A method, electronic device, and storage medium for obtaining target category labels.

Abstract

This invention provides a method, electronic device, and storage medium for obtaining target category tags, relating to the field of APP classification technology. The method obtains a first key tag corresponding to the target APP name based on the category tags of APPs in a preset third-party application download platform. It further obtains a list of second key APP names corresponding to the first key tag. Based on the target category tags of APPs in the target database whose target category tags are not NULL, it obtains a list of second key tags and a list of third key tags from the second key APP name list. The target category tag of the target APP is determined based on the number of second key tags in the second key tag list that are identical to the third key tag. This method reduces the impact of the target APP's download volume and popularity during the target category tag acquisition process, thus improving the accuracy of target category tag acquisition.

Description

Technical Field

[0001] This invention relates to the field of APP classification technology, and in particular to a method, electronic device and storage medium for obtaining target category labels. Background Technology

[0002] In the target database, due to factors such as the rapid upload of new apps, omissions or errors in the automated classification process, delays in manual classification, information loss during data migration or integration, changes in category definitions, inaccurate information submitted by users, and apps whose functions span multiple categories, there are a large number of unclassified apps, i.e., target apps, in the target database. Classifying target apps and obtaining their category tags in the target database, i.e., target category tags, is an effective way to organize and manage target apps. In existing technologies, the method for obtaining target category tags for target apps is to directly reference the category tags in third-party application download platforms, using the category tags of the target app in the third-party application download platform as the target category tags for the target app.

[0003] However, the above method also has the following technical problems:

[0004] For apps with low download numbers or low brand awareness, third-party app download platforms may not have enough attention or resources to provide accurate category tags for the app. Therefore, when the target app has low download numbers or low brand awareness, the accuracy of the target category tags obtained based on the above method is low. Summary of the Invention

[0005] To address the aforementioned technical problems, the technical solution adopted by this invention is as follows:

[0006] According to a first aspect of the present invention, a method for obtaining target category tags is provided, wherein the target category tags are category tags of an APP in a target database, and the method includes the following steps:

[0007] S1. Obtain the target APP name list A = {A1, A2, ..., A...} i A m}, A i Let be the name of the i-th target app, where i ranges from 1 to m, m is the number of target app names, the target app name is the name of the target app, and the target app is the app in the target database whose target category label is NULL.

[0008] S2. Based on the first key APP name list B = {B1, B2, ..., B...} j B n}, obtain the list of first key tags C corresponding to A = {C1, C2, ..., C} iC m}, where, when A i =B j At that time, B j The corresponding third-party category label is C i B j Let C be the name of the j-th primary key app, where j ranges from 1 to n, and n is the number of primary key app names. i For A i The corresponding first key tag, the first key APP name is the name of the first key APP, the first key APP is the APP in the preset third-party application download platform, and the third-party category tag is the category tag of the first key APP in the preset third-party application download platform.

[0009] S3, iterate through B, when B... j The corresponding third-party category tags and C i When they are the same, B j As C i The corresponding second key APP name to obtain C i The corresponding second key APP name list D i ={D i1 D i2 D ie D if(i)}, D ie C i The corresponding e-th second key APP name, where e takes values ​​from 1 to f(i), and f(i) is C. i The number of corresponding second key APP names.

[0010] S4. Based on the intermediate APP name list E = {E1, E2, ..., E...} r , ..., E s}, obtain D i The corresponding second key tag list F i ={F i1 F i2 F ie F if(i)}, and for F i1 F i2 F ie F if(i) Perform deduplication to obtain D i The corresponding third keyword list G i ={G i1 G i2 , ..., G ik , ..., G it(i)}, where, when D ie =E r At that time, Er The corresponding target category label is F ie E r Let F be the name of the r-th intermediate app, where r ranges from 1 to s, s is the number of intermediate app names, the intermediate app name is the name of the intermediate app, and the intermediate app is the app in the target database whose target category label is not NULL. ie D ie The corresponding second key tag, G ik D i The corresponding k-th third key tag, where k takes values ​​from 1 to t(i), and t(i) is D. i The number of corresponding third key tags.

[0011] S5, when G 0 ik =max(G 0 i1 G 0 i2 , ..., G 0 ik , ..., G 0 it(i) When ), G ik As A i The corresponding target category tag for the target app, G 0 ik For F i1 F i2 F ie F if(i) China and G ik The number of identical second key tags, max() is the function to get the maximum value.

[0012] According to a second aspect of the present invention, a non-transitory computer-readable storage medium is provided, wherein a computer program is stored in the storage medium, and the computer program is loaded and executed by a processor to implement the aforementioned method.

[0013] According to a third aspect of the present invention, an electronic device is provided, comprising: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned method.

