CN109145186A - A kind of mobile application data processing method - Google Patents
A kind of mobile application data processing method Download PDFInfo
- Publication number
- CN109145186A CN109145186A CN201810741622.8A CN201810741622A CN109145186A CN 109145186 A CN109145186 A CN 109145186A CN 201810741622 A CN201810741622 A CN 201810741622A CN 109145186 A CN109145186 A CN 109145186A
- Authority
- CN
- China
- Prior art keywords
- data
- comment
- mobile application
- keyword set
- speech
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Entrepreneurship & Innovation (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Marketing (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Data Mining & Analysis (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)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention belongs to mobile application field more particularly to a kind of mobile application data processing methods.A kind of mobile application data processing method of the invention, the huge comment data of data volume can be effectively treated, compress invalid data, it rationally and rapidly screens the evaluation to mobile application and analyzes useful data and carry out respective handling, the present invention sufficiently combines the taxeme of Chinese, and pointedly comment data is handled and stored, the collection processing speed of Chinese comment data can be greatly speeded up, it is convenient that this method reproduces, and tool is simple, has a good application prospect.
Description
Technical field
The invention belongs to mobile application field more particularly to a kind of mobile application data processing methods.
Background technique
With the fast development of mobile application, all kinds of mobile APP have become work indispensable in people's daily life
Tool, it is different from conventional entity tool, move APP background service and concrete operations be it is sightless, user can not or
Be only capable of demand under the freedom degree of very little to oneself and carry out active acquisition and analysis, these work be usually by mobile application after
It actively pushes to user after data acquisition and analysis in platform service or is changed in the way of update etc., this
Kind of mode makes user not need to understand the detailed process of mobile application and internal complicated control mechanism can be by
Dynamic acquisition corresponding information or data, but simultaneously, this passively acquisition modes make user during using mobile application
It was found that encounter problems or cannot be actively changed using when exception, also lack effective way can be by above- mentioned information
Manager's level that purposes has decision-making power is actively transferred directly to generate with the popularization of mobile application in this context
User experience data is just particularly important, and in all types of user experience data, directly delivers and be set out in incoming end by user
User comment on (be primarily referred to as the information such as each application market, forum and collect environment) webpage has most direct referential.It is right
In a long-term stability development and it is expected to obtain for more multi-user obtains mobile application, collects user comment, moved for evaluating
The using effect of dynamic application, analyzes the advantage and disadvantage of mobile application, and carrying out specific aim improvement is a necessary job, but due to
The substantial amounts of user comment, and due to market competition and the unconscious behavior of a large number of users, cause effectively to collect and analyze
User comment is difficult to carry out.
Summary of the invention
The purpose of the invention is, provides a kind of mobile application data processing method, with can be quickly and effectively real
Now to the selection screening of mobile application comment data, mentioned in order to extract the evaluation analysis that effective critical data is mobile application
For foundation.
To achieve the above object, the invention adopts the following technical scheme that.
A kind of mobile application data processing method, including following content,
One, for obtaining mobile application related data the step of, including obtain software mark data, the software mark number
According to including dbase, software classification and software brief introduction；
Two, for obtaining the step of commenting on relevant element data from user comment data, the element data includes
User comment, software version, comment port, comment time, User ID；Specifically comprise the following steps:
Step 1, access comment port, obtain web data, and search needs to grab the flag data of the mobile application of data,
The corresponding webpage of flag data is grabbed；It needs to judge the loading method of webpage when grabbing data, and according to
Different loading methods are using different analysis method crawl data, in particular to using the net of Jsoup analysis static loading mode
Page label data, using the web page tag data of HttpUnit crawl dynamically load mode；
Step 2 judges the corresponding mobile application of the flag data whether web data contains；If having each webpage of crawl simultaneously
It is stored in associated databases；The database is being established respectively according to each mobile application as classification factor with each mobile application phase
The database answered；
Step 3 judges whether there are also subsequent web pages, gos to step two if having, if without going to step 4；
All webpages with crawl are converted to text formatting, positioning element corresponding with comment data is obtained by step 4
Data；Positioning element data corresponding with comment data is obtained specifically includes,
Step 4.1, participle is carried out to user comment data and part-of-speech tagging is handled, extracts noun n, verb v and shape
