WO2019218452A1 - 热词分析方法、计算机可读存储介质、终端设备及装置 - Google Patents
热词分析方法、计算机可读存储介质、终端设备及装置 Download PDFInfo
- Publication number
- WO2019218452A1 WO2019218452A1 PCT/CN2018/096267 CN2018096267W WO2019218452A1 WO 2019218452 A1 WO2019218452 A1 WO 2019218452A1 CN 2018096267 W CN2018096267 W CN 2018096267W WO 2019218452 A1 WO2019218452 A1 WO 2019218452A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- exposure
- text information
- hot word
- frequency
- threshold
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/216—Parsing using statistical methods
Definitions
- the present application belongs to the field of computer technology, and in particular, to a hot word analysis method, a computer readable storage medium, a terminal device and a device.
- Hot word that is, a popular online vocabulary, refers to a vocabulary phenomenon that reflects the problems and things that people in a country and a region pay attention to at a time. Hot words have the characteristics of the times and can be used as a hot topic in a period and a representative of people's death issues.
- the determination of hot words is mainly obtained by network analysts according to the information they browse on the Internet. This method relies on the personal judgment of network analysts, and is subjective and difficult to objectively reflect the reality.
- the embodiment of the present application provides a hot word analysis method, a computer readable storage medium, a terminal device and a device, so as to solve the problem that the determination process of the hot words in the prior art is subjective and the analysis result is directed to the enterprise. A very poor question.
- a first aspect of the embodiments of the present application provides a hot word analysis method, which may include:
- the preferred text information being text information including the hot word
- the degree of association between each enterprise name and the hot word is calculated according to the frequency of exposure of each enterprise name in the preferred text information.
- a second aspect of embodiments of the present application provides a computer readable storage medium storing computer readable instructions that, when executed by a processor, implement the hot word analysis method described above step.
- a third aspect of an embodiment of the present application provides a hot word analysis terminal device including a memory, a processor, and computer readable instructions stored in the memory and executable on the processor, the processor executing The computer readable instructions implement the steps of the hot word analysis method described above.
- a fourth aspect of the embodiments of the present application provides a hot word analyzing apparatus, which may include a module for implementing the steps of the above hot word analyzing method.
- the embodiment of the present application has the beneficial effects that the embodiment of the present application provides a set of objective evaluation criteria for determining the hot words, frees the dependence on the personal ability of the network analyst, and determines the hot words. It is more difficult to objectively reflect the real situation, and after obtaining the hot words, the relationship between the enterprise and the hot words is comprehensively considered, and the analysis results have stronger guiding significance for the enterprise.
- FIG. 1 is a flow chart of an embodiment of a hot word analysis method according to an embodiment of the present application
- FIG. 2 is a schematic flow chart of a setting process of a first exposure threshold and a second exposure threshold in a specific implementation
- FIG. 3 is a schematic flow chart of a setting process of a first exposure threshold and a second exposure threshold in another specific implementation
- FIG. 4 is a structural diagram of an embodiment of a hot word analysis apparatus according to an embodiment of the present application.
- FIG. 5 is a schematic block diagram of a hot word analysis terminal device according to an embodiment of the present application.
- an embodiment of a hot word analysis method in an embodiment of the present application may include:
- Step S101 Crawling a webpage published on the target website in the current statistical period by using a search engine.
- the target website is a website with a page view that is larger than the preset page threshold.
- the threshold of the page view may be set to 100,000 times, 500,000 times, 1 million times, etc. according to actual conditions, and the target website may be Baidu. News (http://news.baidu.com/), Netease News (http://news.163.com/), Tencent News (http://news.qq.com/), Phoenix News (http:/ /news.ifeng.com/) and other news sites or other news sites.
- the statistical period can be set to one day, one week, half a month or one month according to the actual situation.
- Step S102 Perform word-cutting processing on the text information in the webpage to obtain each word segment constituting the text information.
- the word processing refers to dividing a sentence text into a single word, that is, each of the word segments.
- the sentence text can be segmented according to the general dictionary to ensure that the words that are separated are normal words. If the words are not in the dictionary, the words are separated.
- Step S103 Count the frequency of exposure of each participle in the text information.
- the number of occurrences of each participle in the text information is separately counted.
- Step S104 Determine a participle whose exposure frequency in the text information is greater than a preset first exposure threshold as a hot word.
- the first exposure threshold may be set to 10000. If the exposure frequency of the “nineteenth largest” is 20000 in the current statistical period, which is greater than the first exposure threshold, it is directly confirmed as a hot word.
- the participles in the text information whose exposure frequency is less than or equal to the first exposure threshold and greater than the preset second exposure threshold may also be determined as candidate word segments, and then each candidate is obtained from the historical statistical record.
- the frequency of exposure of the word segmentation within T statistical periods before the current statistical period, and the candidate participle satisfying the following conditions are also determined as hot words:
- n is the sequence number of the candidate participle, 1 ⁇ n ⁇ N, N is the total number of the candidate participles, t is the sequence number of each statistical period arranged in chronological order, 1 ⁇ t ⁇ T, T is a positive integer , ExpNum n,t is the frequency of exposure of the nth candidate participle in the tth statistical period, ExpNum n, T+1 is the frequency of exposure of the nth candidate participle in the current statistical period, ln is the natural logarithm
- ThreshRatio is the preset proportional threshold.