[0014] The present invention has at least the following beneficial effects:

[0015] This invention provides a method, electronic device, and storage medium for obtaining target category tags. The method obtains a first key tag corresponding to a target app name based on a third-party category tag corresponding to a first key app name. The first key app is an app on a preset third-party application download platform, and the third-party category tag is the category tag of the first key app on the preset third-party application download platform. Further, it obtains a list of second key app names corresponding to the first key tag. Based on the target category tag of an intermediate app name, it obtains a list of second key tags and a list of third key tags from the list of second key app names. The intermediate app is an app in the target database whose target category tag is not NULL. The target category tag of the target app is determined based on the number of second key tags in the second key tag list that are identical to the third key tag. It can be seen that the above method obtains the target category tag of the target app based on the category tag of the first key app on the preset third-party application download platform and the target category tag corresponding to the intermediate app name. In the process of obtaining the target category tag, the impact of the target app's download volume and popularity is reduced, which helps to improve the accuracy of obtaining the target category tag. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a method for obtaining target category labels according to an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar tasks and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that comprises a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0020] Embodiments of the present invention provide a method for obtaining target category tags, wherein the target category tags are category tags of an APP in a target database, and the method includes the following steps, such as... Figure 1 As shown:

[0021] S1. Obtain the target APP name list A = {A1, A2, ..., A...} i A m}, A i Let be the name of the i-th target app, where i ranges from 1 to m, m is the number of target app names, the target app name is the name of the target app, and the target app is the app in the target database whose target category label is NULL.

[0022] Specifically, the target database includes several apps and app information for each app. The app information includes at least the app name, the name of the corresponding installation package, and the app's target category tag. The above are just some examples of app information. There may be other app information, such as the size of the corresponding installation package and the version of the corresponding installation package, which will not be elaborated here.

[0023] Furthermore, the target category label for the app can be NULL.

[0024] Specifically, category tags are used to describe the type of app, such as entertainment, learning, work, and sports, which will not be elaborated further here.

[0025] S2. Based on the first key APP name list B = {B1, B2, ..., B...} j B n}, obtain the list of first key tags C corresponding to A = {C1, C2, ..., C} i C m}, where, when A i =B jAt that time, B j The corresponding third-party category label is C i B j Let C be the name of the j-th primary key app, where j ranges from 1 to n, and n is the number of primary key app names. i For A i The corresponding first key label, the first key APP name is the name of the first key APP, the first key APP is the APP in the preset third-party application download platform, and the third-party category label is the category label of the first key APP in the preset third-party application download platform. As those skilled in the art know, the preset third-party application download platform is a third-party application download platform that is predetermined by those skilled in the art according to actual needs, and will not be described in detail here.

[0026] S3, iterate through B, when B... j The corresponding third-party category tags and C i When they are the same, B j As C i The corresponding second key APP name to obtain C i The corresponding second key APP name list D i ={D i1 D i2 D ie D if(i)}, D ie C i The corresponding e-th second key APP name, where e takes values ​​from 1 to f(i), and f(i) is C. i The number of corresponding second key APP names.

[0027] S4. Based on the intermediate APP name list E = {E1, E2, ..., E...} r , ..., E s}, obtain D i The corresponding second key tag list F i ={F i1 F i2 F ie F if(i)}, and for F i1 F i2 F ie F if(i) Perform deduplication to obtain D i The corresponding third keyword list G i ={G i1 G i2 , ..., G ik , ..., G it(i)}, where, when D ie =E rAt that time, E r The corresponding target category label is F ie E r Let F be the name of the r-th intermediate app, where r ranges from 1 to s, s is the number of intermediate app names, the intermediate app name is the name of the intermediate app, and the intermediate app is the app in the target database whose target category label is not NULL. ie D ie The corresponding second key tag, G ik D i The corresponding k-th third key tag, where k takes values ​​from 1 to t(i), and t(i) is D. i The number of corresponding third key tags is known to those skilled in the art. Any deduplication method in the prior art falls within the protection scope of this invention, and will not be elaborated here.

[0028] Specifically, if D ie With E1, E2, ..., E r , ..., E s If they are all different, then determine F. ie NULL.

[0029] S5, when G 0 ik =max(G 0 i1 G 0 i2 , ..., G 0 ik , ..., G 0 it(i) When ), G ik As A i The corresponding target category tag for the target app, G 0 ik For F i1 F i2 F ie F if(i) China and G ik The number of identical second key tags, max() is the function to get the maximum value.

[0030] Through the above steps, the first key tag corresponding to the target APP name is obtained based on the third-party category tag corresponding to the first key APP name. The first key APP is an APP in the preset third-party application download platform, and the third-party category tag is the category tag of the first key APP in the preset third-party application download platform. Further, a list of second key APP names corresponding to the first key tag is obtained. Based on the target category tag of the intermediate APP name, a list of second key tags and a list of third key tags are obtained from the list of second key APP names. The intermediate APP is an APP in the target database whose target category tag is not NULL. Based on the number of second key tags in the second key tag list that are the same as the third key tag, the target category tag of the target APP is determined. In the process of obtaining the target category tag, the impact of the target APP's download volume and popularity is reduced, which helps to improve the accuracy of obtaining the target category tag.

