WO2014107989A1

WO2014107989A1 - Method and apparatus for determining hot user generated contents

Info

Publication number: WO2014107989A1
Application number: PCT/CN2013/086839
Authority: WO
Inventors: Yun Yang; Weigang Li
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2013-01-09
Filing date: 2013-11-11
Publication date: 2014-07-17
Anticipated expiration: 2015-07-09
Also published as: US20150161140A1; US10198480B2; CN103914491B; CN103914491A

Abstract

According to an example, at least one hot account is determined for each category according to quality scores and correlation degrees of history user generated content (UGCs); after a UGC newly posted by the hot account is received, if a quality score of the newly posted UGC is higher than a predefined quality score threshold and a correlation degree between the newly posted UGC and the category that the hot account belongs to is higher than a predefined correlation degree threshold, the newly posted UGC is determined as a hot UGC.

Description

METHOD AND APPARATUS FOR DETERMINING HOT USER GENERATED

CONTENTS

PRIORITY STATEMENT

[0001] This application claims the benefit of Chinese Patent Application No. 201310007061.6, filed on January 09, 2013, the disclosure of which is incorporated herein in its entirety by reference.

FIELD

[0002] The present disclosure relates to data processing techniques, and more particularly, to a method and an apparatus for determining a hot user generated content (UGC).

BACKGROUND

[0003] At present, users are both browsers and creators of website contents. The contents created by network users are referred to as user generated content (UGC), e.g., microblogs posted by the users.

[0004] A website system on which user can post UGC is usually referred to as a UGC website system, e.g., microblog system, social network service (SNS) system, social forum system, knowledge sharing system, etc. In the UGC website system, each user may post contents and there may be a large amount of UGCs on the UGC website. Thus, the UGC website system usually selects high quality UGC (also referred to as hot UGC) from the large amount of UGCs and recommends the selected high quality UGC to target users.

SUMMARY

[0005] According to an example of the present disclosure, a method for determining a hot data generated content (UGC) is provided. The method includes: analyzing a history UGC posted by an account in a UGC website system, calculating a quality score of the history UGC posted by the account and a correlation degree between the history UGC and a category, determining a hot account for the category according to the quality score and correlation degree of the history UGC; after receiving a UGC newly posted by the hot account, calculating a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to;

determining whether the quality score of the newly posted UGC is higher than a predefined quality score threshold and whether the correlation degree between the newly posted UGC and the category that the hot account belongs to is higher than a predefined correlation degree threshold of the category; and

determining, if the quality score of the newly posted UGC is higher than the predefined quality score threshold and the correlation degree between the newly posted UGC and the category that the hot account belongs to is higher than the predefined correlation degree threshold, that the newly posted UGC is a hot UGC.

[0006] According to another example of the present disclosure, an apparatus for determining a hot UGC is provided. The apparatus includes:

one or more processors;

a memory;

wherein one or more program modules are stored in the memory and to be executed by the one or more processors, the one or more program modules comprise: a hot account determining module, configured to

analyze a history UGC posted by an account, a quality score of the history UGC and a correlation degree between the history UGC and a category, and

determine, for the category, one or more accounts as hot accounts according to the quality score and the correlation degree of the history UGC; and a hot UGC determining module, configured to

calculate, after receiving a UGC newly posted by the hot account, a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to; determine whether the quality score of the newly posted UGC is higher than a predefined quality score threshold of the category and whether the correlation degree is higher than a predefined correlation degree threshold of the category; and

determine, if the quality score of the newly posted UGC is higher than the predefined quality score threshold of the category and the correlation degree is higher than the predefined correlation degree threshold of the category, the newly posted UGC as a hot UGC in the category that the hot account belongs to.

[0007] According to still another example of the present disclosure, a non-transitory computer-readable storage medium includes a set of instructions for determining a hot UGC is provided, the set of instructions to direct at least one processor to perform acts of:

analyzing a history UGC posted by an account in a UGC website system, calculating a quality score of the history UGC posted by the account and a correlation degree between the history UGC and a category, determining a hot account for the category according to the quality score and correlation degree of the history UGC; after receiving a UGC newly posted by the hot account, calculating a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to;

if the quality score of the newly posted UGC is higher than the predefined quality score threshold and the correlation degree between the newly posted UGC and the category that the hot account belongs to is higher than the predefined correlation degree threshold, determining that the newly posted UGC is a hot UGC.

[0008] Other aspects or embodiments of the present disclosure can be understood by those skilled in the art in light of the description, the claims, and the drawings of the present disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Features of the present disclosure are illustrated by way of example and not limited in the following figures, in which like numerals indicate like elements, in which:

[0010] FIG. 1 is a schematic diagram illustrating an example embodiment of a computer system for implementing a method for determining a hot UGC. [0011] FIG. 2 is a flowchart illustrating a method for determining a hot UGC according to an example of the present disclosure.

[0012] FIG. 3 is a flowchart illustrating a process of determining a hot account in block 201 of FIG. 2 according to an example of the present disclosure.

[0013] FIG. 4 is a schematic diagram illustrating a method for determining a hot UGC according to another example of the present disclosure.

[0014] FIG. 5 is a schematic diagram illustrating a method for determining a hot UGC according to still another example of the present disclosure.

[0015] FIG. 6 is a schematic diagram illustrating a method for determining a hot UGC applied in a microblog system according to an example of the present disclosure.

[0016] FIG. 7 is a schematic diagram illustrating an apparatus for determining a hot UGC according to an example of the present disclosure.

[0017] FIG. 8 is a schematic diagram illustrating an apparatus for determining a hot UGC according to another example of the present disclosure.