- the exposure amount of "artificial intelligence" is 1500 times in the current statistical period, not only less than the first exposure threshold, but less than the second exposure threshold, it is directly confirmed as a common word.
- the determined hot words may also be filtered, that is, the words that cause interference are filtered out from the determined hot words, and some common interference hot words may be preset, for example, like “we”, “ everybody, “this” and so on.
- the frequency of exposure of these interfering hot words has nothing to do with the news content, that is, regardless of the news content, the frequency of exposure of these interfering hot words may remain greater than the first exposure threshold.
- hot word statistics if you do not filter these interference hot words, it will affect the accuracy of the analysis results. Therefore, you need to filter out the hot words from the determined hot words to get the filtered hot words. Get the hot words you really need.
- the preset interference hot words can be obtained from the data list, and then all the determined hot words are compared with all the interference hot words one by one, if a certain hot word and some If the interfering hot words are consistent, they will be filtered out. Otherwise, if a hot word is inconsistent with any interfering hot words, the hot word is retained, and the last retained hot word is the filtered hot word.
- Step S105 counting the frequency of exposure of each enterprise name in the preferred text information.
- the preferred text information is text information including the hot word.
- the current statistical period may include M sub-cycles, where M is a positive integer, and the frequency of exposure of each enterprise name in the preferred text information of each sub-period is calculated.
- Step S106 Calculate the degree of association between each enterprise name and the hot word according to the frequency of exposure of each enterprise name in the preferred text information.
- the degree of association between each enterprise name and the hot word can be calculated according to the following formula:
- q is the serial number of the enterprise name, 1 ⁇ q ⁇ Q
- Q is the total number of enterprise names
- p is the serial number of the hot words
- 1 ⁇ p ⁇ P is the total number of hot words
- m is the time of each sub-period according to time
- the serial number sequentially arranged 1 ⁇ m ⁇ M
- EntExpNum q,p,m is the frequency of exposure of the qth enterprise name in the preferred text information containing the pth hot word in the mth sub-period
- k m is Preset weight coefficient
- Rel q,p is the degree of association between the qth business name and the pth hot word.
- the association threshold can be set to 80% according to the actual situation. 90% or 95% and so on.
- the above is the enterprise corresponding to the hot word association, and the other related angle is the hot word corresponding to the enterprise association.
- the webpage is searched for text information showing the name of the enterprise, and then the hot words are searched for in the text information of the enterprise, the frequency of occurrence of each hot word is counted, and the hot words are sorted according to the frequency from largest to smallest. The higher the ranking of the hot words and the higher the relevance of the enterprise, the lower the relevance of the hot words after the ranking to the enterprise.
- the company name should include its other name.
- the company name should include its other name.
- the company name when analyzing Tencent, not only search for "Tencent”, but also search for "Tencent”, “goose factory”, etc., for Facebook.
- an nickname list of the enterprise name may be preset, and the correspondence between the official name of the enterprise and the other name may be recorded.
- an nickname corresponding to the enterprise is obtained from the list, and the other name is also counted into the pair. The statistical process of the enterprise.
- the setting process of the first exposure threshold and the second exposure threshold may include the steps shown in FIG. 2:
- Step S201 Obtain the frequency of exposure of each historical hot word in each statistical period from the historical statistical record.
- the historical hot words are hot words that have been determined before the current statistical period.
- Step S202 Constructing a first exposure sequence of each historical hot word.
- the first exposure sequence of each historical hot word can be constructed according to the following formula:
- nh is the serial number of the historical hot word, 1 ⁇ nh ⁇ NH
- NH is the total number of historical hot words
- th is the serial number of each statistical period arranged in chronological order
- 1 ⁇ th ⁇ TH nh is the total number of statistical cycle of nh historical hot words
- HsExpNum nh is the total number of statistical cycle of nh historical hot words
- ExpSeq1 nh for the first nh historical hot words of the first exposure sequence .
- Step S203 calculating an average value of each of the first exposure sequences.
- each first exposure sequence can be calculated according to the following formula:
- AvExpSeq1 nh is the mean of the nth first exposure sequences.
- Step S204 Constructing a sequence in which the average values of the respective first exposure sequences are sequentially arranged in descending order.
- sequence in which the average values of the respective first exposure sequences are sequentially arranged in descending order can be constructed according to the following formula:
- AvExpSeq1 nh1 ' is the mean value of the first exposure sequence arranged in the nth position from the largest to the smallest, 1 ⁇ nh1 ⁇ NH.
- Step S205 calculating the first exposure threshold and the second exposure threshold.
- the first exposure threshold may be calculated according to the following formula:
- NMAX floor( ⁇ max ⁇ NH), ⁇ max is a preset coefficient, and 0 ⁇ max ⁇ 1, floor is a downward value function, and Threshold1 is the first exposure threshold;
- the second exposure threshold is calculated according to the following formula:
- NMIN floor( ⁇ min ⁇ NH)
- ⁇ min is a preset coefficient
- 0 ⁇ min ⁇ 1 is the second exposure threshold
- the setting process of the first exposure threshold and the second exposure threshold may include the steps shown in FIG. 3:
- Step S301 Obtain an exposure frequency of each historical hot word in each statistical period from the historical statistical record.
- Step S302 constructing a first exposure sequence of each historical hot word.
- Step S303 calculating an average value of each of the first exposure sequences.
- step S301 to the step S303 is the same as the process of the step S201 to the step S203.
