A user dynamic ranking recommendation method

Abstract

The application relates to the technical field of artificial intelligence, and discloses a user dynamic sorting recommendation method, which comprises the following steps: step one, collecting effective dynamic data from a database; step two, extracting main feature information from the collected dynamic data as a sorting basis; step three, calculating the scores of the dynamic feature information according to a specific feature sorting score algorithm; step four, obtaining the total score of the dynamic feature information; step five, writing all the dynamics into redis in a descending order according to the total score of the feature information; step six, requesting a dynamic list from a dynamic recommendation program by a user; and step seven, sorting and recommending the algorithm to take data from the database to recalculate the sorting at a fixed time. The application realizes the sorting of the dynamic list by collecting multiple feature information of the dynamic and calculating the total score, and has the beneficial effect of preferentially displaying high-quality dynamic content.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a user dynamic sorting recommendation method. BACKGROUND

[0002] The sorting recommendation algorithm refers to selecting the content that the user may be interested in according to a series of characteristic information of the user, and sorting the content according to the characteristic information, and recommending the content to the user according to the sorting result.

[0003] The sorting recommendation algorithm currently used usually only uses a single time characteristic as the sorting basis in the algorithm, that is, the latest content is recommended to the user. Such algorithm does not consider the preference of the user, and is easy to lead to the recommended content not meeting the user's expectation. If multi-dimensional characteristics are used as the sorting basis, a very complex comparison algorithm needs to be used as an auxiliary, and at the same time, such algorithm will increase a large amount of calculation process. When the user requests data content, the data is requested from the database and the characteristic information is calculated again, this process not only occupies a large amount of service resources, but also prolongs the response time of the data, leading to the decrease of user experience. SUMMARY

[0004] The technical problem solved by the present application is:

[0005] In view of the defects of the prior art, the present application provides a user dynamic sorting recommendation method, which has the advantages of preferentially displaying high-quality dynamic content, and the recommended dynamic content is more biased towards the user's preference, thereby improving the user experience and solving the above technical problems.

[0006] TECHNICAL SCHEME

[0007] To achieve the above object, the present application provides the following technical scheme: a user dynamic sorting recommendation method, comprising the following steps:

[0008] Step 1: collecting effective dynamic data from the database;

[0009] Step 2: extracting main characteristic information from the collected dynamic data as the sorting basis, and initializing and assigning the weight of each characteristic information;

[0010] Step 3: calculating the score of each characteristic information of the dynamic according to a specific characteristic sorting scoring algorithm;

[0011] Step 4: performing normalization calculation and processing on the calculated score of each characteristic information of the dynamic to obtain uniform standard score data, and performing weighted summation calculation according to the weight of each characteristic information to obtain the total score of the dynamic characteristic information;

[0012] Step 5: writing all the dynamics into redis in descending order according to the total score of the characteristic information.

[0013] Step six, the user requests a dynamic list to the dynamic recommendation program, the dynamic recommendation program takes the dynamic list from the corresponding bucket of redis according to the feature information and returns it to the user;

[0014] Step seven, the sorting recommendation algorithm takes data from the database for recalculation and sorting at regular intervals.

[0015] Preferably, the effective dynamic data in step one is based on: the time of the dynamic is within 72 hours from the current time, or the dynamic is liked or commented within 24 hours from the current time.

[0016] Preferably, the main feature information extracted in step two includes: the creation time of the dynamic, the influence of the dynamic creator, the interactivity of the dynamic, the specific content type of the dynamic, the language set of the dynamic creator, and the gender of the dynamic creator.