[0018] FIG. 9 is a schematic diagram illustrating an apparatus for determining a hot UGC according to still another example of the present disclosure.

DETAILED DESCRIPTION

[0019] The preset disclosure will be described in further detail hereinafter with reference to accompanying drawings and examples to make the technical solution and merits therein clearer.

[0020] For simplicity and illustrative purposes, the present disclosure is described by referring to examples. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present disclosure. It will be readily apparent however, that the present disclosure may be practiced without limitation to these specific details. In other instances, some methods and structures have not been described in detail so as not to unnecessarily obscure the present disclosure. As used herein, the term "includes" means includes but not limited to, the term "including" means including but not limited to. The term "based on" means based at least in part on. In addition, the terms "a" and "an" are intended to denote at least one of a particular element.

[0021] In a UGC website system, each user may generate contents. Among these contents, there may be erroneous, fake or prejudiced contents. Therefore, the user generated contents should be filtered or selected. Thereafter, hot contents are selected and provided to target users, such that the target users are capable of browsing their interested contents in time.

[0022] In an existing technique, the selected hot contents are provided to users as "hot microblogs". In this technique, a microblog system classifies microblogs into different categories, such as "sports", "finance and economics", "shopping", "news", etc. In each category, one or more accounts are configured as hot accounts by a manager of the microblog system, e.g., according to the number of fans of the account. Microblogs posted by these hot accounts in one category during a period of time are sorted according to forwarding times and number of comments. In other words, for one microblog, the more forwarding times and the number of comments, the higher it ranks.

[0023] In the above technique, the hot account is configured according to the number of fans following this account. If the number of fans of an account exceeds a number, the account is configured as a hot account. However, contents posted by an account having many fans are not always hot contents. Similarly, contents posted by an account having few fans are not necessarily low quality contents.

[0024] In addition, the above existing technique sorts the UGCs according to the forwarding times and the number of comments, but not according to the contents of the UGCs. Thus, the finally selected hot microblog may be less correlated to target users and the category that it belongs to. For example, a hot account in "sports" category may post a hot microblog related to shopping. However, target users of the "sports" category are less interested in shopping.

[0025] Moreover, contents which have more forwarding times and comments are usually posted earlier. Newly posted contents generally have less forwarding times and comments. Therefore, in the above existing technique, newly posted contents have little possibility to be selected as high-quality contents, i.e., hot microblogs.

[0026] In contrast to this, an example of the present disclosure provides a method for determining a hot UGC. In the example of the present disclosure, a UGC website system analyzes history UGCs posted by each account to obtain a quality score of each history UGC and a correlation degree between the history UGC and each category. The UGC website system selects one or more hot accounts in each category according to quality scores and correlation degrees of the history UGCs.

[0027] After receiving a UGC newly posted by a hot account, the UGC website system calculates a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to. The UGC website system determines whether the quality score is higher than a predefined quality score threshold and whether the correlation degree is higher than a predefined correlation degree threshold of the category. If the quality score is higher than the predefined quality score threshold and the correlation degree is higher than the predefined correlation degree threshold of the category, the newly posted UGC is determined as a hot UGC in the category that the hot account belongs to.

[0028] FIG. 1 is a schematic diagram illustrating an example embodiment of a computer system for executing the method for determining a hot UGC. A computer system 100 may be a computing device capable of executing a method and apparatus of present disclosure. The computer system 100 may, for example, be a device such as a server that provides service to users locally or via a network.

[0029] The computer system 100 may vary in terms of capabilities or features. Claimed subject matter is intended to cover a wide range of potential variations. For example, the computer system 100 may include or may execute a variety of operating systems 141. The computer system 100 may include or may execute a variety of possible applications 142, such as a hot UGC determining application 145.

[0030] Further, the computer system 100 may include one or more non-transitory processor-readable storage media 130 and one or more processors 122 in communication with the non-transitory processor-readable storage media 130. For example, the non-transitory processor-readable storage media 130 may be a RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of non-transitory storage medium known in the art. The one or more non-transitory processor-readable storage media 130 may store sets of instructions, or units and/or modules that comprise the sets of instructions, for conducting operations described in the present application. The one or more processors may be configured to execute the sets of instructions and perform the operations in example embodiments of the present application.

[0031] FIG. 2 is a schematic diagram illustrating a method for determining a hot UGC according to an example of the present disclosed hot UGC determining application 145. FIG. 2 is a simplified diagram according to one embodiment of the present invention. This diagram is merely an example, which should not unduly limit the scope of the claims. One of ordinary skill in the art would recognize many variations, alternatives, and modifications. [0032] As shown in FIG. 2, the method includes at least the following.

[0033] At block 201, a UGC website system analyzes history UGCs posted by each account to obtain a quality score of each history UGC and a correlation degree between the history UGC and each category. The UGC website system selects one or more hot accounts in each category according to quality scores and correlation degrees of the history UGCs.

[0034] This block may involve a large amount of calculations. Thus, this block may be performed offline.

[0035] FIG. 3 is a flowchart illustrating a process of obtaining one or more hot accounts in block 201 according to an example of the present disclosure.

[0036] As shown in FIG. 3, the process includes the following. In the example, it is possible to consider only original UGCs posted by each account.

[0037] At block 211, one or more original UGCs posted by each account during a period of time (e.g., last two months) are obtained.

[0038] At block 212, for each original UGC, a quality score of the original UGC and a correlation degree between the original UGC and each category are calculated.