- steps S301 to the step S303 are the same as the process of the step S201 to the step S203.
- Step S304 constructing a second exposure sequence of each historical hot word.
- the second exposure sequence of each historical hot word can be constructed according to the following formula:
- HsExpNum nh, th1 ' ⁇ ExpSeq1 nh , 1 ⁇ th1 ⁇ TH nh , HsExpNum nh, th1 ' ⁇ HsExpNum nh, th1+1 ', ExpSeq2 nh is the second exposure sequence of the nthth historical hot word.
- Step S305 calculating an average value of each of the second exposure sequences.
- each second exposure sequence can be calculated according to the following formula:
- AvExpSeq2 nh is the mean of the nth second exposure sequences
- TH1 nh satisfies the following condition: HsExpNum nh, TH1 ' ⁇ AvExpSeq1 nh and HsExpNum nh, TH1+1 ' ⁇ AvExpSeq1 nh .
- Step S306 constructing a sequence in which the average values of the respective second exposure sequences are sequentially arranged in descending order.
- sequence in which the average values of the respective second exposure sequences are sequentially arranged in descending order can be constructed according to the following formula:
- AvExpSeq2 nh1 ' is the mean value of the second exposure sequence arranged in the nth bit position in descending order.
- Step S307 calculating the first exposure threshold and the second exposure threshold.
- the first exposure threshold may be calculated according to the following formula:
- the second exposure threshold is calculated according to the following formula:
- the embodiment of the present application provides an objective evaluation standard for the determination of hot words, and gets rid of the dependence on the personal ability of the network analyst, and the determined hot words can be more difficult to objectively reflect the real situation, and After getting the hot words, the relationship between the enterprise and the hot words is considered comprehensively, and the analysis results have a stronger guiding significance for the enterprise.
- FIG. 4 is a structural diagram showing an embodiment of a hot word analyzing device provided by an embodiment of the present application.
- a hot word analysis device may include:
- the webpage crawling module 401 is configured to crawl a webpage published on a target website in a current statistical period by using a search engine, where the target website is a website whose browsing amount is greater than a preset browsing threshold;
- the word processing module 402 is configured to perform word segmentation processing on the text information in the webpage to obtain respective word segments constituting the text information;
- a first statistic module 403, configured to count frequency of exposure of each participle in the text information
- a first hot word determining module 404 configured to determine, in the text information, a participle whose exposure frequency is greater than a preset first exposure threshold as a hot word;
- a second statistic module 405, configured to count frequency of exposure of each enterprise name in the preferred text information, where the preferred text information is text information including the hot word;
- the association degree calculation module 406 is configured to calculate the degree of association between each enterprise name and the hot word according to the frequency of exposure of the respective enterprise names in the preferred text information.
- the hot word analysis device may further include:
- a candidate participle determining module configured to determine, in the text information, a participle whose exposure frequency is less than or equal to the first exposure threshold and greater than a preset second exposure threshold as a candidate participle;
