive platforms, merge the first bucket values ​​of each candidate object across different interactive platforms to obtain the second bucket value of each candidate object.

[0122] The value of the second sub-bucket can be determined using the following formula (8):

[0123]

[0124] in, This represents the final merged value, which is the result of integrating the bucket values ​​AvgDataRS, AvgDataRSCate1, and AvgDataRSCate2 from other interactive platforms based on the relevance of the secondary category.

[0125] In one implementation, to ensure compatibility with new objects, historical information can be incorporated based on the above, referencing... Figure 6 As shown, step S230 above may include the following steps:

[0126] Step S610: Obtain the historical information of each candidate object to get the third bucket value.

[0127] Historical information is used to characterize the predictive feature data corresponding to each candidate object as the initial object.

[0128] Historical information refers to characteristic data determined by experts based on experience when a new object is launched, in order to predict the performance of the new object after it is displayed / exhibited.

[0129] Step S620: Based on the first bucket value, second bucket value and third bucket value of each candidate object, determine the target object of each interaction platform from each candidate object.

[0130] Generally, head objects are evaluated based on the first bucket value, potential objects on the second bucket value, and new objects on the third bucket value. The final score for each candidate object is determined based on these three bucket values, and then the target object for each interaction platform is identified from among the candidate objects based on these final scores. In one implementation, refer to... Figure 7 As shown, step S620 above may include the following steps:

[0131] Step S710: Based on the first preset weight, perform a weighted summation of the first bucket values ​​of each candidate object.

[0132] The weighted summation can be achieved using the following formula (9):

[0133]

[0134] in, The first bucket value is calculated by weighting the data representing a certain characteristic of the object.

[0135] Step S720: Based on the second preset weight, perform a weighted summation on the weighted summation result, the second bucket value, and the third bucket value to obtain the final score of each candidate object.

[0136] The final score can be achieved using the following formula (10):

[0137]

[0138] in, This represents the final score of an object on the interaction platform ci, obtained by weighting all feature data of the object and the values ​​of the second and third buckets; W rel and W expert Using fixed values, the weights {W1...........W9} of specific data types Data can be set differently to reflect different target objects and define the target.

[0139] Step S730: Based on the final scores of each candidate object, determine the target object of each interactive platform from among the candidate objects.

[0140] Among these, the candidate objects with the highest final scores can be identified as the target objects.

[0141] Exemplary embodiments of this disclosure also provide a popular object determination apparatus. (See reference...) Figure 8 As shown, the popular object determination device 800 may include:

[0142] The feature data acquisition module 810 is configured to acquire feature data of multiple candidate objects on different interactive platforms;

[0143] The first bucket value determination module 820 is configured to determine the changing trend of the feature data of each candidate object, fit the changing trend, and obtain the first bucket value of each candidate object based on the fitting result.

[0144] The second bucket value determination module 830 is configured to fuse the feature data of each candidate object on different interactive platforms, and to perform cross-platform statistics on the changing trend of each candidate object's second bucket value determination module to obtain the second bucket value of each candidate object.

[0145] The target object determination module 840 is configured to determine the target objects of each interaction platform from the candidate objects based on the first bucket value and the second bucket value of each candidate object.

[0146] In one implementation, the first bucket value determination module 820 is configured to: determine the average value of the feature data of each candidate object within a first preset time period to obtain a recent value; based on the recent value, fit the feature data of each candidate object within a second preset time period and a third preset time period to obtain a first trend value and a second trend value; the second preset time period is less than the third preset time period; convert the recent value, the first trend value and the second trend value into bucket sorting under different dimensions to obtain the first bucket value of each candidate object; the first bucket value is the bucket score of the recent value, the first trend value and the second trend value under different dimensions.

[0147] In one implementation, the first bucket value determination module 820 is configured to: use the least squares method to fit the recent values ​​within the second preset time period and the third preset time period respectively to obtain the first trend value and the second trend value.

[0148] In one embodiment, the aforementioned popular object determination device 800 further includes a post-processing module configured to normalize and extract upper and lower limits from the first trend value.

[0149] In one implementation, the second bucket value determination module 830 is configured to: determine the correlation weight of each candidate object on different interactive platforms based on the feature data of each candidate object on different interactive platforms; and, based on the correlation weight of each candidate object on different interactive platforms, fuse the first bucket values ​​of each candidate object on different interactive platforms to obtain the second bucket value of each candidate object.

[0150] In one embodiment, the target object determination module 840 is configured to: acquire historical information of each candidate object to obtain a third bucket value; the historical information is used to characterize the predictive feature data corresponding to each candidate object as an initial object; and determine the target object of each interaction platform from each candidate object based on the first bucket value, the second bucket value and the third bucket value of each candidate object.

[0151] In one implementation, the target object determination module 840 is configured to: perform a weighted summation of the first bucket values ​​of each candidate object based on a first preset weight; perform a weighted summation of the weighted summation result, the second bucket value, and the third bucket value based on a second preset weight to obtain the final score of each candidate object; and determine the target object of each interactive platform from among the candidate objects based on the final score of each candidate object.