[0039] At block 213, for each account, an average quality score of the account and an average correlation degree between the account and each category are calculated according to the quality scores of the original UGCs and the correlation degrees between the original UGCs and each category, wherein

the average quality score of an account

_ a sum of quality scores of the original UGCs posted by the account ;

the number of the original UGCs posted by the account the average correlation degree between an account and a category

_ a sum of correlation degrees between the original UGCs and the category .

the number of the original UGCs posted by the account

[0040] At block 214, for each account, a category that a highest correlation degree of the account corresponds to is selected as a category that the account belongs to.

[0041] In block 213, for each account, a correlation degree between the account and each category is calculated. Thus, one account may correspond to one correlation degree in each category. Therefore, in block 214, the category that the highest correlation degree of the account corresponds to is selected as the category that the account belongs to. [0042] At block 215, for each account, it is determined whether the average quality score of the account is higher than a predefined average quality score threshold of the category that the account belongs to and whether an average correlation degree between the account and the category that the account belongs to is higher than a predefined average correlation degree threshold. If the average quality score of the account is higher than the predefined average quality score threshold of the category that the account belongs to and the average correlation degree between the account and the category that the account belongs to is higher than the predefined average correlation degree threshold, the account is determined as a hot account in the category that the account belongs to. Otherwise, the account is not a hot account.

[0043] As described above, the quality score of each original UGC and the correlation degree between the original UGC and each category are important parameters for determining a hot account. Based on the above two parameters, i.e., the quality score and the correlation degree, another parameter may be generated and acts as a basis for determining a hot account.

[0044] For example, in block 212, after the quality score of each original UGC and the correlation degree between the original UGC and each category are calculated, it is possible to multiply the quality score of the original UGC by the correlation degree between the original UGC and each category to obtain a reliability degree of the original UGC in each category. The reliability degree is a derived parameter which may be used as a basis for determining the hot account.

[0045] At this time, block 213 further includes: calculating an average reliability degree of each account with respect to each category according to the reliability degrees of original UGCs posted by the account in each category, wherein

the average reliability degree of an account in a category

_ a sum of reliability degrees of the original UGCs posted by the account in the category the number of original UGCs posted by the account

[0046] In addition, block 215 further includes: for each account, after it is determined that the average quality score of the account is higher than the predefined average quality score threshold and the average correlation degree between the account and the category that the account belongs to is higher than the predefined average correlation degree threshold, it is further determined whether the average reliability degree of the account in the category is higher than a predefined average reliability degree threshold. If yes, the account is determined as a hot account. Otherwise, the account is not a hot account.

[0047] Through the above block 201 , one or more accounts may be determined as hot accounts in one category.

[0048] At block 202, after receiving a UGC newly posted by a hot account, the UGC website system calculates a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to. The UGC website system determines whether the quality score is higher than a predefined quality score threshold and whether the correlation degree is higher than a predefined correlation degree threshold of the category. If the quality score is higher than the predefined quality score threshold and the correlation degree is higher than the predefined correlation degree threshold of the category, the newly posted UGC is determined as a hot UGC in the category that the hot account belongs to.

[0049] In one example, the UGC website system may execute block 202 each time it receives a UGC newly posted by a hot account. Alternatively, the UGC website system may also execute block 202 periodically, i.e., after a certain period of time (e.g., every 10 minutes). At this time, the UGC website system executes block 202 to process each UGC newly posted during this period of time.

[0050] In the above blocks 201 and 202, the quality score of a UGC has to be calculated. In block 201, the quality score of a history UGC is calculated. In block 202, the quality score of a newly posted UGC is calculated. The calculation of the quality score in blocks 201 and 202 may be performed following a same manner or different manners. Hereinafter, one exemplary calculation manner is provided. Those with ordinary skill in the art may have other calculation manners to calculate the quality score of the history UGC or the newly posted UGC, which is not restricted in the present disclosure.

[0051] A total text length, number of words, number of filtered words and number of punctuations in a UGC are obtained. The number of filtered words refers to the number of words which match predefined filtering words.

[0052] The number of effective words of the UGC is determined, wherein

the number of effective words

= total number of words - number of filtered words - number of punctuations [0053] A text basic score of the UGC is determined, wherein the text basic score

= w5 x number of effective words + w6 x number of filtered words

w5 and w6 are weight parameters which may be determined based on training data.

[0054] A number of repeated words of the UGC and a word repetition ratio are determined; wherein

number of repeated words

the word repetition ratio= ^£ .

total number of words

[0055] A text score of the UGC is determined; wherein

the text score=

number of effective words _{ΓΛ ίΛ} , . . · _\ , „ text basic score x x / 1 ^χ ( 1 - word repetition ratio ) / w4

total number of words

wherein fl is a predefined function taking the number of punctuations and the total number of words as input parameters, w4 is a weight parameter.

[0056] A posted time of the UGC is obtained and a time score of the UGC is calculated, wherein

posted time of the UGC - predefined reference time the time score=- ,

wl

wherein w7 is a weight parameter.

[0057] The quality score of the UGC is determined, wherein

the quality score = wlx (w2x text score+w3 x time score) , wherein wl, w2 and w3 are weight parameters.

[0058] Now, through the above process, the quality score of each UGC (e.g., a history UGC or a newly posted UGC) is calculated.

[0059] Besides the quality score, in blocks 201 and 202, a correlation degree between the UGC and a category is also required to be calculated. Specifically, in block 201, the correlation degree between the history UGC and each category is calculated. In block 202, the correlation degree between a newly posted UGC and the category that the hot account which posts the new UGC belongs to is calculated. It should be noted that, the correlation degree may be calculated in a same manner or different manners in blocks 201 and 202. One exemplary calculation manner of the correlation degree is described in the following. Those with ordinary skill in the art may have other calculation manners to determine the correlation degree, which is not restricted in the present disclosure. [0060] One exemplary formula is as follows:

Correlation degree = Wl * Fl (weight) + W2 * F2 (rate) + W3 * F3 (rank).