- a third statistic module configured to obtain, from a historical statistical record, an exposure frequency of each candidate participle in T statistical periods before the current statistical period, where T is a positive integer;
- a second hot word determining module configured to determine candidate word segments satisfying the following conditions as hot words:
- n is the serial number of the candidate participle, 1 ⁇ n ⁇ N, N is the total number of the candidate participles, and t is the serial number of each statistical period arranged in chronological order, 1 ⁇ t ⁇ T, ExpNum n, t is the frequency of exposure of the nth candidate participle in the t-th statistical period, ExpNum n, T+1 is the frequency of exposure of the nth candidate participle in the current statistical period, and ln is a natural logarithm function. ThreshRatio is a preset proportional threshold;
- the hot word analysis device may further include:
- a fourth statistic module configured to obtain, from a historical statistical record, an exposure frequency of each historical hot word in each statistical period, where the historical hot word is a hot word that has been determined before the current statistical period;
- a first exposure sequence construction module for constructing a first exposure sequence of each historical hot word according to the following formula:
- nh is the serial number of the historical hot word, 1 ⁇ nh ⁇ NH
- NH is the total number of historical hot words
- th is the serial number of each statistical period arranged in chronological order
- 1 ⁇ th ⁇ TH nh is the total number of statistical cycle of nh historical hot words
- the first exposure sequence mean calculation module is configured to calculate an average value of each first exposure sequence according to the following formula:
- AvExpSeq1 nh is the mean of the nth first exposure sequences
- the first mean sequence construction module is configured to construct a sequence in which the average values of the respective first exposure sequences are arranged in descending order according to the following formula:
- AvExpSeq1 nh1 ' is the mean value of the first exposure sequence arranged in the nh1th position in descending order, 1 ⁇ nh1 ⁇ NH;
- a first exposure threshold calculation module configured to calculate the first exposure threshold according to the following formula:
- NMAX floor( ⁇ max ⁇ NH), ⁇ max is a preset coefficient, and 0 ⁇ max ⁇ 1, floor is a downward value function, and Threshold1 is the first exposure threshold;
- a second exposure threshold calculation module configured to calculate the second exposure threshold according to the following formula:
- NMIN floor( ⁇ min ⁇ NH)
- ⁇ min is a preset coefficient
- 0 ⁇ min ⁇ 1 is the second exposure threshold
- the hot word analysis device may further include:
- a second exposure sequence construction module is configured to construct a second exposure sequence of each historical hot word according to the following formula:
- HsExpNum nh, th1 ' ⁇ ExpSeq1 nh , 1 ⁇ th1 ⁇ TH nh , HsExpNum nh, th1 ' ⁇ HsExpNum nh, th1+1 ', ExpSeq2 nh is the second exposure sequence of the nthth historical hot word;
- a second exposure sequence mean calculation module for calculating an average of each second exposure sequence according to the following formula:
- AvExpSeq2 nh is the mean of the nth second exposure sequences
- TH1 nh satisfies the following condition: HsExpNum nh, TH1 ' ⁇ AvExpSeq1 nh and HsExpNum nh, TH1+1 ' ⁇ AvExpSeq1 nh ;
- the second mean sequence construction module is configured to construct a sequence in which the average values of the respective second exposure sequences are arranged in descending order according to the following formula:
- AvExpSeq2 nh1 ' is the mean value of the second exposure sequence arranged in the nth bit position in descending order