[0152] In one implementation, the feature data includes: user interaction statistics, output data, and input data.

[0153] The specific details of each part of the above-mentioned device have been described in detail in the method section of the implementation plan. For any undisclosed details, please refer to the implementation plan of the method section, and therefore will not be repeated here.

[0154] Exemplary storage media

[0155] Exemplary embodiments of this disclosure also provide a computer-readable storage medium that can be implemented as a program product including program code, which, when run on an electronic device, causes the electronic device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. In an alternative embodiment, the program product can be implemented as a portable compact disc read-only memory (CD-ROM) including program code and can run on an electronic device, such as a personal computer. However, the program product of this disclosure is not limited thereto. In this document, the readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0156] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0157] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0158] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0159] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0160] Exemplary electronic devices

[0161] Exemplary embodiments of this disclosure also provide an electronic device that may include a processor and a memory. The memory stores executable instructions for the processor, such as program code. The processor executes these executable instructions to perform the popular object determination method of this exemplary embodiment, such as... Figure 2 The method and steps.

[0162] The following is for reference. Figure 9 The electronic device is illustrated by way of a general-purpose computing device. It should be understood that... Figure 9 The electronic device 900 shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments disclosed herein.

[0163] like Figure 9 As shown, the electronic device 900 may include: a processor 910, a memory 920, a bus 930, an I / O (input / output) interface 940, and a network adapter 950.

[0164] The memory 920 may include volatile memory, such as RAM 921 and cache unit 922, and may also include non-volatile memory, such as ROM 923. The memory 920 may also include one or more program modules 924, including but not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. For example, program module 924 may include the modules in the aforementioned hotspot object determination device 800.

[0165] Bus 930 is used to connect different components of electronic device 900 and may include data bus, address bus and control bus.

[0166] Electronic device 900 can communicate with one or more external devices 1000 (such as keyboard, mouse, external controller, etc.) through I / O interface 940.

[0167] Electronic device 900 can communicate with one or more networks via network adapter 950. For example, network adapter 950 can provide mobile communication solutions such as 3G / 4G / 5G, or wireless communication solutions such as wireless LAN, Bluetooth, and near-field communication. Network adapter 950 can communicate with other modules of electronic device 900 via bus 930.

[0168] although Figure 9 Other hardware and / or software modules, including but not limited to: displays, microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, may also be configured in electronic device 900.

[0169] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to exemplary embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0170] Those skilled in the art will understand that various aspects of this disclosure can be implemented as systems, methods, or program products. Therefore, various aspects of this disclosure can be embodied in entirely hardware implementations, entirely software implementations (including firmware, microcode, etc.), or implementations combining hardware and software aspects, collectively referred to herein as “circuit,” “module,” or “system.” Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0171] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is defined only by the appended claims.

Claims

1. A method for determining popular objects, characterized in that, include: Acquire feature data of multiple candidate objects on different interactive platforms; The characteristic data of each candidate object is determined to change trend, the change trend is fitted, and the first bucket value of each candidate object is obtained based on the fitting result; the change trend refers to the increase or decrease of the characteristic data of each candidate object in the short term; the fitting refers to determining multiple long-term change trends of the characteristic data of each candidate object based on the short-term change trend of the characteristic data of each candidate object. The feature data of each candidate object on different interactive platforms are fused, and the changing trends of each candidate object are statistically analyzed across platforms to obtain the second bucket value of each candidate object. The feature data fusion is achieved by calculating the relevance weight of the secondary category to which the candidate object belongs across different interactive platforms; Based on the first bucket value and the second bucket value of each candidate object, the target object of each interaction platform is determined from the candidate objects; The step of determining the changing trend of the feature data of each candidate object, fitting the changing trend, and obtaining the first bucket value of each candidate object based on the fitting result includes: Determine the average value of the feature data of each candidate object within a first preset time period to obtain the recent value; Based on the recent values, the feature data of each candidate object within the second preset time period and the third preset time period are fitted to obtain the first trend value and the second trend value; the second preset time period is less than the third preset time period. The recent value, the first trend value, and the second trend value are converted into bucket sorting under different dimensions to obtain the first bucket value of each candidate object; the first bucket value is the bucket score of the recent value, the first trend value, and the second trend value under different dimensions; the dimensions include global, first-level category, and second-level category.

2. The method for determining popular objects according to claim 1, characterized in that, The step of fitting the feature data of each candidate object within a second preset time period and a third preset time period based on the recent value to obtain a first trend value and a second trend value includes: The least squares method is used to fit the recent values ​​within the second preset time period and the third preset time period respectively to obtain the first trend value and the second trend value.

3. The method for determining popular objects according to claim 1, characterized in that, After fitting the feature data of each candidate object within a second preset time period and a third preset time period based on the recent value to obtain the first trend value and the second trend value, the method further includes: The first trend value is normalized and its upper and lower limits are truncated.