[0061] Wl, W2 and W3 are three weight parameters.

[0062] Weight denotes weight of the category.

[0063] Rate denotes a value that the weight of the category is divided by a total weight.

[0064] Rank denotes a ranking position of the category in all categories.

[0065] Fl denotes a function for normalizing the weight to 0-1.

[0066] F2 denotes a function for normalizing the rate to 0-1.

[0067] F3 denotes a function for normalizing the rank to 0-1.

[0068] Through the above blocks 201 and 202, it is possible to determine one or more hot accounts according to quality scores of history UGCs and correlation degrees between the history UGCs and the categories. Compared with the existing technique in which the hot account is determined according to number of fans or other subjective factors (e.g., configured by a network manager manually), the method provided by the example of the present disclosure determines the hot accounts based on the contents of the UGC posted by all accounts. The determination is more objective. In addition, the contents of the hot UGC selected from the UGCs posted by these hot accounts have a high correlation degree with contents that the users are interested in, and also have a high correlation degree with the category that it belongs to. Moreover, the method provided by the example of the present disclosure is capable of performing the selection operation after a newly posted UGC is received. Thus, the hot UGC may be provided to users rapidly.

[0069] FIG. 4 is a flowchart illustrating a method for determining a hot UGC according to another example of the present disclosure.

[0070] As shown in FIG. 4, the method includes the following.

[0071] Block 401 is the same with block 201.

[0072] At block 402, for a newly posted UGC, it is determined that whether the UGC contains a word which is in a predefined blacklist. If yes, the UGC is removed at block 403, i.e., not considered and no further calculation is performed to this UGC. Otherwise, block 404 is performed.

[0073] Through blocks 402 and 403, it is possible to remove UGC containing words which are in the blacklist. The quality of the hot UGC may be increased. The number of candidate UGCs may be reduced, which reduces workload of subsequent calculation.

[0074] Block 404 is the same with block 202.

[0075] It should be noted that, the calculation of the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category in block 202 may be performed each time a newly posted UGC is received or periodically (e.g., every 10 minutes). If the calculation is performed periodically, a repetition removing operation may be performed before the quality score and the correlation degree are calculated.

[0076] FIG. 5 is a flowchart illustrating a method for determining a hot UGC according to still another example of the present disclosure. As shown in FIG. 5, the method includes the following.

[0077] Block 501 is the same with block 201.

[0078] At block 502, it is determined whether at least two UGCs newly posted by a hot account are received. If yes, block 503 is performed. Otherwise, block 504 is performed.

[0079] At block 503, a text similarity degree between the newly posted UGCs is calculated. For UGCs having a text similarity degree higher than a predefined threshold, a UGC which is posted later is removed, or, a UGC which is posted earlier is reserved.

[0080] Thus, the following calculation is only performed for the reserved UGC. The number of candidate UGCs is reduced and the workload of the subsequent calculation is reduced.

[0081] The calculation of the text similarity degree between the newly posted UGCs may be as follows: perform a word segmentation operation to each newly posted UGC to obtain notional words (i.e., words having meanings themselves), calculate a notional word repetition ratio between each two UGCs. The notional word repetition ratio is the word similarity degree. For two UGCs having a notional word repetition ratio higher than a predefined threshold, only the UGC which is posted earlier is reserved for further processing.

[0082] Block 504 is the same with block 202.

[0083] In examples of the present disclosure, the UGC website system may be a microblog system, a social network service (SNS) system, a social forum system, a knowledge sharing system, etc. Hereinafter, the microblog system is taken as an example to describe an implementation of the present disclosure. In the following, the microblog is the UGC described in the above examples. [0084] FIG. 6 is a flowchart illustrating a method for determining a hot UGC applied in a microblog system according to an example of the present disclosure. As shown in FIG. 6, the method includes the following.

[0085] At block 601, one or more hot accounts in each category are determined. This block may specifically include the following blocks 611 to 615.

[0086] At block 611, original microblogs posted by each account within a certain period (e.g., last two months) are obtained.

[0087] For example, microblogs in following table 1 are obtained.

Table 1

[0088] At block 612, a quality score of each original microblog, a correlation degree between each original microblog and each category and a reliability degree of each original microblog in each category are calculated.

[0089] Suppose that a formula for calculating the quality score is as follows:

Quality score = 700000 * (0.5 * text score + 0.4 * (posted time of the microblog - 1293811200) / w7), wherein w7 = 3600 * 87600;

wherein text score = ( total text length + 5 * (total number of words - number of filtered words - number of punctuations) - 20 * number of filtered words) * (total number of words - number of filtered words - number of punctuations) / total number of words * fl (number of punctuations, total number of words) * (1 - number of repeated words / total number of words) / 840.

[0090] The function fl may be obtained through analyzing of training data. An example is as follows.

[0091] The value of fl is 1 in default.

[0092] If the number of punctuations is 0, fl = 0.3 if the total length is larger than 300, fl = 0.6 if the total length is larger than 100, and fl = 0.88 if the total length is larger than 70.

[0093] If the number of punctuations is larger than 40, fl = 0.74.

[0094] If the number of punctuations is larger than 30, fl = 0.82.

[0095] If the number of punctuations is larger than 20, fl = 0.92.

[0096] If a quotient obtained by dividing the number of punctuations by the total length is smaller than 0.03, fl = 0.73.

[0097] If a quotient obtained by dividing the number of punctuations by the total length is smaller than 0.05, fl = 0.9.