[0031] In one specific embodiment, the method further includes the following steps S10-S50 after step S2:

[0032] S10, when B1, B2, ..., B j B n With A i When they are all different, place A i In the corresponding target installation package name string, the substring following the first preset separator from right to left is used as A. i The corresponding first string R i The target installation package name string is a string representing the name of the installation package corresponding to the target APP, containing multiple substrings connected by a preset separator. For example, if the preset separator is ".", and A... i The corresponding target installation package name string is abcd.1234.firewall, then R i For firewall.

[0033] S20. Based on the preset string list set J = {J1, J2, ..., J...} g , ..., J h}, obtain R i The corresponding second string list, where when K i ≤max(J 0 1, J 0 2, ..., J 0 g , ..., J 0 h And K i =J 0 g At that time, J g As R iThe corresponding second string list, J g Let K be the g-th preset string list, where g ranges from 1 to h, and h is the number of preset string lists. Each preset string list contains several preset strings. i For R i The length of the string, J 0 g For J g The corresponding preset string length, J g The length of all preset strings is equal to J 0 g Similarly, the second string list includes several second strings, with the default string being one from which information can be extracted. For example: firewall, talkback.

[0034] S30. Obtain the third string list L = {L1, L2, ..., L...} corresponding to E. r , ..., L s}, L r For E r The corresponding third string, E r The corresponding third string is E r The substring following the first preset separator from right to left in the corresponding intermediate installation package name string is the string that represents the installation package name of the intermediate APP, containing multiple substrings connected by preset separators.

[0035] S40, when R i With R i If any second string in the corresponding second string list is the same, get R. i The corresponding third key APP name list M i ={M i1 M i2 M ix M ip(i)} and M i The corresponding fourth key tag list N i ={N i1 N i2 ,…,N ix ,…,N ip(i)}, where will be with R i The same L r The corresponding E r As R i The corresponding third key APP name, will be the E r The corresponding target category tag serves as the fourth key tag corresponding to the third key APP name, M ix For R iThe corresponding x-th third key APP name, where x takes values ​​from 1 to p(i), and p(i) is R. i The corresponding third key number of APP names, N ix For M ix The corresponding fourth key tag.

[0036] S50, N i1 N i2 ,…,N ix ,…,N ip(i) Perform deduplication to obtain R i The corresponding first intermediate tag list P i ={P i1 P i2 , ..., P iy , ..., P iq(i)} and based on P i Get A i The corresponding target category tag for the target app, where P iy For R i The corresponding y-th first intermediate label, where y takes values ​​from 1 to q(i), and q(i) is R. i The number of corresponding first intermediate tags, when P 0 iy =max(P 0 i1 P 0 i2 , ..., P 0 iy , ..., P 0 iq(i) When P is used, iy As A i The corresponding target category tag for the target app, P 0 iy For N i1 N i2 ,…,N ix ,…,N ip(i) China and P iy The number of identical fourth key tags.

[0037] Through the above steps, when the target APP name is different from any of the first key APP names, the first string corresponding to the target APP name is obtained. Then, a second string list corresponding to the first string is obtained based on a preset string list set. Finally, a third string corresponding to the intermediate APP name is obtained. Based on the first string, the second string list, and the third string, a third key APP name list and a fourth key tag list corresponding to the first string are obtained. A first intermediate tag list corresponding to the first string is obtained based on the fourth key tag list corresponding to the first string. Finally, based on the number of fourth key tags in the fourth key tag list that are identical to the first intermediate tags, the target category tag corresponding to the target APP is obtained. Even if the target APP is not included in the preset third-party application download platform, the target category tag of the target APP can still be obtained, which helps improve the accuracy of obtaining the target category tag.

[0038] In one specific embodiment, the following steps S100-S500 are further included after step S40:

[0039] S100, when K i >max(J 0 1, J 0 2, ..., J 0 g , ..., J 0 h ) or R i With R i If all the second strings in the corresponding second string list are different, proceed to step S200.

[0040] S200. When c(i)≥H, obtain A. i The corresponding list of specified device IDs Q i ={Q i1 Q i2 Q ia Q ic(i)}, where c(i) is A i The corresponding number of specified device IDs, where H is the preset device quantity threshold, and Q is the number of device IDs. ia For A i The corresponding a-th specified device ID, where a ranges from 1 to c(i), A i The corresponding designated device is the one with A installed. i The corresponding target app's device.

[0041] Specifically, the specified device ID is the identity identifier of the specified device.

[0042] S300, according to Q i Get Q i The corresponding list of specified APP names Si ={S i1 S i2 S ia S ic(i)}, where S ia For Q ia The corresponding list of specified APP names includes several specified APP names. The specified APP name is the name of the specified APP, and the specified APP is the APP currently installed on the specified device.