[0017] Preferably, the specific feature sorting score algorithm in step three includes the following aspects:

[0018] S1.1, the creation time score of the dynamic, according to Newton's cooling law, its decay satisfies a certain inverse relationship, the calculation formula is:

[0019]

[0020] Where: t now is the time stamp of the current time, t create is the time stamp of the dynamic creation time, S create is the creation time score of the dynamic, this formula is used to increase the overall score of the newly published dynamic, to avoid the newly published dynamic being sorted at the back due to the lack of interactivity score;

[0021] S1.2, the interactivity score of the dynamic, which focuses on the number of comments, the number of comments and the number of likes are calculated according to the weight ratio of 5:1;

[0022] The calculation formula is: S interaction = 5 x number of comments + 1 x number of likes; Where: S interaction represents the interactivity score of the dynamic, which is the calculation result; the number of comments is a variable, representing the total number of comments received by the dynamic content; the number of likes is another variable, representing the total number of likes received by the dynamic content

[0023] S1.3, the specific content type score of the dynamic: its calculation focuses on the richness of the content, that is, video, language and picture type dynamics will have higher scores than text type dynamics;

[0024] S1.4. Dynamic Creator's Language Set Score: This calculation is tied to the requester's language set. Therefore, in this step, M datasets are created, each representing a possible requester's language set. The highest score is assigned to the language set where the requester's native language is included in the dynamic creator's language set. Subsequently, when the dynamic creator's native language is included in the dynamic creator's language set, different scores are assigned based on the priority of the language set. Each possible scenario is calculated separately and placed into the corresponding dataset.

[0025] S1.5. Gender score of the dynamic creator: A higher score is given when the requester and the dynamic creator are of the opposite sex than when they are of the same sex. In this step, the M datasets mentioned above are further differentiated to generate N datasets containing the requester's language set and the requester's gender as feature information.

[0026] Preferably, the normalization calculation in step 4 is:

[0027]

[0028] Where: S feature is the score of one of the dynamic feature information, S max is the maximum score that can be obtained for the feature information, S min It is the minimum score that can be obtained by the feature information, and S is the score corresponding to the feature information after normalization.

[0029] Preferably, after the dynamic data set in step five is processed by the sorting recommendation algorithm, N data sets distinguished by feature information are generated. The N data sets correspond to N buckets divided in redis. Each bucket is distinguished by the feature information corresponding to its data set. The dynamic lists held by the data sets in each bucket are consistent, and the difference lies in the arrangement order of the dynamic lists in the bucket.

[0030] Preferably, in step 6, the user will carry his or her own characteristic information when requesting the dynamic list. The dynamic recommendation program will find the corresponding bucket based on the user's characteristic information. The dynamic list returned by this method is biased towards the user's preferences.

[0031] Preferably, in step seven, the server will periodically obtain new dynamic data information from the database and repeat the above-mentioned specific algorithm steps to update the dynamic list held in the bucket to ensure that the user obtains the new dynamic list in a timely manner. In step seven, when the user creates a dynamic, the dynamic data will be transmitted to the database, and the server will obtain the user's latest uploaded dynamic. When the user requests the dynamic list, the dynamic list is directly obtained from the bucket that has been pre-calculated and sorted, without the need for recalculation, thereby shortening the data response time.

[0032] Preferably, the comprehensive score calculation formula after considering the user feature matching score in step six is:

[0033]

[0034] Wherein: a and β are weight parameters, respectively representing the relative importance of the interaction score and the matching score; the number of likes represents the number of likes received by the dynamic; sim(u i ,b j ) is the similarity function between feature u i and b j ; u i represents the i-th feature in the user feature vector U; w ij represents an element in the weight matrix, specifically, it represents the importance weight between feature u i and b j ; m represents the number of user features; n represents the number of bucket features.

[0035] Preferably, the server updates the dynamic list and ranking by performing the following steps in a timing task in step seven:

[0036] S2.1, the server obtains the latest uploaded dynamic data from the database, and the obtained dynamic data is D new ;

[0037] S2.2, the new dynamic data D new is added to the existing dynamic list D old to form a new dynamic list D: D = D old ∪D new ;

[0038] S2.3, the interaction score S interaction of each dynamic in the new dynamic list D is recalculated:

[0039]

[0040] S2.4, combined with the user feature U and the bucket feature B, the user feature matching score S match is calculated:

[0041]

[0042] S2.5, the comprehensive score is recalculated: combined with the interaction score and the user feature matching score, the comprehensive score S total is recalculated:

[0043]

[0044] S1.6, updating the dynamic list in the bucket: according to the new comprehensive score S total , reordering the dynamic list and updating into the corresponding bucket: D sre total ).