[0098] Herein, suppose that a formula for calculating the correlation degree between a microblog and a category is as follows.

[0099] Correlation degree = 0.2 * Fl (weight) + 0.6 * F2 (rate) + 0.2 * F3 (rank).

[0100] Fl is defined as follows:

If weight > 3, Fl = 1;

Otherwise, Fl = pow (weight / 3, 0.2).

[0101] F2 is defined as follows:

If rate > 0.5, F2 = 1;

Otherwise, F2 = pow (rate / 0.5, 0.4).

[0102] F3 is defined as follows:

If rank > 10, F3 = 0;

Otherwise, F3 = pow ((11.0 - rank) / 10.0, 1.5).

[0103] Hereinafter, the microblog 1 is taken as an example to describe the calculation of the quality score, the correlation degree and the reliability degree.

[0104] (1) The calculation of the quality score of the microblog 1.

[0105] The total text length of microblog 1 is 134, total number of words is 35, number of punctuations is 9, number of filtered words is 0, and the number of repeated words is 0.

[0106] The text score of the microblog 1 = (134 + 5 * (35 - 0 - 9) - 20 * 0) * (35 - 0 - 9) / 35 * 1 * (1 - 0 / 35) / 840=0.233469.

[0107] The time score of microblog 1 = (1354621754 - 1293811200) / 3600 / 87600 = 0.192829. [0108] The quality score of microblog 1 = 700000 * (0.5 * text score + 0.4 * time score) = 700000 * (0.5 * 0.233469 + 0.4 * 0.192829) = 135706.

[0109] (2) The calculation of the correlation degree between microblog 1 and each category.

[0110] A weight of each word in each category may be obtained through a training method such as term frequency-inverse document frequency (TF-IDF). Then a word classification table with weight is obtained. According to the word classification table, the weight of each word segmented from the microblog in each category may be obtained. For example, the weight of each word segmented from microblog 1 in each category is as shown in table 2.

Table 2

[0111] According to table 2 and the formula of correlation degree 0.2*F1 (weight) + 0.6 * F2 (rate) + 0.2 * F3 (rank), a correlation degree between microblog 1 and each category may be obtained, as shown in table 3.

Table 3

[0112] (3) The calculation of the reliability degree of microblog 1 in each category. A following formula may be used: reliability degree = quality score * correlation degree of the microblog in the category. A calculated result may be as shown in table 4.

category Correlation degree reliability

basketball 0.990000 134348

news 0.306711 41622

military 0.256398 34794

football 0.228975 31073

quotation 0.149597 20301 shopping 0.127356 17282

Table 4

[0113] Based on the above calculations of the quality score, the correlation degree and the reliability degree, a result as shown in table 5 may be obtained.

account index Microblog contents Quality Related Correlation Reliability score category degree degree a 1 Dave noted on Tuesday 135706 basketball 0.990000 134348 that progress has Current 0.306711 41622 seemingly been halted in events

the ongoing labor Military 0.256398 34794 negotiations between the Football 0.228975 31073

NBA and the National Quotation 0.149597 20301

Basketball Players shopping 0.127356 17282 Association. Here's a look

at how some of the players

competing at Impact

Basketball's "Lockout

League" in Las Vegas took

the news. There's

definitely some rising

frustration with the lack of

progress.

2 Spears's aunt Sandra 164149 basketball 0.930000 152658

Bridges Covington, with auto 0.510763 83841 whom she had been very digital 0.483108 79301 close, died of ovarian family 0.261693 42956 cancer in January. In quotation 0.177987 29216

February 2007, Spears telecom 0.155746 25565 stayed in a drug history 0.120424 19767 rehabilitation facility in fun 0.092388 15165

Antigua for less than a day. politics 0.060536 9936 The following night, she Love 0.048842 8017 shaved her head with Current 0.021259 3489 electric clippers at a hair events

salon in Tarzana, Los

Angeles.

b 3 The Su-27 is a highly 130740 military 1.000000 130740 integrated twin-finned politics 0.327131 42769 aircraft. The airframe is Science 0.230002 30070 constructed of titanium work 0.184141 24074 and high-strength Auto 0.143291 18733 aluminum alloys. The Foreign 0.109144 14269 engine nacelles are fitted language

with trouser fairings to IT 0.077400 10119 provide a continuous travel 0.049616 6486 streamlined profile

between the nacelles and

the tail beams. The fins

and horizontal tail

consoles are attached to

tail beams.

4 Scorpios love competition 149081 constellation 0.980000 146099 in both work and play, Love 0.430763 64218 which is why they'll air it quotation 0.397950 59326 out in sports and games. Fun 0.229745 34250 Extreme sports are right up beauty 0.186795 27833

Scorpio's alley, as is most Current 0.127017 18933 anything that will test their events

mettle. They've got to have

an adversary, since it

makes the game that much

more fun. Scorpio's

colors? Powerful red and

serious black. When it

comes to love, though,

Scorpios soften up a bit

and are caring and devoted

with their lovers, even if

they do hold on a bit tight.

5 NBA Players In Vegas 128133 basketball 0.995000 12749

React To Lockout fun 0.470763 60222

Negotiations Stalling Video 0.259470 33186

In NYC Football 0.226073 28958 music 0.148228 18963 dance 0.124404 15888 quotation 0.094198 12044 clothing 0.076465 9738

Table 5

[0114] At block 613, based on the above data, an average quality score of each account, an average correlation degree between the account and each category and an average reliability degree of the account in each category are obtained, as shown in table 6.