[0043] S400: Use the target category label corresponding to the intermediate APP name that is the same as the specified APP name as the second intermediate label corresponding to the specified APP name, to obtain S i The corresponding second intermediate tag list set S 0 i ={S 0 i1 S 0 i2 S 0 ia S 0 ic(i)}, S 0 ia For S ia The corresponding second intermediate tag list, S 0 ia Including S ia The second intermediate tag corresponding to each specified APP name in the file.

[0044] Specifically, if all intermediate APP names in the intermediate APP name list are different from the specified APP name, then the second intermediate tag corresponding to the specified APP name is determined to be NULL.

[0045] S500, for S 0 i1 S 0 i2 S 0 ia S 0 ic(i) Deduplicat all second intermediate tags to obtain S 0 i The corresponding third intermediate tag list T i ={T i1 T i2 ,…,T ib ,…,T id(i)}, and according to S 0 i and T i Get A iThe corresponding target category tag for the target app, where T ib For S 0 i The corresponding b-th third intermediate label, where b takes values ​​from 1 to d(i), and d(i) is S. 0 i The number of corresponding third intermediate tags.

[0046] Specifically, the steps are based on S 0 i and T i Get A i The target category tags for the corresponding target app also include the following steps S501-S504:

[0047] S501, Obtain S 0 ia China T ib The corresponding TGI value U aib U aib The following conditions must be met:

[0048] U aib =V aib / V 0 ia / (∑ c(i) a＝1 V aib / ∑ c(i) a＝1 V 0 ia )×100, where V aib For S 0 ia In and T ib The same number of second intermediate labels, V 0 ia For S 0 ia The number of the second middle tags.

[0049] S502, Obtain S 0 ia China T ib The corresponding target weight W aib W aib The following conditions must be met:

[0050] W aib =V aib / V 0 ia .

[0051] Specifically, the target weight is used to represent the importance of the third intermediate label. The larger the target weight, the more important the corresponding third intermediate label.

[0052] S503, according to U aib and W aib Get T ib The corresponding weighted average TGI value Z ib Z ib The following conditions must be met:

[0053] Z ib =∑ c(i) a＝1 (W aib ×U aib ) / ∑ c(i) a＝1 W aib .

[0054] S504, when Z ib =max(Z) i1 Z i2 , ..., Z ib , ..., Z id(i) When ), T ib As A i The target category tag for the corresponding target app.

[0055] Through the above steps, when the list of second strings corresponding to the first string cannot be obtained, or when the first string is different from all the second strings in its corresponding list, the number of specified device IDs corresponding to the target APP name is obtained. When the number of specified device IDs corresponding to the target APP name is not less than a preset device number threshold, the list of specified APP names corresponding to the list of specified device IDs corresponding to the target APP name is obtained. Based on the list of specified APP names and the target category tags corresponding to the intermediate APP names, a second set of intermediate tag lists is obtained. Further, a third list of intermediate tags is obtained. Based on the TGI value and target weight of the third intermediate tag in the second list of intermediate tags, the weighted average TGI value of the third intermediate tag is obtained. Based on the weighted average TGI value of the third intermediate tag, the target category tag of the target APP is obtained. Even if the target APP is not included in the preset third-party application download platform, the target category tag of the target APP can still be obtained, and the impact of the target APP's download volume and popularity is reduced, which helps to improve the accuracy of obtaining the target category tag.

[0056] In one specific embodiment, the steps S1000-S5000 are further included after step S200:

[0057] S1000, When c(i) < H, A i As the name of the app to be categorized, A iThe corresponding target APP is used as the APP to be classified, so as to obtain a list of APP names to be classified, which includes several APP names to be classified.

[0058] S2000. Obtain the initial user ID list X = {X1, X2, ..., X...} based on the list of APP names to be categorized. u , ..., X v}, X u Let u be the initial user ID, where u ranges from 1 to v, and v is the number of initial user IDs. The initial user is the user using the key device, and the key device is the device that has installed any of the app names in the list of app names to be categorized.

[0059] Specifically, the initial user ID is the unique identifier of the initial user.

[0060] S3000. Obtain the initial user tag list set Y = {Y1, Y2, ..., Y} corresponding to X. u , ..., Y v}, where Y u For X u The corresponding initial user tag list, Y u Includes X u The corresponding initial user tags are as follows: the initial user tag corresponding to the initial user ID is the user tag of the initial user corresponding to the initial user ID. As those skilled in the art know, any method of obtaining user tags in the prior art is within the protection scope of this invention, and will not be described in detail here.

[0061] Specifically, the initial user ID is the unique identifier of the initial user.

[0062] S4000, Based on the K-means clustering algorithm and Y, obtain the total set JL = {JL} of the clustered user tag list. (1) JL (2) , ..., JL (ai) , ..., JL (am)}, JL (ai) ={JL (ai)(1) JL (ai)(2) , ..., JL (ai)(aj) , ..., JL (ai)(an)}, where JL (ai) Let JL be the set of user label lists for the ai-th cluster, where ai ranges from 1 to am, and am is the number of user label list sets for the cluster. (ai)(aj) For JL (ai)The set contains the aj-th clustered user label list, where aj ranges from 1 to an, and an is the number of clustered user label lists in the set. Each clustered user label list contains several clustered user labels.