- a first exposure threshold calculation module configured to calculate the first exposure threshold according to the following formula:
- a second exposure threshold calculation module configured to calculate the second exposure threshold according to the following formula:
- the current statistic period includes M sub-cycles, where M is a positive integer
- the second statistic module may include:
- a sub-period statistical unit for counting the frequency of exposure of each enterprise name in the preferred text information of each sub-period
- the association calculation module may include:
- a first calculating unit configured to calculate a degree of association between each enterprise name and the hot word according to the following formula:
- q is the serial number of the enterprise name, 1 ⁇ q ⁇ Q
- Q is the total number of enterprise names
- p is the serial number of the hot words
- 1 ⁇ p ⁇ P is the total number of hot words
- m is the time of each sub-period according to time
- the serial number sequentially arranged 1 ⁇ m ⁇ M
- EntExpNum q,p,m is the frequency of exposure of the qth enterprise name in the preferred text information containing the pth hot word in the mth sub-period
- k m is Preset weight coefficient
- Rel q,p is the degree of association between the qth business name and the pth hot word.
- FIG. 5 is a schematic block diagram of a hot word analysis terminal device provided by an embodiment of the present application. For convenience of description, only parts related to the embodiment of the present application are shown.
- the hot word analysis terminal device 5 may be a computing device such as a mobile phone, a tablet computer, a desktop computer, a notebook, and a cloud server.
- the hot word analysis terminal device 5 may include a processor 50, a memory 51, and computer readable instructions 52 stored in the memory 51 and operable on the processor 50, such as performing the hot word analysis method described above.
- the processor 50 executes the steps in the embodiments of the various hot word analysis methods described above when the computer readable instructions 52 are executed.
- the functional units in the various embodiments of the present application may be stored in a computer readable storage medium if implemented in the form of a software functional unit and sold or used as a separate product. Based on such understanding, the technical solution of the present application, in essence or the contribution to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium.
- a number of computer readable instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present application.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Description
Claims (20)
- 一种热词分析方法,其特征在于,包括:通过搜索引擎爬取当前统计周期内在目标网站上发布的网页,所述目标网站为浏览量大于预设的浏览量阈值的网站;对所述网页中的文本信息进行切词处理,得到构成所述文本信息的各个分词;统计各个分词在所述文本信息中的曝光频次;将在所述文本信息中的曝光频次大于预设的第一曝光阈值的分词确定为热词;统计各个企业名称在优选文本信息中的曝光频次,所述优选文本信息为包含所述热词的文本信息;根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所述热词之间的关联度。
- 根据权利要求1所述的热词分析方法,其特征在于,在所述统计各个分词在所述文本信息中的曝光频次之后,还包括:将在所述文本信息中的曝光频次小于或等于所述第一曝光阈值且大于预设的第二曝光阈值的分词确定为候选分词;从历史统计记录中获取各个候选分词在所述当前统计周期之前的T个统计周期内的曝光频次,其中,T为正整数;将满足下述条件的候选分词确定为热词:
- 根据权利要求2所述的热词分析方法,其特征在于,所述第一曝光阈值和所述第二曝光阈值的设置过程包括:从历史统计记录中获取各个历史热词在各个统计周期内的曝光频次,所述历史热词为在所述当前统计周期之前已经确定的热词;根据下式构造各个历史热词的第一曝光序列:其中,nh为所述历史热词的序号,1≤nh≤NH,NH为所述历史热词的总数,th为各个统计周期按照时间先后顺序依次排列的序号,1≤th≤TH nh,TH nh为第nh个历史热词的统计周期的总数,HsExpNum nh,th为第nh个历史热词在第th个统计周期内的曝光频次,ExpSeq1 nh为第nh个历史热词的第一曝光序列;根据下式计算各个第一曝光序列的均值:其中,AvExpSeq1 nh为第nh个第一曝光序列的均值;根据下式构造各个第一曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq1 1′,AvExpSeq1 2′,......,AvExpSeq1 nh1′,......,AvExpSeq1 NH′}其中,AvExpSeq1 nh1′为按照从大到小的顺序排列在第nh1位上的第一曝光序列的均值,1≤nh1≤NH;根据下式计算所述第一曝光阈值:其中,NMAX=floor(ξ max×NH),ξ max为预设的系数,且0<ξ max<1,floor为向下取值函数,Threshold1为所述第一曝光阈值;根据下式计算所述第二曝光阈值:其中,NMIN=floor(ξ min×NH),ξ min为预设的系数,且0<ξ min<1,Threshold2为所述第二曝光阈值。