Table 6

[0115] At block 614, for each account, a category that a highest average correlation degree of the user corresponds to is selected as the category that the account belongs to. For example, as shown in table 6, account "a" belongs to category "basketball" and account "b" belongs to category "military".

[0116] At block 615, a hot account is obtained. Herein, suppose a selection criterion of the hot account is that the following three conditions are met:

1) quality score > 70000;

2) correlation degree > 0.3; and

3) reliability degree > 65000.

[0117] According to the above selection criterion, account a is a hot account in category "basketball" and account b is discarded.

[0118] Through the above blocks, a hot account is obtained. After the microblog system receives a microblog posted by the hot account, the following blocks 602 to 606 may be performed.

[0119] Suppose microblogs of three hot accounts A, B and C are received, as shown in table 7.

Table 7

[0120] At block 602, data pre-processing is performed. Suppose that word

"diction" is in the blacklist. Thus, microblog 3 is filtered and other microblogs pass the pre-processing.

[0121] At block 603, data repetition removing operation is performed. The microblogs are segmented to obtain notional words. A notional repetition ratio between each two microblogs is calculated. If the notional repetition ratio is higher than a predefined threshold, it is determined that the two microblogs are similar and the one which is posted earlier is reserved.

[0122] In this example, microblogs 4 and 8 have a repetition ratio higher than the predefined threshold. Therefore, the microblog 4 which is posted later is removed. Subsequent operations are performed to other microblogs.

[0123] At block 604, a microblog correlation evaluation operation is performed.

[0124] According to a correlation degree calculation method similar to block 612, a correlation degree of each microblog is calculated. According to the predefined average correlation degree threshold, it is determined whether the microblog passes the evaluation. If the evaluation is not passed, the microblog is removed. A result is shown in table 8.

Hot category Correlatio Microblo Microblog contents Correlatio Correlatio accou n degree g index n degree n nt threshold evaluatio n

A quotatio 0.85 1 Love means never having 0.83 Not pass n to say you're sorry

2 Your opening shows great 0.83 Not pass promise, and yet flashy

purple patches; as when

describing a sacred grove,

or the altar of Diana, or a

stream meandering

through fields, or the

river Rhine, or a rainbow;

but this was not the place

for them. If you can

realistically render a

cypress tree, would you

include one when

commissioned to paint a

sailor in the midst of a

shipwreck?

B basketba 0.85 4 the last minute kill shot 0.816683 Not pass

11 5 Jeremy Lin scored 28 0.87 pass points and dished out a

career-high 14 assists as

the New York Knicks got

back on the winning

column after defeating the

Dallas Mavericks, winners

of six straight, with a 104-97 victory at the MSG

6 Lin has now surpassed 20 1 pass points for the eighth time

in nine contests on

Sunday, as he shot

ll-of-20 from the floor

and 3-of-6 from the

three -point line. He scored

16 points in the second

half, which included a

couple of timely

three -pointers in the fourth

quarter.

c food 0.86 7 Certain cultures highlight 0.995 pass animal and vegetable

foods in their raw

state. Salads consisting of

raw vegetables or fruits

are common in many

cuisines. Sashimi in Japan

ese cuisine consists of raw

sliced fish or other meat,

and sushi often

incorporates raw fish or

seafood.

9 sales promotion, 0 Not pass

13205593099

Table 8

[0125] After the correlation evaluation operation, a following table 9 is obtained.

Table 9

[0126] At block 605, a quality evaluation operation is performed to each microblog.

[0127] According to a quality score calculation method similar to that of block 612, the quality score of each microblog may be obtained. According to a quality score threshold corresponding to each category, it is determined whether a microblog passes the quality evaluation. If the quality evaluation is not passed, the microblog is removed.

A result may be as shown in table 10.

Table 10

[0128] After the quality evaluation, microblog 6 is selected as a hot microblog in the category "basketball".

[0129] In view of the above, according to the method provided by the examples of the present disclosure, it is possible to find hot microblog contents rapidly and accurately.

[0130] In accordance with the above method examples, an example of the present disclosure further provides an apparatus for determining a hot UGC. As shown in FIG. 7, the apparatus 700 includes: a processor 710 and a memory 720; wherein one or more program modules are stored in the memory 720 and to be executed by the processor 710, the one or more program modules comprise: a hot account determining module 701 and a hot UGC determining module 702.

[0131] The hot account determining module 701 is configured to

for each history UGC posted by each account, calculate a quality score of the history UGC and a correlation degree between the history UGC and each category, and

determine one or more accounts in each category as hot accounts according to quality scores and correlation degrees of the history UGCs.

[0132] The hot UGC determining module 702 is configured to

calculate, after receiving a UGC newly posted by a hot account, a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to;

determine whether the quality score is higher than a predefined quality score threshold and whether the correlation degree is higher than a predefined correlation degree threshold of the category; and

determine, if the quality score is higher than the predefined quality score threshold and the correlation degree is higher than the predefined correlation degree threshold of the category, the newly posted UGC as a hot UGC in the category that the hot account belongs to.

[0133] FIG. 8 is a schematic diagram illustrating an apparatus for determining a hot UGC according to another example of the present disclosure. As show in FIG. 8, the apparatus 800 includes: a processor 810 and a memory 820; wherein one or more program modules are stored in the memory 820 and to be executed by the processor 810, the one or more program modules comprise: a hot account determining module 801, a pre-processing module 802 and a hot UGC determining module 803.

[0134] The hot account determining module 801 is configured to

[0135] The pre-processing module 802 is configured to

determine, after receiving a UGC newly posted by a hot account, whether the newly posted UGC contains a word in a blacklist;

discard the newly posted UGC if the newly posted UGC contains a word in a blacklist; and

provide the newly posted UGC to the hot UGC determining module 803 if otherwise.