[0063] Specifically, the value range of am is [3, 5].

[0064] Specifically, step S4000 includes the following steps S4100-S4300:

[0065] S4100, Y u Perform vectorization to obtain Y u The corresponding first label vector Y 0 u Y 0 u The dimension and Y u The number of initial user tags is the same. As those skilled in the art know, any vectorization processing method in the prior art is within the protection scope of this invention, and will not be elaborated here.

[0066] S4200, Using K-means clustering algorithm to analyze Y 0 1, Y 0 2, ..., Y 0 u , ..., Y 0 v Cluster the corresponding label vectors to obtain the second label vector list Y corresponding to Y. 1 ={Y 1 (1) Y 1 (2) , ..., Y 1 (ai) , ..., Y 1 (am)}, Y 1 (ai) ={Y 1 (ai)(1) Y 1 (ai)(2) , ..., Y 1 (ai)(aj) , ..., Y 1 (ai)(an)}, Y 1 (ai) Let Y be the list of the ai-th second label vectors corresponding to Y. 1 (ai)(aj) For Y 1 (ai) The aj-th second label vector in the cluster is the first label vector after clustering. A list of second label vectors can be understood as a cluster.

[0067] S4300, when Y 0 u =Y 1 (ai)(aj) At that time, Y u As JL (ai)(aj) .

[0068] Through the above steps, the initial user tag list is vectorized to obtain the first tag vectors corresponding to the initial user tag list. All the first tag vectors are clustered to obtain the second tag vector list set. This can cluster the similar first tag vectors into the same cluster. Further obtaining the total set of clustered user tag lists helps to improve the accuracy of obtaining the total set of clustered user tag lists. Based on the initial user tag list set and the total set of clustered user tag lists, the target category tag of the APP to be classified is obtained. Even if the target APP is not included in the third-party application download platform, the target category tag of the target APP can still be obtained, and the impact of the target APP's download volume and popularity is reduced, which helps to improve the accuracy of obtaining the target category tag.

[0069] S5000, based on Y and JL, obtains the target category label of the app to be classified corresponding to the name of all apps to be classified.

[0070] Step S5000 further includes the following steps S5100-S5400:

[0071] S5100, for Y1, Y2, ..., Y u , ..., Y v All initial user tags are deduplicated to obtain an intermediate user tag list ZJ = {ZJ (1) ZJ (2) , ..., ZJ (ae) , ..., ZJ (af)}, ZJ (ae) Let be the 'ae'-th intermediate user tag, where 'ae' ranges from 1 to 'af', and 'af' is the number of intermediate user tags.

[0072] S5200, obtain JL (ai) China ZJ (ae) The corresponding TGI value tgi (ai)(ae) ,tgi (ai)(ae) The following conditions must be met:

[0073] tgi (ai)(ae) =∑ an aj＝1 AA (ae) (ai)(aj) / ∑ an aj＝1 AB (ai)(aj) / (∑am ai＝1 ∑ an aj＝1 AA (ae) (ai)(aj) / ∑ am ai＝1 ∑ an aj＝1 AB (ai)(aj) )×100,

[0074] AA (ae) (ai)(aj) For JL (ai)(aj) Zhong and ZJ (ae) The number of identical clustered user labels, AB (ai)(aj) For JL (ai)(aj) The number of user tags in the cluster.

[0075] S5300, according to TGI (ai)(ae) Get ZJ (ae) The corresponding total TGI value Z-tgi (ae) Z-tgi (ae) The following conditions must be met:

[0076] Z-tgi (ae) =∑ am ai＝1 tgi (ai)(ae) .

[0077] S5400, when Z-tgi (ae) =max(Z-tgi) (1) Z-tgi (2) , ..., Z-tgi (ae) , ..., Z-tgi (af) When ZJ (ae) do

[0078] The target category label for all apps to be categorized.

[0079] Through the above steps, when the number of specified device IDs corresponding to the target APP name is less than a preset device number threshold, the target APP name is taken as the APP name to be classified, and the target APP corresponding to the target APP name is taken as the APP to be classified. An initial user ID list is obtained from the list of APP names to be classified. Further, an initial user tag list set corresponding to the initial user ID list is obtained. Based on the K-means clustering algorithm and the initial user tag list set, the total set of clustered user tag lists is obtained. An intermediate user tag list is obtained from the initial user tag list. Based on the number of clustered user tags in the clustered user tag list that are the same as the intermediate user tags and the number of clustered user tags in the clustered user tag list, the total TGI value corresponding to the intermediate user tags is obtained. Based on the total TGI value corresponding to the intermediate user tags, the target category tag of the APP to be classified is obtained. Even if the target APP is not included in the preset third-party application download platform, the target category tag of the target APP can still be obtained, and the impact of the target APP's download volume and popularity is reduced, which is conducive to improving the accuracy of obtaining the target category tag.