[0045] Compared with the prior art, the application provides a user dynamic sorting recommendation method, which has the following beneficial effects:

[0046] 1. The application realizes the sorting of the dynamic list by collecting multiple feature information of the dynamic and calculating the overall score, which has the beneficial effect of preferentially displaying more high-quality dynamic content.

[0047] 2. The application takes out the dynamic list data from the corresponding bucket according to the partial feature information of the requester, which has the beneficial effect of making the recommended dynamic content more suitable for the user's preferences.

[0048] 3. The application updates the dynamic list by the server without the user, so that the user does not need to perform a large amount of calculation when requesting the dynamic list, but directly obtains the existing list from the bucket, which shortens the response time of the data and has the beneficial effect of improving the user experience. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 is a flowchart of a user dynamic sorting recommendation algorithm according to the application;

[0050] Figure 2 is a schematic diagram of the bucket writing and reading dynamic list in the user dynamic sorting recommendation algorithm according to the application. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0052] Please refer to Figure 1

[0053] S1, collecting valid dynamic data from the database;

[0054] Specifically, in this embodiment, the basis for selecting valid dynamic data is that the publication time of the dynamic is within 72 hours from the current time, or the dynamic is liked or commented within 24 hours from the current time. This basis takes into account the timeliness of the dynamic while taking into account the content richness of the dynamic.​

[0055] S2, extracting main feature information in the collected dynamic data as the basis for sorting, and initializing and assigning the weight occupied by each feature information;

[0056] Specifically, the main feature information extracted in the embodiment includes: dynamic creation time, dynamic creator influence, dynamic interactivity, dynamic specific content type, dynamic creator language set and dynamic creator gender.

[0057] It should be noted that the dynamic interactivity is mainly based on the number of likes and the number of comments of the dynamic; the dynamic specific content type mainly includes: video dynamic, voice dynamic, picture dynamic and text dynamic; the dynamic creator language set mainly includes the mother tongue of the dynamic creator and the three languages to be learned by the dynamic creator.

[0058] S3, calculating the score of each feature information of the dynamic according to the specific feature sorting score algorithm;

[0059] Specifically, the feature sorting score algorithm mainly includes the following aspects:

[0060] 1. Dynamic creation time score, according to Newton's cooling law, its decay satisfies a certain inverse relationship, and the following formula is used for calculation:

[0061]

[0062] t now is the time stamp of the current time, t create is the time stamp of the dynamic creation time, S create is the dynamic creation time score, and the formula is used to increase the overall score of the newly published dynamic, so as to avoid that the newly published dynamic is sorted at the back due to the lack of interactivity score.

[0063] Dynamic interactivity score, which focuses on the number of comments of the dynamic. The number of comments and the number of likes are calculated according to the weight ratio of 5:1.

[0064] Dynamic specific content type score, which focuses on the richness of the content, that is, video, language and picture dynamic will hold higher score than text dynamic.

[0065] The language set score of the dynamic creator is calculated in association with the language set of the requester, so M data sets are created in this step, each of which represents a possible language set of the requester. Specifically, if the language to be learned by the requester contains the native language of the dynamic creator, the highest score is given. Secondly, if the language to be learned by the dynamic creator contains the native language of the requester, different scores are given according to the priority of the language to be learned. According to this algorithm, each possible case is calculated and placed in the corresponding data set.

[0066] The gender score of the dynamic creator, specifically, a higher score is given when the requester and the dynamic creator are of different genders than when they are of the same gender. Therefore, in this step, the above M data sets are further divided to generate N data sets containing the language set of the requester and the gender of the requester as characteristic information.