- 根据权利要求3所述的热词分析方法,其特征在于,在计算各个第一曝光序列的均值之后,还包括:根据下式构造各个历史热词的第二曝光序列:其中,HsExpNum nh,th1′∈ExpSeq1 nh,1≤th1≤TH nh,HsExpNum nh,th1′≥HsExpNum nh,th1+1′,ExpSeq2 nh为第nh个历史热词的第二曝光序列;根据下式计算各个第二曝光序列的均值:其中,AvExpSeq2 nh为第nh个第二曝光序列的均值,TH1 nh满足以下条件:HsExpNum nh,TH1′≥AvExpSeq1 nh且HsExpNum nh,TH1+1′<AvExpSeq1 nh;根据下式构造各个第二曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq2 1′,AvExpSeq2 2′,......,AvExpSeq2 nh1′,......,AvExpSeq2 NH′}其中,AvExpSeq2 nh1′为按照从大到小的顺序排列在第nh1位上的第二曝光序列的均值;根据下式计算所述第一曝光阈值:根据下式计算所述第二曝光阈值:
- 根据权利要求1至4中任一项所述的热词分析方法,其特征在于,所述当前统计周期包括M个子周期,其中,M为正整数,所述统计各个企业名称在优选文本信息中的曝光频次包括:统计各个企业名称在各个子时间段的优选文本信息中的曝光频次;所述根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所述热词之间的关联度包括:根据下式计算各个企业名称与所述热词之间的关联度:
- 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机可读指令,其特征在于,所述计算机可读指令被处理器执行时实现如下步骤:通过搜索引擎爬取当前统计周期内在目标网站上发布的网页,所述目标网站为浏览量大于预设的浏览量阈值的网站;对所述网页中的文本信息进行切词处理,得到构成所述文本信息的各个分词;统计各个分词在所述文本信息中的曝光频次;将在所述文本信息中的曝光频次大于预设的第一曝光阈值的分词确定为热词;统计各个企业名称在优选文本信息中的曝光频次,所述优选文本信息为包含所述热词的文本信息;根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所述热词之间的关联度。
- 根据权利要求6所述的计算机可读存储介质,其特征在于,在所述统计各个分 词在所述文本信息中的曝光频次之后,还包括:将在所述文本信息中的曝光频次小于或等于所述第一曝光阈值且大于预设的第二曝光阈值的分词确定为候选分词;从历史统计记录中获取各个候选分词在所述当前统计周期之前的T个统计周期内的曝光频次,其中,T为正整数;将满足下述条件的候选分词确定为热词:
- 根据权利要求7所述的计算机可读存储介质,其特征在于,所述第一曝光阈值和所述第二曝光阈值的设置过程包括:从历史统计记录中获取各个历史热词在各个统计周期内的曝光频次,所述历史热词为在所述当前统计周期之前已经确定的热词;根据下式构造各个历史热词的第一曝光序列:其中,nh为所述历史热词的序号,1≤nh≤NH,NH为所述历史热词的总数,th为各个统计周期按照时间先后顺序依次排列的序号,1≤th≤TH nh,TH nh为第nh个历史热词的统计周期的总数,HsExpNum nh,th为第nh个历史热词在第th个统计周期内的曝光频次,ExpSeq1 nh为第nh个历史热词的第一曝光序列;根据下式计算各个第一曝光序列的均值:其中,AvExpSeq1 nh为第nh个第一曝光序列的均值;根据下式构造各个第一曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq1 1′,AvExpSeq1 2′,......,AvExpSeq1 nh1′,......,AvExpSeq1 NH′}其中,AvExpSeq1 nh1′为按照从大到小的顺序排列在第nh1位上的第一曝光序列的均值,1≤nh1≤NH;根据下式计算所述第一曝光阈值:其中,NMAX=floor(ξ max×NH),ξ max为预设的系数,且0<ξ max<1,floor为向下 取值函数,Threshold1为所述第一曝光阈值;根据下式计算所述第二曝光阈值:其中,NMIN=floor(ξ min×NH),ξ min为预设的系数,且0<ξ min<1,Threshold2为所述第二曝光阈值。
- 根据权利要求8所述的计算机可读存储介质,其特征在于,在计算各个第一曝光序列的均值之后,还包括:根据下式构造各个历史热词的第二曝光序列:其中,HsExpNum nh,th1′∈ExpSeq1 nh,1≤th1≤TH nh,HsExpNum nh,th1′≥HsExpNum nh,th1+1′,ExpSeq2 nh为第nh个历史热词的第二曝光序列;根据下式计算各个第二曝光序列的均值:其中,AvExpSeq2 nh为第nh个第二曝光序列的均值,TH1 nh满足以下条件:HsExpNum nh,TH1′≥AvExpSeq1 nh且HsExpNum nh,TH1+1′<AvExpSeq1 nh;根据下式构造各个第二曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq2 1′,AvExpSeq2 2′,......,AvExpSeq2 nh1′,......,AvExpSeq2 NH′}其中,AvExpSeq2 nh1′为按照从大到小的顺序排列在第nh1位上的第二曝光序列的均值;根据下式计算所述第一曝光阈值:根据下式计算所述第二曝光阈值:
- 根据权利要求6至9中任一项所述的计算机可读存储介质,其特征在于,所述当前统计周期包括M个子周期,其中,M为正整数,所述统计各个企业名称在优选文本信息中的曝光频次包括:统计各个企业名称在各个子时间段的优选文本信息中的曝光频次;所述根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所 述热词之间的关联度包括:根据下式计算各个企业名称与所述热词之间的关联度:
- 一种热词分析终端设备,包括存储器、处理器以及存储在所述存储器中并可在所述处理器上运行的计算机可读指令,其特征在于,所述处理器执行所述计算机可读指令时实现如下步骤:通过搜索引擎爬取当前统计周期内在目标网站上发布的网页,所述目标网站为浏览量大于预设的浏览量阈值的网站;对所述网页中的文本信息进行切词处理,得到构成所述文本信息的各个分词;统计各个分词在所述文本信息中的曝光频次;将在所述文本信息中的曝光频次大于预设的第一曝光阈值的分词确定为热词;统计各个企业名称在优选文本信息中的曝光频次,所述优选文本信息为包含所述热词的文本信息;根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所述热词之间的关联度。
- 根据权利要求11所述的热词分析终端设备,其特征在于,在所述统计各个分词在所述文本信息中的曝光频次之后,还包括:将在所述文本信息中的曝光频次小于或等于所述第一曝光阈值且大于预设的第二曝光阈值的分词确定为候选分词;从历史统计记录中获取各个候选分词在所述当前统计周期之前的T个统计周期内的曝光频次,其中,T为正整数;将满足下述条件的候选分词确定为热词:
- 根据权利要求12所述的热词分析终端设备,其特征在于,所述第一曝光阈值和所述第二曝光阈值的设置过程包括:从历史统计记录中获取各个历史热词在各个统计周期内的曝光频次,所述历史热词为在所述当前统计周期之前已经确定的热词;根据下式构造各个历史热词的第一曝光序列:其中,nh为所述历史热词的序号,1≤nh≤NH,NH为所述历史热词的总数,th为各个统计周期按照时间先后顺序依次排列的序号,1≤th≤TH nh,TH nh为第nh个历史热词的统计周期的总数,HsExpNum nh,th为第nh个历史热词在第th个统计周期内的曝光频次,ExpSeq1 nh为第nh个历史热词的第一曝光序列;根据下式计算各个第一曝光序列的均值:其中,AvExpSeq1 nh为第nh个第一曝光序列的均值;根据下式构造各个第一曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq1 1′,AvExpSeq1 2′,......,AvExpSeq1 nh1′,......,AvExpSeq1 NH′}其中,AvExpSeq1 nh1′为按照从大到小的顺序排列在第nh1位上的第一曝光序列的均值,1≤nh1≤NH;根据下式计算所述第一曝光阈值:其中,NMAX=floor(ξ max×NH),ξ max为预设的系数,且0<ξ max<1,floor为向下取值函数,Threshold1为所述第一曝光阈值;根据下式计算所述第二曝光阈值:其中,NMIN=floor(ξ min×NH),ξ min为预设的系数,且0<ξ min<1,Threshold2为所述第二曝光阈值。