[0080] In one specific embodiment, step S200 further includes the following steps S210-S250 to obtain H:

[0081] S210. Obtain the second preset quantity value γ corresponding to the first preset quantity value θ, where θ is initially 1 and γ = θ + 1.

[0082] S220. Obtain the candidate device ID list set AC = {AC1, AC2, ..., AC...} corresponding to E. r AC s}, where AC r For E r The corresponding candidate device ID list, E r The corresponding candidate device ID list includes E r The corresponding β candidate device IDs, E r The corresponding candidate device is one that has E installed. r The corresponding intermediate APP device.

[0083] Specifically, the candidate device ID is a unique identifier for the candidate device.

[0084] S220, Let β = θ and use the method described in steps S300-S500 based on Q. i , obtain A i The target category tags of the corresponding target apps are in the same way, according to AC. r Get E r The first candidate category label for the corresponding intermediate app is AD. rTo obtain the first candidate category label AD = {AD1, AD2, ..., AD} corresponding to E. r AD s}; can be understood as Q in step S300 i Replace with AC r And execute steps S300-S500 to obtain E r The first candidate category label for the corresponding intermediate app is AD. r This will not be elaborated upon here.

[0085] S230, Let β = γ and use the method described in steps S300-S500 based on Q. i , obtain A i The target category tags of the corresponding target apps are in the same way, according to AC. r Get E r The corresponding intermediate APP's second candidate category label AD r To obtain the second candidate category label AF = {AF1, AF2, ..., AF3} corresponding to E. r , ..., AF s}

[0086] S240. Obtain the target category label list AE = {AE1, AE2, ..., AE...} corresponding to E. r , ..., AE s}, where AE r For E r The target category tag for the corresponding target app.

[0087] S250, obtain H based on AD, AF, and AE.

[0088] Specifically, step S250 also includes the following steps S251-S253:

[0089] S251. Obtain the accuracy score AG corresponding to θ. θ AG θ The following conditions must be met:

[0090] AG θ =∑ s r＝1 AH r / s, AH r For E r The corresponding first precision weight, where, when AD r =AE r At that time, AH r =1, when AD r ≠AE r At that time, AH r =0.

[0091] S252. Obtain the accuracy score AG corresponding to γ. γ AG γ The following conditions must be met:

[0092] AG γ =∑ s r＝1 AR r / s, AR r For E r The corresponding second precision weight, where, when AF r =AE r At that time, AR r =1, when AF r ≠AE r At that time, AR r =0.

[0093] S253, being AG θ >AJ 0 And |AG γ -AG θ |<AJ 1 If H = θ, then let θ = θ + 1 to update θ and proceed to step S210, where AJ 0 AJ 1 The preset accuracy difference is known to those skilled in the art, and they can set AJ according to actual needs. 0 and A.J. 1 The specific value will not be elaborated here.

[0094] Through the above steps, a candidate device ID list set corresponding to the intermediate APP name list is obtained. The candidate device IDs in the candidate device ID list are set as a first preset quantity value. The first candidate category label corresponding to the intermediate APP is obtained. The candidate device IDs in the candidate device ID list are set as a second preset quantity value. The second candidate category label corresponding to the intermediate APP is obtained. Based on the first candidate category label, the second candidate category label, and the target category label of the intermediate APP, the accuracy score corresponding to the first preset quantity value and the accuracy score corresponding to the second preset quantity value are obtained. The preset device quantity threshold is obtained based on the accuracy score corresponding to the first preset quantity value and the accuracy score corresponding to the second preset quantity value, which helps to improve the accuracy of obtaining the preset device quantity threshold.

[0095] Embodiments of the present invention also provide a non-transitory computer-readable storage medium that can be disposed in an electronic device to store a computer program related to implementing a method in the method embodiments, the computer program being loaded and executed by the processor to implement the method provided in the above embodiments.

[0096] Embodiments of the present invention also provide an electronic device, including: a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method provided in the above embodiments.

[0097] Embodiments of the present invention also provide a computer program product including program code, which, when the program product is run on an electronic device, causes the electronic device to perform the steps of the methods described above in various exemplary embodiments of the present invention.

[0098] This invention provides a method, electronic device, and storage medium for obtaining target category tags. The method obtains a first key tag corresponding to a target app name based on a third-party category tag corresponding to a first key app name. The first key app is an app on a preset third-party application download platform, and the third-party category tag is the category tag of the first key app on the preset third-party application download platform. Further, it obtains a list of second key app names corresponding to the first key tag. Based on the target category tag of an intermediate app name, it obtains a list of second key tags and a list of third key tags from the list of second key app names. The intermediate app is an app in the target database whose target category tag is not NULL. The target category tag of the target app is determined based on the number of second key tags in the second key tag list that are identical to the third key tag. It can be seen that the above method obtains the target category tag of the target app based on the category tag of the first key app on the preset third-party application download platform and the target category tag corresponding to the intermediate app name. In the process of obtaining the target category tag, the impact of the target app's download volume and popularity is reduced, which helps to improve the accuracy of obtaining the target category tag.