Hold word a and constitutes keyword set；Define the keyword set K of corresponding i-th user comment of mobile applicationi, Ki={ w0/f0, w1/
f1..., wk/fk, wherein k=0,1,2......K-1, K are the participle number of i-th comment, wkFor kth in commentiA point
Word, fkFor wkPart of speech；The functions such as wherein participle tool refers to for being segmented, part-of-speech tagging, part of speech identify, words identification
Application tool；
Step 4.2 extracts keyword set KiIn only adjective a comment data, by w all in former keyword setkIt deposits
Keyword set K after entering optimizationNewi, wherein the keyword set K after optimizationNewiIs defined as:
KNewi={ w0, w1......wj, wherein j=0,1,2......J-1, J are the keyword after i-th comment optimization
Number, wjFor the j-th word in comment optimization keyword set；
Step 4.3 extracts keyword set KiIn containing { n+a }, { v+a }, { n+v }, { n+v+a } comment data, will be former
The w of non-a in keyword setkCorresponding part of speech deposit optimization keyword set KNewi, the corresponding crucial part of speech collection of part of speech deposit optimization
FNewi, wherein the keyword set F after optimizationNewiIs defined as:
FNewi={ f0、f1......fj, wherein fjFor the w in comment optimization keyword setjPart of speech；
Three, for extracting keyword set and calculating the step of element is to mobile application scoring weight in keyword set, tool
Body step includes:
Step 1 is established one's own feature dictionary to each mobile application, in particular to is built according to the part of speech of Feature Words
Vertical feature dictionary, including verb feature database, noun feature database, adjective feature database；
Step 2, the keyword w that each mobile application extracting keywords are concentratedj, wjIn corresponding part of speech fjFeature database in
Frequency Tj, include wjTextual data NjAnd text sum N；
Step 3, the weighted score S that each user comment and mobile application scoring are 1. calculated according to formula, and judge that it is obtained
Divide and whether be greater than threshold alpha, is then judged as effective comment if more than threshold value, is otherwise determined as useless comment；
Wherein mjRefer to wjCorresponding part of speech fjAverage characteristics frequency in feature database, J are the key after i-th comment optimization
Word number.
1. a kind of mobile application data processing method according to claim 1, the threshold alpha is one and closes for controlling
To control calculation amount or for rinsing the numerical value for selecting data correlation degree, which is manually specified keyword collection capacity after passing through statistics,
In the case that all data can be effectively treated in computing resource.
2. institute can be effectively treated in computing resource in a kind of mobile application data processing method according to claim 1
In the case where having data, threshold alpha=1.
A kind of mobile application data processing method according to claim 1, for each mobile application to be evaluated,
The user comment sum of acquisition should be not less than 1000, and the acquisition period of comment should be not less than 30 days, and the period refers to
Delivered earliest in database and user comment that is delivering the latest time interval.
The beneficial effect is that: data volume Pang can be effectively treated in a kind of mobile application data processing method of the invention
Big comment data compresses invalid data, rationally and rapidly screens the evaluation to mobile application and analyzes useful data
And respective handling is carried out, the present invention sufficiently combines the taxeme of Chinese, pointedly comment data is handled and stored,
The collection processing speed of Chinese comment data can be greatly speeded up, it is convenient that this method reproduces, and tool is simple, has good application
Prospect.
Detailed description of the invention
Fig. 1 is the flow diagram of mobile application comment data grasping means in the embodiment of the present invention；
Fig. 2 is the schematic diagram of keyword set extracting method；
Fig. 3 is the schematic diagram of keyword and Optimization Steps in the embodiment of the present invention；
Fig. 4 is the flow chart of user comment pertinence analysis method in the embodiment of the present invention.
Specific embodiment
It elaborates below in conjunction with specific embodiment to the invention.
In the present invention, choosing evaluation is rinsed to mobile application user comment to realize, needs to obtain the phase of corresponding mobile application
It closes data and user comment data, the related data of mobile application mainly includes software mark data to answer different movements
With carrying out statistic of classification, while targetedly user comment is selected according to the type of mobile application, it is soft in the present embodiment
Part flag data specifically includes dbase, software classification and software brief introduction.To realize final evaluation purpose, need to be directed to
The user comment data of mobile application are analyzed and determined, and are classified according to information such as the corresponding software versions of user comment
It screens, for the ease of the progress of work such as classifying to comment data, extracting, needs to obtain comment from user comment data
Relevant element data, in the present embodiment, when element data specifically includes user comment, software version, comment port, comment
Between, User ID should include also user's star, in order to as needed on the basis of there is corresponding classification or ranked data
In the case where, excessively huge data are compressed, reject low credit low price Value Data, to improve working efficiency.