[0067] S4, the calculated dynamic characteristic information scores are normalized and processed to obtain uniform standard score data. The characteristic information total score of the dynamic is obtained by weighted summation calculation according to the weight of each characteristic information;

[0068] Specifically, the score normalization uses the following formula:

[0069] In the formula, S feature is the score of one of the characteristic information of the dynamic, S max is the maximum score that can be obtained by the characteristic information, S min is the minimum score that can be obtained by the characteristic information, and S

[0070] S5, all dynamics are written into redis in descending order according to the characteristic information total score;

[0071] Specifically, as shown in Figure 2 , the dynamic data set is processed by the sorting recommendation algorithm to generate N data sets distinguished by characteristic information. The N data sets correspond to the N buckets divided in redis, and each bucket is distinguished by the characteristic information corresponding to its data set. Figure 2 In the formula, B1, B2, B3, B4,..., represent the N buckets.

[0072] It should be noted that the characteristic information used for data set differentiation in the present example includes the language set of the requester and the gender of the requester. The dynamic list held by the data set in each bucket is consistent, and the different is the arrangement order of the dynamic list in the bucket. This approach makes the user prefer to see the dynamic that is biased to his own preference, and does not cause the loss of the remaining dynamic data.

[0073] S6, the user requests a dynamic list to the dynamic recommendation program, and the dynamic recommendation program takes the dynamic list from the corresponding bucket of redis according to the feature information and returns it to the user;

[0074] Specifically, as shown in Figure 2 , the user will carry his own feature information when requesting a dynamic list, and the dynamic recommendation program will find the corresponding bucket according to the user's feature information. Therefore, the dynamic list returned by this method is biased towards the user's preferences.

[0075] S7, the sorting recommendation algorithm takes data from the database for recalculation and sorting at regular intervals.

[0076] Specifically, as shown in Figure 1 , the server will periodically obtain new dynamic data information from the database and repeat the above specific algorithm steps to update the dynamic list held in the bucket, ensuring that the user obtains the new dynamic list in a timely manner.

[0077] It should be noted that, as Figure 1 shown, when creating a dynamic, the user will transmit the dynamic data to the database, and the server obtains the user's latest uploaded dynamic. When the user requests a dynamic list, the dynamic list is directly obtained from the bucket that has been pre-computed and sorted, without the need for re-computation, thus shortening the response time of the data and improving the user experience.

[0078] Embodiment 1

[0079] Dynamic 1:

[0080] Creation time: 2024-11-21 12:00:00 (current time is 2024-11-22 09:00:00)

[0081] Number of comments: 8

[0082] Number of likes: 2

[0083] Content type: video

[0084] Creator language set: Chinese

[0085] Creator gender: male

[0086] User 1:

[0087] Language set: Chinese

[0088] Gender: female

[0089] Embodiment 2

[0090] Dynamic 2:

[0091] Creation time: 2024-11-22 08:00:00

[0092] Number of comments: 5

[0093] Number of likes: 3

[0094] Content type: image

[0095] Creator language set: English

[0096] Creator gender: female

[0097] User 2:

[0098] Language set: English

[0099] Gender: male

[0100] Calculation process

[0101] Calculation of dynamic 1

[0102] Creation time score:

[0103] t now = 2024-11-2209:00:00;

[0104] t create = 2024-11-2112:00:00;

[0105] t now -t create = 21 hours;

[0106]

[0107] Interactivity score:

[0108] S interaction = 5×8 + 1×2 = 42;

[0109] Content type score:

[0110] Video type dynamic score is higher, which is 0.9.

[0111] Language set score:

[0112] The creator's language set matches the user's language set, which is 1.0.

[0113] Gender score:

[0114] The creator and the user have different genders, which is 1.2.

[0115] Normalization processing:

[0116] The maximum score is 1.0 and the minimum score is 0.0.