- 根据权利要求13所述的热词分析终端设备,其特征在于,在计算各个第一曝光序列的均值之后,还包括:根据下式构造各个历史热词的第二曝光序列:其中,HsExpNum nh,th1′∈ExpSeq1 nh,1≤th1≤TH nh,HsExpNum nh,th1′≥HsExpNum nh,th1+1′,ExpSeq2 nh为第nh个历史热词的第二曝光序列;根据下式计算各个第二曝光序列的均值:其中,AvExpSeq2 nh为第nh个第二曝光序列的均值,TH1 nh满足以下条件:HsExpNum nh,TH1′≥AvExpSeq1 nh且HsExpNum nh,TH1+1′<AvExpSeq1 nh;根据下式构造各个第二曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq2 1′,AvExpSeq2 2′,......,AvExpSeq2 nh1′,......,AvExpSeq2 NH′}其中,AvExpSeq2 nh1′为按照从大到小的顺序排列在第nh1位上的第二曝光序列的均值;根据下式计算所述第一曝光阈值:根据下式计算所述第二曝光阈值:
- 根据权利要求11至14中任一项所述的热词分析终端设备,其特征在于,所述当前统计周期包括M个子周期,其中,M为正整数,所述统计各个企业名称在优选文本信息中的曝光频次包括:统计各个企业名称在各个子时间段的优选文本信息中的曝光频次;所述根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所述热词之间的关联度包括:根据下式计算各个企业名称与所述热词之间的关联度:
- 一种热词分析装置,其特征在于,包括:网页爬取模块,用于通过搜索引擎爬取当前统计周期内在目标网站上发布的网页,所述目标网站为浏览量大于预设的浏览量阈值的网站;切词处理模块,用于对所述网页中的文本信息进行切词处理,得到构成所述文本信息的各个分词;第一统计模块,用于统计各个分词在所述文本信息中的曝光频次;第一热词确定模块,用于将在所述文本信息中的曝光频次大于预设的第一曝光阈值的分词确定为热词;第二统计模块,用于统计各个企业名称在优选文本信息中的曝光频次,所述优选文本信息为包含所述热词的文本信息;关联度计算模块,用于根据各个企业名称在所述优选文本信息中的曝光频次计算各个企业名称与所述热词之间的关联度。
- 根据权利要求16所述的热词分析装置,其特征在于,还包括:候选分词确定模块,用于将在所述文本信息中的曝光频次小于或等于所述第一曝光阈值且大于预设的第二曝光阈值的分词确定为候选分词;第三统计模块,用于从历史统计记录中获取各个候选分词在所述当前统计周期之前的T个统计周期内的曝光频次,其中,T为正整数;第二热词确定模块,用于将满足下述条件的候选分词确定为热词:
- 根据权利要求17所述的热词分析装置,其特征在于,还包括:第四统计模块,用于从历史统计记录中获取各个历史热词在各个统计周期内的曝光频次,所述历史热词为在所述当前统计周期之前已经确定的热词;第一曝光序列构造模块,用于根据下式构造各个历史热词的第一曝光序列:其中,nh为所述历史热词的序号,1≤nh≤NH,NH为所述历史热词的总数,th为各个统计周期按照时间先后顺序依次排列的序号,1≤th≤TH nh,TH nh为第nh个历史热词的统计周期的总数,HsExpNum nh,th为第nh个历史热词在第th个统计周期内的曝光频次,ExpSeq1 nh为第nh个历史热词的第一曝光序列;第一曝光序列均值计算模块,用于根据下式计算各个第一曝光序列的均值:其中,AvExpSeq1 nh为第nh个第一曝光序列的均值;第一均值序列构造模块,用于根据下式构造各个第一曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq1 1′,AvExpSeq1 2′,......,AvExpSeq1 nh1′,......,AvExpSeq1 NH′}其中,AvExpSeq1 nh1′为按照从大到小的顺序排列在第nh1位上的第一曝光序列的均值,1≤nh1≤NH;第一曝光阈值计算模块,用于根据下式计算所述第一曝光阈值:其中,NMAX=floor(ξ max×NH),ξ max为预设的系数,且0<ξ max<1,floor为向下取值函数,Threshold1为所述第一曝光阈值;第二曝光阈值计算模块,用于根据下式计算所述第二曝光阈值:其中,NMIN=floor(ξ min×NH),ξ min为预设的系数,且0<ξ min<1,Threshold2为所述第二曝光阈值。
- 根据权利要求18所述的热词分析装置,其特征在于,还包括:第二曝光序列构造模块,用于根据下式构造各个历史热词的第二曝光序列:其中,HsExpNum nh,th1′∈ExpSeq1 nh,1≤th1≤TH nh,HsExpNum nh,th1′≥HsExpNum nh,th1+1′,ExpSeq2 nh为第nh个历史热词的第二曝光序列;第二曝光序列均值计算模块,用于根据下式计算各个第二曝光序列的均值:其中,AvExpSeq2 nh为第nh个第二曝光序列的均值,TH1 nh满足以下条件:HsExpNum nh,TH1′≥AvExpSeq1 nh且HsExpNum nh,TH1+1′<AvExpSeq1 nh;第二均值序列构造模块,用于根据下式构造各个第二曝光序列的均值按照从大到小的顺序依次排列的序列:{AvExpSeq2 1′,AvExpSeq2 2′,......,AvExpSeq2 nh1′,......,AvExpSeq2 NH′}其中,AvExpSeq2 nh1′为按照从大到小的顺序排列在第nh1位上的第二曝光序列的均 值;第一曝光阈值计算模块,用于根据下式计算所述第一曝光阈值:第二曝光阈值计算模块,用于根据下式计算所述第二曝光阈值:
- 根据权利要求16至19中任一项所述的热词分析装置,其特征在于,所述当前统计周期包括M个子周期,其中,M为正整数,所述第二统计模块包括:子时间段统计单元,用于统计各个企业名称在各个子时间段的优选文本信息中的曝光频次;所述关联度计算模块可以包括:第一计算单元,用于根据下式计算各个企业名称与所述热词之间的关联度:
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810456973.4 | 2018-05-14 | ||
| CN201810456973.4A CN108710664B (zh) | 2018-05-14 | 2018-05-14 | 一种热词分析方法、计算机可读存储介质及终端设备 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2019218452A1 true WO2019218452A1 (zh) | 2019-11-21 |
Family
ID=63868099
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2018/096267 Ceased WO2019218452A1 (zh) | 2018-05-14 | 2018-07-19 | 热词分析方法、计算机可读存储介质、终端设备及装置 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN108710664B (zh) |
| WO (1) | WO2019218452A1 (zh) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116155783A (zh) * | 2022-12-29 | 2023-05-23 | 上海哔哩哔哩科技有限公司 | 热点数据的检测方法、装置、存储介质及电子设备 |
| CN116860967A (zh) * | 2023-06-21 | 2023-10-10 | 时趣互动(北京)科技有限公司 | 一种文案热度预估方法、装置、电子设备及存储介质 |
| CN117034904A (zh) * | 2023-10-09 | 2023-11-10 | 北京睿企信息科技有限公司 | 一种获取热度稳定的热词的方法、电子设备及存储介质 |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111310018B (zh) * | 2018-12-11 | 2024-03-01 | 阿里巴巴集团控股有限公司 | 一种时效性搜索词汇的确定方法及搜索引擎 |
| CN110427381A (zh) * | 2019-08-07 | 2019-11-08 | 北京嘉和海森健康科技有限公司 | 一种数据处理方法及相关设备 |
| CN111737553A (zh) * | 2020-06-16 | 2020-10-02 | 苏州朗动网络科技有限公司 | 企业关联词的选取方法、设备和存储介质 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246499A (zh) * | 2008-03-27 | 2008-08-20 | 腾讯科技(深圳)有限公司 | 网络信息搜索方法及系统 |