It is as shown in Figure 1 going on smoothly convenient for data acquisition, the invention also includes the steps of crawl mobile application data
Suddenly, it specifically includes:
One, access comment port obtains web data, and search needs to grab the flag data of the mobile application of data, right
The corresponding webpage of flag data is grabbed；It needs to judge the loading method of webpage when grabbing data, and according to not
Data are grabbed using different analysis method with loading method, in the present embodiment, using the net of Jsoup analysis static loading mode
Page label data, using the web page tag data of HttpUnit crawl dynamically load mode；
Two, judge whether the web data contains respective element data；The webpage is grabbed if having and is stored in database；It is special
It is other, in order to realize the comprehensive analysis or comparative analysis applied, need when storing web data according to each movement more
Related data is handled using database is established respectively in order to targetedly quick；
Three, judge whether there are also subsequent web pages, go to step one if having, if without going to step four；
Four, webpage is converted into text formatting, positioning element data corresponding with comment data is obtained.
Through the above steps, the corresponding comment data of each mobile application in available each comment port, but due to comment
Data are usually to be made of natural language, and computer can not be identified and be handled, and to realize that corresponding judgement is screened, are needed to user
Comment data carry out structuring processing, emphasis refer to in user comment Feature Words or feature phrase carry out Feature Words rinse
It selects, specifically comprises the following steps: in the present embodiment
Step 1: carrying out participle and part-of-speech tagging processing to user comment data, noun n, verb v and adjective are extracted
A constitutes keyword set；
Define the keyword set K of corresponding i-th user comment of mobile applicationi, Ki={ w0/f0, w1/f1..., wk/
fk, wherein k=0,1,2......K-1, K are the participle number of i-th comment, wkFor kth in commentiA participle, fkFor wk's
Part of speech；The extraction process of keyword set is as shown in Figure 2；
Wherein participle tool refer to for being segmented, the identification of part-of-speech tagging, part of speech, the functions such as words identification application work
Tool, in the present embodiment specific to be Chinese data, therefore used herein is ICTCLAS；
Step 2: extracting keyword set KiIn only a comment data, by w all in former keyword setkDeposit optimization
Keyword set K afterwardsNewi, wherein the keyword set K after optimizationNewiIs defined as:
KNewi={ w0, w1......wj, wherein j=0,1,2......J-1, J are the keyword after i-th comment optimization
Number, wjFor the j-th word in comment optimization keyword set；
Step 3: extracting keyword set KiIn containing { n+a }, { v+a }, { n+v }, { n+v+a } comment data, Jiang Yuanguan
Keyword concentrates the w of non-akCorresponding part of speech deposit optimization keyword set KNewi, the corresponding crucial part of speech collection of part of speech deposit optimization
FNewi, wherein the keyword set F after optimizationNewiIs defined as:
FNewi={ f0、f1......fj, wherein fjFor the w in comment optimization keyword setjPart of speech；
Above-mentioned steps detailed process schematic diagram is as shown in Figure 3.
We obtain the user comment data of corresponding mobile application after above-mentioned steps, and obtain to comment data
Selection and separation, obtain can as the critical data for evaluating mobile application, including the corresponding keyword of all kinds of this journeys, but by
A large amount of repeated word (referring in this application with high frequency words generation) and nothing may be included in the comment for belong to natural language
The word of meaning, above-mentioned high frequency words and meaningless word may result in being affected for appraisal, from the angle of statistical property
Degree analysis, the size for the information content that a keyword word provides corresponding data weight coefficient in other words, is that can pass through
The frequency occurred in this class data is expressed, at the same in other data the frequency of occurrences lower keyword for target data
Evaluation for for there are more property of can refer to.Text characteristics based on comment data, can be at this using word frequency statistics
The method of word frequency statistics is utilized as the basis for effectively judging or screening its significance level in the scheme of invention, therefore in the application
Sentence the significance level or relevance of keyword and evaluation result in user comment data.