[0099] While specific embodiments of the invention have been described in detail by way of examples, those skilled in the art should understand that the examples are for illustrative purposes only and are not intended to limit the scope of the invention. Those skilled in the art should also understand that various modifications can be made to the embodiments without departing from the scope and spirit of the invention.

Claims

1. A method for obtaining target category labels, characterized in that, The target category label is the category label of the APP in the target database, and the method includes the following steps: S1. Obtain the target APP name list A = {A1, A2, ..., A...} i A m }, A i Let i be the name of the i-th target app, where i ranges from 1 to m, m is the number of target app names, the target app name is the name of the target app, and the target app is the app in the target database whose target category label is NULL; S2. Based on the first key APP name list B = {B1, B2, ..., B...} j B n }, obtain the list of first key tags C corresponding to A = {C1, C2, ..., C} i C m }, where, when A i =B j At that time, B j The corresponding third-party category label is C i B j Let C be the name of the j-th primary key app, where j ranges from 1 to n, and n is the number of primary key app names. i For A i The corresponding first key tag, the first key APP name is the name of the first key APP, the first key APP is the APP in the preset third-party application download platform, and the third-party category tag is the category tag of the first key APP in the preset third-party application download platform; S3, iterate through B, when B... j The corresponding third-party category tags and C i When they are the same, B j As C i The corresponding second key APP name to obtain C i The corresponding second key APP name list D i ={D i1 D i2 D ie D if(i) }, D ie C i The corresponding e-th second key APP name, where e takes values ​​from 1 to f(i), and f(i) is C. i The number of corresponding second key APP names; S4. Based on the intermediate APP name list E = {E1, E2, ..., E...} r , ..., E s }, obtain D i The corresponding second key tag list F i ={F i1 F i2 F ie F if(i) }, and for F i1 F i2 F ie F if(i) Perform deduplication to obtain D i The corresponding third keyword list G i ={G i1 G i2 , ..., G ik , ..., G it(i) }, where, when D ie =E r At that time, E r The corresponding target category label is F ie E r Let F be the name of the r-th intermediate app, where r ranges from 1 to s, s is the number of intermediate app names, the intermediate app name is the name of the intermediate app, and the intermediate app is the app in the target database whose target category label is not NULL. ie D ie The corresponding second key tag, G ik D i The corresponding k-th third key tag, where k takes values ​​from 1 to t(i), and t(i) is D. i The number of corresponding third key tags; S5, when G 0 ik =max(G 0 i1 G 0 i2 , ..., G 0 ik , ..., G 0 it(i) When ), G ik As A i The corresponding target category tag for the target app, G 0 ik For F i1 F i2 F ie F if(i) China and G ik The number of identical second key tags, max() is the function to get the maximum value.

2. The method for obtaining target category labels according to claim 1, characterized in that, The target database includes several apps and app information for each app. The app information includes at least the app name, the name of the corresponding installation package, and the app's target category tag.

3. The method for obtaining target category tags according to claim 1, characterized in that, If D ie With E1, E2, ..., E r , ..., E s If they are all different, then determine F. ie NULL.

4. The method for obtaining target category tags according to claim 2, characterized in that, Following step S2, the following steps S10-S50 are also included: S10, when B1, B2, ..., B j B n With A i When they are all different, A i In the corresponding target installation package name string, the substring following the first preset separator from right to left is used as A. i The corresponding first string R i The target installation package name string is a string that represents the name of the installation package corresponding to the target APP, which contains multiple substrings and is connected by a preset delimiter. S20. Based on the preset string list set J = {J1, J2, ..., J...} g , ..., J h }, obtain R i The corresponding second string list, where when K i ≤max(J 0 1, J 0 2, ..., J 0 g , ..., J 0 h And K i =J 0 g At that time, J g As R i The corresponding second string list, J g Let K be the g-th preset string list, where g ranges from 1 to h, and h is the number of preset string lists. Each preset string list contains several preset strings. i For R i The length of the string, J 0 g For J g The corresponding preset string length, J g The length of all preset strings is equal to J 0 g Similarly, the second string list includes several second strings, with the default string being a string from which information can be extracted; S30. Obtain the third string list L = {L1, L2, ..., L...} corresponding to E. r , ..., L s }, L r For E r The corresponding third string, E r The corresponding third string is E r The substring following the first preset separator from right to left in the corresponding intermediate installation package name string is the string that represents the installation package name of the intermediate APP, which contains multiple substrings and is connected by preset separators. S40, when R i With R i If any second string in the corresponding second string list is the same, get R. i The corresponding third key APP name list M i ={M i1 M i2 M ix M ip(i) } and M i The corresponding fourth key tag list N i ={N i1 N i2 ,…,N ix ,…,N ip(i) }, where will be with R i The same L r The corresponding E r As R i The corresponding third key APP name, will be the E r The corresponding target category tag serves as the fourth key tag corresponding to the third key APP name, M ix For R i The corresponding x-th third key APP name, where x takes values ​​from 1 to p(i), and p(i) is R. i The corresponding third key number of APP names, N ix For M ix The corresponding fourth key tag; S50, N i1 N i2 ,…,N ix ,…,N ip(i) Perform deduplication to obtain R i The corresponding first intermediate tag list P i ={P i1 P i2 , ..., P iy , ..., P iq(i) } and based on P i Get A i The corresponding target category tag for the target app, where P iy For R i The corresponding y-th first intermediate label, where y takes values ​​from 1 to q(i), and q(i) is R. i The number of corresponding first intermediate tags, when P 0 iy =max(P 0 i1 P 0 i2 , ..., P 0 iy , ..., P 0 iq(i) When P is used, iy As A i The corresponding target category tag for the target app, P 0 iy For N i1 N i2 ,…,N ix ,…,N ip(i) China and P iy The number of identical fourth key tags.