4. The method for determining popular objects according to claim 1, characterized in that, The process of fusing the feature data of each candidate object on different interactive platforms and performing cross-platform statistical analysis on the changing trends of each candidate object to obtain the second bucket value of each candidate object includes: Based on the feature data of each candidate object on different interactive platforms, the relevance weight of each candidate object across different interactive platforms is determined. Based on the relevance weights of each candidate object across different interactive platforms, the first bucket values ​​of each candidate object across different interactive platforms are merged to obtain the second bucket value of each candidate object.

5. The method for determining popular objects according to claim 1, characterized in that, The step of determining the target objects for each interaction platform from the candidate objects based on the first bucket value and the second bucket value of each candidate object includes: The historical information of each candidate object is obtained to obtain the third bucket value; the historical information is used to characterize the predictive feature data corresponding to each candidate object as the initial object. Based on the first bucket value, second bucket value and third bucket value of each candidate object, the target object of each interaction platform is determined from the candidate objects.

6. The method for determining popular objects according to claim 5, characterized in that, The process of determining the target objects for each interaction platform from the candidate objects based on the first bucket value, second bucket value, and third bucket value of each candidate object includes: Based on the first preset weight, the first bucket value of each candidate object is weighted and summed; Based on the second preset weight, the weighted summation result, the second bucket value, and the third bucket value are weighted and summed to obtain the final score of each candidate object; Based on the final scores of each candidate object, the target object of each interactive platform is determined from the candidate objects.

7. The method for determining popular objects according to any one of claims 1 to 6, characterized in that, The feature data includes: user interaction statistics, output data, and input data.

8. A device for determining popular objects, characterized in that, include: The feature data acquisition module is configured to acquire feature data of multiple candidate objects on different interactive platforms. The first bucket value determination module is configured to determine the changing trend of the feature data of each candidate object, fit the changing trend, and obtain the first bucket value of each candidate object based on the fitting result; the changing trend refers to the increase or decrease of the feature data of each candidate object in the short term; the fitting refers to determining multiple long-term changing trends of the feature data of each candidate object based on the short-term changing trend of the feature data of each candidate object. The second bucket value determination module is configured to fuse the feature data of each candidate object on different interactive platforms, and perform cross-platform statistics on the changing trends of each candidate object to obtain the second bucket value of each candidate object. The feature data fusion is achieved by calculating the relevance weight of the secondary category to which the candidate object belongs across different interactive platforms; The target object determination module is configured to determine the target objects of each interaction platform from the candidate objects based on the first bucket value and the second bucket value of each candidate object; The first bucket value determination module is configured as follows: Determine the average value of the feature data of each candidate object within a first preset time period to obtain the recent value; Based on the recent values, the feature data of each candidate object within the second preset time period and the third preset time period are fitted to obtain the first trend value and the second trend value; the second preset time period is less than the third preset time period. The recent value, the first trend value, and the second trend value are converted into bucket sorting under different dimensions to obtain the first bucket value of each candidate object; the first bucket value is the bucket score of the recent value, the first trend value, and the second trend value under different dimensions; the dimensions include global, first-level category, and second-level category.

9. The popular object determination device according to claim 8, characterized in that, The first bucket value determination module is configured as follows: The least squares method is used to fit the recent values ​​within the second preset time period and the third preset time period respectively to obtain the first trend value and the second trend value.

10. The popular object determination device according to claim 8, characterized in that, The popular object determination device further includes a post-processing module, which is configured to: The first trend value is normalized and its upper and lower limits are truncated.

11. The popular object determination device according to claim 8, characterized in that, The second bucket value determination module is configured as follows: Based on the feature data of each candidate object on different interactive platforms, the relevance weight of each candidate object across different interactive platforms is determined. Based on the relevance weights of each candidate object across different interactive platforms, the first bucket values ​​of each candidate object across different interactive platforms are merged to obtain the second bucket value of each candidate object.

12. The popular object determination device according to claim 8, characterized in that, The target object determination module is configured as follows: The historical information of each candidate object is obtained to obtain the third bucket value; the historical information is used to characterize the predictive feature data corresponding to each candidate object as the initial object. Based on the first bucket value, second bucket value and third bucket value of each candidate object, the target object of each interaction platform is determined from the candidate objects.

13. The popular object determination device according to claim 12, characterized in that, The target object determination module is configured as follows: Based on the first preset weight, the first bucket value of each candidate object is weighted and summed; Based on the second preset weight, the weighted summation result, the second bucket value, and the third bucket value are weighted and summed to obtain the final score of each candidate object; Based on the final scores of each candidate object, the target object of each interactive platform is determined from the candidate objects.

14. The popular object determination device according to any one of claims 8 to 13, characterized in that, The feature data includes: user interaction statistics, output data, and input data.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 7.

16. An electronic device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the method of any one of claims 1 to 7 by executing the executable instructions.

Images