In the present invention, in order to realize the weight judgement judgement of keyword and mobile application in user comment, need
Keyword set K after obtaining optimizationNewiIn wjCorresponding part of speech fjFrequency T in feature databasej, include wjTextual data NjWith
And text sum N etc..And according to all wjTj、Nj, N solve corresponding words and concentrate the weight coefficient that scores mobile application of element,
The influence for feedback high frequency words to user comment, the frequency of keyword also must be incorporated into numerical procedure simultaneously.
To achieve the above object, it would be desirable to carry out the following contents, include the following steps:
Step 1: establishing one's own feature dictionary to each mobile application, in particular to according to the part of speech of Feature Words
Establish feature dictionary, including verb feature database, noun feature database, adjective feature database；
Step 2: the keyword w concentrated to each mobile application extracting keywordsj, wjIn corresponding part of speech fjFeature database in
Frequency Tj, include wjTextual data NjAnd text sum N；
Step 3: 1. calculating the weighted score S of each user comment and mobile application scoring according to formula, and judge that it is obtained
Divide and whether be greater than threshold alpha, is then judged as effective comment if more than threshold value, is otherwise determined as useless comment；
Wherein mjRefer to wjCorresponding part of speech fjAverage characteristics frequency in feature database, J are the key after i-th comment optimization
Word number；Formula 1. in introduce Laplce's smoothing algorithm to avoid null events appearance；
Its specific method process is as shown in Figure 4；
Above scheme process, which realizes primary user comment, rinses and selects work, and judges in a complete user comment
Cheng Zhong, it will usually be related to the multistage continuous acquisition judgement of multiport, therefore, will do it continuous updating in feature dictionary to repair
Correction data improves accuracy, when update, carries out different disposition according to situation difference, comprising:
If the Feature Words whole that updates or to reach a certain higher proportion (higher proportion be more than or equal to 80%) be shape
Hold word, then Feature Words are stored in adjective feature dictionary and increase frequency, be otherwise stored in character pair dictionary and increase frequency；If
Feature Words are not present in feature dictionary, then feature dictionary are added in the specific word, and set 1 for its frequency.
Based on above-mentioned basic scheme, its correlation has been obtained by analyzing us to a certain mobile application A progress Simulation evaluation
The following table 1 of each feature dictionary, table 2, shown in table 3,
Certain mobile application A verb feature dictionary of table 1
Certain mobile application A noun feature dictionary of table 2
Certain mobile application A adjective feature dictionary of table 3
Similarly, threshold value is gone by taking the user comment in a few money mobile application A, B, C, D, E as an example according to above-mentioned same steps
It is 1, chooses special characteristic word and calculate the pertinence score of user comment and judge whether related, and carried out pair in handmarking
Than obtaining the following table 4
4 pertinence score of table calculates (selection)
By it is found that due to having more key feature words in Article 3 comment, corresponding to comment on score higher in table 4.
It is also easier to be recognized by the system, by taking mobile application A as an example, carries out pertinence judgement from different incoming ends crawl user comment,
When commenting on item number less than 110, biggish randomness is presented in accuracy rate, until accuracy rate just tends to be steady directly when more than 110 or more
To stable in some stationary value, while in actual application, due to the short-term internal fault of mobile application, incoming end atmosphere shadow
The influence of the prominent situations such as loud or even artificial control, the concentration that will lead to particular community user comment in the period occur, for
For the evaluation of mobile application, if may be caused to the validity of evaluation very big when choosing without respective handling
Therefore the difficulty for influencing to have aggravated data processing significantly simultaneously in the specific implementation process, is answered in the acquisition scheme of user comment
When following following principle, comprising: user comment data are obtained from multiple incoming ends, between each incoming end and inside incoming end
The schemes such as time segment or random acquisition are taken to carry out rejecting processing to the abnormal data largely to happen suddenly if necessary.
It is evaluated for common mobile application, the user comment sum of acquisition should be not less than 1000, the acquisition of comment
Period should be not less than 30 days.The abnormal data occurred to avoid mass-sending and short-term concentration comment.
Finally it should be noted that above embodiments are only to illustrate the technical solution of the invention, rather than to this hair
It is bright create protection scope limitation, although being explained in detail referring to preferred embodiment to the invention, this field it is general
Lead to it will be appreciated by the skilled person that can be modified or replaced equivalently to the technical solution of the invention, without departing from this
The spirit and scope of innovation and creation technical solution.