[0136] The hot UGC determining module 803 is configured to

calculate, after receiving a newly posted UGC from the pre-processing module 802, a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to;

determine, if the quality score is higher than the predefined quality score threshold and the correlation degree is higher than the predefined correlation degree threshold of the category, the newly posted UGC as a hot

UGC in the category that the hot account belongs to.

[0137] FIG. 9 is a schematic diagram illustrating an apparatus for determining a hot UGC according to another example of the present disclosure. As show in FIG. 9, the apparatus 900 includes: a processor 910 and a memory 920; wherein one or more program modules are stored in the memory 920 and to be executed by the processor 910, the one or more program modules comprise: a hot account determining module 901, a repetition removing module 902 and a hot UGC determining module 903.

[0138] The hot account determining module 901 is configured to

[0139] The repetition removing module 902 is configured to

determine whether at least two UGCs newly posted by a hot account are received;

calculate, if at least two UGCs newly posted by the hot account are received, a text similarity ratio of each two newly posted UGCs;

if the two newly posted UGCs have a text similarity degree higher than a predefined threshold, discard a UGC which is posted later and Provide a UGC which is posted earlier to the hot UGC determining module 903; if the two newly posted UGCs have a text similarity degree not higher than the predefined threshold, provide the two newly posted UGCs to the hot UGC determining module 902.

[0140] The hot UGC determining module 903 is configured to

calculate, after receiving a newly posted UGC from the repetition removing module 902, a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to;

[0141] The processor 910 may include one or more processors for executing the sets of instructions stored in the memory 920. The processor 920 is a hardware device, such as a central processing unit (CPU) or a micro controlling unit (MCU). The memory 920 is a non-transitory processor-readable storage media, such as a RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of non-transitory storage medium known in the art.

[0142] What has been described and illustrated herein is a preferred example of the disclosure along with some of its variations. The terms, descriptions and figures used herein are set forth by way of illustration only and are not meant as limitations. Many variations are possible within the spirit and scope of the disclosure, which is intended to be defined by the following claims— and their equivalents— in which all terms are meant in their broadest reasonable sense unless otherwise indicated.

Claims

1. A method for determining a hot User Generated Content (UGC), comprising: analyzing a history UGC posted by an account in a UGC website system, and calculating a quality score of the history UGC posted by the account and a correlation degree between the history UGC and a category;

determining a hot account for the category according to the quality score and correlation degree of the history UGC;

after receiving a UGC newly posted by the hot account, calculating a quality score of the newly posted UGC and a correlation degree between the newly posted UGC and the category that the hot account belongs to;

determining whether the quality score of the newly posted UGC is higher than a predefined quality score threshold, and whether the correlation degree between the newly posted UGC and the category that the hot account belongs to is higher than a predefined correlation degree threshold of the category;

2. The method of claim 1, further comprising:

before calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to, determining whether the newly posted UGC contains a word in a blacklist;

if the newly posted UGC does not contain the word in the blacklist, performing the process of calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to.

3. The method of claim 1 or 2, further comprising:

before calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to, determining whether at least two newly posted UGCs are received;

if at least two newly posted UGCs are received, calculating a text similarity ratio between each two newly posted UGCs, if the text similarity ratio between two newly posted UGCs is not higher than a predefined threshold, performing, for each of the two newly posted UGCs, the process of calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to.

4. The method of claim 1, wherein the process of calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to is performed after the newly posted UGC is received, or is performed periodically for each newly posted UGC received during a period of time.

5. The method of claim 1, wherein the analyzing the history UGC posted by the account in the UGC website system, and calculating the quality score of the history UGC posted by the account and the correlation degree between the history UGC and the category and determining the hot account for the category according to the quality score and correlation degree of the history UGC comprises:

obtaining one or more history UGCs posted by the account during a period of time;

for each history UGC, calculating the quality score of the history UGC and the correlation degree between the history UGC and each of a plurality of categories; calculating an average quality score of the account and an average correlation degree between the account and each category according to the quality score and correlation degree of each history UGC;

determining a category that a highest correlation degree of the account corresponds to as the category that the account belongs to;

determining, for the account, whether the average quality score of the account is higher than a predefined average quality score threshold and whether the average correlation degree between the account and the category that the account belongs to is higher than a predefined average correlation degree threshold;

if the average quality score of the account is higher than the predefined average quality score threshold and the average correlation degree between the account and the category that the account belongs to is higher than the predefined average correlation degree threshold, determining that the account is a hot account.

6. The method of claim 5, further comprising:

before calculating the quality score of the history UGC and the correlation degree between the history UGC and each category,

multiplying the quality score of the history UGC with the correlation degree between the history UGC and each category to obtain a reliability degree of the history UGC in each category; and

calculating an average reliability degree of the account according to the reliability degree of the account in each category;

after determining that the average quality score of the account is higher than the predefined average quality score threshold and the average correlation degree between the account and the category that the account belongs to is higher than the predefined average correlation degree threshold,

determining whether the average reliability degree of the account is higher than a predefined average reliability degree threshold, if the average reliability degree of the account is higher than the predefined average reliability degree threshold, determining that the account is a hot account.