5. The method for obtaining target category labels according to claim 4, characterized in that, The steps S100-S500 are included after step S40: S100, when K i >max(J 0 1, J 0 2, ..., J 0 g , ..., J 0 h ) or R i With R i If all the second strings in the corresponding second string list are different, proceed to step S200; S200. When c(i)≥H, obtain A. i The corresponding list of specified device IDs Q i ={Q i1 Q i2 Q ia Q ic(i) }, where c(i) is A i The corresponding number of specified device IDs, where H is the preset device quantity threshold, and Q is the number of device IDs. ia For A i The corresponding a-th specified device ID, where a ranges from 1 to c(i), A i The corresponding designated device is the one with A installed. i The corresponding target app device; S300, according to Q i Get Q i The corresponding list of specified APP names S i ={S i1 S i2 S ia S ic(i) }, where S ia For Q ia The corresponding list of specified APP names includes several specified APP names. The specified APP name is the name of the specified APP, and the specified APP is the APP currently installed on the specified device. S400: Use the target category label corresponding to the intermediate APP name that is the same as the specified APP name as the second intermediate label corresponding to the specified APP name, to obtain S i The corresponding second intermediate tag list set S 0 i ={S 0 i1 S 0 i2 S 0 ia S 0 ic(i) }, S 0 ia For S ia The corresponding second intermediate tag list, S 0 ia Including S ia The second intermediate tag corresponding to each specified APP name in the file; S500, for S 0 i1 S 0 i2 S 0 ia S 0 ic(i) Deduplicat all second intermediate tags to obtain S 0 i The corresponding third intermediate tag list T i ={T i1 T i2 ,…,T ib ,…,T id(i) }, and according to S 0 i and T i Get A i The corresponding target category tag for the target app, where T ib For S 0 i The corresponding b-th third intermediate label, where b takes values ​​from 1 to d(i), and d(i) is S. 0 i The number of corresponding third intermediate tags.

6. The method for obtaining target category labels according to claim 5, characterized in that, If all intermediate app names in the intermediate app name list are different from the specified app name, then the second intermediate tag corresponding to the specified app name is determined to be NULL.

7. The method for obtaining target category labels according to claim 5, characterized in that, Steps according to S 0 i and T i Get A i The target category tags for the corresponding target app also include the following steps S501-S504: S501, Obtain S 0 ia China T ib The corresponding TGI value U aib U aib The following conditions must be met: U aib =V aib / V 0 ia / (∑ c(i) a＝1 V aib / ∑ c(i) a＝1 V 0 ia )×100, where V aib For S 0 ia In and T ib The same number of second intermediate labels, V 0 ia For S 0 ia The number of second middle tags; S502, Obtain S 0 ia China T ib The corresponding target weight W aib W aib The following conditions must be met: W aib =V aib / V 0 ia ; S503, according to U aib and W aib Get T ib The corresponding weighted average TGI value Z ib Z ib The following conditions must be met: WITH ib =∑ c(i) a＝1 (IN aib ×U aib ) / ∑ c(i) a＝1 IN aib ; S504, when Z ib =max(Z) i1 Z i2 , ..., Z ib , ..., Z id(i) When ), T ib As A i The target category tag for the corresponding target app.

8. The method for obtaining target category labels according to claim 5, characterized in that, The specified device ID is the identifier of the specified device.

9. A non-transitory computer-readable storage medium, characterized in that, The storage medium stores a computer program, which is loaded and executed by a processor to implement the method for obtaining target category labels as described in any one of claims 1-8.

10. An electronic device, comprising: A processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements the method for obtaining a target category label as described in any one of claims 1-8.

Images