Claims (4)
1. a kind of mobile application data processing method, which is characterized in that it, which contains, has the following steps,
Step A, for obtaining mobile application related data the step of, including obtain software mark data, the software mark number
According to including dbase, software classification and software brief introduction；
Step B, for obtaining the step of commenting on relevant element data from user comment data, the element data includes using
Family comment, software version, comment port, comment time, User ID；Specific step is as follows:
One, access comment port obtains web data, and search needs to grab the flag data of the mobile application of data, to mark
The corresponding webpage of data is grabbed；It needs to judge the loading method of webpage when grabbing data, and is added according to difference
Load mode is using different analysis method crawl data, in particular to using the web page tag of Jsoup analysis static loading mode
Data, using the web page tag data of HttpUnit crawl dynamically load mode；
Two, the corresponding mobile application of the flag data whether web data contains judged；If having each webpage of crawl and being stored in phase
Answer database；The database is that established respectively according to each mobile application as classification factor corresponding with each mobile application counts
According to library；
Three, judge whether there are also subsequent web pages, go to step two if having, if without going to step 4；
Four, all webpages with crawl are converted into text formatting, positioning element data corresponding with comment data is obtained；
Positioning element data corresponding with comment data is obtained specifically includes,
Step 4.1, participle is carried out to user comment data and part-of-speech tagging is handled, extracts noun n, verb v and adjective a
Constitute keyword set；Define the keyword set K of corresponding i-th user comment of mobile applicationi, Ki={ w0/f0, w1/
f1..., wk/fk, wherein k=0,1,2......K-1, K are the participle number of i-th comment, wkFor kth in commentiA point
Word, fkFor wkPart of speech；The functions such as wherein participle tool refers to for being segmented, part-of-speech tagging, part of speech identify, words identification
Application tool；
Step 4.2 extracts keyword set KiIn only adjective a comment data, by w all in former keyword setkIt is stored in excellent
Keyword set K after changeNewi, wherein the keyword set K after optimizationNewiIs defined as:
KNewi={ w0, w1......wj, wherein j=0,1,2......J-1, J are the keyword number after i-th comment optimization, wj
For the j-th word in comment optimization keyword set；
Step 4.3 extracts keyword set KiIn containing { n+a }, { v+a }, { n+v }, { n+v+a } comment data, by former keyword
Concentrate the w of non-akCorresponding part of speech deposit optimization keyword set KNewi, the corresponding crucial part of speech collection F of part of speech deposit optimizationNewi,
Keyword set F after middle optimizationNewiIs defined as:
FNewi={ f0、f1......fj, wherein fjFor the w in comment optimization keyword setjPart of speech；
Step C, specific for extracting keyword set and calculating the step of element is to mobile application scoring weight in keyword set
Step includes:
Step 1 is established one's own feature dictionary to each mobile application, in particular to is established according to the part of speech of Feature Words special
Levy dictionary, including verb feature database, noun feature database, adjective feature database；
Step 2, the keyword w that each mobile application extracting keywords are concentratedj, wjIn corresponding part of speech fjFeature database in frequency
Number Tj, include wjTextual data NjAnd text sum N；
Step 3, the weighted score S that each user comment and mobile application scoring are 1. calculated according to formula, and judge that its score is
It is no to be greater than threshold alpha, then it is judged as effective comment if more than threshold value, is otherwise determined as useless comment；
Wherein mjRefer to wjCorresponding part of speech fjAverage characteristics frequency in feature database, J are the keyword number after i-th comment optimization.
2. a kind of mobile application data processing method according to claim 1, which is characterized in that the threshold alpha is a use
In control keyword set capacity to control calculation amount or for rinsing the numerical value for selecting data correlation degree.
3. a kind of mobile application data processing method according to claim 1, which is characterized in that computing resource can be effective
In the case where handling all data, threshold alpha=1.