7. The method of claim 1, wherein the calculating the quality score of the history UGC comprises:

obtaining a total text length, a total number of words, a number of filtered words and a number of punctuations of the history UGC;

determining a number of effective words of the UGC, wherein the number of effective words

= total number of words - the number of filtered words - the number of punctuations ' determining a text basic score of the history UGC, wherein

the text basic score

, w5 and w6

= w5 x number of effective words + w6 x number of filtered words

represent weight parameters;

calculating a number repeated words and a word repetition ratio of the history

_{T T^^} , . , , . . . the number of repeated words

UGC, wherein the word repetition ratio= ^£ ;

the total number of words

determining a text score of the history UGC, wherein the text score=

, . the number of effective words -.„ .. , , . . . _\ , „ the text basic score x x / 1 ^χ ( 1 - the word repetition ratio ) / w4 the total number of words

, fl represent a predefined function taking the number of punctuations and the total number of words as input parameters, w4 represent a weight parameter;

determining a posted time and a time score of the history UGC, wherein

the posted time of the UGC - a predefined reference time

the time score= , w7 wl

represent a weight parameter;

determining the quality score of the UGC, wherein the quality score = wlx (w2x text score+w3x time score) , wl, w2 and w3 represent weight parameters.

8. The method of claim 1, wherein the calculating the correlation degree between the history UGC and the category comprises:

calculating the correlation degree according to a following formula:

the correlation degree = Wl * Fl (weight) + W2 * F2 (rate) + W3 * F3 (rank);

Wl, W2 and W3 represent three weight parameters;

weight denotes a weight of the category;

rate denotes a value that the weight of the category is divided by a total weight;

rank denotes a ranking position of the category in all categories;

Fl denotes a function for normalizing the weight to 0-1;

F2 denotes a function for normalizing the rate to 0-1;

F3 denotes a function for normalizing the rank to 0-1.

9. An apparatus for determining a hot user generated content (UGC), comprising: one or more processors;

a memory;

wherein one or more program modules are stored in the memory and to be executed by the one or more processors, the one or more program modules comprise:

a hot account determining module, configured to

analyzing a history UGC posted by an account, a quality score of the history UGC and a correlation degree between the history UGC and a category, and determine, for the category, one or more accounts as hot accounts according to the quality score and the correlation degree of the history UGC; and a hot UGC determining module, configured to

10. The apparatus of claim 9, further comprising:

a pre-processing module, configured to

determine, after receiving the UGC newly posted by the hot account, whether the newly posted UGC contains a word in a blacklist;

provide the newly posted UGC to the hot UGC determining module if otherwise.

11. The apparatus of claim 9 or 10, further comprising:

a repetition removing module, configured to

determine whether at least two UGCs newly posted by the hot account are received;

if the two newly posted UGCs have a text similarity degree higher than a predefined threshold, provide the one which is posted earlier to the hot UGC determining module; if the two newly posted UGCs have a text similarity degree not higher than the predefined threshold, provide the two newly posted UGCs to the hot UGC determining module.

12. The apparatus of claim 9, wherein the hot account determining module is further configured to:

obtain one or more history UGCs posted by the account during a period of time;

calculate, for each history UGC, the quality score of the history UGC and the correlation degree between the history UGC and each of a pluraliyt of categories;

calculate, for the account, an average quality score of the account and an average correlation degree between the account and each category according to the quality scores and correlation degrees of the one or more history UGCs;

determine, for the account, a category that a highest correlation degree of the account corresponds to as the category that the account belongs to; determine, for the account, whether the average quality score of the account is higher than a predefined average quality score threshold and whether the average correlation degree between the account and the category that the account belongs to is higher than a predefined average correlation degree threshold; and

determine that the account is a hot account if the average quality score of the account is higher than the predefined average quality score threshold and the average correlation degree between the account and the category that the account belongs to is higher than the predefined average correlation degree threshold.

13. The apparatus of claim 9, wherein the hot account determining module is further configured to:

before calculating the quality score of the history UGC and the correlation degree between the history UGC and the category,

multiply the quality score of the history UGC with the correlation degree between the history UGC and the category to obtain a reliability degree of the history UGC in the category; and

calculate an average reliability degree of the account according to the reliability degree of the account in the category;

determine whether the average reliability degree of the account is higher than a predefined average reliability degree threshold; and determine that the account is a hot account if the average reliability degree of the account is higher than the predefined average reliability degree threshold.

14. A non-transitory computer-readable storage medium comprising a set of instructions for determining a hot user generated content (UGC), the set of instructions to direct at least one processor to perform acts of:

determining whether the quality score of the newly posted UGC is higher than a predefined quality score threshold and whether the correlation degree between the newly posted UGC and the category that the hot account belongs to is higher than a predefined correlation degree threshold of the category;

15. The non-transitory computer-readable storage medium of claim 14, further comprising: before calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to, determining whether the newly posted UGC contains a word in a blacklist;

16. The non-transitory computer-readable storage medium of claim 14 or 15, further comprising:

17. The non-transitory computer-readable storage medium of claim 14, wherein the process of calculating the quality score of the newly posted UGC and the correlation degree between the newly posted UGC and the category that the hot account belongs to is performed after the newly posted UGC is received, or is performed periodically for each newly posted UGC received during a period of time.

18. The non-transitory computer-readable storage medium of claim 14, wherein the analyzing the history UGC posted by the account, and calculating the quality score of the history UGC posted by the account and the correlation degree between the history UGC and the category and determining the hot account for the category according to the quality score and correlation degree of the history UGC comprises: obtaining one or more history UGCs posted by the account during a period of time;

determining, for the account, whether the average quality score of the account is higher than a predefined average quality score threshold and whether the average correlation degree between the account and the category that the account belongs to is higher than a predefined average correlation degree threshold, if the average quality score of the account is higher than the predefined average quality score threshold and the average correlation degree between the account and the category that the account belongs to is higher than the predefined average correlation degree threshold, determining that the account is a hot account.