| CN103106227A (zh) * | 2012-08-03 | 2013-05-15 | 人民搜索网络股份公司 | 一种基于网页文本的新词查找系统及方法 |
| CN103186675A (zh) * | 2013-04-03 | 2013-07-03 | 南京安讯科技有限责任公司 | 一种基于网络热词识别的网页自动分类方法 |
| CN105045882A (zh) * | 2015-07-21 | 2015-11-11 | 无锡天脉聚源传媒科技有限公司 | 一种热词处理方法及装置 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101727494B (zh) * | 2009-12-29 | 2012-03-28 | 华中师范大学 | 特定区域内网络热词生成系统 |
| US20170169062A1 (en) * | 2015-12-14 | 2017-06-15 | Le Holdings (Beijing) Co., Ltd. | Method and electronic device for recommending video |
-
2018
- 2018-05-14 CN CN201810456973.4A patent/CN108710664B/zh active Active
- 2018-07-19 WO PCT/CN2018/096267 patent/WO2019218452A1/zh not_active Ceased
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101246499A (zh) * | 2008-03-27 | 2008-08-20 | 腾讯科技(深圳)有限公司 | 网络信息搜索方法及系统 |
| CN103106227A (zh) * | 2012-08-03 | 2013-05-15 | 人民搜索网络股份公司 | 一种基于网页文本的新词查找系统及方法 |
| CN103186675A (zh) * | 2013-04-03 | 2013-07-03 | 南京安讯科技有限责任公司 | 一种基于网络热词识别的网页自动分类方法 |
| CN105045882A (zh) * | 2015-07-21 | 2015-11-11 | 无锡天脉聚源传媒科技有限公司 | 一种热词处理方法及装置 |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116155783A (zh) * | 2022-12-29 | 2023-05-23 | 上海哔哩哔哩科技有限公司 | 热点数据的检测方法、装置、存储介质及电子设备 |
| CN116860967A (zh) * | 2023-06-21 | 2023-10-10 | 时趣互动(北京)科技有限公司 | 一种文案热度预估方法、装置、电子设备及存储介质 |
| CN117034904A (zh) * | 2023-10-09 | 2023-11-10 | 北京睿企信息科技有限公司 | 一种获取热度稳定的热词的方法、电子设备及存储介质 |
| CN117034904B (zh) * | 2023-10-09 | 2023-12-08 | 北京睿企信息科技有限公司 | 一种获取热度稳定的热词的方法、电子设备及存储介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108710664B (zh) | 2023-04-18 |
| CN108710664A (zh) | 2018-10-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11003726B2 (en) | Method, apparatus, and system for recommending real-time information | |
| WO2019218452A1 (zh) | 热词分析方法、计算机可读存储介质、终端设备及装置 | |
| US11710054B2 (en) | Information recommendation method, apparatus, and server based on user data in an online forum | |
| US9595053B1 (en) | Product recommendation using sentiment and semantic analysis | |
| CN107451199B (zh) | 问题推荐方法及装置、设备 | |
| CN111651666B (zh) | 用户主题推荐方法、装置、计算机设备及存储介质 | |
| US20170242919A1 (en) | Analysis of Unstructured Computer Text to Generate Themes and Determine Sentiment | |
| US20170018030A1 (en) | System and Method for Determining Credit Worthiness of a User | |
| US20180181667A1 (en) | System and method to model recognition statistics of data objects in a business database | |
| CN112559747A (zh) | 事件分类处理方法、装置、电子设备和存储介质 | |
| CN104424308A (zh) | 网页分类标准获取方法、装置及网页分类方法、装置 | |
| CN107885783B (zh) | 获取搜索词高相关分类的方法和装置 | |
| CN112559895A (zh) | 一种数据处理方法、装置、电子设备及存储介质 | |
| CN110321437A (zh) | 一种语料数据处理方法、装置、电子设备及介质 | |
| CN108763272B (zh) | 一种事件信息分析方法、计算机可读存储介质及终端设备 | |
| CN113961810A (zh) | 新闻推送方法、装置、计算机设备和存储介质 | |
| US12118322B2 (en) | System and method for generating subjective wellbeing analytics score | |
| CN112818221B (zh) | 实体的热度确定方法、装置、电子设备及存储介质 | |
| JP2017188004A (ja) | ソーシャルメディアにおける特定テーマの投稿の時系列変化を追跡的に分析するコンピューティング | |
| CN111723201A (zh) | 一种用于文本数据聚类的方法和装置 | |
| CN110059725B (zh) | 一种基于搜索关键词的检测恶意搜索系统及方法 | |
| US20160314477A1 (en) | Identifying entities trending in a professional community | |
| US20140324524A1 (en) | Evolving a capped customer linkage model using genetic models | |
| CN114331789B (zh) | 一种廉洁知识智能推荐方法、装置、设备及存储介质 | |
| US20140324523A1 (en) | Missing String Compensation In Capped Customer Linkage Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18918834 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 18918834 Country of ref document: EP Kind code of ref document: A1 |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 01.03.2021) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 18918834 Country of ref document: EP Kind code of ref document: A1 |












