[0117] The normalized score is:

[0118] The comprehensive score is:

[0119] S total = α · 42+ β · (0.171 + 0.9 + 1.0 + 1.2);

[0120] Wherein α = 0.7, β = 0.3;

[0121] The comprehensive score is: total = 0.7 · 42+ 0.3 · 3.271 ≈ 29.4 + 0.9813 ≈ 30.3813;

[0122] The calculation of dynamic 2

[0123] The creation time score is:

[0124] t now = 2024-11-22 09:00:00;

[0125] t create = 2024-11-22 08:00:00;

[0126] t now -t create = 1 hour;

[0127]

[0128] The interaction score is:

[0129] S interaction = 5 × 5 + 1 × 3 = 28;

[0130] The content type score is:

[0131] The picture type dynamic score is higher, which is 0.8.

[0132] The language set score is:

[0133] The creator language set matches the user language set, which is 1.0.

[0134] The gender score is:

[0135] The creator and the user gender are different, which is 1.2.

[0136] The normalization processing is:

[0137] Wherein the maximum score is 1.0, and the minimum score is 0.0.

[0138] The normalized score is:

[0139] The comprehensive score is:

[0140] Stotal = a - 28 + β - (0.913 + 0.8 + 1.0 + 1.2);

[0141] wherein: a = 0.7, β = 0.3;

[0142] S total = 0.7 - 28 + 0.3 - 3.913 = 19.6 + 1.1739 = 20.7739;

[0143] New embodiments

[0144] Embodiment 1

[0145] Dynamic 1:

[0146] Creation time: 2024-11-21 12:00:00

[0147] Number of comments: 8

[0148] Number of likes: 2

[0149] Content type: video

[0150] Creator language set: Chinese

[0151] Creator gender: male

[0152] Overall score: 30.3813

[0153] User 1:

[0154] Language set: Chinese

[0155] Gender: female

[0156] Embodiment 2

[0157] Dynamic 2:

[0158] Creation time: 2024-11-22 08:00:00

[0159] Number of comments: 5

[0160] Number of likes: 3

[0161] Content type: picture

[0162] Creator language set: English

[0163] Creator gender: female

[0164] Overall score: 20.7739

[0165] User 2:

[0166] Language set: English

[0167] Gender: male

[0168] Through the above calculation, we get two new sets of embodiments, each set of embodiments with a dynamic overall score of 30.3813 and 20.7739, respectively, for dynamic ranking and recommendation to better match the user's interests and preferences.

[0169] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A user dynamic ranking recommendation method, characterized in that: The following steps are involved: Step 1: Collect valid dynamic data from the database; Step 2: Extract the main feature information from the collected dynamic data as the basis for sorting, and initialize the weight of each feature information; Step 3: Calculate the scores of each dynamic feature information according to a specific feature ranking and scoring algorithm; Step 4: normalize the calculated scores of each dynamic feature information to obtain a unified standard score data, and perform weighted sum calculation based on the weight of each feature information to obtain the total score of the dynamic feature information; Step 5: Write all dynamic data into Redis in buckets in reverse order of feature information total score; Step 6: The user requests a dynamic list from the dynamic recommendation program. The dynamic recommendation program retrieves the dynamic list from the corresponding bucket in Redis based on the feature information and returns it to the user. Step 7: The ranking recommendation algorithm periodically retrieves data from the database and recalculates the ranking; The main feature information extracted in step 2 includes: the creation time of the dynamic, the influence of the dynamic creator, the interactivity of the dynamic, the specific content type of the dynamic, the language set of the dynamic creator, and the gender of the dynamic creator; The specific feature ranking and scoring algorithm in step 3 includes the following aspects: S1.

1. The dynamic creation time score, according to Newton's law of cooling, has a decay that satisfies a certain inverse relationship. The calculation formula is: in: The timestamp of the current time, The timestamp of the dynamic creation time, The creation time score of the post. This formula is used to increase the overall score of newly published posts to prevent them from being ranked low due to a lack of interactivity. S1.