4. a kind of mobile application data processing method according to claim 1, which is characterized in that each movement to be evaluated
Using the user comment sum of acquisition should be not less than 1000, and the acquisition period of comment should be not less than 30 days, the period
Refer to deliver earliest in database and user comment that is delivering the latest time interval.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810741622.8A CN109145186A (en) | 2018-07-09 | 2018-07-09 | A kind of mobile application data processing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810741622.8A CN109145186A (en) | 2018-07-09 | 2018-07-09 | A kind of mobile application data processing method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109145186A true CN109145186A (en) | 2019-01-04 |
Family
ID=64799944
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810741622.8A Pending CN109145186A (en) | 2018-07-09 | 2018-07-09 | A kind of mobile application data processing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109145186A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111125193A (en) * | 2019-12-23 | 2020-05-08 | 精硕科技（北京）股份有限公司 | Multimedia abnormal comment identification method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105975487A (en) * | 2016-04-26 | 2016-09-28 | 昆明理工大学 | Method for judging correlativity of user comments of APP software |
| CN106547748A (en) * | 2015-09-16 | 2017-03-29 | 中国移动通信集团公司 | The creation method and device of a kind of APP index databases, the method and device of search APP |
| CN107451116A (en) * | 2017-07-14 | 2017-12-08 | 中国地质大学（武汉） | Raw big data statistical analysis technique in a kind of Mobile solution |
-
2018
- 2018-07-09 CN CN201810741622.8A patent/CN109145186A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106547748A (en) * | 2015-09-16 | 2017-03-29 | 中国移动通信集团公司 | The creation method and device of a kind of APP index databases, the method and device of search APP |
| CN105975487A (en) * | 2016-04-26 | 2016-09-28 | 昆明理工大学 | Method for judging correlativity of user comments of APP software |
| CN107451116A (en) * | 2017-07-14 | 2017-12-08 | 中国地质大学（武汉） | Raw big data statistical analysis technique in a kind of Mobile solution |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111125193A (en) * | 2019-12-23 | 2020-05-08 | 精硕科技（北京）股份有限公司 | Multimedia abnormal comment identification method, device, equipment and storage medium |
| CN111125193B (en) * | 2019-12-23 | 2023-08-29 | 北京秒针人工智能科技有限公司 | Method, device, equipment and storage medium for identifying abnormal multimedia comments |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106874378B (en) | Method for constructing knowledge graph based on entity extraction and relation mining of rule model | |
| Panigrahi et al. | Word2Sense: Sparse interpretable word embeddings | |
| CN104268197B (en) | A kind of industry comment data fine granularity sentiment analysis method | |
| CN106528599B (en) | A kind of character string Fast Fuzzy matching algorithm in magnanimity audio data | |
| Thakkar et al. | Graph-based algorithms for text summarization | |
| CN107193801A (en) | A kind of short text characteristic optimization and sentiment analysis method based on depth belief network | |
| CN110162630A (en) | A kind of method, device and equipment of text duplicate removal | |
| CN105630890B (en) | New word discovery method and system based on intelligent Answer System conversation history | |
| CN110362678A (en) | A kind of method and apparatus automatically extracting Chinese text keyword | |
| CN112347223B (en) | Document retrieval method, apparatus, and computer-readable storage medium | |
| CN102184256A (en) | Clustering method and system aiming at massive similar short texts | |
| CN110362601A (en) | Mapping method, device, equipment and the storage medium of metadata standard | |
| CN109918648B (en) | Rumor depth detection method based on dynamic sliding window feature score | |
| CN104679731B (en) | Extract the method and device of keyword in the page | |
| CN113076735B (en) | Target information acquisition method, device and server | |
| CN107247739A (en) | A kind of financial publication text knowledge extracting method based on factor graph | |
| CN109522416A (en) | A kind of construction method of Financial Risk Control knowledge mapping | |
| CN112905800A (en) | Public character public opinion knowledge graph and XGboost multi-feature fusion emotion early warning method | |
| CN104346382B (en) | Use the text analysis system and method for language inquiry | |
| CN113590810A (en) | Abstract generation model training method, abstract generation device and electronic equipment | |
| CN107526721A (en) | A kind of disambiguation method and device to electric business product review vocabulary | |
| CN116362243A (en) | Text key phrase extraction method, storage medium and device integrating incidence relation among sentences | |
| Elfida et al. | Enhancing to method for extracting Social network by the relation existence | |
| Fharook et al. | Are you a hero or a villain? A semantic role labelling approach for detecting harmful memes. | |
| CN110413985B (en) | Related text segment searching method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190104 |