2. The dynamic interactivity score is calculated based on the number of dynamic comments, with the number of comments and likes weighted at a 5:1 ratio. The calculation formula is: ;in: Represents the dynamic interactivity score, which is the calculation result; is a variable representing the total number of comments received by the dynamic content; is another variable representing the total number of likes received by the dynamic content; S1.

3. Content-Type Score for Posts: This score is calculated based on the richness of the content. For example, videos, text, and images will have higher scores than text. S1.

4. Dynamic Creator's Language Set Score: This calculation is tied to the requester's language set. Therefore, in this step, M datasets are created, each representing a possible requester's language set. The highest score is assigned to the language set where the requester's native language is included in the dynamic creator's language set. Subsequently, when the dynamic creator's native language is included in the dynamic creator's language set, different scores are assigned based on the priority of the language set. Each possible scenario is calculated separately and placed into the corresponding dataset. S1.

5. Gender score for the dynamic creator: A higher score is given when the requester is of the opposite sex than when the requester is of the same sex. In this step, the M datasets are further segmented to generate N datasets containing the requester's language set and gender as feature information. After the dynamic data set in step 5 is processed by the sorting recommendation algorithm, N data sets distinguished by feature information are generated. These N data sets correspond to N buckets divided in Redis. Each bucket is distinguished by the feature information corresponding to its data set. The dynamic lists held by the data sets in each bucket are consistent, and the difference lies in the arrangement order of the dynamic lists in the bucket.

2. The method for ranking and recommending user dynamics according to claim 1, characterized in that: The basis for valid dynamic data in step 1 is: the dynamic post was published within 72 hours of the current time, or the dynamic post was liked or commented on within 24 hours of the current time.

3. The method for ranking and recommending user dynamics according to claim 2, characterized in that: The normalization calculation in step 4 is: in: is the score of one of the dynamic feature information, is the maximum score that can be obtained for this feature information, is the minimum score that can be obtained for this feature information. is the score corresponding to the feature information after normalization.

4. The method for ranking and recommending user dynamics according to claim 3, characterized in that: In step six, when the user requests the dynamic list, he or she will carry his or her own characteristic information. The dynamic recommendation program will find the corresponding bucket based on the user's characteristic information. The dynamic list returned by this method is biased towards the user's preferences.

5. The method for ranking and recommending user dynamics according to claim 4, characterized in that: In step seven, the server will periodically obtain new dynamic data information from the database and repeat the above-mentioned specific algorithm steps to update the dynamic list held in the bucket to ensure that the user obtains the new dynamic list in a timely manner. In step seven, when the user creates a dynamic, the dynamic data will be transmitted to the database, and the server will obtain the user's latest uploaded dynamic. When the user requests the dynamic list, the dynamic list is directly obtained from the bucket that has been pre-calculated and sorted, without the need for recalculation, thereby shortening the data response time.

6. The user dynamic ranking recommendation method according to claim 5, characterized in that: The formula for calculating the comprehensive score after considering the user feature matching score in step 6 is: in: and are weight parameters, indicating the relative importance of interactivity score and matching score respectively; likes indicates the number of likes received by the dynamic; It is a feature and The similarity function between Represents user feature vector The Features Represents an element in the weight matrix, specifically, it represents the feature and The importance weight between them; Indicates the number of user features; Indicates the number of bucket features.

7. The method for ranking and recommending user dynamics according to claim 6, characterized in that: In step 7, the server executes the following steps in a scheduled task to update the dynamic list and sort: S2.

1. The server obtains the latest uploaded dynamic data from the database. The dynamic data obtained is ; S2.2, the new dynamic data Add to existing dynamic list In the form of a new dynamic list : ; S2.

3. New dynamic list Recalculate the interactivity score for each post in : ; S2.

4. Combining User Characteristics and barrel characteristics , calculate the user feature matching score : ; S2.

5. Recalculate the comprehensive score: Combine the interactivity score and the user feature matching score to recalculate the comprehensive score : ; S2.

6. Update the dynamic list in the bucket: based on the new comprehensive score , reorder the dynamic list and update it to the corresponding bucket: .

Images